TEXT CROSS
WITHIN THE
BOOK ONLY
168160
= CD
> m
OSMANIA UNIVERSITY LIBRARY
Call No. GT1 A<^\O*5vtC. Accession No. (J1\\I3V1T-
Author OVVr^ £tvv*\ CXVvck
Title C-tlA^Vo
This book should be returned on or before theuate last marked below.
HIGH POLYMERS
Vol. I: Collected Papers of W. H. Carothers on High
Polymeric Substances
Edited by H. Mark and G. S. Whitby
Vol, II: Physical Chemistry of High Polymeric Systems
Second Edition. By H. Mark and A. V. Tobolsky
Vol. Ill: Mechanism of Polymer Reactions
By G. M. Burnett
Vol. IV: Natural and Synthetic High Polymers
Second Edition. By K. H. Meyer
Vol. V: Cellulose and Cellulose Derivatives
Second Edition, In three parts. Edited by Emil Ott,
H. M. Spurlin, and M. W. Grafflin
Vol. VI: Mechanical Behavior of High Polymers
By Turner Alfrcy, Jr.
Vol. VII: Phenoplasts: Their Structure, Properties, and
Chemical Technology
By T. S. Carswell
Vol. VIII: Copolymerization
By Turner Alfrcy, Jr., John J. Bohrcr, and H. Mark
Vol. IX: Emulsion Polymerization
By F. A. Bovcy, I. M, Kolthoff, A. Mcdalia, and E. J.
Meehao
Other volumes in preparation
HIGH POLYMERS
A SERIES OF MONOGRAPHS ON THE CHEMISTRY, PHYSICS
AND TECHNOLOGY OF HIGH POLYMERIC SUBSTANCES
Editorial Board
H. MARK, Brooklyn, New York H. W. MELVILLE, Birmingham, England
C. S. MARVEL, Urbana, Illinois G. S. WHITBY, Akron, Ohio
VOLUME V
Cellulose and Cellulose Derivatives
Second Completely J&vtscd and Augmented Edition
Edited by
Emil Ott and Harold M. Spurlin, Cocditors
Mildred W. Grafflin, Assistant Editor
Part II
CELLULOSE
AND
CELLULOSE DERIVATIVES
Second Completely Revised and Augmented Edition
Prepared under the Editorship of
EMIL OTT HAROLD M. SPURLIN
MILDRED W. GRAFFLIN
Research Department, Hercules Powder Company
Wilmington, Delaware
P ART II
INTERSCIENCE PUBLISHERS, INC., NEW YORK
INTERSCIENCE PUBLISHERS LTD., LONDON
LIBRARY OF CONGRESS CATALOG CARD NUMBER 5>7161
Copyright, 1954, by
INTERSCIENCE PUBLISHERS, INC.
All Rights Reserved. This book or any
part thereof must not be reproduced in any
form without permission of the publisher
in writing. This applies specifically to
photostat and microfilm reproductions.
INTERSCIENCE PUBLISHERS, INC.
250 Fifth Avenue, New York 1, N. Y.
For Great Britain and Northern Ireland:
INTERSCIENCE PUBLISHERS LTD.
88/90 Chancery Lane, London W. C. 2
PRINTED IN THE UNITED STATES OF AMERICA BY
MACK PRINTING COMPANY, E ASTON, PENNSYLVANIA
CONTENTS
Parti
I. Introduction (On and TENNENT)
II. Occurrence of Cellulose (WARD)
A. Significance of Terms
B. Formation of Cellulose
C. Identification of Cellulose
D. Sources of Cellulose
E. Natural Occurrence of Combined Cellulose
III. Chemical Nature of Cellulose and Its Derivatives
A. Historical Survey (PURVES)
B. Chain Structure ( PURVES)
C. Degradation of Cellulose (McBuRNEY and Siu)
D. End Groups (SOOKNE and HARRIS)
E. Base-Exchange Properties (SOOKNE and HARRIS)
IV. Structures and Properties of Cellulose Fibers
A. Interaction and Arrangement of Cellulose Chains (MARK)
B. Submicroscopic Structure (HOWSMON and SISSON)
C. Microscopic Structure (HOCK)
D. Structure-Sorption Relationships (HOWSMON)
V. Properties of Substances Associated with Cellulose in Nature
A. The Physical and Chemical Nature of Wood (LEWIS and RITTER)
B. Noncellulosic Carbohydrates (NORMAN)
C. Lignin and Other Noncarbohydrates (BRAUNS)
Part II
VI. Preparation of Cellulose from Its Natural Sources 511
A. Wood Pulp (HOLZER) 511
1. General Considerations 513
2. Physical and Chemical Factors in Pulping 520
3. The Sulfite Process. . . 524
4. The Soda Process 534
5. The Kraft (Sulfate) Process. . 537
6. Other Pulping Processes 545
7. Properties of Pulps and Their Constituents Affecting End Use 547
VI CONTENTS
B. Cotton Lint and Linters (MARTIN). . 550
1. Types of Cottonseed Fibers 550
2. Economics of Utilization of Cottonseed Fibers in the United States. 553
3. Cottonseed Fibers in Other Countries. . . 555
4. Linters Purification 556
5. Uses for Chemical Cotton . . 563
C. Rags (LEWIS) 566
D. Bast Fibers, Fibrovascular Elements, Cereal Straws, and Grasses (WELLS) 573
1. Bast Fibers 574
2. Cellulose Pulps from Bast Fibers 578
3. Fibrovascular Elements 580
4. Cereal Straws.. . 581
5. Esparto Grass . 584
6. Bamboo 585
7. Bagasse and Cornstalks . 586
8. Value of Annual Plants 586
VII. Bleaching and Purification of Wood Cellulose (HATCH) 589
A. General Principles 589
B. Specific Bleaching Treatment 590
1. Chlorination 590
2. Alkaline Extraction 597
3. Hypochlorite Bleaching 597
4. Chlorine Dioxide and Sodium Chlorite 600
5. Peroxide Bleaching 602
6. The Washing Operation.. 603
C. Equipment for Commercial Chlorination, Bleaching, and Alkaline
Extraction . ... 603
1. Chlorination Equipment 604
2. Bleaching Equipment. .. . 607
3. Alkaline Extraction Equipment . 612
D. Quality Requirements for Specific Products. . 613
1. Nonpermanent Papers 613
2., Permanent Papers 613
3. Purified Pulp for Cellulose Derivatives 614
4. Ground wood and Semichemical Pulps. 617
E. New Trends , . 617
1. Use of New Reagents 617
2. Continuous versus Batch Processing 619
3. Chemical Control 619
VO. Properties and Treatment of Pulp for Paper (CLARK). . 621
A. Physical Properties of Paper Pulp 621
1. The Strength of Paper 623
2. Common Tests for Paper 625
3. Pulp Testing 626
CONTENTS Vll
B. Beating . . 633
1. Influence of Moisture on Pulp. . 634
2. Beating Equipment . 635
3. The Action of Beating Equipment 638
4. Measurement of Beating. . . . 641
C. Theories of Beating 643
1. Chemical Theory of Beating. . 644
2. Physical Theories of Beating. . 645
3. Composite Theory of Beating. 654
D. Results of Beating 657
1. Effects of Beating on the Fibers. 657
2. Effects of Beating on the Pulp . . . 658
3. Effects of Beating on the Paper. . 661
4. Factors Affecting the Rate of Beating . . 662
5. Phenomena Relating to Beaten Pulp 663
E. Pulp Chemistry and Papermaking Properties. 665
1. Degree of Polymerization. . . 666
2. Hemicellulose Content. . 668
3. Lignin Content 670
4. Cooking Reactions. . 671
IX. Derivatives of Cellulose 673
A. Reactivity and Reactions of Cellulose (SPURLIN) 673
1. Uniformity of Substitution. . 674
2. Reaction of Cellulose Fibers. 691
3. Chemical Factors Influencing Reactivity 702
4. Conclusion 712
B. Inorganic Esters (BARSHA). 713
1. Nitrocellulose 713
2. Cellulose Sulfate . . 755
3. Cellulose Phosphate . . 760
4. Cellulose Esters of Other Inorganic Acids . 762
C. Organic Esters (MALM AND HIATT) . . 763
1. Aliphatic Fatty Acid Esters. . 766
2. Other Aliphatic Esters. . 809
3. Miscellaneous Esters . 815
D. Alkali and Other Metal Derivatives (NICOLL, Cox, AND CONAWAY). . . . 825
1. Alkali Cellulose .825
2. Mercerization of Cellulosic Textiles . 863
3. Metal Alcoholates of Cellulose. . . 871
4. Cuprammonium-Cellulose Complexes. . 874
5. Cellulose-Organic Base Complexes 879
E. Ethers (SAVAGE, YOUNG, AND MAASBERG) ... . . 882
1. History . 882
2. Chemistry of the Etherification Reaction. . 883
3. Properties of Cellulose Ethers. .. 905
4. Ethyl Cellulose . 913
Vlll CONTENTS
5. Methyl Cellulose 930
6. Carboxyraethyl Cellulose . . 937
7. Hydroxyethyl Cellulose. . . . 945
8. Benzyl Cellulose 949
9. Addition to Cellulose of Olefins Activated by Polar Substituent Groups 954
F. Xanthates (KLINE) . .... 959
1. Preparation of Viscose . 960
2. Mechanism of Cellulose Xanthate and Viscose Formation 975
3. Characteristics and Reactions of Viscose 987
G. Degradation of Cellulose Derivatives (McBuRNEv) 1019
Bf. Water-Soluble Cellulose Derivatives 1020
2. Organosoluble Cellulose Derivatives. . 1026
Part III
X. Physical Properties of Cellulose and Its Derivatives in Solution
A. Solubility (SPURLIN)
B. Thermodynamic Properties of Solutions of Long-Chain Com-
pounds (HUGGINS)
C. Behavior of Cellulose Xanthate as a Polyelectrolyte (SWANSON)
D. Determination of Molecular Weight and Molecular- Weight Distri-
bution (Doxy and SPURLIN)
E. Theory of the Viscosity of Dilute Solutions of Long- Chain Com-
pounds (HUGGINS)
F. Flow Properties (DAVIS and ELLIOTT)
XL Mechanical Properties of Cellulose and Its Derivatives (WAKEHAM)
A. Mechanical Behavior of Cellulose Substances
B. Molecular and Morphological Considerations of Extension
C. Ultimate Strength
XII. Tests for Cellulose and Its Derivatives
A. Cellulose Tests (MARTIN)
B. Cellulose Derivatives Tests (KLUG and GENUNG)
Appendixes (GLOOR and KLUG)
A. Some Properties of Commercial Chemical Celluloses
B. Degree of Substitution Relationships
C. Viscosity
D. Solubility
E. Properties of Cellulose Derivatives
F. Identification of Cellulose Derivatives
G. Trade Names and Generic Names for Cellulose and Cellulose
Derivatives
H. Definitions and Constants
Author Index, Parts I-III
Subject Index, Parts I-III
CELLULOSE
AND CELLULOSE DERIVATIVES
Part II
Chapter VI
PREPARATION OF CELLULOSE FROM ITS
NATURAL SOURCES
A. WOOD PULP1
WALTER F. HOLZER
The commercial pulping of wood as it is known today is based on proc-
esses less than one hundred years old. Watt and Burgess2 in 1854 pat-
ented the pulping of wood in hot aqueous sodium hydroxide solutions under
pressure. From the cooking chemical the name "soda" process is derived.
Tilghman3 was granted a patent in 1867 on his discovery that wood is
pulped in aqueous sulfur dioxide solutions in which part of the sulfur diox-
ide is combined with calcium as the bisulfite. The sulfite process produces
a white pulp, and once the corrosive behavior of the cooking chemicals was
overcome, it became for several decades the leading method for preparing
pulp. Dahl4 obtained a patent in 1889 on a modification of the soda proc-
ess in which sodium sulfate replaced sodium carbonate as make-up in the
recovery of cooking chemicals. The sulfate is reduced to sulfide when the
organic material in the spent liquor is burned. This pulping process has
been called "sulfate" from the make-up chemical, or "kraft" from the Ger-
man word meaning "strength*" in recognition of the strength of its paper
products. Because it is less critical of wood species, because of the strength
of its products, and because of the advantages of chemical recovery, the
kraf t process has displaced sulfite by a wide margin in amount of pulp pro-
duced. Many other methods have been proposed for pulping wood, but
1 In this section has been incorporated pertinent material from C. M. Koon's section
on "Wood Pulp" which appeared (pp. 475-518) in the original edition of this book.
2 C. Watt and H. Burgess, U. S. Patents 1448 and 1449 (1854); E. Hagglund, Chem-
istry of Wood, 3d ed., Academic Press, New York, 1951, p. 414.
* B. C. Tilghman, U. S. Patent 70,485 (Nov. 5, 1867); J. D. Rue, Paper Trade /.,
81, 54 (Oct. 15, 1925).
4 Wf O. Hisey, in L. E. Wise, editor, Wood Chemistry. Reinhold, New York, 1944, p.
716.
511
512
CELLULOSE
O
3 C
« 2
co d
t!
JS* o
fif
W 'S
si
n . o
^^ «*-^
8 -d
*j "p<
CO d
•d §
5
N- i-H O
a> oo *-•
CO <M 00
s
3
^ 22 P
C^ CO Is*
rt< I>- b-
<N 00 10
CM a>
CS|
rH CO
rf *O
O5 rH O5 t*» C5 ^C
r-H C^ T— I CM ^O O^
O (N 00 t> h- t^
esf o ic eJ CD o
i-i CO iO CO <N <N
CD lO CO CO 00 CM
CO O^ CM iC CO
»O b- Cfc> 00 CO
"ss
s
CO Is* O3 »"H
CO CO 00 CO
o >o t^ t^
O »-O
CO CO
CM CM
lO CO
Oi rH
00 O
sss
VI. PREPARATION FROM NATURAL SOURCES 513
few have developed beyond the laboratory stage, and only one, the neutral
sulfite process, in which a liquor containing sodium sulfite and sodium bi-
carbonate is used, has achieved any commercial importance.
Softwoods or conifers (gymnospenns) are by far the more important type
of trees in pulping and subsequent papermaking because of their longer
fibers (about 3-4 mm.)- Hardwoods or deciduous trees (angiosperms)
with their shorter finer fibers (1 mm.) are used in special applications where
paper properties such as surface smoothness or softness are of value. In
some areas where softwoods are becoming scarce, or to achieve better
balance in the use of the wood, the utilization of hardwoods is increasing,
especially through application of the neutral sulfite process.
The size of the pulp industry and its growth in the last quarter of a cen-
tury are illustrated in Table I.6 The production of pulp in 1950 was over
three and one-half times that in 1925. In that period the production of
soda pulp remained static, close to the 500,000-ton level. Groundwood
(a pulp produced mechanically by defibering wood on a grindstone) in-
creased less than 40%. Sulfite production doubled. The miscellaneous
small processes (mostly high-yield pulps for paperboard and wallboard)
increased from almost nothing to 1,700,000 tons. The most phenomenal
growth was in kraft which started the period with a production of only
400,000 tons or 10% of the total, and ended it with 7,500,000 tons or a little
over 50% of the total.
1. General Considerations
(a) COMPOSITION OF WOOD AND PULP
The composition of a softwood and of a hardwood in terms significant
in pulping are diagrammed in Figure 1A. The values given are considered
generally typical. Cellulose is the alpha fraction of a holocellulose residue.
This fraction usually constitutes slightly over half the total wood sub-
stance, and hardwoods are generally richer in it than are softwoods. Hemi-
celluloses comprise the balance of the carbohydrate fraction and usually
amount to about 20%. Softwoods contain just under this figure, and hard-
woods just over. A typical lignin content of a softwood is 28% with a range
of 25-32%. The same fraction of a hardwood is 22% with a range of about
18-24%. The extractives, including resins, gums, fats, waxes, and coloring
matters, will average 2-3%, but, of course, vary widely with the species.
The sulfite or the kraft process will produce a pulp suitable for average
6 U. S. Bureau of the Census, Pulp & Paper, 25, No. 7, 152 (1951).
514
CELLULOSE
paper-making in about 45% yield based on the original wood. This yield
is represented by the block graph in Figure IB. On the basis of the pulp,
the cellulose fraction now amounts typically to 85%, the hemicelluloses to
11%, the lignin to 3%, and the extractives to 1% or less. These fractions
will vary from softwood to hardwood pulps in about the same way that
they did in the woods. Cooking has altered the properties and composition
of all fractions. The degree of polymerization (D.P.) of the cellulose has
a
52%
b
17%
c <t
28% 3%
SOFTWOOD
a
54%
b
21%
c d
22%3%1'
HARDWOOD
SOFTWOOD PULP
— 50% •-)-* 50%-
HARDWOOD PULP
Fig. 1. The approximate chemical composition of softwood and hard-
wood, and of pulps prepared from them. (A) Wood; (B) pulp; (a) cellu-
lose, (b) hemicelluloses, (c) lignin, (d) extractives, (e) loss in pulping. All
percentages in A based on total wood, those in B on pulp.
been lowered and has been Widened in range. The nonresistant hemicel-
luloses have been largely removed along with some of the resistant ones,
and the hemicellulose fraction now consists of the balance of the resistant
hemicelluloses and degradation products of cellulose. The amount will
vary on either side of the figure given as much as 4%, depending on the
wood species and the cook. The residual lignin will vary from less than 1%
for a well-cooked sulfite pulp to 5% in a raw kraft pulp. Further changes
in the pulp on bleaching will be virtually complete removal of lignin, and
reduction of the hemicellulose fraction. For a paper pulp this reduction is
small. For a dissolving pulp this may amount to removal of all but the
most resistant fraction.
VI. PREPARATION FROM NATURAL SOURCES
515
(b) STRUCTURE OF THE CELL WALL
Some familiarity with the microstructure of wood is necessary to under-
stand the processes of pulping. The arrangement of the fibers in wood and
their relation to each other have been discussed in Chapter V-A. The
wood technologists' conception6-7 of the structure of the cell wall is shown
in Figure 2.
/^
/
f\
4-*
j
- • &
'
£
1
^
fyfr
'• :. '•'••'..'•• ' .' •'•':
i •;.:-:::>::vi ' v.V
1
I
: •:•;.• >::^::V-:..'-i
B
Fig. 2. Thfe compound cell wall of wood fibers (Jahn and Holrnberg7). Cell wall
parts according to I. W. Bailey.
(A) Cross section of fiber
(B) Section of two adjacent cell walls
(a) Intercellular substance
(b) Primary (cambial) wall
(c) Outer layer secondary wall
(d) Central layer secondary wall
(e) Inner layer secondary wall
(/) Compound middle lamella
(g) Complete secondary wall
(b-e) Complete cell wall of one fiber
For some time it has been thought from observation of stained wood sec-
tions that a large portion of the lignin is concentrated in the compound
middle lamella (the region represented in Figure 2 by the portion labeled/),
6 I. W. Bailey, Ind. Eng. Chem., 30, 40 (1938).
7 E, C. Jahn and C. V. Holmberg, Paper Trade J., 109, 30 (Sept. 28, 1939).
516
CELLULOSE
that is, the intercellular membrane plus the primary wall of the fiber on
each side. Bailey 8 gave quantitative support to this observation by devis-
ing a microtechnique for isolating a small quantity of middle lamella. In
this region of Douglas fir he reported the composition as 71% lignin, 4%
Fig. 3. The surface of an undercooked pine tracheid at a bordered pit showing the
lignin sheath and underlying primary fibrillar structure. Note the spot where the lignin
sheath is mechanically broken. No description of the membrane under the pit aper-
ture has been advanced (Horio, Kobayashi, and Komagata10). Magnification X 7,500.
cellulose, and 14% pentosans. Lange9 obtained similar results by optical
absorption methods which gave a picture of the composition of the entire
8 A. J, Bailey, Paper Ind., 18, 379 (1936); Ind. Eng. Chew., Anal. Ed., 8, 52, 389
(1936).
9 P. W. Lange, Svensk Pappcrstidn., 48, 241 (1945); Chent. Abstracts, 39, 3661 (1945).
VI. PREPARATION FROM NATURAL SOURCES 517
cell wall. He also found 70% lignin in the middle lamella. There was a
sharp decrease in lignin content in the secondary wall with further decrease
toward the lumen. Hemicellulosic material was found to be more concen-
trated near the middle lamella than adjacent to the lumen. Horio,10
using chromium shadow technique, has recently obtained striking optical
confirmation in an electron micrograph of the surface of a very raw fiber
(Fig. 3). The surface appeared very smooth in the picture, but on one
spot the surface layer had been chipped much as an enamel coating might
be, and underneath the fibrillar structure of the cellulose fiber was exposed.
(c) DEFINITION OF AND REMARKS ON PULPING
Chemical pulping of wood may be defined as a process in which the fiber-
cementing material (lignin) is dissolved by reasonably specific reagents,
and the individual fibers are liberated.
In the laboratory the reagents used to isolate holocellulose11 are quite
specific in their removal of lignin. In commercial operation none of today's
practical pulping agents can even approximate this quantitative removal
of lignin, and one of the major challenges to the industry is a more efficient
recovery of cellulose. A more complete removal of lignin in the cook would
entail an unreasonable loss of cellulose. On the other hand, increasing
yield by cooking less results in a greater proportion of lignin to cellulose.
Since the lignin is concentrated on the surface of the fiber, papermaking
properties are then impaired.
Lewis and Richardson12 prepared from a single sample of wood a series
of raw alkaline pulps in which the various pulps had a progressively smaller
amount of lignin. The beating characteristics and the swelling behavior
in cuprammonium hydroxide of these pulps were determined. As the
lignin content of the pulps decreased from 22.2 to 8.6%, the pulps beat
more readily, felted more satisfactorily into handsheets, and attained higher
physical tests. Cuprammonium hydroxide had almost no effect on the
pulp with 22.2% lignin. As the lignin content of the pulps decreased, a
"ballooning" type of swelling began to occur on contact of the fibers and
reagent. It appeared that swelling was restricted by a resistant sheath
surrounding the fiber. Wherever this sheath was absent or weakened, the
fiber would swell to several times its diameter giving the appearance of a
10 M. Horio, K. Kobayashi, and O. Komagata, unpublished work.
11 W. G. VanBeckum and G. J. Ritter, Paper Trade /., 105, 127 (Oct. 28, 1937).
12 H. F. Lewis and C. A. Richardson, II, Paper Trade J., 109, 48 (Oct. 5, 1939).
OJ.S CELLULOSE
series of balloons. This study supported the hypothesis that wood fibers,
as they naturally occur, are surrounded by a sheath of high lignin content.
(See also Chapter I V-C for a discussion of swelling and ballooning of cellulose
fibers.)
Commercial pulping of wood, then, is always a compromise. Although
lignin is largely concentrated in the middle lamella, its distribution through-
out the secondary wall makes it difficult to remove quantitatively. On
the other hand, the lower carbohydrates are rather easily soluble in pulp-
ing chemicals, and under conditions necessary to meet commercial pulping
schedules, even the resistant cellulose is attacked. The common pulping re-
agents in use today in order of decreasing specificity in rendering lignin
soluble are:
Most specific : Chlorine
Sodium sulfite plus sodium bicarbonate
Sulfurous acid plus a bisulfite
Sodium hydroxide plus sodium sulfide
Least specific : Sodium hydroxide
Although it is the best of the commercial pulping agents, chlorine cannot
be used on wood because of its cost. Its use is confined to pulping grasses
in which the lignin content is low, or to bleaching wood pulp after the bulk
of the lignin has been removed. The mild action of sodium sulfite in the
mildly alkaline range is too slow for general pulping, and its use is generally
limited to production of high-yield pulps. The specific action of the sulfite
ton in sulfurous acid plus a bisulfite is partially nullified by the hydrolytic
action of the acid. The net result is that acid sulfite is less specific than
neutral sulfite. In general, strong alkali is the least specific of the com-
mercial pulping agents because of its rather slow action on lignin and its
tendency to attack carbohydrates. Sodium sulfide greatly increases the
rate of dissolution of ligninf and thus improves specificity. Although
sodium hydroxide was the first successful chemical pulping agent, its use
is now limited to cooking hardwood.
One of the most significant points evident from the above list is the im-
portance of sulfur in commercial pulping. Virtually all of the wood pulp
is cooked with some form of sulfur — alkaline sulfide in kraft, acid sulfite
in the sulfite process, and mildly alkaline sulfite in the neutral sulfite process.
The reactivity of these forms of sulfur with lignin, and the cheapness of
these chemicals indicate that sulfur will remain of prime importance in
pulping for at least the foreseeable future.
VI. PREPARATION FROM NATURAL SOURCES 519
(d) STEPvS IN CHEMICAL PULPING COMMON TO ALL PROCESSES
For a better understanding of the ensuing discussion it is desirable at
this time to describe the steps in wood pulping in a very general way.
More specific remarks will be made on each process in separate sections.
(7) Wood Preparation
Bark must be removed from the wood since in the cook it would be a
source of dark color and dark fibers in pulp. Barking is sometimes done by
hand-peeling the logs in the woods while the bark is loose in the early rush
of growth in spring. More commonly it is done mechanically at the mill.
The logs are reduced to 4-foot or sometimes 8-foot lengths, and fed into
one end of a large, horizontal drum which is constructed from steel bars.
The drum is rotated on its longitudinal axis and the bark is knocked off by
the bolts of wood falling against each other and the steel drum. Bark
drops from the drum between the bars, and the wood being fed in one end
of the drum forces the flow from the other end. The most recent method
of barking is to direct a high-pressure jet of water perpendicularly to the
axis of the log, and either by turning the log and moving the jet along its
length, or by rotating the jet and moving the log past it, the bark is re-
moved. In some installations the log is passed through a ring of water
jets which remove the bark. Water pressures up to 1400 Ib./sq. in. are
used.
The wood must be reduced to units small enough so that the cooking
liquor can penetrate completely and uniformly in a reasonable time. This
is done by cutting chips which have the preferred dimensions: b/s to 7/s
inch in length, Vi6 to l/$ inch in thickness, and 1/2 to 1 inch in width.
(2} Digestion
Chips are fed into large steel pressure vessels known as digesters, which
have capacities ranging from 1500 to 3500 cu. ft. for kraft, and from 3500
to 12,000 cu. ft. for sulfite. Cooking liquor is added, and the contents of the
digester are heated with steam and under pressures ranging from 75 to
125 Ib./sq. in. according to a predetermined cooking curve. When the
wood has been cooked, the softened chips and spent liquor are discharged
from the digester by blowing under pressure into pits or tanks.
(5) Washing and Screening
The pulp is freed from spent liquor by washing, and is screened to remove
uncooked wood. A coarse screen or knotter first removes uncooked chips
520 CELLULOSE
and knots from the dilute water suspension of pulp, and then fine screens
remove the uncooked fiber bundles. The pulp can then be used in the un-
bleached state, or further purified by bleaching.
2. Physical and Chemical Factors in Pulping
(a) PENETRATION OF WOOD BY LIQUIDS
One of the first considerations in pulping is the need to bring the cooking
liquor into intimate contact with all portions of the chip so that delignifica-
tion can proceed uniformly. The most extensive studies in this field were
made by Maass and coworkers whose work has been summarized by
Beazley, Johnston, and Maass.13 The principal path of penetration in the
chip is longitudinal, that is, parallel to the fiber length. As a rough aver-
age, penetration in this direction is 100 times as rapid as in either radial or
tangential direction. The one exception to this is sodium hydroxide solu-
tion or an equivalent strong alkali, which can penetrate wood with almost
equal rapidity from any direction. The flow through sap wood is much
greater than through heartwood of the same species. Rate of flow in-
creases with temperature faster than can be explained on the basis of vis-
cosity-temperature relations. Air in the fiber lumens greatly hinders
penetration. Contrary to popular belief, resins have only a minor retard-
ing effect. Penetration of jack pine increased only 20% on removal of
resins. McGovern and Chidester14 have demonstrated that wetting agents
as a class do not facilitate penetration of wood by cooking liquor. Wetting
agents promote surface wetting through lowered surface tension of the
liquid, but since capillary rise is important in penetration, and since it
decreases with decrease in surface tension, wetting agents actually decrease
penetration. Poorer penetration was observed in cooks to which wetting
agents were added.
(b) EFFECT OF TEMPERATURE
Temperature increases the rate of digestion of wood. Maass and co-
workers16 found that the rate of delignification approximately doubles for
a rise in temperature of 10°C. In commercial practice, high temperatures
18 W. B. Beazley, H. W. Johnston, and O. Maass, The Penetration into Wood of Cooking
Liquors and Other Media, Canada Dept. of Mines and Resources, Lands, Parks, and
Forests Branch, Dominion Forest Service, Bull. No. 95, Ottawa, 1939.
14 J. N. McGovern and G. H. Chidester, Paper Trade J., Ill, 35 (Dec. 12, 1940).
18 A. J. Corey and O. Maass, Can. J. Research. 14B, 336 (1936); J. M. Calhoun, F. H.
Yorston, and O. Maass, Can. J. Research, 17B, 121 (1939).
VI. PREPARATION FROM NATURAL SOURCES 521
are frequently used to increase the rate of cooking. However, this prac-
tice is not without its drawbacks, since cellulose suffers degradation from
heat, and this effect also increases with temperature. McGovern and
Chidester16 found that in the range of 130-150°C. pulp yields suffered a
loss of 0.5% for an increase of 10°C., and physical properties of the pulps
were degraded.
(c) EFFECT OF CHEMICAL CONCENTRATION
An increase in the concentration of cooking chemical will increase the
rate of delignification.17"20 Since in sulfite pulping one of the chemicals,
sulfur dioxide, is a gas, an increase in the concentration of the cooking
liquor must be accompanied by an increase in digester pressure in order to
maintain the concentration. In general, it has been found that both yield
and physical properties of pulp are less sensitive to an increase in chemical
concentration than to an increase in temperature. Within reasonable limits
it is preferable to increase the rate of cooking with chemical concentration.
(d) EFFECT OF WOOD PROPERTIES
The cellulose in a fiber can be said to have its optimum properties while
it is still a unit in sound wood. Any pulping reaction subtracts from this
optimum condition. It seems obvious then that the properties of wood
have a basic relation to the properties of pulp.
(1) Sapwood and Heartwood
Sap wood is the ring of light-colored wood adjacent to the bark. Al-
though no longer alive and growing, this wood is still taking part in the life
process by transporting water from the roots to the crown of the tree.
Heartwood is the darker wood in the center of the trunk. This wood no
longer takes part in the life process and contributes only strength to the
trunk. Its darker color is caused by the deposition of coloring matters,
resins, etc. McGovern and Chidester21 reported that there was relatively
little difference in pulps from pulping of sap wood and heartwood. The
drier condition and extractives in heartwood made pulping slightly more
16 J. N. McGovern and G. H. Chidester, Paper Trade J.t 106, 39 (June 2, 1938).
17 J. N. McGovern, Paper Trade J., 103, 29 (Nov. 12, 1936).
18 M. W. Bray, J. S. Martin, and S. L. Schwarz, Paper Trade J., 105, 39 (Dec. 9,
1937).
19 W. F. Holzer, Paper Trade J., 118, 35 (Apr. 20, 1944).
20 W. Pittam, Tech. Assoc. Papers, 29, 613 (1946); Pulp & Paper, 21, 76 (Nov., 1947).
21 J. N. McGovern and G. H. Chidester, Paper Trade /., 107, 34 (Oct. 6, 1938).
522 CELLULOSE
difficult. Pulps from sapwood tended to have a lower burst and higher
tear. In some species where the extractives have a specific effect on the
pulping reaction, there is marked difference in pulping of heartwood and
sapwood. This will be discussed later.
(2} Springwood and Summerwood
The alternate light and dark rings visible in the cross section of a log
mark the annual growth of the tree. Each year one light-colored ring and
one dark-colored ring are formed. The first, springwood, consists of large-
diameter, thin-walled fibers which are formed when there is need for a
large volume of water for growth in spring. The second, summerwood, has
smaller diameter, thick-walled fibers formed during summer. Probably
because of the large difference in the wall thicknesses, the relative amounts
of springwood and summerwood have a marked effect on pulp properties.
Holzer and Lewis22 separated springwood and summerwood of Douglas fir
by hand methods and cooked them separately. Springwood cooked with
more difficulty, gave lower yields, had a higher burst but lower tear than
summerwood. After beating, the springwood fibers were badly cut and
fibrillated; the summerwood fibers showed little visible effect. Hammond
and Billington23 found that burst, fold, and sheet density increased with
the percentage of springwood while tear decreased. McGovern and Chi-
dester24 reported an increase in springwood from butt to top of the tree.
There was a decrease in yield, a small decrease in burst, and a large de-
crease in tear in pulps in the same direction. These same investigators25
in a study of the southern pines found a greater difference from a 10% in-
crease in percentage of springwood, than among the various species studied.
The ratio of springwood to summerwood appears to be one of the most
critical physical characteristics of wood in determining pulp properties.
Since the ratio will change with rate of growth, with the age of the tree, and
from bottom to top, it is readily realized that very little control can be
exercised by the pulp producer. This is one excellent example of the
heterogeneity of wood, and serves to emphasize that any pulp property
can only be an average of the properties of the individual fibers.
22 W. F. Holzer and H. F. Lewis, Tappi, 33, 110 (1950).
28 R. N. Hammond and P. S. Billington, Tappi, 32, 563 (1949).
24 J. N. McGovern and G. H. Chidcster, Paper Trade /., 106, 37 (June 9, 1938).
26 G. H. Chidester, J. N. McGovern, and G. C. McNaughton, Paper Trade J., 107,
36 (July 28, 1938).
VI. PREPARATION FROM NATURAL SOURCES 523
(.?) Species
The principal effect of species is that of fiber properties as determined by
growth habit. The largest difference due to species is between hardwoods
and softwoods. The former have fibers to the order of 1 mm. long and 10-
25 microns in diameter. The latter have fibers from 2.5 to 4.5 mm. long
with diameters of 25-50 microns. Aside from this there are characteristic
differences in various softwood species. For example, fibers in species
characteristic of eastern Canada and northern United States from the
Great Lakes eastward, are shorter and narrower than those found on the
Pacific Coast or in the South. Papers made from the finer fibers tend to be
better formed and smoother; those from the coarser fibers will have poorer
formation, rougher surface, but better tearing strength. Space will not
permit specific comment possible on nearly eVfcry species.
Other variations within the species include pigments which affect color
and bleachability of the pulps; extractives which well may cause trouble
in cooking from pitch deposits in the system, or reduced pulp yields;
density which will affect pulp yield and production; lignin content which
can alter the ease of cooking; and hemicellulose content which has direct
relation to physical pulp properties.
(4) Decay in Wood
When wood was plentiful and demand was small, only the best and
soundest was used. Now that supplies are limited and demand is sharply
up, all the wood must be used, and each forest industry must use down
through the lowest quality it can tolerate. Since lower quality wood can
be used for pulping than for lumber, much decayed wood finds it way to
the pulp mill. Holzer26 described the effect of wood decay on pulp proper-
ties. Pulp becomes noticeably darker with incipient decay, and suffers
further loss in brightness as decay progresses. The acid formed in the
process of decay renders lignin less soluble and noticeably slows the rate
of cooking. Fiber properties begin to deteriorate as decay reaches an
intermediate stage. Wood substance is lost in such proportions in ad-
vanced decay that pulp yield and production suffer. In practice the use
of decayed wood must be so programmed that only small percentages are
encountered at any time and that the full effects as described are not felt.
(e) UNIFORMITY
The question of uniformity has been purposely left to the last. Pro-
duction of a uniform pulp requires a uniform raw material, uniform condi-
26 W. F. Holzer, Proc. Forest Products Research Soc., 4, 134 (1950).
524 CELLULOSE
tions within each chip, uniform conditions within a digester, and uniform
conditions from digester to digester. With the variables in the process,
some of which have been indicated, and the variable properties of wood, it
is easily imagined that the skill and ingenuity of the operator is constantly
being taxed. Although to a certain extent pulp manufacture is an art of
balancing these many variables, and will remain so, our more detailed
knowledge, better control through instruments, and better equipment are
more and more averaging or controlling these variables and reducing the
art to a scientific practice.
3. The Sulfite Process
(a) DESCRIPTION OF PROCESS
It is now desirable to supplement the earlier very general description of
pulping by giving certain specific details of the sulfite process. A more com-
plete description may be found in the literature.27'28
(1) The Cooking Acid
Sulfite cooking acid is a solution of sulfur dioxide and a bisulfite. The
latter may be an alkali or alkaline earth, but usually for economic reasons
it is calcium or dolomite. In recent years magnesium, sodium, or ammo-
nium bisulfite are being used in commercial processes. The industry ex-
presses the chemical concentration in terms of ' 'total/' "free," and "com-
bined" sulfur dioxide according to the following definitions:
Total: The total concentration of sulfur dioxide whether present as sulfurous acid,
bisulfite, or dissolved sulfur dioxide.
Free: The amount of sulfur dioxide in excess of the theoretical amount to form mono-
sulfite.
Combined: The amount of sulfur dioxide required theoretically to form monosulfite.
In each case the concentration is reported as "per cent" 862, but the figure
as determined is actually grams per 100 ml. The total and free are most
commonly determined by the method of Palmrose.29 Combined is deter-
mined by difference.
Acid is prepared by burning sulfur to form sulfur dioxide, cooling the gas
27 G. H. McGregor, in J. N. Stephenson, editor, Pulp and Paper Manufacture, Vol. 1,
McGraw-Hill, New York, 1950, Chapter 4; J. P. Casey, Pulp and Paper, Vol. I, Inter-
science, New York-London, 1952, pp. 74-132.
28 Chemipulp Sulfite Mill Operation, Chemipulp Process, Inc., Watertown, N. Y.,
1939, 191 pp.
29 G. V. Palmrose, Paper Trade J., 100, 38 (Jan. 17, 1935).
VI. PREPARATION FROM NATURAL SOURCES 525
rapidly, and absorbing the sulfur dioxide in water in the presence of the
base. Usually the base is limerock, and conditions are controlled to dis-
solve just the required amount. The sulfur dioxide concentration is raised
to the desired level by the absorption of return or ' 'relief " gas from the
digesters. About five times as much sulfur dioxide as is actually used is
put into the digester. The excess, being a gas, tends to escape, and is re-
turned to the acid through a relief system. A typical cooking acid con-
tains 6.5% total SO2, 5.3% free, and 1.2% combined.
(2} The Cooking Cycle
The sulfite cooking cycle is divided into three main parts: the penetra-
tion period, the cooking period, and the recovery period. Time must be
allowed for the chemical, both free and combined, to penetrate the chip
completely. The combined, being the slower, is the determining factor.
Temperature in the digester is raised slowly over a period of about 4 hrs. to
110°C., at which level the rate of cooking becomes significant. Penetra-
tion becomes rapid from about 80°C. on. Following the penetration period
the temperature is raised to a maximum, usually between 135° and 145°C.
The pressure is allowed to rise until it reaches about 80 lb., after which it is
maintained constant by relieving gas. When the cook reaches the proper
stage at about the end of the seventh hour from the start, pressure is re-
duced to 30 lb. during a period of about 1.5 hrs. in order to recover chemical.
The cook finishes during the recovery period, and at the end is blown into a
pit by the remaining digester pressure. The total cooking time is about
8.5 hrs.
(5) The Pulp
Sulfite pulp as discharged from the digester has a grayish-white color.
Yields of pulp on the basis of dry wood are 46-48%. The pulp is of medium
strength and very versatile in its papennaking qualities. It is mixed with
groundwood to make newsprint; it will make soft papers such as napkins
and the various sanitary tissues; and it will produce excellent bond papers.
Further, by adjustment of the cooking, and with further chemical purifica-
tion, it is the raw material for cellulose derivatives, as a replacement for or
supplement to cotton linters.
(b) PROGRESS OF THE COOKING PROCESS WITH TIME
The removal of lignin from the cell wall during the sulfite cook has been
followed photographically. Bixler30 pulped wood sections and studied the
80 A. L. M. Bixler, Paper Trade /., 107, 29 (Oct. 13, 1938).
526
CELLULOSE
removal of lignin with staining techniques. The intercellular material was
removed first, but attack was started on the secondary wall before the re-
moval of the intercellular material was complete. The primary or cam-
bial wall persisted through the entire cook. Lange9-31 made similar studies
160
4 6
Cooking time in hours
10
Fig. 4. The concentration of chemical components and coloring matter as
a function of reaction time in sulfite pulping fSimerl32).
using ultraviolet absorption methods. The lignin in the middle lamella
was attacked in the early stages of the cook. However, the ultraviolet
31 P. W. Lange, Svensk Papperstidn., 50, 130 (1947); through Chem. Abstracts, 41,
7112 (1947).
32 L. E. Siincrl, Tech. Assoc. Papers, 23, 114 (1940).
VI. PREPARATION FROM NATURAL SOURCES 527
absorption spectrum did not change, indicating that no deep-seated reac-
tion with the benzene nucleus of the lignin took place. Lange observed
dichroism in the intercellular material which indicated a degree of orienta-
tion in the lignin. The dichroism weakened during the cook and had disap-
peared by the time the fibers were separating. This was interpreted as
strongly suggesting that the lignin was not only the cementing material
between the fiber, but that there was some sort of bond between the lignin
and cellulose. It was also suggested that the path of penetration of the
cooking liquor was through the middle lamella, and was not by diffusion
through the cell wall from the lumen.
The appearance of dissolved materials in the digester liquor during cooks
of several species of wood was determined by Simerl.32 Figure 4 presents
typical curves for the total sulfur dioxide, total solids, methoxyl, sugars,
and light absorption coefficient as a function of reaction time of a cook on
black spruce (Picea mariana) made under the following conditions:
Sulfur dioxide concentration :
Total . . 6 2%
Free... . . 5.0%
Combined... 1 2%
Liquor: oven-dry wood ratio. ... 3.6 ml./g.
Time to 110°C. 3.0 hrs.
Time i 10-140 °C .4.0 hrs.
Time at 140 °C... . ... 3.0 hrs.
Maximum pressure . . 75 Ib./sq. in.
Linear relief to blowing pressure of 50 Ib./sq. in. during last hour.
The color development of the liquor during the cook could be divided into
three stages. During the first 3 hrs. the color was light yellow due to extrac-
tion of water solubles and to small amounts of lignosulfonic acid. About
the third hour it changed gradually to a reddish brown with appearance of
larger quantities of lignosulfonic acid. This color deepened and during the
last 2 hrs. turned to a coffee brown due to decomposition of carbohydrates.
These changes can be traced in the curves. Starting with the third hour
the methoxyl content, an indication of the lignin present in the liquor, and
the total solids increased in parallel fashion. Sugars started from zero at
the fourth hour and increased to 30 g. per liter at the eighth hour, after
which there was little further change. The absorption coefficient at 620
millimicrons, an indication of the depth of color of the liquor, did not
change until the sixth hour and then developed at a progressively acceler-
ated rate. Because the absorption coefficient curve was not parallel to any
of the curves of the chemical constituents, Simerl suggested that the rapid
528 CELLULOSE
color change during the last 2 hrs. was due to degradation of carbohydrates.
This is corroborated in commercial practice by the appearance of carbon
dioxide in relief gases toward the end of the cook.
(c) THEORY OF THE SULFITE PROCESS
Modern theories of the reaction in the sulfite cook had their inception
when Pedersen38 and Lindsey and Tollens34 first showed that pulping was
due to the reaction of lignin and sulfite to form a soluble lignosulfonic acid.
The development of the subject has been controversial and a detailed dis-
cussion would be only of historical value. The most widely accepted
theory today is the one evolved by Hagglund35 from work extending over
the past thirty years. The removal of lignin from wood takes place in two
steps, first the sulfonation of lignin in the solid state, and second the dissolu-
tion of this solid lignosulfonic acid.
The formation of the primary solid lignosulfonic acid can be demonstrated
by cooking wood with a sodium bisulfite solution at a pH of 6. Sulfona-
tion takes place rapidly until one sulfur atom has been added for every
three to four methoxyl groups. At this level sulfonation almost ceases,
and practically no lignin dissolves. If the pH is lowered to 4.5 the degree
of sulfonation reaches one sulfur atom per two methoxyl groups, but still
solution of lignin is very low. Lignosulfonic acid can be dissolved from
the sulfonated wood by heating in an acid buffer solution. The rate of
dissolution increases with lower pH. Then, too, if sulfonated wood is
treated with a strong mineral acid, the base associated with the lignosul-
fonic acid is removed leaving the solid free acid. This acid, known as Kull-
gren acid, can be dissolved merely by heating the wood in water. Hydro-
gen ions are furnished by the lignosulfonic acid itself.
The hydrogen-ion catalysis of the dissolution of solid lignosulfonic acid
appears to be a hydrolytic action. It can be assumed that the lignin is
joined to some carbohydrate material or polymerized into large insoluble
lignin molecules through acetal-like bonds. The hydrolytic splitting of
these bonds by hydrogen ions breaks the lignosulfonic acid into simpler
units which are soluble.
At the International Congress of Pure and Applied Chemistry held in New York in
September 1951, the consensus was that these acetal bonds were between low lignin
polymers and not between lignin and a carbohydrate.
88 N. Pedersen, Papier-Ztg., 15, 422, 787 (1890); E. Hagglund, Chemistry of Wood,
3d ed.f Academic Press, New York, 1951, p. 415.
84 J. B. Lindsey and B. Tollens, Ann., 267, 341 (1892); E. Hagglund, Chemistry of
Wood, 3d ed., p. 415.
* E. Hagglund, Chemistry of Wood, 3d ed., pp. 215, 415; Tappi, 33, 520 (1950).
VI. PREPARATION FROM NATURAL SOURCES 529
Erdtman36 and his coworkers37 describe the reaction with model sub-
stances which can be simplified as follows :
A — Lignin — B
"A" is a group which will sulfonate readily in neutral or mildly acid solu-
tion. It occurs in the lignin molecule, probably as a benzyl hydroxyl
group. ' 'B " will sulfonate only in strongly acid solution. This is probably
an acetal group linking lignin into larger molecules, either to carbohydrate
material or into a larger polymer of lignin.
Calhoun, Yorston, and Haass38 claim that the second step in the solution
of lignosulfonic acid as described by Hagglund is controlled not alone by
pH, but by a combination of pH and concentration of bisulfite ions. They
prepared three solid lignosulfonic acids in wood with sulfur contents, on
the basis of the lignin, of 4.2, 5.1, and 6.2%. The first sample in a buffer
at pH 2 delignified only slightly, the second to a level of 1.5% lignin, and
the third to 0.3% lignin. The first sample cooked more slowly than in
sulfite liquor, the second about the same as in sulfite liquor, and the last
much more rapidly. When a portion of the third sample was cooked in a
buffer at pH 3, it cooked about as it did in sulfite liquor. However, this
third sample in sulfite liquor at a constant pH cooked more rapidly as the
concentration of the base was increased. It was concluded that the dis-
solution of the solid lignosulfonic acid was controlled both by the hydrogen
and bisulfite ions present.
Brauns and Brown39 methylated sprucewood meal with diazomethane.
This wood, methylated under very mild conditions, could not be cooked
with sulfite acid. The lignin became partially sulfonated, but could not be
solubilized. Apparently the first stage of the sulfonation as postulated
by Hagglund took place at least to some extent, but the second stage of the
reaction was blocked by the methoxyl group.
(d) EFFECT OF PRETREATMENTS ON THE SULFONATION OF LIGNIN
Several investigators have studied the effects of pretreatments on the
sulfite reaction. Corey and Maass40 have heated wood in water and in
various buffers for varying lengths of time and at different temperatures.
Such treatments retard subsequent delignification. The effect is greater
36 H. Erdtman, Tappi, 32, 75 (1949).
87 B. O. Lindgren and U. Saeden, Svensk Papperstidn., 54, 795 (1951).
88 J. M. Calhoun, F. H. Yorston, and O. Maass, Can. J. Research, 15B, 457 (1937).
39 F. E. Brauns and D. S. Brown, Ind. Eng. Chem., 30, 779 (1938).
40 A. J. Corey and O. Maass, Can. J. Research, 13B, 149 (1935); 13B, 289 (1935).
530 CELLULOSE
at higher temperatures and longer times. With regard to pH the inhibit-
ing action is least at pH 4.6. Richter41 carried on a series of two- and three-
stage cooks. If the first stage subjected the wood to high temperatures, the
lignin became almost inert to later sulfonation. However, if the wood
were first sulfonated, treatments which, when carried on initially, rendered
lignin insoluble, now had no effect. It seems reasonable to suggest that
lignin under the influence of heat or acid in the presence of moisture poly-
merizes to an insoluble form. If the lignin is first sulfonated, the group
through which polymerization takes place is blocked, and the balance of
the sulfonation and dissolution can proceed without change.
A practical example of the above is that of * "red-centered" chips in cook-
ing. The sulfur dioxide of the cooking acid penetrates the chip faster
than the calcium base. If the temperature gets too high before the base
reaches the center of the chip, the sulfonation reaction starts, but the ligno-
sulfonic acid is not neutralized. The acidity and heat renders the unsul-
fonated lignin insoluble, and the chip remains uncooked. In a ' 'burnt"
cook the base is exhausted before the cook is completed, the liquor becomes
extremely acid, and the lignin is reprecipitatecl on the fiber. The dark
pulp from such a cook is almost impossible to bleach. It is interesting to
note that when ammonia is substituted for calcium as the base, it is almost
impossible to raise the temperature fast enough to obtain red-centered
chips. This indicates that the ammonium ion penetrates the chip with
approximately the same speed as the bisulfite, and much faster than cal-
cium.
(e) KINETICS OF THE SULFONATION OF LIGNIN
Several studies have been made on the kinetics of the sulfite reaction at
the Pulp and Paper Research Institute of Canada.42 This work demon-
strated that the sulfonation of lignin approximates a first-order reaction,
that is, the rate at which lignin was dissolved from wood was approximately
proportional to the lignin remaining. Deviation was most pronounced
at the beginning of the cook when the rate of solution was slower than would
be predicted. Goldfinger,43 using these data,42 made a theoretical study of
the reaction and found that the energy of activation of lignin solution
started at 16,000 calories per mole and increased to 22,500 calories per mole
41 G. A. Richter, Tappi, 32, 553 (1949)
42 F. H. Yorston, Proc Tech. Sect., Can Pulp and Paper Assoc., 1935, 91; A. J
Corey and O. Maass, Can J. Research, 14B, 330 (1930); J. M Calhoun, F. H. Yorston,
and O. Maass, Can. J. Research, 15B, 457 (1937).
43 G. Goldfmger, Paper Trade J., 112, 29 (June 12, 1941); 113, 27 (Oct. 9, 1941).
VI. PREPARATION FROM NATURAL SOURCES 531
at the end of the cook. Equations, describing the reaction as a surface
adsorption, followed by an activated adsorption, and finally the true reac-
tion, fitted the data12 reasonably well.
Bryde44 studied the kinetics of the sulfite reaction by making short cooks
in which ammonia was gassed into an evacuated bomb filled with chips,
sulfur dioxide solution was added, and the temperature was raised rapidly.
He found three characteristic subdivisions in the series, namely: the region
where lignin removal was in the range of 0-7%, where it was 7-25%, and
where it was 25-28%. The sulfur in the lignin increased rapidly until it
reached a ratio of 1 S:4 CgH^OCH^ (first stage); the sulfur content then
increased more slowly until it reached a ratio of 1 S:2 C9H9OCH3 (second
stage) ; a further increase to liquor blackening was followed by a decrease
(third stage). These stages were related to: first, the breaking-down of
the middle lamella and splitting of the presumed bond between lignin and
carbohydrates; second, the dissolution of lignin; and third, the attack on
the carbohydrates.
(f) LIMITATIONS OF THH VSULFITK PROCESS
Certain species of wood can be pulped by the sulfite process only with
difficulty or under special conditions. The most common of these are the
pines and Douglas fir. It was first believed that the resins in the wood
hindered penetration. Following preliminary work by Hagglund and
Schwalbe, Erdtman4f) isolated and identified a phenol from pine heartwood
which he showed to have the power of inhibiting the sulfite cook. This
compound, which he named pinosylvin, is a stilbene derivative:
The' monomethyl ether was also isolated. Further studies showed that
many phenols including pyrogallol, resorcinol, phloroglucinol, catechin,
naphthols, and the like will strongly inhibit the sulfite cook. Pinewood
was cooked successfully by sulfonating first in slightly alkaline sodium
sulfite, and then cooking normally.
" 0. Brydc, Finnish Paper Timber J., 29, 296 (1947); through Chem. Abstracts, 42,
5220 (1948).
45 H. Erdtman, Ann., 539, 116 (1939); through Chem. Abstracts, 33, 7098 (1939);
Tappi, 32, 303 (1949).
532 CELLULOSE
This behavior is explained on the basis of the presence of two kinds of
groups in lignin :
(1) Those which will react with sulfite easily, but with phenols only
under very acid conditions. These are the groups which sulfonate in the
first stage of the process.
(2) Those which will react with sulfite under acid conditions, but not
with phenols. These groups are the ones involved in the hydrolytic stage
of the reaction.
If the first groups are blocked by sulfite before the phenols have the
opportunity of reaction, the process proceeds normally. Douglas fir
(Pseudotsuga taxifolia) also has a compound in its heartwood which inhibits
the sulfite cook. Pew46 has isolated the compound and identified it:
OH O
Taxifolin
This flavanone has been given the name taxifolin. According to chemical
structure it is dihydroquercetin. Its behavior is apparently the same as
pinosylvin in pines.
(g) ACID-SUSCEPTIBLE WOOD
In a study of some unexpected variations of pulp strength found in
different wood samples, Green and Yorston47 found that certain areas in
wood became brittle on treatment with SC>2 or other dilute acids, while the
balance of the wood retained its soft fibrous character. The wood which
became brittle they termed "acid susceptible/' When isolated and cooked,
such wood produced a pulp with inferior strength and a shorter fiber length.
It is usually found around compression wood (wood subjected to compres-
sive strains in growth such as on the outside of the curve of a tree which
has righted itself in growth after being bent), or around compressive in-
juries. They extended their studies to chips and found that the second
cut, in which the bevel of the knife crushed the wood, caused acid suscepti-
bility. The strength loss from acid-susceptible wood was much lower from
an alkaline cook. Grondal48 also reported damage to cell walls and to
46 J. C. Pew, Tappi, 32, 39 (1949).
47 H. Green and F. H. Yorston, Pulp & Paper Mag. Can., 40, 244 (1939); 41, 123
(1940).
48 B. L. Grondal, Pacific Pulp Paper Ind., 13, 12 (July, 1939).
VI. PREPARATION FROM NATURAL SOURCES 533
pulp strength from compressive injuries. He observed microscopic longi-
tudinal cracks in the walls. An observation by Hagglund49 may be a
partial explanation. He found that the pH of a cooking liquor increased
from 2.3 to 3-4 as it diffused into the cell wall, and stated that this decrease
in acidity protected the cellulose from serious degradation. The cracks
observed by Grondal would allow penetration of the acid without decrease
of pH. As further evidence, chips made by sawing to avoid compressive
damage yielded a sulfite pulp with a burst equal to kraft.60 It seems prob-
able that the difference in burst between normal sulfite and kraft pulps is
caused primarily by this abnormal penetration and concomitant hydrolytic
degradation.
(h) VARIOUS BASES USED IN SULFITE COOKING
Several bases in the alkali and alkaline earth groups are being used com-
mercially. Calcium or the natural calcium-magnesium mixture, dolomite,
is the commonest because of low cost. Hatch51 has described the use of
magnesium, stating that there is very little difference between magnesium-
base and calcium-base pulps. Similar conclusions were reached by Schur
and Baker52 when comparing sodium base to calcium base. Chidester
and Billington63 took advantage of the solubility of sodium salts and studied
the effect of varying combined SO2 from 0.9 to 6.0%. They found that
yields and strength values went through a maximum between 2 and 3%
combined and dropped to a minimum at 5%. LaFond and Holzer54 re-
ported on a commercial trial with ammonia base in which it was found that
pulp strength and yield improved while cooking time decreased, as com-
pared to a dolomite base. A patent55 covering a controlled pH, sodium-
base pulping has been issued recently. According to this patent, pulps of
the sulfite type can be prepared with any desired alpha-cellulose content
up to 98% in one pressure cycle. A system for recovery of both sodium
ion and sulfur has been developed.
49 E. Hagglund, Svensk Papperstidn., 39, 95 (1936); through Ghent. Abstracts, 30, 6560
(1936).
w W. F. Holzer, unpublished data.
51 R. S. Hatch, Pulp & Paper Mag. Can., 47, 80 (Aug., 1946); Tech. Assoc. Papers,
29,485(1946).
52 M. O. Schur and R. E. Baker, Paper Trade J., 112, 38 (May 15, 1941); 115, 33
(Sept. 17, 1942); M. O. Schur and E. G. Ingalls, Paper Trade J., 117, 34 (Sept. 16,
1943).
58 G. H. Chidester and P. S. Billington, Paper Trade J., 104, 39 (Feb. 11, 1942).
64 L. A. LaFond and W. F. Holzer, Tappi, 34, 241 (1951).
66 G. Sivola, Canadian Patent 480,404 (Jan. 22, 1952).
534 CELLULOSE
The interest in bases other than calcium in the last few years, and, in
particular the soluble bases, has been caused by the need for stream im-
provement through disposal of the spent liquor in ways other than by
dumping. The most practical method for general application is to evap-
orate and burn the liquor. Any heat and chemical recovery will help
defray the cost of the operation. The insolubility of calcium monosulfite
makes scaling of evaporators a major problem in evaporating calcium-base
liquor. It has been done in vSweden in special evaporators in which the
flow of steam and liquor can be reversed periodically so that the acid con-
densate can remove the lime scale.56 There is no recovery of chemical pos-
sible, and the disposal of the calcium ash is a problem. One mill in Sweden
is cooking with sodium-base sulfite with recovery of heat and chemicals
in a complicated process, the details of which have not been published.
Magnesium lends itself to a unique cyclic process57 since the magnesium
salts break down to MgO and SO2 and both chemicals can be recovered
along with large quantities of heat. This process is now in iise at the Long-
view plant of the Weyerhaeuser Timber Company.58 Ammonia-base
sulfite is being used in several mills but no commercial recovery systems
have been installed. Pilot plant results59 have indicated good recovery of
heat and St>2, but the ammonia breaks down to nitrogen and water in the
combustion of the liquor.
The manufacture of by-products from spent sulfite liquor has been the
subject of extensive research for several decades. The suggested products
are numerous. A partial list of commercial products being made today
includes ethyl alcohol, vanillin, yeast, tannin extracts, dispersing agents,
dye levelling agents, glue extenders, and foundry core binders. In spite
of the many types of products possible, there has been no widespread chemi-
cal utilization of sulfite spent liquor, primarily because the production cost
is not competitive or the volume of product from even one mill would
exceed the demand. As yet .there are not enough products to utilize all
the mill wastes.
4. The Soda Process
Although the soda process was the first historically and although it is
66 F. W. Grewin and S. G. Lindberg (to Rosenblad Corp.), U. S. Patent 2,490,750
(Dec. 6, 1949); Chem, Abstracts, 44, 1706 (1950); A. E. Tyden (to Rosenblad Corp.),
U. S. Patent 2,490,759 (Dec. 6, 1949); Chem. Abstracts, 44, 2154 (1950).
57 G. H. Tomlinson and L. S. Wilcoxson, Paper Trade J., 110, 31 (Apr. 1 1, 1941).
68 R. E. Baker and L. S. Wilcoxson, Tappi, 33, 187 (1950).
69 J. H. Hull and G. V. Palmrose, Tappi, 35, 19,3 (1952).
VI. PREPARATION FROM NATURAL SOURCES 535
still in use, it has never reached the prominence of either sulfite or kraf t, and
its total production has been static since the beginning of the twentieth
century in spite of the enormous expansion of the industry in general.
The reason for this has been the superiority of both sulfite and kraft,
especially the latter, in yield and strength of pulp, and ease of cooking
Production of soda pulp is confined almost entirely to cooking hardwoods
In fact, soda pulp in the trade is understood to be made from hardwood
unless otherwise specified. The more drastic action of the sodium hydrox-
ide produces a soft, limp fiber which is peculiarly adapted to the furnish
of printing papers.
(a) INSCRIPTION OF THE PROCEvSS
(1) The Cooking Liquor
Soda liquor is made up in concentrations averaging 80 g. of NaOH per
liter. It contains about 10% soda ash from the uncausticized portion of
the chemical from the recovery process. The soda ash serves no useful
purpose in the cook, but, since its reduction would be expensive, it is car-
ried as an inert load.
(2} The Cooking Cycle
The cook is carried on in unlined steel digesters. The amount of liquor
added is kept at a minimum volume to facilitate liquor recovery, but suffi-
cient chemical must be added to complete the reaction. The charge of
chemical will vary with the species, but will amount to 25-28% NaOH on
the weight of the dry wood (20-22% Na2O). While bringing the cook to
temperature, air is relieved from the digester several times to prevent false
pressure, and to prevent degradation of the cellulose by oxygen in presence
of alkali. Cooks are usually controlled by pressure with the top ranging
from 80 to 110 Ib./sq. in. Top pressure will be reached in 1 to 2 hrs. and
maintained for 4 to 6 hrs. depending on the level maintained. Pulp is
blown into tanks and washed in trays or over cylinder washers to recover
the liquor at the highest possible concentration.
(5) Recovery of Chemicals
The liquor is evaporated to 50-60% solids, at which concentration it will
burn if properly atomized in a hot firebox. The organic fraction is con-
sumed in the combustion, steam is recovered from the furnace, and the in-
organic chemical as sodium carbonate is run from the furnace in a molten
536 CELLULOSE
state. The liquor is regenerated to sodium hydroxide with lime, and reused
in cooking. Chemical losses in the process, amounting to about 200 Ib. of
Na2COs per ton of pulp produced, are made up with commercial soda ash
before the liquor is recausticized.
(4) The Pulp
The pulp as blown from the digester is a light grayish-brown color.
It is almost always bleached before using, since the paper grades in which
it is customarily used are white.
(b) THEORY OF THE SODA PROCESS
The theories of the alkali delignification of wood, as proposed by Brauns
and Grimes60 and by Larocque and Maass,61 are essentially in agreement.
The steps in the process are : first, an absorption at the liquor-lignin inter-
face of sodium hydroxide by the acidic hydroxyl groups on the lignin;
second, a chemical combination as the temperature rises between the
lignin and absorbed alkali; and third, an alkaline hydrolysis of the assumed
lignin-carbohydrate bond at elevated temperature, and diffusion of the
sodium lignate from the wood.
Brauns and Grimes60 point out that "the carbohydrates dissolve very
rapidly at the beginning of an alkaline cook and only after about 20% of
the nonligneous part of the wood has been dissolved does the lignin start
to disperse at an effective rate." They also partition the alkali consumed
in the cook as follows: for a total alkali consumption of 16%, about 1.5%
is used to neutralize such acid groups as fonnyl and acetyl, about 4% is
used to dissolve the lignin, and the remaining 10.5% is used in dissolution
of the carbohydrates except for a small portion held in the pulp by adsorp-
tion.
Larocque and Maass61 found that the alkali dissolution of lignin follows
the course of a monomolecular reaction, except for the removal of the last
2% which is much slower. The last fraction is probably the lignin distrib-
uted through the cell wall and is less available to the liquor. Through
studies of the effect of liquor concentration the reaction rate is believed to
be controlled by the absorption of the alkali at the lignin-liquor interface.
There is no indication as to whether the combination of the alkali and lignin
and the dissolution of the lignin are simultaneous or whether one is slower
« F. E. Brauns and W. S. Grimes, Paper Trade /., 108, 40 (Mar. 16, 1939).
61 G. L. Larocque and O. Maass, Can. J. Research, 19B, 1 (1941).
VI. PREPARATION FROM NATURAL SOURCES 537
than the other. The energy of solution of the lignin is 32,000 calories per
mole.
5. The Kraft (Sulfate) Process
(a) DESCRIPTION OF THE PROCESS
Certain features of the kraft or sulfate process are discussed below. A
complete description of this process and also of the equipment may be
found in the literature.62'63
(jf) The Cooking Liquor
Kraft cooking liquor differs from soda liquor primarily in its sodium sul-
fide content. The amount of sulfide is expressed in terms of "sulfidity"64
which is defined as the Na2S content divided by the sum of the NajCOa,
NaOH, Na2S contents, all expressed as Na2O. (In this discussion all cook-
ing chemicals will be expressed in terms of Na2O according to the usage of
the kraft industry.) A typical kraft cooking liquor with all chemicals
expressed as grams of Na2O per liter would contain :
NaOH 70 g./liter
Na2S 30 g./liter
Na2CO3 20 g./liter
In the terms of the industry this would be :
Total 120 g./liter (NaOH + Na2S + Na2CO8)
Active 100 g./liter (NaOH + Na2S)
Sulfide 30 g./liter (Na2S)
The sodium carbonate fraction, as in the soda liquor, is present as an inert
load. It is kept to an economic minimum in the recausticization of liquor.
The sulfidity of the liquor is most generally maintained between 25 and
28%. This is a decrease from 33% formerly recommended.
(2} The Cooking Cycle
Since an alkaline liquor will penetrate into wood faster than an acid one,
and since higher temperatures can be used with alkali without damaging
cellulose, kraft cooks can be much shorter than sulfite. Pulps for paper-
82 G. H. Tpmlinson, II, and J. N. Swartz, in J. N. Stephenson, editor, Pulp and Paper
Manufacture, Vol. 1, McGraw-Hill, New York, 1950, Chapter 5; J. P. Casey, Pulp and
Paper, Vol. I, Interscience, New York-London, 1952, pp. 133-177.
M F. G. Sawyer, W. F. Holzer, and L. D. McGlothlin, Ind. Eng. Chem., 42, 756 (1950)
64 Tech. Assoc. Pulp Paper Ind., Standards, O 400 p-44 (Aug., 1944).
538 CELLULOSE
board can be cooked in as little as 90 inin. Pulps for paper are cooked in
2 to 6 hrs. In this country cooks are usually on the short side of the range;
those in Europe on the long side.
In the kraft cook only enough chemical is added to complete the reaction.
This requires about 15% active chemical (NaOH + Na2S) based on the
dry weight of the wood. Limited chemical addition acts as a cooking con-
trol and also minimizes the load in the chemical recovery. If more volume
of liquid is required to distribute the chemical and the heat, a sufficient
quantity of black or spent liquor is added.
Even with the more rapid penetration of alkaline liquors there should be
a penetration period of at least 1.5 hrs. If cooking rate becomes appre-
ciable before penetration is complete, the outside of the chips will consume
more chemical then necessary, and the centers will not have enough. The
resulting pulp would be a mixture of overcooked and undercooked wood.
Air is relieved from the digester during the penetration period as in the soda
cook.
Maximum cooking temperatures for kraft are most commonly in the range
of 170-1 75°C. (100-1 15 Ib./sq. in.). The cook is retained at this maximum
until the desired degree of pulping is attained, pressure is relieved quickly
to about 80 lb., and the cook is blown into a tank.
(3) The Recovery of Chemicals
The pulp is separated from the weak black liquor on cylinder washers and
is then screened before use, as in the case of sulfite pulp. The weak black
liquor containing about 10% solids (60% organic, 40%) inorganic) is con-
centrated in multiple-effect evaporators to about 58% solids. At this con-
centration it is steam-atomized into a hot furnace, where it flash dries and
then burns. Air is carefully controlled so that all sulfur compounds are
reduced to the sulfide. The molten chemicals run from the furnace and are
dissolved in water to make "green" liquor containing Na2COs and Na2S.
The green liquor is treated with lime to causticize the carbonate. The
lime mud is settled out leaving "white" liquor which is returned to the
digester. In modern kraft mills the lime mud is reburned to regenerate
the CaO. Chemical make-up is with salt cake (Na2SO4) added to the fur-
nace with the black liquor so that the sulfate is reduced to the sulfide. De-
pending on the equipment used and general efficiency, the make-up will
run from 100 to 250 lb. of salt cake per ton of pulp made. Efficiencies
expected in a well-run kraft mill are :
Chemical recovery . . . . 90%
Reduction of sulfur compounds 90—95%
Causticization 85-90%
VI. PREPARATION FROM NATURAL SOURCES 539
Steam recovered in modern equipment is sufficient to cook the pulp, evap-
orate the black liquor, supply the heat necessary in the liquor making, and
have a small excess.
(4) The Pulp
Kraft pulp is brown in color -the familiar brown of grocery bags. It
is the strongest pulp made from wood, and now that it can be bleached suc-
cessfully it is displacing sulfite pulp where more strength is desirable in
white papers. Probably the largest use for kraft in point of tonnage is in
the paperboard field for manufacture of corrugated or solid fiber cartons.
Here its lower cost and lighter weight have brought about the displace-
ment of practically all wooden boxes. Kraft pulp, being tough, utilizes
more power in the beating in preparation for papermaking, and the fibers
tend to be less broken down so that special care must be taken in better
paper grades to eliminate a coarse surface. Kraft pulp is not yet as satis-
factory for dissolving pulps as is sulfite, because removal of hemicelluloses
by purification is more difficult. An acid prehydrolysis prior to cooking
has produced pulps with much higher alpha-cellulose and in the next few
years kraft pulp may invade even the dissolving pulp field.
(b) PROGRESS OF THE COOKING PROCESS WITH TIME
The work of Bixler30 on pulping thin wood sections and following the
lignin removal with staining techniques included kraft. It was noted
that the intercellular material was removed completely before the lignin
in the secondary walls was attacked. This seems confirmed by the fact
that although sulfite will easily produce a pulp with only 1% lignin, reduc-
tion of lignin to as far as 2% with kraft is done only at the expense of ruinous
loss in pulp yield. {Since the sulfite process attacks the secondary cell wall
before all lignin in the middle lamella is removed, this process is sometimes
considered less specific in lignin removal than kraft. However, high-yield
cooks of kraft show that up to 20% of the carbohydrate material will be
removed before appreciable quantities of lignin dissolve. As an explana-
tion of the behavior of kraft liquor, the theory is proposed that the well-
known swelling effect of alkali on cellulose blocks the penetration of the
liquor into the cell wall and retards the removal of the last traces of lignin.
Kimble65 has followed the development of color and changes in chemical
composition of kraft liquor during the cook. Figure 5 taken from Kimble's
work illustrates these changes during a cook with the specifications:
w G. C. Kimble, Paper Trade /., 115, 37 (July 16, 1942).
540 CELLULOSE
Wood Spruce (Picea mariana)
Active alkali 20%
Sulfidity 33%
Water : oven-dry wood ratio 5.5:1
Maximum temperature 170 °C.
Time to maximum temperature 1.5 hrs.
Time at maximum temperature 4 hrs.
The following yield data were obtained :
Total yield 46.0%
Screened yield 45. 7%
Lignin in screened pulp 3 . 5%
123456
Cooking time m hours
Fig. 5. The concentration of chemical components and coloring matter in sulfate
black liquor as a function of cooking time (Kimble65).
Kimble found that by far the greatest proportion of the color in kraft
black liquor was developed from the lignin. The carbohydrates contrib-
uted nothing to the color. The rate of extraction of carbohydrates during
'VI. PREPARATION FROM NATURAL SOURCES
541
the first 2 hrs. of the cook, however, is much faster than the extraction of the
lignin.
(c) THEORY OF THE KRAFT PROCESS
When sodium sulfide is added to sodium hydroxide the rate of pulping
with the resulting liquor is greatly increased. Figure 6 from the work of
o
o
o
l/>
<20
o
O
o
o
z
o
4 6
COOKING TIME, HOURS
10
Fig. 6. Increase in rate of alkaline digestion of wood with the sulfidity
of the liquor (Hagglund66). (a) 0% Na2S, (b) 5.25% Na2S, (c) 15.6%
Na2S, (d ) 31 % Na2S. Maximum temperature, 160°C.
Hagglund66 shows quantitatively the increase in rate of lignin removal.
The greatest change in rate is found in the range of low sulfidities. Reduc-
tion of time required to reduce lignin in wood to 10% is very close to 50%
in going from a soda liquor to a kraft liquor with 31% Na2S.
Further work by the same investigator compares rate of lignin and carbo-
hydrate removal from wood in comparable cooks (Fig. 7). The rate of
carbohydrate removal by the two liquors may be considered identical. In
the case of lignin, however, the action of the kraft liquor is more rapid from
the start, and even after three-quarters of the lignin has been removed, the
reaction is proceeding at an only slightly diminished rate. The dissolution
of lignin by soda liquor, on the other hand, has almost stopped after less
than half has been removed. These two graphs are very significant in the
discussion of the theory of the reaction.
86 E. Hagglund, Tappi, 32, 241 (1949).
542
CELLULOSE
The earliest theory for the behavior of the sulfide in kraft liquor was that
it acted as a buffer, hydrolyzing to form NaOH and NaHvS as the NaOH
was consumed. In this way the wood was never subjected to the full
concentration of NaOH. Although it has been demonstrated that this
hydrolysis takes place,67 it has long since been recognized that this buffer
2468
COOKING TIME, HOURS
Fig. 7. Comparative rate of dissolution of carbohydrates and lignin during
a soda and a kraft cook (Hagglund66). (a) Carbohydrate, kraft cook; (b)
carbohydrate, soda cook; (c) lignin, kraft cook; (d) lignin, soda cook Max-
imum temperature, 140°C., attained in 1.75 hrs.
action has slight effect on the kraft process. The comparative rate of solu-
tion of carbohydrates in a kraft and a soda cook is supporting evidence.
The presence of sulfur in kraft lignin was recognized early, but it was not
until later that its significance was recognized. Hanson68 found sulfur
fi7 G. E. Martin, Tappi, 33, 84 (1950).
68 F. S. Hanson, Paper Trade J.f 112, 32 (Jan. 9, 1941).
VI. PREPARATION FROM NATURAL SOURCES 543
contents of thiolignin (kraft lignin) up to 10%, and proposed that sulfide
speeds up the kraft cook by rendering the lignin molecule more acid and
thus more soluble in alkali. Ahlm69 postulated that sulfur in thiolignin
was present in the mercaptan form, and that it had added at a carbonyl
group on the lignin. Hagglund66 has objected to both of these theories on
the basis that he cannot find any mercaptan sulfur in thiolignin, and that
there are not enough carbonyl groups in lignin to account for the sulfur con-
tent found in thiolignin. As a result of a long series of researches he pro-
poses the hypothesis that lignin first takes up sulfur in the solid phase, pre-
sumably by replacing a hydroxyl group with a mercaptan group. This is
not stable in alkali, and is converted to a sulfide by reacting with another
hydroxyl either in the same or another lignin molecule. The second
step consists of a hydrolytic splitting under the effect of alkali to form free
phenolic hydroxyl groups, thus rendering the lignin soluble in the alkali.
The effect of the sulfur is apparently twofold. It is probable that the
hydrolytic splitting of the phenolic groups takes place much more easily
if sulfur has first reacted with lignin; and second, it is also probable that
the sulfidization blocks a condensation -sensitive group, thus preventing
the formation of larger, less soluble lignin complexes. Both of these ac-
tions of sulfur probably assist in the more rapid dissolution of lignin as
shown from the start of the cook (see Fig. 7), and the blocking of lignin
condensation accounts for the sustained rate of reaction of the kraft cook as
compared to the decrease in ligniri removal in the soda cook after condensa-
tion has had an opportunity to take place.
(d) COLOR OF KRAFT PULP -
Since the first sulfur dyes were produced by heating organic material
with sodium sulfide, it was originally assumed that the color of kraft pulp
was caused by sulfur dye formation in the cook. vSchwartz, McCarthy,
and Hibbert70 found that the color was clue to lignin and its degradation
products. Bard71 produced colors similar to kraft pulp by cooking alpha
pulp in kraft liquor with tannins and short-chain carbohydrates. Since
black liquors from digestions with and without sodium sulfide gave similar
spectral absorbencies, sulfur dye formation, if any, cannot be the cause of
kraft pulp color. Kirnble65 found in studies of black liquor that 80-90%
69 C. K. Ahlm, Paper Trade /., 113, 115 (Sept. 25, 1941).
70 H. vSchwartz, J. L. McCarthy, and H. Hibbert, Paper Trade /., Ill, 30 (Oct. 31
1940).
71 J. W. Bard, Paper Trade J., 113, 29 (Sept 18, 1941)
544 CELLULOSE
of the color was due to lignin. Pigman and Csellak72 pointed to lignin as
the primary source of kraft pulp color. The brightness of the pulp was a
direct function of lignin content.
(e) THE EFFECT OF COOKING VARIABLES ON KRAFT PULPS
There have been many publications describing the effect of one or more
variables of the kraft cook on the resulting pulp. Most of these results
were obtained by keeping all conditions constant except the single variable
studied and, as a result, producing series of pulps with varying degree of
delignification. Most comparisons of pulps must be made at the same
degree of delignification to be valid, and therefore most published data on
this subject must be studied with extreme care. Hart and Strapp73 made
a comprehensive survey of the variables in the kraft process, and from their
data pulps may be compared at constant degrees of delignification.
The Effect of Sulfidity. At constant effective alkali (NaOH -f Y2 Na2S)
there were marked changes from 0 to 20% sulfidity, slight changes from
20 to 40% sulfidity, and little change thereafter. The changes noted were :
cooking time and yield were decreased, physical strength was increased,
and chemical constants showed little if any change.
Variations with Maximum Temperature. . Screened pulp yield decreased
from 48 to 42% on increasing maximum temperature from 160° to 180°C.
Burst and tensile were slightly higher and tear lower for the lower tempera-
ture. Alpha-cellulose and pentosan showed no change with lower cooking
temperature, but viscosity was higher.
Chemical Charge. Screened yield was higher at higher chemical charge,
but the reverse was true of burst. Tear showed no consistent trend. Alpha-
cellulose content was slightly higher at higher charges, but pentosans and
viscosity were lower.
Chemical Concentration. Bray, Martin, and Schwartz74 have made a
study of the effect of chemical concentration over a wide range, and found
best results at 50 g./liter. Most commercial liquors are made at double
this concentration, but in use are diluted nearly to the preferred range by
black liquor or condensate from direct steam, or both.
(f) BY-PRODUCTS FROM THE KRAFT PROCESS
Probably a major share of the kraft pulp produced is made from one or
72 W. W. Pigman and W. R. Csellak, Tech. Assoc. Papers, 31, 393 (1948).
78 J. S. Hart and R. K. Strapp, Pulp & Paper Mag. Can., 49, No. 3, 151 (1948).
74 M. W. Bray, J. S. Martin, and S. L. Schwartz, Paper Trade J., 109, 29 (Nov. 2,
1939).
VI. PREPARATION FROM NATURAL SOURCES 545
another of the various species of pines. The large amounts of extractives
in these woods yield recoverable amounts of by-products.
Turpentine in amounts of 2 gal. per ton of pulp can be recovered by pass-
ing the relief gases through a condenser, running the condensate through a
decanter, and collecting the turpentine from the overflow.
The extractives in pine will dissolve in the alkaline liquor, and will cream
from black liquor after it has been concentrated to half its volume. The
skimmings can be collected by decantation in amounts of 150-200 Ib. per
ton of pulp. The crude soaps are treated with sulfuric acid to form a dark
oil known in the trade as "tall oil." Tall oil consists of a mixture of resin
acid, fatty acids, and unsaponifiables.75 The first two comprise over 90%
of the total, and are present in fairly equal amounts. The resin acids are
substantially the same as those in American wood rosin76; the fatty acids
are linoleic, linolenic, and oleic; and the unsaponifiables are lignoceric
alcohol and sterols, principally phytosterols.
During World War II the shortage of fats led to extensive studies on
tall oil and to its introduction into many industries some of which are:
soap, emulsion, detergent, flotation, adhesives, paint and varnish, and
printing ink.77
6. Other Pulping Processes
(a) NEUTRAL SULFITE
The neutral sulfite process has come into some prominence in the last ten
years as a method of producing high-yield pulps from hardwoods. This
process in its present form was first proposed by Rue and coworkers78 in
1927. It consisted of cooking wood with about 15% of its weight of Na2SO3
with sufficient NaHCO3 present to neutralize the acids formed during cook-
ing. The action was so mild that delignification was incomplete, and best
results were obtained on hardwoods, since their lignin content was lower
than that of softwoods.
The process was very little used for several years, but the growing scar-
city of softwoods in the Great Lakes and New England sections, together
with increasing stands of hardwoods, forced the use of the latter in pulping,
and brought the neutral sulfite method to the fore.79 At first its use was
76 T. Hasselstrom, Paper Trade /., 118, 30 (Apr. 20, 1944).
7C G. C. Harris, Tappi Monograph Series, No. 6, 167 (1948).
77 National Southern Products Co., Tappi, 33, 76A (Jan., 1950); 58A (Feb., 1950).
78 J. D. Rue, S. D. Wells, F. G. Rawlings, and J. A. Staidl, Chem. & Met. Eng., 34, 611
(1927); Tech. Assoc. Papers, 10, 90 (1927).
w M. W. Phelps, Northeastern Wood Utilization Council Bull, 14, 59 (Jan., 1947).
546 CELLULOSE
to make a high-yield pulp for board, but now several mills arc bleaching the
high-yield hardwood pulp, and are obtaining a papermaking fiber. Bleached
neutral sulfite hardwood pulp has strength nearly equal to softwood sulfite
pulp, and replaces the latter in many papers. Yields of pulp run from 55
to 60% for a bleached papermaking fiber to 80% for board.
(h) PRKHYDROLYSIS OF WOOD
During World War II the need for dissolving pulps in Central Europe
became acute, but no cotton linters and very little softwood pulps were
available. Beechwood, which like all hardwoods is high in pentosans,
became the principal raw material. A process of prehydrolysis80 was de-
vised to render these pentosans soluble. The wood chips were subjected
to an acid treatment at high temperatures (160-1(SO°C.) for a short time
(15-60 min.), and this pretreatment was followed by an alkaline cook.
The pretreatment could be merely a cook in water, the acids in the wood
doing the hydrolyzing; or a small amount of mineral acid, hydrochloric
or sulfuric, might be added. After bleaching, pulps of 95% alpha-cellulose
or higher were obtained. Pulps made by the prehydrolysis process are
already a considerable factor in the American dissolving pulp market.81
It is anticipated that this process will be used to an even greater extent in
the near future.82
(c) OTHER PROCESvSEvS
Suggestions for methods of pulping of wood have been legion, but few
have survived outside of the laboratory. Many modifications of the es-
tablished processes, principally multistage cooks, have been tried, but the
time consumed, heat lost, chemicals used, or equipment required in the
complications have made them impractical. Many chemicals have been
tried ranging from strong acids, such as nitric, to bases both organic and
inorganic, to neutral organic solvents, such as alcohols and glycols, to
hydrotropic solvents, such as xylene sulfonates, to name a few. None of
these has as yet shown sufficient merit in quality of pulp or in cost to com-
pete with the established processes. Some have been of theoretical in-
terest, but aside from the two mentioned in preceding paragraphs, none
has achieved any commercial production.
80 G. Sirakoff, Holz Roh- u. Werkstojf, 4, 205 (1941); through Chem. Abstracts, 38,
2201 (1944).
«l Anon., Pulp & Paper, 24, 66, 92 (Nov., 1950); Paper Trade J., 132, 11 (Apr. 6,
1951).
82 Anon., Paper Trade /., 134, 41 (Feb. 1, 1952); 11 (Mar. 14, 1952).
VI. PREPARATION FROM NATURAL SOURCES 547
7. Properties of Pulps and Their Constituents Affecting End Use
The many variables in the pulping processes plus the inherent differences
in various wood species can be combined in an almost infinite number of
combinations to produce a varied assortment of pulps. The properties of
these pulps depend on their content of lignin, cellulose, and hemicelluloses,
on the condition of these constituents, and on the dimensions of the fibers
themse ves.
The principal requirement for a pulp as a chemical raw material is a high
content of alpha-cellulose with almost complete removal of lignin and ex-
tractives, and substantial reduction of hemicelluloses. This will be dis-
cussed fully elsewhere in this volume (see Chapter VII).
The properties of papers will depend on the degree to which the fibers
will adhere to each other, the ultimate strength of the fibers, the dimensions
of the fibers, and on the relative stiffness or softness of the fibers.
Lignin. The stated purpose of pulping was to remove lignin, and generally
a low lignin content is necessary. In a very raw pulp the lignin remains as
a sheath enveloping the fiber. Since it does not swell in water itself, hinders
the swelling and hydration of the cellulose in the fiber, and makes the fiber
very stiff, lignin is usually detrimental to paper quality. vSheets made from
such fibers are harsh and low in strength. On the other hand, it is desirable
to leave as much lignin in the fiber as can be tolerated since its complete
removal entails considerable loss and degradation of cellulose. In certain
instances the presence of lignin is desirable. In paperboard it contributes
stiffness to the product, and in hardboard (a wallboard finished at high
temperatures under extremely high pressures) it acts as a resin, making the
product extremely hard and dense.
Hemicelluloses. This fraction is of considerable importance in papermaking.
In the process of beating, where the fibers arc bruised and rubbed in a
water suspension, the hemicelluloses absorb water, swell, and become gelat-
inous. During the beating this gel coats the fibers; and when the fibers
are formed into a web, and the web is pressed and dried, the hemicellulose
gel acts as an adhesive, and cements the fibers into a strong sheet. As a
general rule, the greater the hemicellulose content is, the faster the pulp
will respond to beating, the harder and denser will be the resultant paper,
the higher will be its bursting strength and tensile strength, and the greater
will be its transparency. The extremes of paper grades illustrating these
differences are facial tissue having fibers low in hemicellulose, with soft feel
and little bonding; and glassine and greaseproof papers needing the highest
possible amount of hemicellulose to produce a very dense, hard sheet with
548
CELLULOSE
good transparency. In between are such grades as paper towels, blotting,
and soft-type printing papers on the soft side, wrapping papers in the
medium range, and strong bond papers on the hard side.
Cellulose. Some of the discussion of cellulose has already been anticipated.
The presence of cellulose itself in a fiber in undegraded form is no assurance
that a paper will be strong. Pure cellulose is quite soft and has poor
B
2400
02000
£1600
o 800
UJ
u
cr
400
X
0 10 20 30 40 50 60 70
% OF ORIGINAL WOOD
2HUU
2
opooo
/
/
SE
fsi
CE i Ron
L
' /
X
UJ II:>UU
2E
i
O 1 POO
/
/
Ou
U.
o on
//
u
UJ
a:
0 400
J
/
m HUU
0
0
/
0 10 20 30 40 50 60 70
% OF ORIGINAL WOOD
LYMERIZATION
— ro o
0 <fi O 4
3 0 0 C
3 0 0 C
/
/
/
/
o
a.
u.
O BOO
y
/
X
/
x '
f*\
-
UJ
UJ
cr
o 400
/
/
/
/
UJ U
0
o
/
/
D
DEGREE OF POLYMERIZATION
— — ro i>
4* a> ro o> O 4
0 0 0 0 0 C
o o o o o o c
X
/
/
1
/
X
^**
//
X
) 10 20 30 40 50 60 7C
% OF ORIGINAL WOOD
Fig. 8. Degree of polymerization of the carbohydrate fraction of (A) sprucewood
holocellulose, (B) sulfite pulp, (C) kraft pulp, and (D) soda pulp (Atchison83). Curves
are shown for both unbleached (solid line) pulps and bleached (dotted line) pulps. Holo-
cellulose and all pulps were prepared from the same wood sample.
VI. PREPARATION FROM NATURAL SOURCES
fiber-to-fiber bonding qualities. This will produce paper of the soft type
such as facial tissue. On the other hand good bonds from hemicellulose
cannot make a strong sheet unless they are complimented in the fiber with
strong cellulose. In general, cellulose contributes tearing resistance to
paper, and when present in nearly pure form will make a permanent sheet.
The condition of the carbohydrate fraction after various cooking and
bleaching treatments makes an interesting comparison. Atchison,88 by
fractional solution methods, developed D.P. curves of the holocellulose of
sprucewood, and of the sulfite, kraft, and soda pulps, both bleached and un-
bleached. His data are summarized in Figure X. The curve for wood holo-
cellulose shows an appreciable fraction of very short-chain hemicelluloses,
relatively little of the medium-length chains, and the bulk of the material
with chain lengths of D.P. 2000 and over. The curves for sulfite and kraft
show the great decrease of the short-chain material, the presence of con-
siderable amount of medium-length chains resulting from degradation
of the longer chains, and the very appreciable loss of long-chain material.
Greatest loss in short-length chains is found in the soda pulp, and the ex-
treme degree of degradation is evident. The relative yield of each carbo-
hydrate product is shown by basing the abscissas of the graphs on the ex-
tractive free wood. By comparing the yields along with the shape of the
distribution curve, the magnitude of the over-all degradation becomes
apparent. The further degradation on bleaching is shown in the respective
dotted curves.
Fiber Dimensions. The approximate fiber sizes and variations in cell walls
have already been described. The length of softwood fibers, about 3-4
mm., is generally too long for good forming properties, and reduction of this
length is one of the objects of the physical treatment a pulp receives before
the papermaking process. The small fiber fragments mat together into a
smoother sheet, and being smaller, present more surface for contact with
each other and therefore more bonded strength. If the reduction process
is carried too far the ultimate strength of the fragments will be less than that
of the bonds, and the paper strength will drop correspondingly.
Pulping of wood and making paper from these pulps has long been looked
upon as an art. The discussion of the field has indicated the degree to
which the technical men have reduced it to a science. However, with the
variations of a natural raw material, plus the infinite combinations among
the variables of the processes, and the gaps of our knowledge, it seems that
the complete success of science is still some distance in the future.
83 J. E. Atchison, Paper Trade /., 116, 23 (June 3, 1943).
B. COTTON LINT AND LINTERS
A. F. MARTIN
Cotton is the only plant seed fiber or seed hair to attain major commercial
importance. It is one of the most important raw materials for textiles and
is an important source of chemical cellulose for conversion into derivatives.
Raw cotton consists almost entirely of cellulose, with minor amounts of
waxes and pectins being the chief contaminants.1 Lignin is not present in
appreciable amounts. Other seed hairs such as kapok and milkweed floss
are discussed briefly in Chapter II-D-2.
1. Types of Cottonseed Fibers
Two types of hair, lint fibers (or staple cotton) and linters, are found on
the usual varieties of cotton seed (Fig. 9). These fibers are removed from
Fig, 9. Cotton seeds with fibers: (A) seed with all fibers (lint and linters);
(B) seed with linters after removal of lint; (C) seed with hull fiber after removal
of linters. Courtesy of Hercules Powder Company.
1 J. H. Kettering and C. M. Conrad, hid. Ent>. Chem., Anal. Ed., 14, 432 (1942).
VI. PREPARATION FROM NATURAL SOURCES 551
the seed in two separate operations. The staple cotton is the major prod-
uct and comprises about half the weight of the seeds which are taken from
the boll of the matured cotton plant. This staple cotton is removed from
the seeds in mills or gins of which there are about 8000 in the United
States. The machines used in this operation are also called gins and consist
primarily of a bank of rotating circular saws. The seeds after the ginning
operation are sent to cottonseed oil mills which delint (that is, remove
linters from) the seeds and crush them to recover cottonseed oil and meal.
There are about 400 oil mills in the United States.
The most important difference between lint fibers and linters is in fiber
length. Nevertheless, the two types of hair differ also in diameter, degree
of collapsing of the central canal, and other features (see Chapter IV-C).
Lint fibers have a length of over 2 cm. and are used primarily in textiles
because they can be spun into thread possessing high tensile strength. The
shorter linter fiber are used form attresses, upholstery, and similar prod-
ucts and especially for the preparation of pure cellulose ("chemical cot-
ton") which is used in the manufacture of derivatives.2"5
Several types of linters can be produced by modification of the operation
of the cottonseed oil mill. Mill runs are made by removing the majority
of the linters from the seed in one operation. They have a number-average
fiber length of about 3 mm. and are quite nonuniform. More frequently
the linters are removed in two or more passes through the delinting equip-
ment6 (Fig. 10A). The first-cut linters average 6-12 mm. in length and
the second-cuts average about 2 mm. A small amount of fiber remains on
the ends of the cotton seed after the usual delinting operation and is some-
times recovered from the hulls after the seeds are crushed. This product,
known as hull fiber, differs from the other cottonseed fibers in that it is us-
ually produced by abrading the fibers from the pieces of hull in machines
such as the Reynell-Ware defibrator7 (Fig. 10B). Hull fiber may have a
length equal to second-cut linters but is usually somewhat shorter. An
additional small amount of fiber known as ' 'delint" is obtained from some of
the seeds which are to be planted to furnish the next season's cotton crop.
2 W. D. Munson, 2nd. Eng. Ckern., 22, 467 (1930).
3 G. D. Bieber, Chem. & Met. Eng.t 48, 92 (Jan., 1941).
4 E. F. Hinner, Chemurgic Digest, 4, 179 (1945).
6 J. Barsha and P. VanWyck, in R. E. Kirk and D. F. Othmer, editors, Encyclo-
pedia of Chemical Technology, Vol. 3, Interscience, New York-London, 1949, pp. 352-357.
6 W. R. Woolrich and E. L. Carpenter, Mechanical Processing of Cottonseed, Eng.
Expt. Sta., Univ. of Tennessee, Knoxville, 1935, p. 51.
7 C. H. Reynell and A. J. V. Ware (to Reynell-Ware Inc.), U. S. Patent 2,004,731
(June 11, 1935).
552
CELLULOSE
In addition to fiber length, color and degree of contamination are two im-
portant factors in determining the utility of cottonseed fibers. The color of
lint fibers will vary from white to yellow to gray, depending on the species
of cotton and the degree of exposure of the fibers after the boll has opened.
Raw linters are usually olive or buff, although a large proportion of the ap-
parent color may come from the contamination which is present. The for-
eign matter in both lint and linters will include seedcoat fragments; pieces
Fig. 10. Linters production equipment. (A) Delinter6: Linters stripped from the
cotton seeds by the saws are doffed by the more rapidly moving brush roll. (B) De-
fibratdr7 : Linters are abraded from the hulls against the vertical screen by centrifugal
force set up by the main rotor. The heavier hulls fall through the bottom screen, while
the lighter linters are carried inward and upward by the air stream.
of stalk, leaf, and boll from the cotton plant; other plant materials; and
sometimes sand and dust. Both the cotton gins and the cottonseed oil
mills commonly use seed -cleaning equipment as well as fiber-cleaning
equipment in order to upgrade the fibers. The seeds are usually cleaned
with shaker screens; the fibers, with machines called cards or beaters.
Standards have been set up for grading and classifying cottonseed fibers
on the basis of fiber length, color, and extent of contamination. The
official U. S. Department of Agriculture methods involve subjective com-
parison of the fiber samples with standards.8 For grading American up-
8 U. S. Dept. Agr., Misc. Pub. 310 (1938).
VI. PREPARATION FROM NATURAL SOURCES 553
land cotton, there are nine classifications which relate to degree of contam-
ination and amount of waste, and six classifications which relate to color.
The fiber length is described by one of twenty classifications between 0.75
and 1.5 inches. The linters standards are similarly divided into seven
grades, with chief emphasis on fiber length and color.9'10 Grades 1 to 4 rep-
resent the longer first-cut and mill-run linters; Grades 5 to 7 represent
second-cuts and hull fiber.
For linters fibers which are to be used to make chemical cotton, the sub-
jective grading methods of the Department of Agriculture are not sufficient.
The development of additional objective tests is a goal of the American Oil
Chemists' Society. Of those developed to date, the most widely used is
the "pot yield" method11 which determines the amount of cellulose which
can be obtained from a given weight of raw cotton linters. The results of
this test are used in determining the price of second-cuts and hull fiber,
with premiums being paid for yields above 73% and deductions made for
yields below 73%. A test for cotton linters contaminants is also under con-
sideration.12 Fiber length is relatively unimportant in chemical cotton
except as it affects the design of equipment to process the cellulose into
derivatives.
2. Economics of Utilization of Cottonseed Fibers in the
United States
In the last ten years, the cotton production in the United States has
varied from about 9 to about 16 million bales per year13 (Table 2). Sales
prices in this ten-year period have ranged from 11 to 40^/lb., so that U. S.
cotton is frequently a billion-dollar annual crop.14 Cotton is grown in all
of the southern states from North Carolina to California. The center of
production is shifting from the Southeast, where small farms are the rule, to
Texas and California, 15 where the economies of large-scale production mean
9 U. S. Dept. Agr., Bur. Agr. Econ., Service and Regulations Announcement, 94
(1925).
10 G. S. Meloy, U. S. Dept. Agr., Misc. Pub. 242 (1936).
11 L. N. Rogers, Oil & Soap, 14, 199 (1937); 22, 24 (1945).
12 T. L. Rettger, Oil & Soap, 22, 7 (1945).
18 U. S. Dept. Agr., Bur. Agr. Econ., Statistical Bull. 99 (1951), p. 5; Supplement
(1952), p. 11.
18a U. S. Dept. Commerce, Bur. Census, Cotton Production in the United States — Crop
of 1952, Washington (1953), p. 2.
14 U. S. Dept. Agr., Bur. Agr. Econ., United States Cotton Statistics, Washington
(1951), p. 2.
16 Ibid., p. 1.
554 CELLULOSE
greater profits. Nevertheless, acreage controls, support prices, and govern-
ment loans are all used to protect the marginal high-cost cotton producer.
These controls have modified the operation of the economic laws of supply
and demand.13
TABLE 2
Production of Cotton in the United States18' 13a
Year
beginning Aug. 1
Acreage harvested
(in thousands of acres)
Production
(as thousands of
500-lb. bales)
1940
23,861
12,566
1941
22,236
10,742
1942
22,602
12,820
1943
21,610
11,429
1944
19,617
12,230
1945
17,029
9,016
1946
17,584
8,640
1947
21,330
11,857
1948
22,911
14,868
1949
27,439
16,128
1950
17,843
10,012
1951
26,687
15,144
1952
—
15,137
The high price of cotton has been a large factor in the growth of the
synthetic fiber industry, which in turn has furnished intense competition
for cotton. In 1949 the textile industry consumed 28% as much synthetic
fiber as it did cotton.16 Both acetate and viscose rayon as yarn can now be
sold at prices below that for cotton yarn.17 This situation can exist only
when the price of chemical cellulose is far below the price of cotton textile
fibers. Second-cuts and other cottonseed fibers are used in rayon to the
extent that availability and price will allow.
Cotton linters are recovered in amounts of about 180 pounds per ton18 of
the seeds processed in cottonseed oil mills. They are a by-product in that
their sales value is less than that of cottonseed oil and meal19 (Table 3).
16 Textile Organon, 23, 40 (1952).
17 Textile Organon, 23;, 37 (1952).
18 U. S. Dept. Agr., Production and Marketing Administration, Weekly Cotton Linters
Review, 22, No. 5 (Aug. 31, 1951).
19 Compiled from figures given in: U. S. Dept. Agr., Bur. Agr. Econ., Statistical Bull.
99 (1951), p. 368; U. S. Dept. Agr., Agricultural Statistics, Washington (1951), p. 125;
various issues of U. S. Dept. Agr., Production and Marketing Administration, Weekly
Cotton Linters Review.
VI. PREPARATION FROM NATURAL SOURCES 555
TABLE 3
Gross Sales Value of Cottonseed Products in the United States19
Year
beginning Aug. 1
Sales value in millions of dollars
Cottonseed
oil
Cottonseed
meal
Linters
1940
93.2
52.1
27.4
1941
153.4
64.2
33.5
1942
178.6
75.4
37.6
1943
157.6
89.0
31.7
1944
168.8
94.8
34.2
1945
129.8
79.0
—
1946
241.1
161.2
57.8
1947
334.9
164.7
52.8
1948
262.8
151.4
39.9
1949
229.9
160.7
58.6
1950
242.2
128.7
115.8
Nevertheless, since the profit margin in oil and meal may be comparatively
small, linters sales are of great importance to the oil mills. The amount of
linters, being a function of the size of the cotton crop, will vary from year to
year. In addition, the demand for chemical cellulose has shown an unus-
ually large growth rate. The result is that in several periods the total
amount of chemical cellulose including wood pulp has been insufficient to
supply the demand, and linters prices have risen sharply.20 Because of
planned expansions in the wood pulp industry and an apparent leveling-off
in demand for rayon, it is expected that prices will remain fairly stable at
reasonably low levels for the next several years.
The bleaching establishments which transform second-cut linters into
chemical cotton follow the policy of purchasing raw linters only upon re-
ceipt of a binding contract for finished chemical cotton. The operators of
the bleaching establishments have a threefold function. As bankers, they
finance the purchase of raw linters and are repaid on delivery of the chemi-
cal cotton. As warehousers, they store up to a full year's supply of raw
linters. Finally, as manufacturers, they transform raw linters into the
custom-made types of chemical cotton.
3. Cottonseed Fibers in Other Countries
The United States produces about one-half of the world's supply of cot-
20 U. S. Dept. Agr., Production and Marketing Administration, Weekly Cotton Linters
Review, 19, No. 29 (Feb. 18, 1949); 21, No. 38 (Apr. 20, 1951); 23, No. 13 (Oct. 23,
1952).
556 CELLULOSE
ton.21 As shown in Table 4, other countries producing more than a million
bales a year are U.S.S.R., India, Egypt, China, Brazil, and Pakistan. Of
these countries, China, India, and Pakistan produce only small quantities of
TABLE 4
Production pf Cotton in Principal Cotton-Growing Countries21 for Year Beginning
August 1, 1949
Production
(as thousands of
Country 480-lb bales)
United States 16,800
U.S.S.R.. . . 2,700
India. . . 2,300
Egypt. 1,790
China (including Manchuria) 1 ,700
Brazil... . . ... 1,385
Pakistan. . . . 1,000
World total, including all other countries . 31 ,190
linters, primarily because transportation is insufficient to support a cotton-
seed oil industry. The strains of cotton grown in Egypt have no linters on
the seed, that is, the seeds are "bald."22 The production of linters outside
the United States is therefore limited chiefly to Brazil, Mexico, U.S.S.R.,
East Africa, and Paraguay.
Linters purification plants have in the past been installed close to the
large users of chemical cotton. Thus, England, Germany, France, Italy,
and Japan have bleaching establishments which depend entirely upon im-
portation of linters. The trend, however, is toward purification of linters
in the countries in which they are grown. Currently, the consumption of
chemical cotton in Europe and Japan is so great that the United States
bleaching establishments export sizable quantities to these areas.23
4. Linters Purification
All the cottonseed fibers may be purified by similar means, even though
the equipment for carrying out the purification will vary depending on the
fiber length. The bales of lint cotton are broken up and mechanically
21 U. S. Dept. Agr.f Bur. Agr. Econ., Statistical Bull. 99 (1951), pp. 128-129.
22 American Cotton Handbook, American Cotton Handbook, Inc., New York, 1941,
p. 128.
23 U. S. Dept. Commerce, Bur. Census, U. S. Exports of Domestic and Foreign Mer-
chandise, Report No. FT410, Part 1, 13 (Nov., 1951).
VI. PREPARATION FROM NATURAL SOURCES
557
cleaned at the textile mills but usually undergo no other purification until
after they have been spun into yarn or woven into cloth. The subsequent
purification process usually consists of a clesizing operation, an alkaline
boil, and a hypochlorite (or peroxide) bleach with thorough washing to re-
move the remaining bleaching solution.24
The transformation of linters into chemical cotton demands an unusual
amount of quality control and will be described in detail. 25~29 The chemi-
cal operations involved in purification of raw linters are pressure digestion
STlONt
soumoi
TANK
VASN TANK
ILCACH TUB
BLEND AND STORAGE TUBS
IALEP
RAW
, BALE Of EH El AND
LINTCM t LINTEHS CLEANEI
nT 1 II
0
ISESTEI
S
^
f
'
J_
0
o
'
•
TO
ORYIN
OPERA
TIONS
O
c=>
<0
«o
0
o
(=-
~=0
1
{.
A
Fig. 11. Linters batch purification process. Courtesy of Hercules Powder Company.
in alkaline liquors, and multistage bleaching. The sequence of steps is
similar to that used in wood pulping by the soda process (see Chapters
VI -A and VII), but the exact conditions employed are different because of
the differences in the types and amounts of impurities which must be re-
moved and in the physical form of the raw cellulosic material. Linters
purification steps are primarily designed to remove wax, pectins, and small
amounts of coloring matter which are distributed on or in the fiber, and at
the same time to destroy almost completely the contaminants which are
physically dispersed in but separate from the linters. In addition, the
purification steps regulate such important chemical cotton properties as
viscosity and reactivity.
24 J. H. Kettering and R. M. Kraemer, U. S. Dept. Agr., Tech. Bull. 941 (1947).
26 J. A. Lee, Chem. & Met. Eng., 48, 90 (Apr., 1941).
26 Chem. fir Met. Eng., 48, 108 (Apr., 1941).
27 Hercules Chemical Cotton, Hercules Powder Co., Wilmington, Del., 1947, 35 pp.
28 Bukipulp, Buckeye Cotton Oil Co., Memphis, Tenn., 1949, 40 pp.
29 How Chemcot Is Custom Made to Your Specifications, Southern Chemical Cotton
Co., Chattanooga, Tenn., 24 pp.
558 CELLULOSE
The raw linters contaminants include particles of boll, stalk, leaf, and the
palisade-cell layers of hull which are quite resistant to the usual chemical
treatments. In addition, somewhat larger quantities of the epidermal
layer of the cottonseed hull (called "hull pepper" or 'hull bran") are pres-
ent in raw linters but are more easily removed in the purification steps.
The loss of hull pepper is a major reason for the drop in yield during purifica-
tion.
Batch equipment is commonly used to perform the purification steps
(Fig. 11), but recently a completely continuous process30 has been installed
in one of the bleaching establishments.
(a) Selection of Raw Linters
The amount of contamination in raw linters is one of the major factors
influencing the quality of chemical cotton and the uses for which it may be
sold. Both the quantity and type of contamination will vary widely in
linters from different cottonseed oil mills and even in the production from a
single mill. The bleaching establishments rely upon the experience of
highly skilled inspectors to purchase enough high-quality raw linters to
make chemical cotton within specification.
Representatives of the bleachers go to almost every cottonseed oil mill to
sample the production and to approve individual bales for shipment. Often
the inspectors advise the oil mills as to the quality needed for the chemical
industry and suggest methods of obtaining larger quantities of satisfactory
linters from the seeds. The response of the oil mills to the need for im-
proved quality has been good, and present quality levels are high compared
with those of a few years ago.
Each carload of inspected linters is further tested at the purification
plant. In many cases, this testing may include a complete laboratory-
scale purification. After testing, the linters are segregated and stored ac-
cording to the grades of chemical cotton in which they may be used. As
expected, the larger the stock of raw linters, the better is the opportunity
for selection so that uniform satisfactory quality levels may be maintained
in all grades of chemical cotton.
The quality level of raw linters can be improved to some extent by me-
chanical cleaning operations in the bleaching establishment. Equipment
which separates contaminants from linters by centrifugal action on a water
80 W. E. Segl (to Hercules Powder Co.), U. S. Patent 2,673,690 (Mar. 30 1954);
Chem. Eng., 61, 116 (1954).
VI. PREPARATION FROM NATURAL SOURCES 559
slurry31 is now in commercial operation. Beaters, screens, and other ap-
paratus may be used to clean dry linters.32'33
(b) Digestion
The bales of raw linters are usually disintegrated by mechanical opening
equipment as a first step in the purification process. When mechanical
cleaning treatments are used, they are applied after the bale opening and
prior to digestion. For digestion, the opened linters are wetted with
sodium hydroxide solution of the required strength (usually 2 to 4%) and
are transported to the digesters. Digestion factors such as time, tempera-
ture, and concentration of alkali must be balanced to get the desired degree
of cellulose purification and the desired level of final viscosity. Wetting
agents are often added in small amounts to the digestion liquors to aid in
removing impurities.34 Continuous digestions are reported to be carried
out in times as short as 10 min. and at temperatures as high as 185°C.30
Batch digestion, either in vertical stationary digesters or in rotating or
tumbling digesters, is usually carried out at 135-170°C. for 2 to 6 hrs. Even
though the pressure treatment is carried out in the absence of air, cellulose
is degraded by the contact with hot alkali. Severe digestion conditions
result in loss of yield and loss of viscosity.
After digestion, the spent liquor ("black liquor") must be removed by
efficient washing operations. In batch processing this is done by displace-
ment in false-bottom tubs, an operation which is feasible because of the
rapid draining characteristics of the cotton linters. Countercurrent con-
tinuous vacuum washers35 are rapidly coming into favor because they use
less water and consequently minimize dilution of the black liquor.
The black liquor contains most of the soluble organic materials which
are removed from linters during the purification process. This liquor may
be evaporated and burned as is done in the wood pulping industry (see
Chapter VI -A). Recovery of the caustic soda used in digestion is a part of
this operation. The economics of black liquor recovery require minimum
dilution for profitable operation.
81 J. D. Atkinson (to Buckeye Cotton Oil Co.), U. S. Patent 2,504,944 (Apr. 18, 1950)
82 A. J. V. Ware, U. S. Patent 2,210,016 (Aug. 6, 1940).
38 A. K. Schwartz and F. J. Walker (to South Texas Cotton Oil Co.), U. S. Patent
2,239,059 (Apr. 22, 1941); A. K. Schwartz, E. Bradshaw, and F. J. Walker (to South
Texas Cotton Oil Co.), U. S. Patent 2,274,385 (Feb. 24, 1942); Chem. Abstracts, 36, 4354
[1942).
14 E. K. Bolton (to E. I. du Pont de Nemours & Co.), U. S. Patent 2,048,775 (July
28, 1936); Chem. Abstracts, 30, 6210 (1936).
36 Filters for the Process Industries, Bull. 214, Oliver United Filters, Inc., New York,
5pp.
560 CELLULOSE
4.
(c) Bleaching
As a result of the digestion step, the linters have been freed from most of
the contaminating substances but are still tan in color and generally re-
quire bleaching treatments. Viscosity, reactivity, and other properties
may be further adjusted during bleaching. Time, temperature, pH, and
concentration of bleach are the important factors in the bleaching opera-
tion. It is customary to use multistage treatment with a sequence of steps
similar to that employed in wood pulp purification (see Chapter VII). The
active bleaching agents may be chlorine, hypochlorite, chlorine dioxide,
peroxides, permanganate, and other oxidizing agents, although the first
two are most frequently used. The bleach requirements for cotton linters
are so low that excess reagents must generally be added in order to get a
sufficiently high concentration to exert a bleaching action. It is, therefore,
not usually possible to employ the common wood pulping practice of using
less than the bleach demand in order to prevent attack on the cellulose.
Special precautions must be taken in linters bleaching to prevent or to
regulate cellulose degradation.36
The bleaching treatments must be carefully controlled and coordinated
with selection of raw linters and with digestion87 so that the finished product
specifications for such factors as color, viscosity, and cleanliness may be
met. Because these specifications vary from grade to grade, the bleaching
procedures cannot be standardized. Custom bleaching is the usual prac-
tice.
Batch bleaching is generally done in large, well-agitated, stainless steel-
lined tubs. Interstage washing is carried out in the same tubs, and the
liquors are drained off through false bottoms. Cotton bleaching is also
carried out in continuous equipment similar to that of the wood pulp in-
dustry. Both low-consistency and high-consistency towers are used,
while washing is accomplished on rotary table filters. All the bleaching
equipment is constructed of corrosion -resistant materials because contami-
nation of the purified cellulose must be avoided.
The bleaching operation also includes in many cases a treatment with
sulturic or other acidsjtojict as souring agents ana to reduce the ash content
^Sequestering agents and chgrpjfalff si|r>h.as oxalic acid are also often used tc
remove metal ions in the final
36 A. M. Dodsoii (to Hercules Powder Co.), U. S. Patent 2,478,379 (Aug. 9, 1949);
Chew. Abstracts, 43, 9447 (1949).
37 L. M. Sheldon (to Cellulose Research Corp.), U. S. Patent 2,190,274 (Feb. 13,
1940); Chfm. Abstracts, 34, 4267 (1940).
VI. PREPARATION FROM NATURAL SOURCES 501
(d) Drying
Before drying, the purified chemical cotton is usually submitted to fur-
ther mechanical treatments such as riffling38 or centrifuging, in order to re-
move siliceous39 or other foreign matter which was not removed by chemical
treatment.
Chemical cotton is usually dried and packaged in sheet form for the vis-
cose, cellulose ether, and paper trades, and in loose form for most other uses
(Fig. 12). The temperature sfnd time of drying have in general a great in-
fluence on the physical form of the finished product and its reactivity.
Careful drying is, therefore, of the utmost importance.
The first step in drying in loose form is dewatering the slurry by screens
and rubber-covered squeeze rolls to a moisture content of approximately
J502o^ The pads of chemical cotton from the squeeze rolls are fed through
pickers revolving at high speed which put the linters in a fluffy form.40 A
layer of fluffed linters on a metal apron is carried through a tunnel dryer.
Air of carefully controlled temperature and moisture content is blown
through the cellulose layer. The dried cotton is baled, weighed, and
wrapped in kraft paper in readiness for release for shipment. Loose pulp
bales usually average about ISO Ib. with a moisture content of about 5%.
Conventional fourdrinier papermaking equipment is used to dry chemi-
cal cotton in sheet form. Jordan engines are used to reduce the fiber
length so that a smooth, strong sheet can be obtained. Careful adjustment
of this cutting treatment is necessary to get uniform sheet properties. As
in the case of loose pulp, sheet pulp is dried at controlled temperatures.
The thickness, density, formation, and porosity of the sheet may be reg-
ulated over wide ranges to fit the requirements of each customer. For
some uses, the continuous sheet from the machine is cut, stacked, and
baled ; for other uses, the sheet is wound onto rolls of specified width. The
bales generally weigh 400 Ib. ; the rolls may weigh as much as 600 Ib.27
Chemical cotton for acetylation use is generally dried at low temperature
(60° to 80°C.) in order to retain reactivity. Overdrying must be partic-
ularly avoided. Cotton for papermaking is often dried to a high moisture
content. The other grades of chemical cotton are usually less critical*" in
regard to drying conditions, but a uniform product must always be ob-
tained.
38 W. E. Henry (to Hercules Powder Co.), U. S. Patent 2,394,378 (Feb. 5, 1946);
Chem. Abstracts. 40, 6815 (1946).
39 A. Langmeier (to Hercules Powder Co.), U. S. Patent 2,576,464 (Nov. 27, 1951);
Chem. Abstracts, 46, 1257 (1952).
40 W. E. Segl (to Hercules Powder Co.), U. S. Patent 2,516,262 (July 25, 1950).
562
CELLULOSE
a
I
a
I
Cu
O
bJO
s
'£
•s
I
.+_>
8
S
(M
T-H
tub
S
VI. PREPARATION FROM NATURAL SOURCES 563
(e) Analysis
The different grades of chemical cotton and the narrow specifications for
these grades require careful analytical control at every stage in the purifica-
tion process.27 The analytical work begins with raw linters and includes
"pot yield" determinations, small-scale purifications, and visual inspection.
The digestion liquors are analyzed for alkali strength, and the digested
linters are tested for such properties as viscosity and cleanliness. At each
stage in the bleaching operation, viscosity and cleanliness will again be
measured and other tests will be employed. As might be expected, a com-
plete analysis is made of the finished, dried chemical cotton so as to make
sure that the requirements of the customer are met. In many cases, the
final analysis includes the small-scale preparation of derivatives such as the
acetate and viscose. Such use tests are especially valuable in predicting
how the chemical cotton will behave in the customer's manufacturing
operations. (The details of the commonly used cellulose tests may be
found in Chapter XII.)
5. Uses for Chemical Cotton
Chemical cotton has achieved outstanding success in the preparation of
those derivatives in which good clarity, freedom from color, and high
strength are of importance. 2~6'27-28 This success is due in large part to the
exceptional purity of cotton cellulose as distinguished from cellulose ob-
tained from other sources. In addition, chemical cotton is almost always
the standard cellulose which is used for the preparation of new cellulose
derivatives. However, since chemical cotton is not available in large
enough quantities to fill all needs, wood pulp is frequently substituted in
the manufacture of established derivatives up to the extent that quality
requirements will permit. Chemical cotton, therefore, finds widest uses
when the good strength, color, clarity, and fiber properties which it can con-
tribute are essential to the finished product.41
(a) Viscose and Cuprammonium Rayon
The largest present use for chemical cotton is in the manufacture of high-
tenacity rayon by the viscose process (see Chapter IX-F). The strength
of tire cord from chemical cotton, especially under the severe heat and stress
encountered in motor vehicle operation, is outstanding. The recent com-
mercial development of the prehydrolysis-sulfate wood pulping process42
41 C. J. Malm, Svensk Papperstidn., 50, No. 11B, 135 (1947).
42 Pulp & Paper, 24, 66, 92 (Nov., 1950); Paper Trade J.f 132, 11 (Apr. 6, 1951)
564 CELLULOSE
(see Chapter VI-A) is expected to cause keen competition for chemical cot-
ton in high-tenacity rayon. This development is an indication of the dy-
namic nature of the chemical cellulose industry and of the constant need for
improvement if present markets are to be kept or new markets obtained.
Chemical cotton is also the preferred raw material for conversion by the
viscose process into extruded sausage casings. High wet strength is the
necessary property in this use. In the viscose textile rayon industry,
chemical cotton was displaced by wood pulp almost 20 years ago because of
the availability of the latter raw material.
Chemical cotton finds wide use in the cuprammonium rayon industry,
which specializes in strong, fine yarns. The spinning process, which de-
pends on drawing down thick (0.5-1.0 mm. diameter) filaments to low
deniers, can take advantage of the high strength contributed by chemical
cotton.
(b) Cellulose Esters
The exceptionally good color obtainable from chemical cotton has made it
the base for cellulose acetate and nitrate which are to be used in clear or
pastel plastics. Photographic film, whether acetate or nitrate, is generally
made from chemical cotton in order to have good clarity. In rayon uses,
chemical cotton, when used as a portion of the cellulose furnish, is said to
contribute to ease of spinning and to strength of yarn.
Chemical cotton is widely used in the manufacture of nitrocellulose ex-
plosives. It is the preferred raw material for rocket powder. Because it
can be furnished at very high viscosity levels, it is the only chemical cellu-
lose used in the manufacture of dynamite.
(c) Cellulose Ethers
As in the case of cellulose esters, chemical cotton is used in the manufac-
ture of ethers where color, clarity, and high viscosity are important. In
these alkaline manufacturing processes, chemical cotton with its low caus-
tic-soluble content has distinct advantages in yield as compared with wood
pulp.
(d) Paper and Miscellaneous Uses
The uses for rags in the manufacture of paper which have developed as a
result of the strength, permanence, and brightness of the cotton fibers are
described in Section C of this Chapter VI. The diminishing supply of
high-quality rags has created interest in the use of purified cotton linters in
VI. PREPARATION FROM NATURAL SOURCES 565
the paper field.43 Purified linters do not usually give strengths equivalent
*to rags when conventional beating conditions arc employed, but perma-
nence and brightness are excellent. In many cases, the combination of
chemical cotton with bleached sulfate wood pulp will give papers entirely
equivalent to the conventional bond papers made from rags and bleached
sulfite. Chemical cotton is often added as a part of the paper furnish to
give absorbency, bulking value, and good formation.
The high porosity of linters sheets has resulted in the development of
numerous specialty uses. Laboratory and commercial filter papers are
generally made from chemical cotton. Other chemical cotton papers are
impregnated with resins for the manufacture of plastic laminates, floor
coverings, and lubricating-oil filters. These specialty applications are ex-
pected to be of growing importance in the future.
43 W. H. Jones, Paper Trade J.t 121, 145 (Sept. 27, 1945); 123, 29 (Aug. 29, 1946).
C. RAGS
HARRY F. LEWIS
Cotton and linen rags have long been an important source of cellulose
for all uses. In the United States their use has been limited to paper and
fiber stock. However, rags have been used in other countries as raw mate-
rial for cellulose derivatives as well as for paper; for example, various types
of new cotton waste were used, in the immediate past, in nitration processes.
This discussion will be confined to the utilization of cellulose from rags in
this country.
In the earliest period in American history, rags represented almost the
only raw material available for the manufacture of paper, and it might be
said that the amount of paper produced was controlled by the amount of
rags thus collected. Wood pulps produced by the sulfite and sulfate proc-
esses have gradually supplanted rags in one type of paper after the other.
Today rags are going principally into high-grade bond and writing paper
and into ledgers where permanence and durability are of importance, and
into blottings, fiberboards, and felts, where absorbency and porosity must
be obtained.
As the result of these developments, the amount of rag stock produced
and converted into fine paper at first fell off considerably. During the
past few years, however, the production of rag-content paper in the United
States has again shown an increase — 170,110 tons in 1947, as against an
estimated 100,000 tons in 1899. These production figures do not mean,
however, that rag stock production has increased 70%. The modern
papers average 40% rag content because common practice is to dilute rag
stock with other papermaking fibers. The earlier papers were 100% rag
content.
The rags used for conversion into bleached rag stock, one of the purest
Forms of cellulose produced technically, may be either new rags or old rags.
New rags include bleached cuttings from the textile field, such as light
prints, white and fancy shirt cuttings, shoe cuttings, and blue overalls.
The old rags are sold under a variety of classifications, being termed old
whites, thirds and blues, blue overalls, and the like.
Naturally, these different materials show considerable variation in the
666
VI. PREPARATION FROM NATURAL SOURCES 567
noncellulosic impurities which must be removed in the cooking and bleach-
ing processes; hence, there is some variation in processing required to re-
move these different components. The rags before cooking are sorted, cut,
and dusted.
New rags may contain as sizing agents starch, certain synthetic resins,
and saponifiable and nonsaponifiable fats and waxes. They may also con-
tain synthetic fibers, including besides the derived cellulose fibers (rayons)
such newer fibers as those made from polyamides, polyesters, and acryloni-
triles. New rags may be white or dyed. Faster dyes, which are continually
being developed, complicate the problem of dye removal. Various resins
such as the ureas, melamines, acrylates, and polyvinyls may have been
added to impart stiffness, freedom from wrinkling, or other special effect.
Rubber and rubber derivatives are often present in knit goods. Old rags
contain similar contaminants and frequently also contain inorganic and
organic dirt. Many of these noncellulosic substances are removed during
the normal pulping processes; however, a sufficient number of the materials
cause trouble to warrant avoiding use of rags containing them. In addi-
tion, the synthetic fibers are undesirable constituents of rags intended for
conversion to paper in that these fibers contribute nothing in the way of
strength and may detract from the appearance of the sheet and degrade its
physical characteristics.
The cooking chemicals generally used for removing the various impurities
or rendering them susceptible to bleaching include lime, lime and soda ash,
and caustic soda. Grimm1 has studied the action of the different alkalies
on vegetable and animal fibers and concluded that sodium carbonate in
excess works well on vegetable but not on animal fibers, and that sodium
hydroxide attacks vegetable fibers, destroys animal fibers, and saponifies
waxes, fats, oils, etc.; in contrast, the use of lime results in less degradation
of vegetable fibers and at the same time destroys animal fibers and colors.
In general, the effect of lime on cellulose is milder than that of an analogous
amount of caustic soda; hence, with stocks where the minimizing of deg-
radation is an important factor, lime is used; in this case, the cooking
period must be extended. When the pulping agent is either sodium hydrox-
ide or sodium carbonate, the majority of the fatty impurities present go into
solution as soluble salts of fatty acids following saponification. Non-
saponifiable hydrocarbons are first emulsified but may recoagulate on the
fibers in the beater to appear in the finished paper.
In the presence of vat dyes such as indigo and the indanthrenes, the
1 H. Grimm, Zellstof u. Papier.h, 7, 32 (1921X
568 CELLULOSE
action of the cooking agent may be extended by the use of strong re-
ducing agents, for example, sodium hydrosulfite (Na2S2O4) or the modified
hydrosulfites, or milder reducing agents such as the modified starches and
simple carbohydrates. Thorough mixing of the hydrosulfite with the rags
in the presence of the least possible amount of air, followed by a washing
operation to remove the reduced dye, is essential to good stripping of such
pigments. These conditions are particularly important in the case of
indanthrene-dyed rags.
Laughlin2 has studied the effect of variables such as time, temperature,
and concentration of cooking chemicals on the degradation of cellulose.
He showed that a cooking process using 3.5% NaOH or 10% lime carried
out for 3 hrs. at temperatures corresponding to 100 to 150 Ib. steam would
not degrade the cellulose too greatly; conditions in excess of these proved
to be harmful, with pressure and chemical concentration having more
effect than time.
It is almost impossible to outline any set of conditions as being standard
for the processing of these various grades of rags. The amount of cooking
chemical used depends on the rag being cooked ; white rags take less chemi-
cal and time than do colored ones, clean rags less than dirty ones. In the
United States, the cooking operation is generally carried out in horizontal
boilers having capacities from 5000 Ib. to 16,000 Ib. or more. Rarely does
the caustic used exceed 10% of the weight of the rags nor the lime 20%.
The product obtained from the rag boiler after cooking and washing is
generally far from the brightness desired ; hence, it is necessary to remove
the cooking residues by bleaching. The cooked rags are first washed,
generally in a beater equipped with a washing cylinder, in order to remove
soluble colored substances or loosened dirt. Lime-cooked rags are generally
washed with cold water to take advantage of the greater solubility of lime
in cold water. After washing is complete, the roll is lowered, and the cut-
tings are drawn out to separate them into threads and the threads into
fibers. When this stage is achieved, bleach is added. Most rag mills use
calcium hypochlorite for the purpose, although other bleaching agents have
been proposed, and at least one of the newer ones, sodium chlorite, is said
to yield a product of satisfactory color with no degradation of cellulose.
A number of investigations have been made on the effect of the variables
of bleaching on the quality of the rag stock produced. Grain3 has con-
sidered in detail the effect of variables such as pH, temperature, and time
of bleaching on the chemical constants and physical properties of the
2 E. R. Laughlin, Paper Trade J., 97, 39 (Oct. 26, 1933).
3 R. C. Grain, Paper Trade /., 103, 37 (Dec. 10, 1936).
VI. PREPARATION FROM NATURAL SOURCES 569
bleached pulp. He has shown that, for hypochlorite bleaches between pH
4.85 and 9.5, the maximum degradation as measured by viscosity occurs
at or near the neutral point. Degradation also occurs more rapidly with
increase in temperature; this is particularly true above 45°C. Grain's
work confirms the experiments made by Birtwell, Clibbens, and Ridge,4
by Clibbens and Ridge,6 and by Davidson6 on the action of bleaching
agents on cotton to be used for textile purposes. (A detailed discussion
of factors affecting bleaching is given in Chapter VII.)
After a satisfactory color has been achieved, the stock is dropped into a
drainer with or without washing. Here it evens up in brightness and softens
somewhat, possibly due to the degradation which may occur. Inasmuch
as in the manufacture of the various grades of rag papers a mixed furnish
of stocks of different sources and cooking processes is used, the drainers
both improve the color and at. the same time provide storage facilities for
the different stocks.
The physical and chemical characteristics of the drainer stock will depend
upon the quality of the rag before cooking, the severity of cooking and
bleaching, thoroughness of washing, and conditions existing in the drainer.
The three chemical constants most commonly used to characterize rag
stocks are the cuprammonium viscosity, copper number, and alpha-cellu-
lose content (see Chapter XII). Standard methods for carrying out the
determinations have been established by the Technical Association of the
Pulp and Paper Industry.7'8-9 Bleached stocks in a good mill will vary in
cuprammonium viscosity from 30-40 to 200-400 centipoises, in copper num-
ber from 0.1 to 1 .5, and in alpha-cellulose from 90 to 98%. Few rag stocks
will be obtained having the optimum values in these various ranges; it is
not uncommon, however, to see bleached stocks from new rags having a
viscosity of 250, copper number of 0.1 , and alpha-cellulose content of 97.5%.
High-grade new rag stock of this quality will contain principally alpha-
and beta-cellulose; rag stock from old rags may contain small amounts of
gamma-cellulose.
In view of the fact that high-quality, rag-content papers are widely used
where permanence is desired, the nature of the term ' 'permanence" in this
4 C. Birtwell, D. A. Clibbens, and B. P. Ridge, /. Textile Inst., 16, T13 (1925).
6 D. A. Clibbens and B. P. Ridge, /. Textile Inst., 18, T136 (1927).
6G. F. Davidson, J. Textile Inst., 24, T185 (1933); 25, T174 (1934); 29, T195
[1938); 31, T81 (1940).
7 Tech. Assoc. Pulp & Paper hid., Standards, T 206 m-37 (Sept. 15, 1937).
8 Tech. Assoc. Pulp & Paper Ind., Standards, T 215 m-38 (Jan., 1938).
!> Tech Assoc. Pulp & Paper Ind., Standards, T 203 m-40 (Jan. 15, 1940).
570 CELLULOSE
connection is of interest. The term is associated with the concept of life
expectancy as applied to paper. The life expectancy of a paper may de-
pend on the complex chemical system of the sheet as well as on the external
conditions encountered by the sheet during its use or storage. The direct
determination of this life expectancy has not been done. It is unfortunate
that actuarial statistics are not as easily obtained with paper as with humans
and that they do not have the same wide interest; otherwise, permanence
might be estimated with some sureness. To make the collection of data
more difficult and less probable, the life expectancy of a permanent sheet
extends for several centuries and many generations of investigators.
The next best approach is to devise means for accelerating the more im-
portant processes of natural degradation so as to obtain relative figures
which will in themselves be without absolute value but will enable an esti-
mate of the relative permanence of a number of sample sheets under condi-
tions which are readily reproducible. One such accelerating agent is an
increase in the temperature at which the sheet is maintained. The ac-
celerated aging test of the Bureau of Standards10 is carried out for 72 hrs. in a
current of moving air at 100°C. Changes in the chemical composition and
physical characteristics of the sheet establish a measure of the permanence
of the sheet. Ultraviolet light has been used by a number of investigators
as an accelerated aging test, although possibly its greatest use is in obtaining
a measure of the color stability of a sheet. (This may be related to the
fastness of the dyes used in the sheet furnish.) Wood pulps show wide
variation in ultraviolet-light color stability. Any attempt to establish a
q'ukntitative relationship between the results obtained by an accelerated
aging test and the life expectancy of a paper is not warranted except in
terms of the deteriorating conditions specified in the particular accelerated
test employed.
The conditions within the sheet which influence its permanence include
the acidity of the sheet, the nature and amount of additives and impurities,
and the quality of the pulp stock used in the furnish. The pH of the water
extract11 gives a measure of the acidity; this is related to the amounts of
papermaker's alum and rosin size used in the manufacturing process. In
general, the lower the extract pH, the less permanent is the sheet in terms
of the Bureau of Standards test; this is supported by experience with com-
mercial papers. Sheets having an extract pH of 3.0-3.5 may be expected
to undergo degradation in a relatively few years even where high-grade
rag stocks are used in the furnish. The same papers having an extract pH
10 R. H. Rasch, Bur. Standards J. Research, 7, 1 (1925).
11 Tech. Assoc. Pulp & Paper Ind., Standards, T 435 m-42 (July, 1942).
VI. PREPARATION FROM NATURAL SOURCES 571
of 5.0-5.5 will last many more years. This effect is strikingly illustrated
by Hanson,12 who was intrigued by the fact that certain sections in a book
printed on rag paper in 1576 were badly discolored and weakened but others
were white and sensibly strong. The strong white sheets all had higher
calcium carbonate ash (2.5-3.0%) and extract pH value. The weaker
sheets had less ash (less than 1%) and showed lower extract pH's. The
gradation from white to brown was in direct order with the increase in
carbonate ash and the diminution in strength and extract pH of the sheets.
The quality of the stock going into the sheet may be measured by the
alpha-cellulose content, the cuprammonium viscosity, and the copper
number; it may be concluded from studies at the Bureau of Standards and
The Institute of Paper Chemistry that the higher the alpha-cellulose con-
tent and viscosity of the stock and the lower its copper number, the longer
will be the life expectancy of a sheet made from that stock, other things
being the same.18'14 Again, any attempt to establish a quantitative rela-
tionship between the results of any of these tests and the life expectancy of
the sheet is unwarranted since these results together with the pH do not
represent all of the factors in the sheet influencing its permanence. The
mechanism of degradation and the resistance of the sheet to degradation
are complex and at best are not too clearly understood.
Among the other factors which have been established as causing a sheet
to degrade are the sulfur dioxide in the atmosphere,15-16 metallic resi-
dues which catalyze the oxidation of sulfur dioxide to sulfur trioxide,17
acids and other corrosive material in the inks,18 fungus attack under
favorable conditions of heat and humidity, and the action of hot sunlight
at high humidity. Under any of these conditions, the best-made sheet will
deteriorate rapidly. The better the sheet, the longer will it withstand such
deterioration. Under normal conditions, a sheet of permanent paper will
last a long time. For permanent storage of valuable papers, precautions
must be taken to maintain optimum conditions of storage with the elimina-
12 F. S. Hanson, Paper Ind. and Paper World, 20, 1157 (1939).
18 H. F. Lewis, Paper Trade J., 95, 29 (Nov. 24, 1932); 96, 41 (May 11, 1933).
14 R. H. Rasch and B. W. Scribner, /. Research Natl. Bur. Standards, 11, 727 (1933);
23, 405 (1939).
16 A. E. Kimberly, J. Research Natl. Bur. Standards, 8, 159 (1932); U. S. Bur. Census,
Vital Statistics, Special Reports 3, No. 33, 153 (1937).
16 M. S. Kantrowitz and R. H. Simmons, Proc. Graphic Tech. Conference, 1936, 3.
17 W. H. Langwell, Tech. Bull., Brit. Paper Board Mfg. Assoc., 29, No. 1, 21 (1952);
No. 2, 52 (1952).
18 M. E. Whalley, "Abstract of Report to League of Nations on the Permanence of
Paper/1 Paper Trade J., 97, 32 (July 20, 1933).
572 CELLULOSE
tion of degrading light waves and undesirable atmospheric constituents.
Often associated with permanence is the term durability. Not all per-
manent papers are also durable but paper such as currency which will be
handled extensively must be both permanent and durable. A permanent
paper which is initially strong will generally be classified as durable.
Apart from the use of rag cellulose in the fine papers, large amounts of
rags go into the manufacture of felt for later impregnation with asphalt
for roofing purposes or impregnation, coating, and printing for floor cover-
ings. The paper used is generally soft and very porous; because a free
sheet is desired, the beating process is carried out with a minimum of
hydration.
Quantities of rags also find their way into the manufacture of vulcanized
fiber, which operation involves the treatment of waterleaf paper, usually
made from old rags, with a solution of zinc chloride or with sulfuric acid.
For many types of fiber, old rags work better than new rags or wood pulp,
although the latter is used to a considerable extent. The virtues of old
rags may well lie in their characteristic combination of degradation and
oxidation. A number of plies of the treated sheet are combined, and the
laminated sheets are passed through successive baths of ever weaker zinc
chloride and finally into fresh water. After drying, pressing, and calender-
ing, the material is ready for use in electrical insulation, in the manufacture
of luggage and trunk coverings, and for other related applications.
Although for years the principal source of cotton fiber for the manufac-
ture of high-grade paper has been rags, within recent years there has been
an increasing use of cotton linters and even the long staple lint cotton (see
Section B of this Chapter VI). In the case of the latter, some attention has
been paid to the development of cotton types which would yield fibers show-
ing improved papermaking qualities. Promising results have been obtained.
D. BAST FIBERS, FIBROVASCULAR ELEMENTS, CEREAL
STRAWS, AND GRASSES1'2
SIDNEY D. WELLS
This section will deal primarily with the use of bast, stem, and leaf fibers
for the preparation of cellulose pulps and not with the use of these fibers
for textiles or cordage. Regardless of the end use of the fibers, the isola-
tion of the fibrous constituents of the plant material from the nonfibrous
constituents is largely mechanical. The filamentous character of the
fibrous elements is retained as much as possible when use for spinning and
weaving is contemplated, whereas reduction to the ultimate fiber, usually
too short for textile purposes, is practiced when use for papermaking or
chemical cellulose is the objective.
The bast fibers form fiber bundles between the outer bark and the woody
portion of the stems of plants. Their function is to give strength and flexi-
bility to the stem. Bast fibers are also called "soft" fibers. Before they
can be used in the chemical cellulose, paper, or textile industries, they must
be separated from the wood of the stem and to a greater or less extent from
the gums and other materials which hold them together to form the inner
bark. The ultimate fibers of which these bundles are composed are gener-
ally short, so that, except for chemical conversion, the bundles themselves
are not broken down this far.
The fibrovascular elements of plants are the veins of the leaves and stems.
Their function is not only to give rigidity, as is the case with the bast fibers,
but also to transport water and plant foods. These vessels with their pro-
tective tissues form fibers that are larger and stiflfer than the bast fibers.
Hence, they are often known as "hard" fibers.
1 This section is in part a revision of that on "Bast Fibers" prepared by Kyle Ward,
Jr., for the first edition, pp. 539-549.
2 In Bibliography Series No. 176 of The Institute of Paper Chemistry, Part II by C.
J. West, is given an annotated bibliography of bast fibers up to August, 1950; in Part I,
Series No. 171, is given an annotated bibliography of cereal straws up to October, 1949.
Additional information can be found in the Bibliography of Papermaking published by
the Technical Association of the Pulp and Paper Industry. The availability of these
bibliographies makes unnecessary the inclusion of many literature references in this
contribution.
573
574 CELLULOSB
The cereal straws, cornstalks, sugar cane, and esparto are all derived
from plants that are classified as grasses and are monocotyledons. Me-
chanical separation of the fibrovascular bundles is rarely practiced because
of the futility of the operation; pulping is accomplished by digestion of the
material as harvested without any effort to remove the nonfibrous constit-
uents before cooking. Separation after pulping can be accomplished but
usually paper or board made therefrom contains both the nonfibrous and
the fibrous elements in much the same proportions that they occur in the
original plant material as harvested.
1. Bast Fibers
The principal bast fibers are flax, hemp, jute, ramie, and paper mul-
berry. Other less important sources are sunn, kenaf, cadillo, baobab,
nettle, hops, okra, milkweed, lespedeza, soybean, kudzo, sweet clover, and
alfalfa.8 Rarely can the value of the bast fiber, for papermaking or chemi-
cal cellulose, warrant the cost of cultivation and collection for those pur-
poses alone. Although much work has been done in studying these fibers
primarily for papermaking, the information gained has been of academic
interest and the instances of verification through commercial use are ex-
ceedingly rare. When the flow of commerce is interrupted by war, interest
in new fibers and sources of fiber becomes active, but when normal exchange
of products in international commerce is restored, the well-established
sources regain their pre-eminence. Much hand labor is usually required to
produce an adequate supply of these substitute fibers for further processing
for dissolving pulps, papermaking, textiles, and cordage. This can be
accomplished only in those sections of the world where such labor is avail-
able at rates much lower than in more industrially advanced areas.
Bast fibers were the principal source of cellulose for papermaking for
many centuries before the cotton fiber occurred in any important quantity
as rags available for paperma4dng purposes. Linen, from the flax plant,
was by far the main standby and it still excells in its adaptability for use
in the highest grades of paper. When the flax plant is grown for fiber, the
seed is sown more densely so that the plants crowd each other and the stems
are comparatively free from branching until reaching the crown ; the pro-
duction of seed is a minor consideration. When raised for seed for linseed
oil, as is largely the case in the United States and most flax-producing
countries in the temperate zones of the world, the seed is sown more widely
and the plant is more branchy in character. For many years the straw
8 J. M. Mathews, Textile Fibers, 5th ed., Wiley, New York, 1947.
VI. PREPARATION FROM NATURAL SOURCES 575
obtained as a by-product from the production of flaxseed has been burnt or
allowed to rot in the fields to the extent of several million tons each year in
the United States and in comparable amounts in Canada, Argentina, Russia
and India, which are important areas in the production of flaxseed.
The actual plant fiber from which the bast fibers are derived usually
amounts to from 10 to 20% by weight of the stalks or stems. The re-
mainder is waste or, at best, a source of fuel to operate the necessary equip-
ment. The decortication of flax straw exemplifies the processing and
general principles involved in all bast fiber production. Mechanical
treatments which remove the nonbast woody fragments of the stem from
the fiber include the rippling or separation of the seed, decortication of the
fiber, and scutching; these operations require considerable hand labor.
Preliminary mild enzymatic treatments (called retting) decompose the pec-
tins and gums and thus facilitate the mechanical separation without deg-
radation of the cellulose. However, retting processes are, in general, too
expensive for production of bast fiber for papermaking.
When rags or textile wastes are available at the price of rags or cuttings,
it is uneconomical to start with the original flax plant for the manufacture
of paper. Rags and cuttings have benefited by all the treatments through
which the preparation of the textile product has proceeded.
The straw from seed flax, on treatment with brakes, is commercially
useful as upholstery tow and insulation quilts. In the United States alone,
however, several million tons of flax straw are produced in an average year
and only an insignificant quantity is used for industrial purposes. The
manufacture of cigaret paper from flax straw utilizes part of this agricul-
tural residue which would otherwise be wasted. During the years leading
up to the 1930's, the manufacture of cigaret paper occurred to the largest
extent in France and other European countries. A few mills existed in the
United States but not nearly enough to supply the domestic needs. Linen
rags supplied the largest proportion of the fiber used. With the threaten-
ing conditions occurring in the late thirties, it became apparent that the
supply of foreign cigaret paper or of imported linen rags would probably
be seriously restricted. Consequently, the domestic manufacturers in the
field studied the factors involved in producing cigaret paper from flax tow
obtained from seed flax straw and to a remarkable extent were successful.
The Forest Products Laboratory had been engaged during the 1920's in
pulping flax tow with various cooking liquors.4 Those composed of caustic
soda and sulfur or caustic soda and sodium sulfide mixtures (similar to
4 E. R. Schafer and C. E. Curran, U. S. Forest Products Laboratory, Mimeographed
Report R1159 (1938).
576 CELLULOSE
those used in kraft wood pulp practice) showed considerable promise when
applied to flax tow. Multistage chlorination, caustic extraction, and mild
hypochlorite bleaching had also been developed in the early 1930's, so that
the tools became available to produce pulps from flax tow that made very
satisfactory cigaret papers. The newest and largest mill in the field was
built in North Carolina, making the United States self-sufficient in its total
production of cigaret paper. In the United vStates 200,000 to 300,000 tons
of flax straw were utilized annually for making cigaret paper in 1948 to
1951. The yields of bleached fiber amounted to approximately 8% of the
weight of the original flax straw processed.
Flax pulp must produce paper having wearing properties impossible of
attainment with cotton fiber or wood pulp to justify consideration. The
needs for the great majority of paper products and dissolving pulps can be
met by cotton or purified wood pulps. Under these circumstances it is
very difficult, if not impossible, to organize the collection, transportation,
and purification of agricultural residues on a substantially competitive
basis.
Although the utilization of flax straw in the manufacture of cigaret
paper has indicated a" limited solution of the use of this agricultural by-
product, there still remains 85 to 90% of the total annual production which
finds no commercial use. In the manufacture of cigaret paper, a certain
amount of degradation of the cellulose in processing can be permitted. If a
large portion of the annual crop is to be used, the original degree of poly-
merization of the cellulose in the bast must be preserved. At the same
time, a higher purity of the cellulose is necessary if utilization is to be ex-
tended to the manufacture of banknote or currency papers or if the fine
paper manufacturers are to be induced to pay a price greater than the pre-
vailing prices of high-grade cotton rags or purified wood pulps. The actual
flax fiber that withstands all the various stages of pulping and purification
and which, under the best conditions, occurs in the final paper product,
amounts to about one ton per fifteen tons of flax straw harvested for the
purpose. The value of the final product must justify the large volumes of
raw material entering the system. Textile fibers have long been sold at
prices that can justify such costs. Closer integration of the production
of the so-called "line" fiber for textiles and cordage, with the utilization of
the combings as tow for papermaking, may be possible and, to whatever
extent it is, the economics of the system can be benefited. The demand for
textile fiber must be balanced by the demand for papermaking fiber. When
lot in proper balance, the benefits of integration may cease to exist.
Hemp for textile purposes undergoes a type of retting process similar to
VI. PREPARATION FROM NATURAL SOURCES 577
that employed for flax. The fiber thus obtained is used in ropes and cordage
and as a substitute for flax in the cheaper linens. Hemp is grown widely
in Europe and Asia, and, to a lesser extent, in the United States. In the
last country its growth is controlled by law, since the narcotic marijuana
is produced from the same plant.
Jute is also retted, usually in pools of stagnant water. When retting is
complete (from 10 to 20 days), the bast fiber is easily separated from the
woody stem. As a textile it is used in twines and cordage and in rough
fabrics, such as carpet backing or burlap bags.5 India produces practically
the entire world supply of jute.
Ramie finds much less commercial utilization than flax, hemp, and jute
fibers. The reason for this lies, in part at least, in the difficulty of purify-
ing and spinning the fiber. Simple retting is not sufficient for removing
the bast fiber, as in the case of the preceding crops, and none of the many
mechanical methods suggested for decortication has established itself in-
dustrially. In China, the bast fibers are stripped off by hand in long rib-
bons. The bark and some of the gums are scraped off, also by hand. The
fiber thus obtained must be further degummed before spinning, which is
difficult to do without injury to the fiber. Retting is not satisfactory, and
chemical treatments, especially alkaline cooks, seem to be the most promis-
ing. Ramie is grown in quantity only in the Far East,6 although small
plantings,7 mostly for experimental purposes, have been made at many
places in the southern part of the United States.
The bast fiber of the paper mulberry is unusual in that the fiber is proc-
essed into a fabric without either spinning or weaving. The clean fibers
are laid out wet in several layers and allowed to dry overnight. The next
morning they will be found to have adhered to each other to form a single
layer which is beaten with a wooden mallet until it forms a smooth strong
cloth.3 The paper mulberry is grown in India and Japan and on the islands
of the Pacific.
Sunn is prepared by a retting process similar to that used for true hemp
and finds its main use in the preparation of nets and cordage. This plant
is grown extensively in Southern Asia.
Kenaf is used like jute for cordage and sacking but, in general, is inferior.
Cultivation and processing are also similar to that of jute. Kenaf is said
6 N. C. Chaudhury, Jute and Substitutes, 3d ed., W. Newman & Co., Calcutta, 1933,
249 pp.
6 G. L. Carter and P. M. Horton, Ramie, Louisiana State Univ. Studies No. 26, L. S.
U. Press, Baton Rouge, 1936, 100 pp.
7 L. H. Dewey, U. S. Dept. Agr., Misc. Pub. 518 (1943), 95 pp.
578 CELLULOSE
to have been introduced into India from Africa. Today the commercial
crop is almost entirely obtained from India, and one frequently used name
for the fiber is Deccan hemp, the name coming from the section of India in
which kenaf is grown. Gambo hemp and ambari hemp are other names
for the same material.
Cadillo or urena fiber6 is the bast fiber from certain tropical shrubs which
include Urena lobata and Urena sinuata. These fibers, with those from
some other shrubs, are now being designated as Cuban jute and are recom-
mended as substitutes for true jute.
The bast fiber of the baobab, or monkey's bread tree (Adansonia digitata),
is known as Adansonia fiber and finds some use as cordage. The tree is a
native of Africa, but is now grown in both the East and the West Indies.
The nettle is reported to be an excellent textile fiber and received a great
deal of attention in Germany during the war years of 1914-18 and 1939-
45. The fiber content of the wild nettle is about 6%, but this has been in-
creased by cultivation to 12-13%. The fiber may be separated from the
stalk by retting, as with flax, or by mechanical decortication, as with ramie.
In either case, strips of fiber are obtained which must be chemically de-
gummed.
The hop fiber can be separated by retting, but the process is time con-
suming and not very practicable.
The bast fiber of the potato plant has been studied in Germany for tex-
tile purposes, but it* does not appear to be economically practical under
normal circumstances.
The bast fibers of certain trees, such as the willow or sequoia, have found
some very limited applications in textiles, usually as cordage or felting ma-
terials. The bast of the castor bean plant has also been recommended for
this purpose.
2. Cellulose Pulps from Bast Fibers
The bast fibers have never had any industrial importance in the prepara-
tion of rayon or of chemical cellulose except when increased demand and
interruption in the supply of cotton and wood pulp may have caused a
search for other sources. There is considerable patent literature on the
subject.2 The following is a brief summary of the methods proposed for the
utilization of these materials.
There are three possibilities to be evaluated if any bast fiber plant is to
be considered as a source of cellulose. First, the separated bast fiber may
be used. However, the quality of these long fibers is very high, and they
VI. PREPARATION FROM NATURAL SOURCES 579
are usually too valuable for textile purposes to compete with cotton or
chemical wood pulp in the cellulose industry, except in the form of rags
or similar textile wastes. Second, the woody residues from the bast fiber
separation may be pulped. These residues have received considerable
attention in the case of the important textile materials, flax and hemp.
Third, the entire stem may be treated. This is not simple, for a process
capable of completely pulping the woody portion is so drastic that the bast
fibers lose the ability to resist wear for which the bast plant was chosen.
Cotton rags or wood pulp can usually be obtained in a free market at a cost
lower than that of collecting and processing fibrous plants; if the special
properties possessed by the bast cellulose are lost in the processing, there is
no point in trying to utilize them. Flax straw is decorticated by breaking
between rolls to "fine tow" amounting to about 20% of the original straw.
It is then cooked with approximately 15% caustic soda and 5% sodium
sulfide. The ultimate yield of paper is one ton from fifteen tons of straw.
Flax straw is now the source of much of the fiber in the long-established
mills engaged in the manufacture of cigaret paper in western Massachusetts,
New Jersey, and Pennsylvania and in a new plant in North Carolina.
When the decorticated fiber is used for papermaking, the mechanical treat-
ment may be carried to the extent that the tow is too short for the purposes
other than paper or chemical cellulose manufacture. More complete re-
moval of the nonbast constituents is possible, however, without degrading
the cellulose of the bast, as is invariably the case when chemical methods
are used. It has recently been found possible to produce bleached pulps
of high viscosity satisfactory for the manufacture of banknote paper and
for use with cotton in the manufacture of high-grade ledger and other rag
papers. The mechanical disintegration of the woody matter, when about
95% dry, proceeds much faster than that of the bast fiber when the tow or
fluff is beaten with rods; the bast fibers can be isolated by means of screens
and by pneumatic separation, and are then treated at room temperature
with dilute caustic soda. Finally, with multistage chlorination, caustic
extraction, and bleaching, a shive-free pure white pulp is secured having a
TAPPI viscosity greater than 100 centipoises.8
Where the special wearing characteristics of the bast fibers are desired,
primarily because cotton and wood pulp fibers are incapable of such develop-
ment, it is obvious that the woody shive from the flax stem or branch struc-
tures must be removed as completely as possible before any chemical pulp-
8 S. D. Wells (to The Institute of Paper Chemistry), U. S. Patent 2,298,994 (Oct. 13,
1942); Chem. Abstracts, 37, 1868 (1943); U. S. Patent 2,452,533 (Oct. 26, 1948); Chem.
Abstracts, 43, 3198 (1949); and patent applied for.
580 CELLULOSE
ing occurs. Otherwise, the cellulose fibers and cells from the pulping of the
shive will contaminate the bleached bast pulp and the paper made there-
from will appear, on microscopic examination, to be adulterated with wood
pulp. With complete decortication of the straw, the yield of bast should
not exceed 20% of the original chaff -free straw. On subsequent pulping and
bleaching, the final products, free from cellulose derived from the shive but
containing all the bast, cannot exceed 12.5% of the original agricultural
residue. It is consequently apparent that only manufacturers of the papers
that command exceedingly high prices can afford to use it. The proper-
ties possessed by the properly prepared fiber, however, are such that the
high cost is justified.9-10
3. Fibrovascular Elements
Many fibers used in commerce and industry consist of filaments of indefi-
nite length rather than individual fibers. These filaments are obtained by
separating the fibrovascular elements of stems and leaves from the paren-
chyma tissue by means of scraping, beating, scutching, and combing,
usually by hand. The product is valued principally for its strength, either
wet or dry, and its ability to be spun or twisted into cords or coarse threads
which can be used for rope, cordage, or twine and in coarse textiles, such
as burlap, bagging, netting, matting, and hammocks. The filaments have
essentially the same chemical composition as the original plant, and the
presence of noncellulosic matter is not important in so far as it does not
affect the durability, strength, or other physical properties of the product
made therefrom. Their value for cordage is usually so much more than
the prevailing price of paper stock that they are usually not considered as
sources of cellulose. Among the fibers of this class may be mentioned the
following: henequen, sisal, abaca, phormium, and caroa. In addition are
latona, mescal, zapupe, cantala, ixtle, pita, cabuya, fique, cocuiza, pitre,
yucca, banana, palm, palmetto, zray, and toquilla which have local im-
portance or which enter occasionally into world commerce. They are all
generally classified as leaf fibers and comprise from 10 to 20% of the leaf
and stem substance from which they are derived.7
When these materials are subjected to the usual alkaline cooking treat-
ments, the individual ultimate fibers are liberated and the noncellulosic
constituents are dissolved. The lengths of the ultimate fibers bear no rela-
tion to the lengths of the filaments from which they were derived. Sisal,
for instance, yields a fiber not much longer than short-fibered hardwoods,
* H. S. Spencer, Pulp & Paper Mag. Can., 47, No. 10, 95 (Sept., 1946).
10 G. H. Lafontaine, Pulp & Paper Mag. Can., 52, No. 7, 142 (June, 1951).
VI. PREPARATION FROM NATURAL SOURCES 581
such as poplar, beech, birch, or maple. Manila hemp and caroa, on the
other hand, yield a very long and uniform fiber, capable of use in the manu-
facture of high-priced papers suitable for special uses such as tags, sand-
paper, flour sacks, electric insulation, tea bags, stencils for mimeographing,
and lens paper. These fibers do not fibrillate or hydrate on beating to the
same extent that bast fibers, wood pulp, or rag fibers do, and papers made
therefrom are characterized by an unusual combination of porosity, resist-
ance to wear, and wet and dry strength (see Chapter VIII).
4. Cereal Straws
Rye, wheat, oat, and barley straws have been an important source for
cellulose fiber since the invention of paper between A.D. 25 and A.D. 58 by
Tsai L'un, secretary in the court of Honaug-Han, emperor of China. In
fact, the use of straw antedates linen and cotton rags as well as wood pulp,
the most important source of cellulose at the present time. At the time of
and prior to about 1 860, straw was the principal source for cheaper papers,
and most of the newspapers of that period were manufactured from a mix-
ture of rag and soda straw pulp. Newspaper files of the Civil War period
in our large public libraries are in much better condition than those of the
First World War because of the greater permanency of straw pulp as com-
pared with groundwood pulp.
With the advent, during the last half of the last century, of the soda,
sulfite, and sulfate processes for pulping wood (in addition to the mechani-
cal process which was invented in 1S55) (see Section A of this Chaptei VI),
straw was replaced by wood pulp on account of the lower cost of producing
the latter, and straw pulp is now used only in papers where it is desired to
impart properties such as most even formation not possessed by papers
made from the more recent competitors. Advances in the art of pulping,
however, have greatly enlarged the possibility of reducing the cost of pulps
from cereal straws. Some of the more recent processes, such as multistage
bleaching with chlorination and caustic extraction, have so improved the
quality and so enlarged the range of properties which can be obtained that
it seems probable that the use of straw may attain a relatively more im-
portant role in the future. In the older and longer established industrial
countries, straw has always retained its position as a source of cellulose.
It seems reasonable to expect that this position will be regained to an im-
portant extent in the United States as the supply of virgin timber becomes
more limited and the dependence upon second or third growth cordwood
becomes more common.
Most of the straw used for paper pulp in the United Stated is cooked with
582 CELLULOSE
milk of lime, dolomite lime being preferred. About 10% of calcium oxide
or 13% of burnt dolomite is required. The cooking is carried out in spheri-
cal rotary digesters at a temperature of about 1 15°C. f or 8 to 10 hrs. There
is usually considerable false pressure, so that the pressure as shown by the
pressure gage will usually be at least 45 Ib./sq. in. After cooking, the pres-
sure is relieved (this is accompanied by the escape of considerable ammonia
produced during the digestion), and the contents are dumped on a conveyoi
and piled in heaps for drainage and a further softening of the knots and more
resistant portions through seasoning in the presence of the spent liquors.
The material is then washed in beaters equipped with drum washers, or the
spent liquor is pressed out by means of screw or roll presses, and the fibers
separated by passage through rod mills. The product may be further
washed on vacuum filters or decker washers and is then suitable for further
treatment with jordans for the manufacture of corrugated paper, capstock,
egg-case filler board, and ordinary stiff cardboard.
The yield of board obtained is usually between 65 and 70% of the weight
of straw used; the mineral matter present varies from 10 to 15%, of which
about half is silica and the other half magnesium and/or calcium compounds.
When the straw is reasonably free from weeds and has been baled while
dry and stored under cover, the pulp obtained as described above can be
bleached with chlorine, caustic extraction, and hypochlorite in several
stages to produce a superior white pulp which is suitable for high-grade
uses. Only since 1932 has availability of suitable equipment made this
procedure possible and then only in locations where stream pollution from
the spent cooking liquors is not objectionable. Ultimate yields up to 50%
of the weight in the dry straw are obtained ; this fiber is softer and not so
easily slowed down in draining properties as when the older conventional
cooking methods are used.
Among the foremost common cereal straws, rye straw is preferred because
of its somewhat longer fiber and higher cellulose content; wheat straw
comes next. Oat straw is used where the supply of the other two is inade-
quate; the yield of cellulose obtained therefrom is noticeably lower, and
the pulps produce a paper or board inferior in strength. Barley straw is
objectionable on account of the beards which are not readily reduced in the
cooking operation.
The soda and sulf ate processes are well established in continental Europe
and for the last decade have been used in Great Britain. The silica present
in the straw interferes with settling in the causticization of the recovered
soda, so that recoveries of 65 to 80% are more common as compared with
80 to 90% for soda and kraft pulp mill processes in which wood is used.
VI. PREPARATION FROM NATURAL SOURCES 583
In recent years, several pulp and paper mills have been built in South
America, South Africa, Europe, and the Phillipines which use the Pomilio
process11 for the pulping of cereal straw and bagasse. The process consists
of digestion of vegetable fibrous material with caustic soda at moderate
temperature and concentration, washing, treatment with chlorine, extrac-
tion with dilute caustic soda solution, washing, and bleaching with cal-
cium hypochlorite. It differs from the multistage bleaching of mildly
cooked soda pulps in that a much greater part of the pulping and purifica-
tion of the plant material is accomplished by the chlorine than by the caus-
tic soda. The electrolytic production of chlorine and caustic soda is an im-
portant part of the process, and sodium chloride is considered the primary
chemical rather than caustic soda or chlorine.
World patent rights to the Pomilio continuous process have been ac-
quired by the Cellulose Development Corporation of Hatch End, Middle-
sex, England. With the experience gained by the operation of a pilot
plant of a daily capacity of about three tons, this firm has designed and
built over twenty commercial plants in various parts of the world, among
which is the Cie Nord-Africaine de Cellulose recently built and placed in
operation at Baba-Ali in Algiers. The process used in these plants is com-
pletely continuous and is known under the name of Celdecor-Pomilio.
Straw and esparto grass are the only fibrous raw materials processed in the
last-named plant, and production is reported as 35 tons of high-grade white
paper per day.
Caustic soda will pulp wheat and rye straw12 if the chopped material, in
the presence of five to six times its weight of water, is passed through two
rod mills in series at temperatures between 95 and 98°C. The pulped
material can be further purified by digestion under 60-lb. pressure in rotary
cookers with an excess of caustic soda. The partially spent cooking liquors
can be used as the source of caustic soda in the preliminary treatment with
rods, so that a two-stage cooking system results with countercurrent flow
of the cooking chemical. Unusually effective utilization of the chemical is
thus attained, with higher yields and more effective use of equipment be-
cause of the reduction in bulk of the straw through the rod mill action.
Cereal straw may be pulped effectively by means of sodium carbonate
and sulfur or sodium sulfite. Both processes have been adopted by Ameri-
can strawboard manufacturers in the production of superior corrugated
paper under the name of "Strawkraft" and odorless egg-case fillers under
11 T. G. L. Becker, "Twenty-five Years' Pulping Developments of Agricultural Resi-
dues," Paper Trade J., 132, 16 (Mar. 23, 1951).
12 S. D. Wells and P. A. Forni, Paper Trade /., 112, 32 (June 12, 1941).
584 CELLULOSE
the name of "Nuprocess." Yields of 65 to 70% of fiber are realized, and
products are obtained which are from 50 to 100% stronger than correspond-
ing products cooked with lime. The extension of the use of Hydrapulpers,
Dyno Pulpers, Pulpmasters, and similar equipment to straw pulping has
recently shown considerable merit when active reagents such as caustic
soda and sodium sulfide are used. The Northern Regional Laboratory has
designated the process as the mechanicochemical process.18 Recent devel-
opments have shown that superior pulps can be made from cereal straws
by using dilute caustic for soaking the straw from 48 hours to 3 days at
room temperature.14 The Cartiera F. A. Marsoni at Villorba in Italy has
used for several years a cold caustic process in the production of very satis-
factory greaseproof and glassine papers from cereal straw.
5. Esparto Grass
Esparto was first used as a source of papermaking cellulose fiber by
Thomas Routledge in Great Britain in 1856. It has since become one of
the major sources for fiber in Great Britain and is imported from Spain
and Northern Africa. Bales of the grass are used as return cargo in steamers
transporting British coal to countries bordering the Mediterranean.
The bales of grass on receipt at the paper mill are opened, dusted in a
conical duster or willow, and charged into a vertical digester larger in
diameter in proportion to height than is customary in cooking wood by
either the soda or kraft processes. The digester is provided with a per-
forated false bottom through which the cooking liquor is withdrawn and
pumped to above an annular perforated distributing plate just below the
top of the digester. This plate distributes the liquor uniformly over the
charge. In the course of 2.5 to 3 hrs. the digestion is completed. In filling
a digester, several additions of the loose grass are necessary to make a com-
plete charge (in much the same manner as when straw is charged into the
globe rotaries used in American strawboard mills) since the material packs
more closely as cooking proceeds. Rotaries are not used, however, because
of the objectionable effect of the rotation on the fiber structure. Vessels
holding as much as seven tons of esparto grass have been reported in use.
The liquor charged will test 45 grams per liter of NaOH and the temperature
used may reach 148°C. with a gage pressure of 50 Ib./sq. in. On comple-
tion of the cook, the steam is blown off to recover the heat, and the strong
18 S. I. Aronovsky in J. Newell Stephenson, editor, Pulp and Paper Manufacture,
Vol. 2, McGraw-Hill, New York, 1951, pp. 67-69, 76, 77.
14 S. D. Wells (to Mine & Smelter Supply Co.), U. S. Patent 1,769,811 (July 1, 1930);
Chem. Abstracts, 24, 4634 (1930).
VI. PREPARATION FROM NATURAL SOURCES 581
black liquor is allowed to drain. The charge of chemical and the volum
of liquor may be changed to meet variations in the quality and conditioi
of the grass; these modifications are a matter which depends upon the judg
ment of the supervisors. A more extensive description is given by Bevei
idge.15
After the strong black liquor is drained off, the pulp is flushed with ho
washings from earlier cooks and finally with hot water. The pulp is the
treated in much the same manner as that followed in wash pan or diffuse
operation in soda and sulfate mills operating on wood.
Papers made with esparto pulps as the major portion of the furnish hav
excellent bulking properties. The faithfulness with which they reproduc
the tones and impressions of type and plates accounts for the distinctive
ness of printing in England. Although the length of the fibers is less thai
that obtained from coniferous woods, the fiber diameter is much less an<
the strength of the paper obtained is ample for printing purposes. Th
retention of china clay, with which Great Britain is abundantly supplied, i
unexcelled and the over-all cost of the furnish compares favorably with th
cost of furnishes based upon wood pulps made by the sulfite and soda proc
esses.
Furnishes containing esparto pulps in considerable proportion reniaii
dispersed to a remarkable degree in the formation of the web on the wir<
of the paper machine and resist agglomeration better than most othe
fibers. For this reason the formation of the sheet is usually better. Ther<
is probably a close relationship between this property and the retention o
clay with which the use of esparto is usually associated.
Considerable success is also reported by users of the Celdecor-Pomili<
process in pulping a wide range of fibrous agricultural residues and fibrou
materials derived from the grasses, and the organization of harvesting tech
niques and machinery can be expected to follow the more extended use o
such materials.
6. Bamboo
Bamboo is now an important source of cellulose in India. The sod;
process as modified by Raitt16 is used; it involves the preliminary crushin]
of the bamboo nodes and counter current use of the alkali in a two-stagi
digestion. Higher yields of cellulose are obtained than from cereal straws
15 J. Beveridge in J. Newell Stephenson, editor, Pulp and Paper Manufacture, Vol. 2
McGraw-Hill, New York, 1951, pp. 85-91.
16 W. Raitt, The Digestion of Grasses and Bamboo for Papermaking, Crosby, Lockwooc
& Son, London, 1931, 116 pp.
586 CELLULOSE
but the operation must bear the cost of the collection, whereas with cereal
straw this cost is borne by the grain. In India, where the supply of bamboo
is enormous and labor exceedingly cheap, an industry of considerable mag-
nitude has become established.
7. Bagasse and Cornstalks
Bagasse and cornstalks have been the subject of papermaking develop-
ments on numerous occasions since the middle of the last century. These
efforts have all failed because of the bulkiness of the raw, material, the large
proportion of nonfibrous cellulose which is less resistant to pulping than
the fibrous material, and the fact that most attempts to utilize new mate-
rials are in the hands of promoters rather than under the guidance of estab-
lished operators. For the future, however, with advances in pulping tech-
nology, it is conceivable that bagasse pulping operations can be made
feasible where the economics of competitive materials will allow it.17 The
procedure referred to for use on wheat straw, which consists of a counter-
current two-stage cook with the rod mill as a continuous digester for the
first stage, has been found to be particularly applicable to bagasse and
cornstalks. The nonfibrous cellulose material produced is used as a stiffen-
ing agent in paperboard manufacture, and bleached cellulose fiber is used
for light-weight and high-grade papers for numerous uses. Four tons of
cornstalks or bagasse have been found to yield one ton of high-grade
bleached cellulose fiber and one ton of the nonfibrous material mentioned
above.18
8. Value of Annual Plants
In the manufacture of viscose rayon and staple fiber, purified straw pulp
found use in Germany in the 1930's. For cellulose esters, such as the
nitrate and acetate, and cellulose ethers, such as the methyl and ethyl
ethers, cotton linters, and purified wood cellulose low in pentosan content
are preferred. The pulps obtained from annual plants, especially cereal
straws which comprise the largest source of cellulosic material collected
as a step for the harvesting of a more valuable product, have a high pento-
" Chem. Eng. News, 30, 2708 (1952).
18 S. D. Wells and J. E. Atchison, Paper Trade J., 112, 34 (Mar. 27, 1941); S. D. Wells,
U. S. Patent 2,029,973 (Feb. 4, 1936); Chem. Abstracts, 30, 2000 (1936); U. S. Patent
2,181,556 (Nov. 28, 1939); Chem. Abstracts, 34, 2174 (1940); Ind. Eng. Chem., 21, 275
(1929); S. D. Wells and R. Steller, Paper Trade J., 116, 45 (Apr. 15, 1943).
VI. PREPARATION FROM NATURAL SOURCES 587
san content; where this constituent is unobjectionable, an enormous
source of material is available.
The degree to which plant fibers can be used in paper manufacture de-
pends upon the dimensions of the fibers, the proportion of fibrous to non-
fibrous cells, and the physical properties of the fiber, rather than upon the
chemical properties of the pulp. With the exception of the bast fibers of
flax, hemp, ramie, and the like, the pulps from annual plants are high in
pentosans and hemicelluloses. The fibers "hydrate," that is, become more
resistant to drainage of water, on mechanical treatment more easily than
wood pulps or rag fibers; in general, this property limits their use to papers
in which the amount of mechanical treatment given the pulps is compara-
tively mild. The retardation of free drainage on mechanical treatment (in
some instances even pumping) interferes with the subsequent bleaching and
washing operations, with the formation and drainage of the sheet on the
fourdrinier wire or cylinder mold, and with the removal of the water from
the sheet on the presses. Consequently, weak brittle paper often results,
or the rate of production of the paper may be retarded. With due con-
sideration of these properties, however, annual plant fibers may contribute
to the quality of the product or reduce the cost of fabrication. They are
not suitable, however, as general-purpose pulps and should be used only
in papers where they contribute definitely desired characteristics.
Cereal straws have been used as a source of cellulose fiber for paper and
other products throughout the Christian era. The feasibility of their use
depends entirely upon economic factors. Recent improvements in pulp
purification will probably extend the use of straw, particularly for paper
products to which its fiber will impart the superior sheet-forming proper-
ties. The same may be said of grasses such as bamboo, bagasse, and corn-
stalks. Utilization of the nonfibrous cells as a stiffening agent in ordinary
paperboard further contributes to the feasibility of the use of annual
plants as a source of cellulose fiber. Of the agricultural fibrous residues
occurring each year, probably cornstalks comprise the largest single item
in the United States; the annual volume amounts to probably over 40
million tons. Much study has been applied to the utilization of corn-
stalks, and several industrial ventures have been attempted. The per-
fection of systems of good roads in the corn belt has greatly contributed to
the feasibility of utilization. The possibility of employment during the
winter months is always attractive except during war periods when labor
is unusually scarce. The storage of cornstalks has been found to be
feasible through periods of several years duration at Ames and Dubuque,
Iowa, and Danville, Illinois; the losses have been no more and to some
588 CELLULOSE
extent less than in the storage of baled straw. All that seems necessary
for solving the problem of utilizing cornstalks is the assurance of a depend-
able outlet to justify the attention of agricultural engineers and manufac-
turers of farm equipment. With the integration of paper and board manu-
facture, the assurance of a supply of baled shredded cornstalks, and an out-
let when the supply is assured, there seems to be no valid reason why much
of this fibrous residue annually occurring cannot be profitably utilized.
CHAPTER VII
BLEACHING AND PURIFICATION OF WOOD
CELLULOSE
R. S. HATCH
The bleaching of wood cellulose, frorn one point of view, represents a
continuation of the pulping process. The objective of pulping is the re-
moval of the maximum amount of noncellulosic constituents in wood
(lignins, fats, waxes, tannins, watcr-extractable material, and carbo-
hydrates related to cellulose in nature) through the use of relatively low-
priced chemical reagents, with a maximum yield of crude cellulose. At-
tempts to remove all the noncellulosic material by such methods result in
severe degradation of the cellulose itself and subsequent loss of yield.
The bleaching processes applied to this crude wood cellulose effect further
purification under relatively mild, controllable conditions. A more ob-
vious function of bleaching, as the name implies, is the actual whitening of
the pulp. In some instances, this is simply a manifestation of the particular
state of purity achieved. In other cases, the whitening action involves
destruction of colored contaminants without appreciably raising the degree
of refinement.
A. GENERAL PRINCIPLES
The techniques employed in bleaching of wood cellulose are dependent
on the nature of the pulp being processed and on the anticipated end use.1
Processing details fall into one of three general categories in accordance
with the following objectives *
1. For certain types of pulp such as groundwood or semichemical pulps,
the objective is the partial removal, or alteration, of the colored noncellu-
losic residues to provide a product of satisfactory brightness or "whiteness"
1 For an excellent discussion of the bleaching of pulps, see also J. P. Casey, Pulp and
Paper, Vol. I, Interscience, New York-London, 1952, Chapter V; F. Kraft in J. N.
Stephenson, editor, Preparation & Treatment of Wood Pulp (Pulp & Paper Manufac-
ture, Vol. I), McGraw-Hill, New York, 1950, Chapter 7.
589
590 CELLULOSE
to serve as a background for printing and illustrations, without materially
reducing the yield of these pulps with respect to the original wood.
2. The bleaching of chemical pulps, commonly designated as sulfite,
sulfate, and soda pulps involves substantially complete removal of non-
cellulosic impurities as well as the production of a finished product having
a satisfactory degree of brightness. The extent of brightness desired is
dependent on end use.
3. Pulps intended for the manufacture of cellulose derivatives are fur-
ther purified during the bleaching operations through the removal of car-
bohydrates (such as pentosans, hexosans, and uronic acids) which normally
accompany the pure cellulose as it exists in wood.
It is possible by specific oxidation or reduction treatments to bleach the
colored noncellulosic contaminants (chiefly lignin) in a crude wood cellulose
furnish. Such techniques are ordinarily applied to groundwood or semi-
chemical pulps (process category 1). The removal of lignin can be ac-
complished by more severe oxidative methods which result in profound
degradation of the aromatic system. At the same time there occurs con-
siderable oxidative damage to other noncellulqsics and to cellulose itself.
A more selective and less expensive method for removing the bulk of the
lignin is by means of chlorination, following which the chlorinated lignin
can be washed out by appropriate methods. Further purification and
simultaneous whitening of the pulp is normally effected by controlled oxi-
dative treatments. Alkaline extraction processes are also applied if a
further reduction in the level of noncellulosic carbohydrates is desired.
The process involving chlorination and mild oxidative bleaching with or
without alkaline extraction is customarily applied to the so-called chemical
pulps (process category 2). The thorough refining of chemical pulp for
cellulose derivative applications involves the use of more drastic alkaline
extractions and preferably multistage oxidative bleaches (process category
3).
B. SPECIFIC BLEACHING TREATMENT
1. Chlorination
The first chemical wood pulp prepared in this country was made by the
soda process which was brought here from England by Watt and Burgess
in 1854. These pioneers proposed to use chlorination in the bleaching of
soda pulp and obtained a patent2 for a bleaching process, the first step of
which was chlorination. The process was not a commercial success because
1 C. Watt and H. Burgess, U. S. Patent 11,343 (July 18, 1854).
VII. BLEACHING AND PURIFICATION 591
of the high cost of elemental chlorine at that time and because of the lack
of suitable acid-resisting equipment for commercial operation. The
use of elemental chlorine as a step in the bleaching process was not con-
sidered again until the end of World War I when liquefied chlorine became
a cheap and readily available article of commerce.
For the bleaching of sulfate pulp in which chlorination was the first step,
de Vains3 obtained a German patent in 1913 and a IT. S. patent in 1914.
The de Vains patent was followed by a patent to Drewsen4 in 1918 in which
a very similar process was used. Cataldi5 was granted a patent in 1916
for the use of chlorination in the bleaching of straw and other lignin-con-
taining pulps.
In considering the action of chlorine on crude wood cellulose, there will
be frequent occasions to use the term "bleachability." This represents a
measure of the amount of chlorine in the form of an oxidizing agent which
a given sulfite pulp will consume under certain standard conditions. Many
different methods have been suggested for this purpose, but the one most
used in this country is the TAPPI permanganate test.6 The test measures
the number of milliliters of 0.1 N KMnO4 consumed in acid solution by one
oven-dry gram of pulp under standard conditions of concentration, time,
and temperature. Multiplication of this value by 0.355/x gives the so-
called "chlorine bleachability," which is a rough approximation of the chlo-
rine in the form of hypochlorite which would be consumed by 100 grams of
pulp when bleached to a standard brightness.7 The factor x varies with
permanganate number between about 0.8 and 0.5 for the usual sulfite
pulps. The method has been extended to sulfate pulps for which the factor
x is generally 0.9 or higher. A recent article by Alander8 illustrates con-
cisely the relationships between the various bleachability numbers.
When chlorine is dissolved in water, the following equilibrium exists :
C12 + H2O , HOC1 + HC1 (1)
3 A. R. de Vains and J. F. T. Peterson, German Patent 283,006 (Feb. 26, 1913);
A. R. de Vains, U. S. Patent 1,106,994 (Aug. 11, 1914).
4 V. Drewsen (to West Virginia Pulp & Paper Co.), U. S. Patent 1,283,113 (Oct. 29,
1918); Chem. Abstracts, 13, 187 (1919).
6 B. Cataldi, Brit. Patent 101,475 (Sept. 11, 1916); Chem. Abstracts, 11, 209 (1917);
French Patent 482,222 (1916).
8 Tech. Assoc. Pulp Paper Ind., Standards, T 214 m-50.
7 The term brightness is a measure of the light reflected from the surface of a sheet
of pulp at a wavelength of approximately 457 millimicrons. This reflectance is measured
with a spectrophotometer and is referred to the reflectance at the same wavelength of a
surface of pure magnesium oxide (see Chapter XII-A-3-f).
8 P. Alander, Finnish Paper Timber J.t 33, No. 6, 201 (1951).
592 CELLULOSE
It is obvious that the composition of this equilibrium mixture, and therefore
its effect on organic matter, is highly dependent on pH. Thus, under very
acidic conditions, pH of 2 or less, the solution contains chiefly dissolved
undissociated chlorine. Although some oxidation can occur under these
conditions because of the presence of some hypochlorous acid, the environ-
ment is conducive to chlorination as can be expressed by the equation (R
designates the organic residue) :
RH + C12 > RC1 + HC1 (2)
As the acidity of the chlorine-water system is decreased, the formation of
hypochlorous acid is favored with resultant increase in the rate of oxidative
attack as represented below :
RH + C12 + H2O > ROM + 2HC1 (3)
RCH3 + 3C12 + 2H2O > RCOOH + 6HC1 (4)
Increasing dissociation of hypochlorous acid to hypochlorite ion follows at
higher pHs (above 5), and the rate of oxidation is thereby decreased since
the latter has a somewhat lower oxidation potential than does HOC1.
In alkaline environments, the reaction is slow but entirely oxidative.
Although there may be a brief preliminary oxidative attack of aqueous
chlorine on cellulose and noncellulosic materials in a typical chlorination
process (equations 3 and 4), the pH of the slurry quickly drops because of
the hydrochloric acid liberated, and substitution of lignin hydrogens by
chlorine (equation 2) then becomes the predominant reaction. It should
be observed that the chlorination reaction maintains a high level of acidity
because of simultaneous hydrochloric acid formation. When the lignin
in the wood pulp is saturated with chlorine, a slow oxidation reaction
then proceeds until all the chlorine is consumed. According to equation 2,
one-half of the chlorine added should appear as hydrochloric acid when true
chlorination is complete, but in the case of equations 3 and 4 all the chlorine
added will appear as hydrochloric acid.
The course of a chlorination is illustrated by the curve in Figure 1.
A sample of unbleached sulfite pulp having a test bleachability of 4.9%
chlorine was suspended in water and thoroughly agitated. Chlorine gas
was bubbled in at a fixed rate and samples were withdrawn at the points
indicated on the curve. The ordinates on this plot represent the test
bleachability after washing the samples withdrawn ; the abscissas show the
percentage of the chlorine consumed, based on the original test bleach-
ability. This curve is a straight line to a point representing 30% of the
test bleachability after which the slope begins to change, and at point A,
VII. BLEACHING AND PURIFICATION
593
which represents 40% of the test bleachability, the slope changes sharply
and the curve again assumes a straightline character. Thus it would ap-
pear that true chlorination or substitution is dominant approximately as
far as point A on the curve but from this point on oxidation as shown in
equations 3 and 4 dominates. The ratio of hydrochloric acid formed to
chlorine added will vary somewhat depending on how long the process is
3
0
5
U .
10 20 30 40 50
PER CENT CHLORINE
BASED ON BLEACHABILITY
oO
Fig. 1. Residual bleachability of a sulfite pulp versus the chlo-
rine requirements for single-stage bleaching. True chlorination
reaction is approximately complete at point A. From this point
on, oxidation begins to dominate.
allowed to continue after point A is reached. In any case, at or shortly
after the end of chlorination one would expect on the basis of stoichiometry
that the ratio would be nearer to 1 :2 than to 1 : 1. As confirmation of this
view regarding the nature of the process, a number of investigators9""12
9 E. Heuser and R. Sieber, Z. angew. Chem., 26, 801 (1913).
10 P. Waentig, Papier- Fabr., 25, Tech.-wiss. XL, 340 (1927).
11 L. Rys, Papier -Fabr. , 26, Xech.-wiss. XL, 256 (1928).
18 O. Kress and E. H. Voigtman, Paper Trade J.t 97, 29 (Aug. 17, 1933).
594
CELLULOSE
have found that between 50 and 60% of the chlorine added finally appears
as hydrochloric acid.
The actual practice of pulp chlorination is illustrated by the following
account which specifically concerns sulfite pulp. Experiments have shown
that if a sulfite pulp is chlorinated to point A in Figure 1, and then the prod-
ucts of chlorination are washed out and the washed pulp is bleached with
250
TIME IN HOURS
Fig. 2. Viscosities of sulfite pulps treated with chlorine and with
hypochlorite. The bleachability of the pulp was 5.3%; the chlorine
used was 3.45% in each case. A, chlorination; B, hypochlorite
bleaching.
hypochlorite, the minimum amount of total chlorine will be required to
bleach the pulp to a standard brightness. With sulfite pulp it is customary,
however, to chlorinate to approximately 65% of the test bleachability
and wash before the final bleach. This practice is followed because any
additional degradation which may be involved in chlorinating to 65%
rather than to 40% of the test bleachability is relatively small, and the over-
all cost of bleaching is less because chlorine in the form of hypochlorite
costs about 50% more than elemental chlorine.
The curves in Figure 2 illustrate the advantage, as far as viscosity re-
tention is concerned, of chlorination over hypochlorite oxidation as the
first operation in the bleaching process. The illustration involves two
samples of the same pulp, one treated with elemental chlorine in an amount
equal to 65% of the test bleachability, and the other with hypochlorite.
It should be noted that in the case of chlorination (curve A), the viscosity
of the chlorinated pulp has become nearly constant at the end of 1 hr.,
VII. BLEACHING AND PURIFICATION
595
whereas the viscosity of the hypochlorite-treated pulp (curve B) is not only
lower at the end of 1 hr. but is decreasing rapidly.
01234
TIME IN HOURS
Fig. 3. Comparison of action of chlorine and hypochlorite on sulfite pulp. The
points on these two curves were determined on pulp samples taken at the same
points indicated on the curves in Figure 2. A, chlorination; B, hypochlorite
bleaching.
TIME IN HOURS
Fig. 4. Residual bleachability of sulfite pulp. The points on these two curves
were determined on pulp samples taken at the same points indicated on the curves
in Figure 2. A, chlorination ; B, hypochlorite bleaching.
The curves in Figures 3 and 4 demonstrate other technical and economic
advantages achieved by proper chlorination in the first bleaching stage.
596 CELLULOSE
Curve A in Figure 3 indicates that at the end of 1 hr. approximately 95%
of the chlorine added has been consumed. If hypochlorite is used instead
of chlorine, only about 65% is consumed at the end of 1 hr. (curve B).
In Figure 4, curve A (chlorination) denotes a residual bleachability of
approximately 0.5% chlorine after 1 hr., whereas curve B (hypochlorite
treatment) shows a residual bleachability of 2.0% at the end of this time.
The chief reason why the chlorination of sulfite pulp brings about such
a great reduction in bleachability is that the chlorinated lignin residue in
sulfite pulp is soluble in the acid solution resulting from chlorination.
Thus, these readily oxidized materials are mostly removed and do not con-
sume large quantities of reagent in the hypochlorite bleaching stage.
With sulfate and soda pulps these considerations do not apply. The
lignin residue of these chlorinated pulps is almost entirely insoluble in the
acid solution and must be removed by alkaline extraction. If a sulfate
pulp having a test bleachability of 6.0% chlorine is chlorinated with 3.9%
chlorine (65% of the test bleachability), the chlorinated pulp after washing
will still have a residual bleachability of nearly 3.0% chlorine.
Sulfate pulps made from coniferous woods are generally chlorinated to
about 65% of the test bleachability. Exceptions are made in the case of
some sulfate pulps prepared from hardwoods where it has been found that
chlorination cannot be carried beyond approximately 40% of the test
bleachability without severe degradation. In Figure 1 it has been shown
that oxidation under acid conditions begins to dominate after about 40%
of the chlorine, based on the test bleachability, has been consumed. It
appears that the thin-walled vessels prevailing in certain hardwoods are
more easily degraded by oxidation in acid solution than are fibers from
coniferous woods.
Dodson18 obtained a patent in 1949 which is said to minimize greatly
the drop in viscosity which occurs when pulps are chlorinated under acid
conditions. According to this patent, the addition of small amounts of
amines, substituted amines, and salts of the amines, as well as chloramines,
suppresses the drop in the viscosity of pulps undergoing chlorination. An
examination of curve A in Figure 2 shows that sulfite pulp undergoing chlo-
rination suffered a viscosity drop of about 25% in the course of 1 hr. Ac-
cording to the Dodson patent much of this viscosity drop will be prevented
by the addition of amines or amine derivatives amounting to from 0.03
to 6.0% of the chlorine used.
Although chlorination is considered a definite part of the bleaching treat-
18 A. M. Dodson (to Hercules Powder Co.), U. S. Patent 2,478,379 (Aug. 9, 1949);
Chem. Abstracts, 43, 9447 (1949).
VII. BLEACHING AND PURIFICATION 597
ment, it should be understood that under normal conditions chlorination
is in no sense a bleaching process but rather an extension of the pulping
process, lie object of which is to render the lignin remaining in the fiber
soluble either in acid or alkaline solutions.
2. Alkaline Extraction
In the discussion of the chlorination of sulfite pulp, it was stated that
most of the chlorinated sulfite lignin was readily soluble in the acid chlorinat-
ing solution. However, the lignin and coloring matter in the medullary
ray fiber is not soluble and, though it represents only a small portion (5 to
10%) of the total fiber, it has been found that a mild alkaline extraction
following chlorination will remove at least a portion of this material. By
washing the alkaline-extracted pulp prior to final hypochlorite bleaching,
improved brightness may be realized. The alkaline extraction of sulfite
pulps intended for paper manufacture is carried out at either high or low
consistency,14 at prevailing temperatures, and under very mild alkaline
conditions. Usually 10 to 15 Ib. of caustic per ton of pulp is used for this
purpose. For the manufacture of sulfite pulps for cellulose derivatives,
more drastic caustic extraction is employed as will be discussed later (see
Section D-3 of this Chapter VII) .
Since the chlorinated lignin residues in sulfate and soda pulps are not
soluble in the acid chlorinating solution, it is necessary to remove as much
of this material as possible by alkaline extraction to avoid high concen-
trations of oxidants in the final bleaching. Cold alkaline extraction is of
little value for the maximum removal of the chlorinated lignins of sulfate
and soda pulps. The general practice is to extract the chlorinated pulp
at a temperature of 65° to 70°C. and at a consistency of 10 to 15% for a
period of about an hour. The amount of caustic soda is usually 20 to 40 Ib.
per ton of pulp. Hot alkaline extraction will, after washing, reduce the
bleachability of the extracted pulp enough to permit hypochlorite bleach-
ing in one or two stages without excessive degradation.
3. Hypochlorite Bleaching
The chlorination of sulfite pulp and the chlorination plus alkaline ex-
traction of sulfate and soda pulps result in the removal of most of the non~
cellulosic incrustants which tend to protect the cellulose against oxidative
degradation. However, chlorinated pulp is still colored and must be
14 The consistency of a pulp is defined as the percentage of pulp solids in a measured
volume of suspension.
598 CELLULOSE
submitted to some type of oxidation to destroy the color adsorbed on the
fiber. For the purpose of destroying this color, many oxidizing agents
have been tried, but those which have been adopted commercially are
limited by their specific action and their price. The cheapest of all oxidiz-
ing agents are the hypochlorites in the form of sodium or calcium hypo-
chlorite, and these are the most widely used in the bleaching of pulp. For
special qualities in the finished product, the oxidizing agents sodium per-
oxide, hydrogen peroxide, sodium chlorite, and chlorine dioxide have been
used in increasing amounts in recent years. These special bleaching agents
will be considered after the use of the hypochlorites is discussed.
The role of different oxidizing agents in the degradation of cellulose has
been covered in Section C of Chapter III. There it is shown that the deg-
radation encountered in hypochlorite bleaching is largely influenced by
the pH and temperature of the treatment. The action of hypochlorite
at pH 7 brings about a sharp reduction in viscosity and a corresponding in-
crease in the copper number. The reagent becomes less degradative with
increasing pH. Under any conditions, hypochlorite bleaching causes some
carbonyl and carboxyl group formation in the cellulose molecule. The
end products in exhaustive hypochlorite bleaching are carbon dioxide and
low molecular weight organic acids.15 It is evident from the foregoing
that careful control of bleaching variables is necessary for the production
of bleached pulps of maximum physical and chemical quality.
The final bleaching of sulfite pulp is usually done in one hypochlorite
stage, especially if the chlorinated pulp has been given a mild, cold alkaline
extraction after chlorination. The consistency is usually held at 14 to
16% and the temperature at 30° to 35°C. If no alkali is present, the pH
will droy rapidly to a point where the cellulose will suffer marked degrada-
tion (see Section B-l of this Chapter VII). To avoid degradation it is
customary to add enough caustic soda at the start of the bleaching to keep
the pH not less than 8.0 during the entire bleaching cycle. An excess of
alkali over that necessary to neutralize the acids formed will act to slow
down the bleaching action. To avoid this, various buffering agents such
as magnesium oxide or alkaline phosphates have been suggested but caustic
soda is most generally employed. A slight excess of hypochlorite over that
necessary to complete the bleaching is always used because complete
exhaustion of the bleach will cause color reversion.
In the case of sulfate and soda pulps, chlorination fails to reduce the
bleachability as much as is the case with sulfite pulp. Furthermore, even
11 H. Rashback and F. H. Yorston, Quart. Rev.t Forest Products Laboratory of Can-
ada, No. 7, 12 (July-Sept., 1931).
VII. BLEACHING AND PURIFICATION 599
after hot alkaline extraction, the bleachability is still considerably higher
than that for chlorinated sulfite pulp. It is also true that pulps produced
by alkaline cooking processes are more easily degraded by oxidizing agents
especially if the concentration of the oxidizing agent is high. To avoid
high concentration of oxidant, it is customary to divide the bleaching of
sulf ate and soda pulps into two or more stages with washing between stages.
Some mills subject the pulp to a mild alkaline extraction between stages.
The temperature, pH, consistency, and excess bleach at the end of the proc-
ess are the same as in the bleaching of sulfite pulps. It has also been
found that treatment of the pulp after the final bleaching stage with a
solution of sulfur dioxide or other acid to a pH of 5 to 6 will increase bright-
ness and prevent later color reversion. Furthermore, the acidification of
the bleached pulp with sulfur dioxide reduces ferric iron to the ferrous
state and lowers the ash content of the pulp.
Although chlorination, hot caustic extraction, and final hypochlorite
bleaching yield pulps of satisfactory brightness from most wood species,
there are pulps made from some wood species (specifically Douglas fir)
which are very difficult to bleach to satisfactory brightness because of the
dark color of the pulp after caustic extraction. A modification of the nor-
mal process has been found to give very satisfactory results.
The pulp is first chlorinated in the normal manner. Following this it is
washed and neutralized at low consistency to give a pH slightly over 7,
and then is washed and thickened to a consistency of 10 to 12%. A cal-
culated amount of hypochlorite, equivalent to approximately 80% of the
test bleachability of the chlorinated and neutralized pulp, is added together
with enough caustic to keep the pH above 8 during the bleaching cycle
which, at a temperature of 30°C., is from 1 to 1.5 hrs. The partially
bleached pulp is then washed and extracted with 1.5 to 2.0% caustic,
based on the pulp, at 65° to 70°C. after which it is washed and given a final
bleach in one or two stages. The partial hypochlorite bleach after chlorina-
tion gives a pulp, after caustic extraction, of a light buff color instead of a
dark brown color, and the final bleaching may be much less drastic.
A similar modification is proposed by Rue and Sconce.16 According to
their method, an excess of chlorine is added to the unbleached pulp and
allowed to react for only a short time (5 to 10 min.) after which lime slurry
is added to neutralize the hydrochloric acid formed and to convert the ex-
cess chlorine to hypochlorite. The hypochlorite is allowed to exhaust in a
retention tower or tank, then the pulp is washed, extracted hot with caustic,
16 J. D. Rue and J. S. Sconce, Tech. Assoc. Papers, 16, 503 (1933).
600
CELLULOSE
and bleached. This process is in use in some mills and is said to give satis-
factory results. It should be noted, however, that the hydrochloric acid
formed during the chlorination must be neutralized with lime whereas, if
the correct amount of chlorine is used for chlorination and the pulp is
washed in the acid condition, no alkali is consumed for neutralization.
Furthermore, lime salts of chlorinated lignin are much less soluble than
sodium salts and are more difficult to bleach.
4. Chlorine Dioxide and Sodium Chlorite
During the past ten years, chlorine dioxide and sodium chlorite have
come into extensive use especially in the final bleaching of alkaline-cooked
Fig. 5. Oxidation potentials of different oxidizing agents
used in pulp bleaching as a function of pH.19 Courtesy of
Solvay Process Division.
pulps.17 Sodium chlorite and chlorine dioxide have lower oxidation poten-
tials than hypochlorous acid, and can act as efficient oxidizing agents for
the destruction of color while having a minimum degrading action on cellu-
lose when used under proper conditions of pH and temperature.18 Figure 5
shows the oxidation potentials of sodium hypochlorite, chlorine dioxide,
17 G. Jayme and S. Mo, Papier-Fabr., 39, No. 33, 193 (Aug. 16, 1941).
18 O. Samuelson and C. Ramsel, Svensk Papperstidn., 53, 155 (1950).
VII. BLEACHING AND PURIFICATION 601
sodium chlorite, and sodium peroxide.19 It should be noted that in the
case of chlorite, at no ordinary pH level does the oxidation potential (i.e.,
oxidative driving force) exceed that of a mild (pH 10) hypochlorite bleach.
The use of chlorite under mild acidic conditions permits a selective attack
on readily oxidized colored noncellulosic contaminants without damage to
the cellulose structure itself. Chlorine dioxide is somewhat less selective
but, as the potential curves show, it is considerably safer than hypochlorite
in the acid range.
Sodium chlorite was first produced commercially in this country by pass-
ing chlorine dioxide into a caustic soda solution in the presence of a suitable
reducing agent 20 The product was sold originally under the name "Tex-
tone" and was recommended for the bleaching of cotton textiles to high
brightness values. Later the name was changed to "C2" and the product
was recommended for the final bleaching of pulps which were difficult to
bleach to high brightness without the severe degradation resulting when
hypochlorite was used.21
Sodium chlorite is a yellowish, readily soluble salt which may be stored
and shipped without danger22 provided it is kept out of contact with
organic material. When acidified, sodium chlorite yields chlorous acid
and chlorine dioxide which are the active bleaching agents. If only small
quantities of chlorine dioxide are needed, sodium chlorite offers a conveni-
ent means of storage although it is more expensive than chlorine dioxide
generated as it is used.
The most rapid development of processes for the generation of chlorine
dioxide on a commercial scale occurred in Sweden.23 Sodium chlorate
in acid solution is reduced by diluted sulfur dioxide gas, and the chlorine
dioxide gas is absorbed in water. The chlorine dioxide solution must be
used promptly to avoid loss of oxidizing power by decomposition. Chlorine
dioxide obtained by this process contains some chlorine, the ratio of chlorine
to chlorine dioxide being about 1 to 20. A recent synthesis involves reduc-
19 The Oxidation Potentials of Common Bleaching Agents, Research Report 1949,
Alkali Section, Solvay Process Division, Allied Chemical & Dye Corp., Syracuse, N. Y.,
1949.
20 G. P. Vincent, Chem. Inds., 47, No. 3, 280 (Sept., 1940); G. P. Vincent (to Mathie-
son Alkali Works), U. S. Patents 2,092,944-5 (Sept. 14, 1937); Chem. Abstracts, 31,
8134 (1937).
21 G. P. Vincent, Mfrs.' Record, 109,' No. 12, 34 (1940).
22 M. C. Taylor, J. F. White, G. P. Vincent, and G. L. Cunningham, Ind. Eng. Chetn.,
32, 899 (1940); J. F. White, M. C. Taylor, and G. P. Vincent, Ind. Eng. Chem., 34, 782
(1942)
28 G. Hoist, Svensk Papperstidn., 50, 472 (1947).
602 CELLULOSE
tion of an acid solution of sodium chlorate with methanol vapor. This
process is said to yield chlorine dioxide free of chlorine.
Chlorine dioxide bleaching is carried out at relatively high consistency
and at temperatures of 50° to 60°C. in closed bleachers to prevent loss of
reagent during the bleaching cycle. Then the bleached pulp is treated
with sulfur dioxide, which destroys excess chlorine dioxide, and is then
washed.
Since the chlorine equivalent of chlorine dioxide costs approximately
four times that of elemental chlorine, the use of chlorine dioxide is largely
limited to final bleaching only where its unique properties make it of special
value in producing high brightness with minimum cellulose degradation.
5. Peroxide Bleaching
Peroxidic materials (hydrogen and sodium peroxides) are of obvious
interest as oxidative bleaching agents. However, because of poor avail-
ability and lack of suitable techniques, peroxides have not until recently
been employed in the chemical purification of wood cellulose. The re-
agents are now used very successfully in the bleaching of groundwood and
semichemical pulps.24 The low oxidation potential for peroxide (see Fig.
5) denotes a high degree of selectivity. It has actually been established
that under ordinary bleaching conditions, very little degradation of
cellulose occurs.25 At the same time, the peroxide potential is sufficiently
high to destroy readily oxidized, colored lignin residues. Thus, although
the relative cost of these reagents is high, this is offset by the fact that they
permit a high degree of whitening of crude wood cellulose furnishes without
serious yield loss.
In actual practice certain precautions must be taken to prevent rapid
peroxide decomposition. In the presence of large amounts of iron and
copper, such as may enter through service water, peroxide decomposition
far exceeds the rate at which the furnish may be oxidized. Normally,
sodium silicate and magnesium sulfate are added to inhibit this catalyzed
decomposition.26 The additives incidentally have a desirable buffering
action.
Although the chief use of peroxidic bleaching is in the instances referred
to above, it is known that the technique can be applied as a finishing treat-
24 J. P. Casey, Pulp and Paper, Vol. I, Interscience, New York-London, 1952, pp.
296-302.
M H. Staudinger and J. Jurisch, Papier-Fabr., 35, Tech. Tl., 459 (1937).
26 J. S. Reichert, D. J. Campbell, and R. T. Mills, Paper Trade /., 118, 45 (Apr. 13,
1944).
VII. BLEACHING AND PURIFICATION 603
ment for previously chlorinated and hypochlorite-bleached chemical pulps.
Thus, peroxides are to some extent interchangeable with chlorites and chlo-
rine dioxide.
6. The Washing Operation
One of the most important operations in the bleaching process is washing
after the different chemical treatments. Soluble residues from any of the
operations, if left in the pulp, will interfere with the steps which follow,
causing increased consumption of chemicals. Pulp is now washed almost
universally on continuous vacuum filters by a combination of dilution
washing and displacement washing. Pulp from any of the bleaching steps
must be diluted to a consistency of 1 to 1.5% before entering the filter.
Modern vacuum filters are capable of discharging a pulp cake having a
consistency of 10 to 20%, depending on whether or not press rolls are used.
Thus, when a pulp slurry of 1% consistency, having the soluble solids evenly
distributed in the aqueous phase, is thickened to a consistency of 10%
in passage over a vacuum filter, nearly 90% of the soluble material will
pass out with the filtrate. Showers are provided to complete the removal
of the soluble solids by displacement washing. The wood cellulose fiber
possesses a complex physical structure which shows strong tendencies to
adsorb or to hold impurities tenaciously in the minute fiber canals; con-
sequently the washing of pulp is not as simple as the theory of washing
would make it appear. Each step in multistage bleaching requires large
volumes of water for washing, so countercurrent techniques are employed
as far as is practical. The most highly contaminated filtrates, such as
those from chlorination and caustic extraction, are sent to the sewer.
Fresh water is used for showers on the filters handling bleached pulp, and
the effluent from these filters is used for dilution on the next filters in line.
By utilizing countercurrent washing, it is possible to accomplish efficient
purification in multistage bleach plants with the use of 30,000 to 40,000 gal.
of water per ton of pulp bleached. Water for use in the bleach plant must
be very low in color, of high clarity, and as free as possible from metal ions,
especially iron, copper, and manganese, which, if present, will act as cata-
lysts to degrade the cellulose in the bleaching steps.
C. EQUIPMENT FOR COMMERCIAL CHLORINATION,
BLEACHING, AND ALKALINE EXTRACTION
Equipment designed for the chlorination of pulp must be strictly acid
resistant since the pH of the pulp slurry, shortly after chlorination is started,
604
CELLULOSE
drops rapidly to a value between 1 and 2. For alkaline extraction, vessels
of plain steel or tile-lined concrete are quite satisfactory. Equipment for
hypochlorite bleaching is usually constructed of tile-lined concrete or steel
and need not be highly resistant to acid because the bleaching action is
normally carried out at a pH greater than 7.
1. Chlorination Equipment
Chlorination at the customary 3 to 4% consistency level may be per-
formed in batch or continuous chlorinators. As pointed out previously,
Fig. 6. Type VL batch chlorinator: (A) tank; (B) central
draft tube; (C) constricted throat of draft tube; (D) propeller.
This chlorinator has been widely adopted in this country for the
Chlorination of both sulfite and sulfate pulps. Courtesy of Pulp
Bleaching Co., Seattle, Wash.
the action of chlorine on either sulfite or sulfate pulp is rapid, and no at-
tempt is made to control temperature. As temperature is increased, the
reaction rate increases, but it is necessary only to provide equipment suffici-
ently large to accomplish the exhaustion of the chlorine in a reasonable
time at water temperatures attained during the cold season.
VII. BLEACHING AND PURIFICATION
605
In simple batch chlorination, a known weight of pulp is introduced into
a suitable vessel, and a weighed or metered amount of chlorine is bubbled
into the slurry over a relatively short period of time. Following this, the
chlorine is allowed to react until it is exhausted, and the chlorinator con-
tents are then pumped over an acidproof washer. Figure 6 is an elevation
of a commonly used batch chlorinator. This chlorinator consists of a
Fig. 7. Kamyr batch chlorinator: (A) tank; (B) external pipe for return
of pulp; (C) inlet for chlorine gas or chlorine water; (D) impeller. Courtesy
of Sandy Hill Iron & Brass Works, Hudson Falls, N. Y.
tile-lined concrete or rubber-covered steel tank A and a central draft tube B
with a constricted throat C. A propeller D is located in the throat of the
draft tube and is designed to circulate the pulp slurry downward through
the draft tube and upward around the outside. Chlorine gas from a suit-
able liquid chlorine evaporator is piped to a point directly above the intake
of the impeller or a solution of chlorine in water may be piped to the same
point. Chlorine gas may be measured with a flowmeter or it may be de-
livered from a weighing tank. The chlorinator is filled with the unbleached
pulp slurry of known consistency and bleachability. The amount of chlo-
606
CELLULOSE
rine calculated from the test bleachability is then run in as rapidly as it will
be absorbed. At the end of about 1 hr. the chlorinated pulp is pumped
to an acidproof vacuum washer and washed free of acid.
The Kamyr batch chlorinator is shown in Figure 7. This chlorinator
depends upon a pump instead of an impeller for circulation. The pump
Fig. 8. Continuous chlorinator: (A) point at which pulp
suspension mixed with chlorine is introduced; (B) point at
which the chlorinated pulp suspension is discharged. Cour-
tesy of Pulp Bleaching Co., Seattle, Wash.
circulates the pulp upward through the tank A and the return is through
an external pipe B. Chlorine gas or chlorine water is injected at C.
The present trend in chlorination of both sulfite and sulf ate pulp is toward
continuous methods. Continuous chlorinators are usually vertical towers
of sufficient capacity to allow the proper retention time for the completion
of chlorination at system rate of flow. They may be constructed of tile-
lined concrete or rubber-covered steel. If good mixing of chlorine and pulp
is assured, the continuous process is quite satisfactory, and considerable
VII. BLEACHING AND PURIFICATION
607
building space as well as power input may be saved. One form of continu-
ous chlorinator is shown in Figure 8. Good mixing of the chlorine with
the pulp is secured by injecting the chlorine into the suction of the pump A
which delivers the pulp to the bottom of the tower. Intermittent agita-
tion may be supplied by agitating arms located at different levels in the
tower. The chlorinated pulp is discharged at the top of the tower through
B to a suitable acidproof washer.
Fig. 9. Kamyr continuous chlorinator: (A) flow
box for introduction of pulp; (B) inlet for chlorine;
(C) channel for mixing. Courtesy of Sandy Hill
Iron & Brass Works, Hudson Falls, N. Y.
The Kamyr continuous chlorinator is shown in Figure 9. This chlorin-
ator has a special agitating zone located in the base of the tower to secure
good mixing of chlorine and pulp. Chlorine is injected into the discharge
of the pump which feeds the tower, and the mixture of pulp and chlorine
is thoroughly agitated in the channel C before rising through the tower
and discharging to an acidproof washer.
2. Bleaching Equipment
After either acid- or alkaline-cooked pulps have been chlorinated, they
are washed on acidproof vacuum washers. These washers consist of a
cylinder, usually of rubber-covered cast iron, and a stainless steel wire
covering. The cylinder, with its wire covering, operates either in a rubber-
covered steel vat or an acidproof tile-lined vat. After the pulps have been
washed, they are caustic-extracted in the case of alkaline-cooked pulps
(see topic 3 below) or, in the case of acid-cooked pulps, they may be
608
CELLULOvSE
bleached immediately or given a very mild cold alkaline treatment. In
either case, pulps are usually washed again after alkaline treatment. The
washing after alkaline treatment is done on vacuum washers which need
not be acidproof .
The hypochlorite bleaching stage (or stages) is carried out at a relatively
high consistency (usually 16-18%) and as before by a batch or continuous
OPERATING FLOOR
Fig. 10. Type VW batch bleacher with dilution chest: (A)
plow designed to sweep close to bottom of tank; (B) screw con-
veyor. Courtesy of Pulp Bleaching Co., Seattle, Wash.
process. One of the widely used batch bleachers is the so-called Type VW,
as illustrated in Figure 10. This bleacher was originally developed and
patended by Fletcher.27 The bleacher consists of a cylindrical tank with a
hemispherical bottom constructed of concrete or steel and tile lined. A
central shaft, on which is mounted a heavy screw conveyor B and a plow
A designed to sweep close to the bottom of the tank, is the means of cir-
culation. Openings are provided for charging the bleacher with pulp and
chemicals and discharging the bleached pulp into a dilution chest. This
27 P. K. Fletcher, U. S. Patent 1,466,499 (Aug. 28, 1923); Chem. Abstracts, 17, 3790
(1923).
VII. BLEACHING AND PURIFICATION 609
bleacher is capable of circulating 6 to 8 tons of pulp at 16 to 18% consist-
ency in the following manner. The plow A forces the pulp at the bottom
of the tank into the screw conveyor which elevates it to the top of the tank
where it discharges and returns to the bottom by gravity. Pulp at this
consistency loses fluid when mechanical force is applied, thus increasing the
solids ratio. Liquid is squeezed out of the pulp as it is gathered by the
plow and forced into the screw conveyor. Then, when the pulp is dis-
charged at the top of the conveyor, it absorbs fluid from the surrounding
mass and reaches an equilibrium moisture content as it descends again to
the plow. Thus, it will be seen that there is a constant turnover of the
fluid in contact with the pulp as the circulation proceeds. Moreover, the
plow constantly removes a cross section of the pulp as it returns by gravity
to the bottom. The net result is an intimate mixing of bleaching fluid
with the fiber, with internal friction tending to flex and open up the fiber
bundles and expose them to the bleaching action. This method of process-
ing provides a cleaner bleached pulp. The bleachers described operate on a
batch cycle of 2 to 4 hrs. and many of them are installed as part of modern
multistage bleaching systems.
For batch processing, the bleachability of the chlorinated and washed
pulp is determined. The bleacher is filled with a definite tonnage, and
bleach liquor, slightly in excess of that required as determined by the
bleachability, is added together with sufficient caustic and hot water or
steam to complete the bleaching action at a pH not less than 8 and at a
temperature of 30° to 35°C.
Shortly after the introduction of the batch bleacher just described,
Thorne28 proposed a continuous unit for high-consistency bleaching. Figure
11 is an elevation of a modern Thorne-type bleacher. This bleacher con-
sists of a tower A built of tile-lined concrete or rubber-lined steel. A
double-shaft mixer B, in which pulp, bleach solution, and hot water or
steam are intimately blended feeds into the top of tower A, and a scraper
C feeds the bleached pulp into a twin discharge screw D and a dilution
:hest E. The tower A is designed to give sufficient retention time for
completion of the bleaching reaction at system rate of flow. After the
tower is filled from the mixer B, the scraper and twin discharge devices
are started and bleaching proceeds continuously. In a continuous bleacher
}f this type, no provisions are made for intermittent mixing as in the case
Df the batch units; consequently uniform processing depends on the inti-
mate mixing of pulp and bleaching solution as they enter the tower.
The Kamyr Machine Works of Sweden has designed a continuous,
28 C. B. Thorne, U. S. Patent 1,656,765 (Jan. 17, 1928).
610
CELLULOSE
high-consistency bleacher which does away with the dilution chest and
also provides a zone where intimate mixing at low consistency is accom-
plished before the treated pulp is discharged. An elevation of this bleacher
is shown in Figure 12. Pulp, steam, and bleaching chemicals pass into
the tower A through the double-shaft mixer B at consistencies of 18 to 16%.
Fig. 11. Thome-type continuous bleacher: (A) tower; (B) double-
shaft mixer; (C) scraper; (D) twin discharge screw; (E) dilution chest.
Courtesy of Improved Paper Machinery Corp., Nashua, N. H.
When the tower has been filled, dilution water enters the bottom of the
vessel through nozzles C and reduces the consistency to 2 or 3% ; a circu-
lating pump D circulates this low-density slurry around the channel E
wrhile a discharge pump F withdraws the bleached pulp at system rate of
flow. The high-consistency pulp undergoing bleaching floats on the low-
consistency bleached pulp in channel E.
The bleachers described in the preceding paragraphs, in one form or
another, are used as the hypochlorite stage or stages of modern multistage
VII. BLEACHING AND PURIFICATION
611
bleaching systems in which hypochlorite or chlorine dioxide bleaching is
preceded by chlorination.
When solutions of chlorine dioxide are used for bleaching, it is necessary
that all parts of the bleaching equipment be acidproof because bleaching is
carried out at a pH of 4 to 6. Temperatures used in chlorine dioxide bleach-
Fig. 12. Kamyr continuous bleacher: (A) tower; (B) double-shaft mixer;
(C) nozzles; (D) circulating pump; (E) channel; (F) discharge pump; (G)
cylindrical bottom piece with conical top. Percentages represent consistency
of the pulp. Detail of bottom is shown at the left. Courtesy of Sandy Hill
Iron & Brass Works, Hudson Falls, N. Y.
ing are usually 50° to 60°C. All metal parts of bleachers used for bleach-
ing with chlorine dioxide should be covered with a rubber composition cap-
able of withstanding the operating temperatures. Furthermore, it is
desirable to carry out chlorine dioxide bleaching at pressures slightly above
atmospheric because of the relatively poor solubility of chlorine dioxide in
water.
612
CELLULOSE
3. Alkaline Extraction Equipment
The equipment for alkaline extraction is generally of the same type as
used for bleaching. For batch extraction, the Type V W bleacher as shown
in Figure 10 or one of similar design is used. For continuous extraction,
Fig. 13. Kamyr continuuua ^ausi.^ cAu.actor: (A) vacuum washer; (B) mixing
trough; (C) pump; (D) pipe; (E) steam mixer; (F) screw press; (G) tower.
Courtesy of Sandy Hill Iron & Brass Works, Hudson Falls, N. Y.
the tower shown in Figure 11 is used to a large extent. The Kamyr Ma-
chine Works has developed an ingenious extractor for this purpose as shown
in Figure 13.
The operation of this extractor is as follows : Washed pulp from vacuum
washer A is delivered into a mixing trough B, to which both fresh caustic
solution and returned caustic are added in predetermined amounts; then
VII. BLEACHING AND PURIFICATION 613
the pulp is concentrated to give a consistency of about 6%. The pump C
delivers the mixture through pipe D to a steam mixer E and a screw press F
at the bottom of tower G. The screw press thickens the hot mixture to
approximately 20% consistency. This thickened pulp rises through tower
G and is discharged at the top to a washer. The effluent from the screw
press is pumped back to trough B for dilution of the pulp from washer A
to the desired 6% consistency. Fresh caustic is added here to compensate
for the caustic contained in the 20%-consistency pulp rising through
tower G. This design of tower is recommended in cases where it is desir-
able to extract at temperatures above 100°C. This may be done by con-
structing tower G high enough to give a static head of more than one
atmosphere.
D. QUALITY REQUIREMENTS FOR SPECIFIC PRODUCTS
1. Nonpermanent Papers
For most papermaking purposes (i.e., nonpermanent papers), pulps
bleached by the methods outlined in the preceding paragraphs are entirely
satisfactory as raw material. The chief requirement for such pulps is
that they provide papers of adequate physical strength. Minor proportions
of retained noncellulosic materials are of no great concern provided that
their color contribution has been eliminated by simple and inexpensive
bleaching techniques.
2. Permanent Papers
In the so-called permanent papers, the presence of pentosans, hexosans,
and short-chain carbohydrate material is undesirable because such con-
taminants eventually induce discoloration and embrittlement. These
papers have been made, in the past, entirely from cotton and linen rags
(see Chapter VI-C). The rapid development of synthetic fibers and the
increasing use of these fibers in mixture with cotton for a great variety of
textiles have greatly restricted the availability of rags as raw material for
permanent papers. At the present time there is increased use of bleached
cotton linters as well as specially purified wood pulps for this purpose.
The higher degree of sulfite pulp refinement which appears to be necessary
in the permanent paper application can be realized by suitable alkaline
extraction. In the bleaching of sulfite pulp for most papermaking pur-
poses, the alkaline extraction after chlorination, if used at all, is very mild
(e.g., 0.75% alkali based on the pulp at prevailing mill temperature).
614 CELLULOSE
For the production of paper-making high-alpha pulp (for permanent papers),
the caustic soda concentration is usually 1 or 2% and the extraction tem-
perature is increased to 75-1 10°C. The conditions of alkaline extraction
are ordinarily kept sufficiently mild to avoid excessive yield loss and any
adverse effects on beating characteristics. Under certain conditions,
however, yield must be sacrificed in the interests of quality.
The equipment used for hot alkaline extraction is the same as that used
for normal alkaline extraction. If the extraction is carried out by a batch
operation as in a Type V W bleacher, a closed version of this bleacher must
be used for operation at temperatures exceeding 100°C. The Kamyr-
type caustic extractor previously described (Fig. 13) is capable of operating
at temperatures above 100°C. at the bottom of the retention tower because
of the static head.
Many attempts have been made to increase the yield of hot-alkaline-
extracted sulfite pulp by adding reducing agents such as sodium sulfite
or sodium sulfide on the theory that the shrinkage may be due, in part,
to the presence of atmospheric oxygen which will attack the resistant cellu-
lose in the presence of alkali at elevated temperature. The alkaline re-
fining of wood pulps has been extensively studied by many investigators.
Excellent reviews of this subject have been presented by Jayme29 and by
RysandBonish.30
3. Purified Pulp for Cellulose Derivatives
Pulp for the manufacture of cellulose derivatives calls for special tech-
niques in the preparation of the unbleached pulp as well as in the bleaching
and purification process. One of the most important properties of pulp for
cellulose derivatives is reactivity or the ease with which the cellulose under-
goes esterification and etherification reactions. Other important considera-
tions involve viscosity, alpha-cellulose content, reducing groups, carboxyl
content, and the presence of metal ions. These so-called dissolving pulps
are usually cooked to a lower bleachability than is customary for paper-
making pulps. The viscosity of the unbleached pulp is also controlled
within as narrow limits as practical in the cooking process. In the bleach-
ing process, chlorination follows normal practice. Alkaline extraction is
more or less drastic depending on the end use of the pulp. The final vis-
cosity of the bleached pulp depends, first, on the viscosity of the unbleached
pulp and, second, on the conditions of the final bleaching. After chlorina-
n G. Jayme, Paper Trade J., 106, 37 (May 26, 1938).
30 L. Rys and A. Bonish, Paper Trade J.t 108, 31 (May 11, 1939).
VII. BLEACHING AND PURIFICATION 615
tion, the pulp must have a viscosity appreciably higher than that required
in the finished product. The final viscosity is controlled by manipulation
of the bleaching variables : temperature, pH, time, and bleach concentra-
tion. Acidification of the bleached pulp to a pH of 5 to 6 aids in reducing
metal-ion contamination to an acceptable level.
Pulps for nitration or for the manufacture of rayon or cellophane are
usually given a hot alkaline extraction at temperatures not exceeding
100°C. In actual practice, 3 to 4% caustic based on the pulp will provide
in a 100°C. extraction a refined cellulose of the desired alpha-cellulose con-
tent (90 to 94%) at over 90% yield.31 Because of the low caustic proportion
and the fact that extensive dilution occurs in washing, caustic recovery
is not attempted in an extraction process such as specified above.
Sulfite pulps intended for the manufacture of high-tenacity rayon and for
cellulose esters must be purified to an even greater extent. To attain the
required high alpha-cellulose content, usually 94.5 to 96.5%, it is necessary
to increase the amount of caustic in the extraction process. Caustic
proportions of 5 to 12% based on the pulp will provide a wood cellulose of
the above purity in a hot extraction (temperature range, 100° to
120°C.).31*32 However, this caustic concentration at the temperatures
specified generally results in a poor yield (70-80% for purification of 86-
88% alpha pulp).81 On the other hand, the alkali proportion is not high
enough to necessitate recovery measures.
Practically quantitative yields of refined pulp are obtained by cold ex-
traction with higher caustic concentration. Solutions of mercerizing
strength, approximately 17.5% NaOH, will raise the alpha-cellulose con-
tent of chlorinated sulfite pulps to above 97%, 31 but under these conditions
extensive swelling occurs and in subsequent washing and drying, the
fibers collapse and become unreactive. It is therefore necessary to select
a caustic solution of concentration such that swelling of the fibers will not
exceed a certain maximum value, while at the same time the alpha-cellulose
content will be raised to the desired percentage. For example, experi-
ments have shown that a chlorinated sulfite pulp, extracted with an 8%
caustic solution at 30°C., will not have been swollen enough to render it
nonreactive after washing and drying. Because of the sensitivity to oxi-
dation of cellulose in high-strength alkali, the low-temperature caustic
extractions are usually performed in the substantial absence of atmospheric
oxygen. Closed vessels are used and air is displaced beforehand by steam
injection. It is necessary on economic grounds to recover caustic from
81 G. A. Richter, Ind. Eng. Chem.t 33, 1518 (1941).
32 N. W. Coster and R. Vincent. Paper Trade J., 119, 27 (Sept. 21, 1944).
616 CELLULOSE
such treatments, since an 8% concentration for 10% slurry density rep-
resents 1440 Ib. of ingredient per ton of pulp. The caustic washed from
the purified pulp must be evaporated to the proper concentration and re-
used until the organic solids content becomes too great. At this point,
the material is subjected to dialysis or is evaporated and burned to remove
organic contamination. Sulfate white liquor used for cooking in the sul-
fate process may be utilized to raise the alpha-cellulose content of sulfite
pulp to the desired level, and the washings may be sent to a sulfate recovery
system in those localities where a sulfate mill is adjacent to the sulfite mill.
Manufacturers of cellulose derivatives have shown increasing interest
in the use of purified pulps manufactured by the kraft process for conver-
sion into cellulose derivatives. Sulfate pulp cooked by the normal kraft
process, even when drastically purified, is not suitable for the manufacture
of cellulose derivatives because it is not reactive. The sulfate process,
as normally operated, may bring about some type of cross-linkage between
cellulose and residual carbohydrates.
It was discovered in 1931 that if wood was first subjected to an acid
hydrolysis prior to the sulfate cook, the alpha-cellulose content of the re-
sulting pulp was much higher, and the pulp when purified and bleached
was sufficiently reactive for the production of cellulose derivatives.33
Somewhat later, continued research in Germany led to a practical procedure
for this operation.34 During World War II the prehydrolysis process
was operated on an extensive scale in Germany with use of dilute sulfuric
acid, solutions of sulfur dioxide, or water at elevated temperatures. The
solutions of the hydrolyzate separated from the wood were subjected to
fermentation for the production of either alcohol or food yeast.
Prehydrolyzed sulfate pulp is purified by the same general processes as
are employed in the bleaching of normal sulfate pulp. The pulp is first
chlorinated, then washed and extracted with caustic soda at elevated tem-
peratures. When pulps of-high purity are required, the strength of the
caustic used in hot alkaline extraction may be as high as 10 to 15%, based
on the pulp, with temperatures as high as 120°C.
The alkaline-extracted pulp is then bleached in one or more stages either
with hypochlorite alone or with hypochlorite as one stage and chlorine di-
oxide as a final stage. It is then acidified with sulfur dioxide and washed.
This bleaching process will result in a bleached pulp having an alpha-
83 G. A. Richter (to Brown Co.), U. S. Patents 1,787,953-4 (Jan. 6, 1931); Chem.
Abstracts, 25, 816 (1931).
34 G. Sirakoff, Hoh Roh- u. Werkstoff, 4, 205 (1941); through Chem. Abstracts, 38, 2201
(1944).
VII. BLEACHING AND PURIFICATION 617
cellulose content of over 94% and sufficiently reactive for the production
of various cellulose derivatives.
4. Groundwood and Semichemical Pulps
For years groundwood made from eastern wood species such as spruce
and balsam was sufficiently bright when combined with unbleached sulfite
to produce a newsprint of satisfactory brightness. As the newsprint
industry moved westward and wood species such as western hemlock
were used for the production of groundwood, the resulting newsprint was
not as bright as that produced from eastern wood species. Manufacturers
therefore attempted to raise the brightness of western groundwood first
through the use of reducing agents. The first reducing agents tried were
bisulfite solutions. These raised the brightness somewhat but the improve-
ment was only temporary and was not entirely satisfactory. The manu-
facturers then turned to the use of hydrosulfites, chiefly in the form of zinc
hydrosulfite, produced by the action of SO2 on a zinc-dust slurry. This
gave a much greater and more permanent increase in brightness. Zinc
hydrosulfite is now being used in some of the western newsprint mills.
In the meantime, large-scale production of sodium and hydrogen peroxides
has resulted in extensive use of these reagents for the bleaching of ground-
wood. These compounds produce a much higher brightness and one which
is more permanent without materially reducing the yield of groundwood
on the basis of the original wood.
The considerations in the bleaching of semichemical pulps are similar
to those outlined above, in that it is desirable to achieve a whitening action
without serious yield loss. Accordingly, peroxides are used widely in
purification of semichemical pulps. Some work has been done on the
bleaching of such materials with a combination of mild chlorination
followed by peroxide bleaching. Under these conditions the yield suffers to
some extent but higher brightnesses are obtained.
E. NEW TRENDS
1. Use of New Reagents
It is doubtful if any fundamentally new lower cost methods of producing
chemical pulps will be developed for some time. Both the sulfite and alka-
line cooking processes have been advanced to the point where maximum
recovery of the cooking chemicals may be accomplished. Also these chemi-
618 CELLULOSE
cals are low in initial cost. As an additional economy feature, the organic
material dissolved from the wood can be utilized as fuel for production of
steam and electricity. A similar situation prevails with regard to semi-
chemical pulps since in their preparation the same chemicals are used ex-
cept in reduced proportions.
The bleaching process has until comparatively recently been confined
to the use of hypochlorites, which are still the cheapest oxidizing agents
available. During the past decade, the development of successful methods
for production of chlorites and chlorine dioxide has led to increasing use
of these oxidizing agents because of their unique properties of destroying
colored material in highly refined wood cellulose without inflicting severe
degradation. Under the most favorable conditions, the cost of these oxi-
dizing agents is much higher than the cost of hypochlorites. Nevertheless
chlorites and chlorine dioxide have found, and will continue to find, a place
in the industry in instances where exceptional brightness with minimum
degradation is the objective. These two reagents are logically applied in
finishing techniques for pulps which have already been bleached under
mild hypochlorite conditions. It is improbable that these reagents will
be applied in the near future to relatively crude wood celluloses because
of their high cost. It is known that certain mixtures of chlorite and
hypochlorite35 or mixtures of chlorine and chlorine dioxide36 can be used
under slightly alkaline conditions to provide efficient and relatively non-
degradative bleaches. The mechanism of interaction of reagents in the
two instances referred to is not completely understood. Insofar as carefully
controlled dilution of the more expensive reagents with hypochlorite or
chlorine gives about the same effect as use of chlorite or chlorine dioxide
alone, there is an important economic advantage to be gained by these
techniques. Such processes will certainly be used to an increasing extent
in the years to come.
Peroxide bleaching of semichemical or groundwood pulps has been de-
veloped extensively in recent years.24 The peroxides, as reagents new to
wood pulp technology, have proved to be excellent bleaching agents in
these applications. Peroxide bleaching of these relatively crude forms
of wood cellulose will continue to be the subject of active investigation.
Also it is likely that peroxides, together with chlorine dioxide and chlorites,
will find increasing use as final-stage bleaching agents in multistage puri-
fication processes.
* G. P. Vincent, L. E. Russell, and V. Woodside, Paper Trade J.t 121, 25 (Nov. 15,
1945).
«• G. P. Vincent, Paper Trade /., 124, 53 (June 26, 1947).
VH. BLEACHING AND PURIFICATION 619
2. Continuous versus Batch Processing
The tendency of modern chemical engineering practice is constantly
away from batch operation toward continuous operation. This trend is
very marked in wood pulp processing. Chlorination is now almost uni-
versally carried out as a continuous process. The treatment is always
performed at a fluid consistency, and with efficient mixing there is sufficient
time for very uniform chlorination of the lignin remaining in chemical pulps.
Caustic extraction may also be accomplished by a continuous process, and
satisfactory results are obtained by this means. When the bleaching proc-
ess was confined to the use of hypochlorites alone, intimate mixing of
fibers and bleaching agent was essential for a uniform result. The chlori-
nation process has, to a great extent, reduced the differences from fiber to
fiber, and continuous bleaching processes are being more universally
adopted. If there is sufficient mixing and flexing of the fiber bundles dur-
ing the bleaching operation, entirely satisfactory results may be obtained
at high slurry density in a continuous process, particularly if more than
one bleaching stage is employed.
Continuous bleaching systems can be operated with lower power input
and require less building space. On the other hand, batch systems have
been so designed that the batch bleachers serve as building columns and
offer ample room for installation of the necessary facilities for multiple
washing. In the final analysis, there is little doubt that continuous
bleaching systems will completely supplant batch bleaching.
3. Chemical Control
The chemical control of bleaching systems is relatively simple and rapid,
and accurate methods for the determination of bleachability throughout
the process are well established. The maintenance of temperature and pH
through the use of automatically controlling instruments has greatly simpli-
fied bleach plant operation. Rapid methods for the determination of
viscosity throughout the bleaching process have been developed and
enable the bleach plant operator to control the final viscosity of a given
pulp through the use of suitable charts or nomographs.
As a result of all of these modern chemical and engineering tools, the proc-
esses involved in the bleaching and purification of wood cellulose have
been simplified to the point where product quality is readily controllable.
Chapter VIII
PROPERTIES AND TREATMENT OF PULP FOR
PAPER
JAMES D'A. CLARK
Almost all cellulose that is produced for the market appears in sheet
form from which it is converted mainly to paper or board. Purified pulps
and cotton linters intended for chemical purposes are also conveniently
handled as sheets. For most chemical purposes, and after conversion into
paper and board, the strength and behavior of the sheet both dry and wet
are of interest. The marked changes occurring in pulp during "beating"
(i. e., the process in which natural cellulose fibers in the presence of watcy,
are pounded, rubbed, or otherwise subjected to mechanical action) and
during preparation for sheet formation, are discussed in considerable de-
tail in this chapter. Supplementing this discussion are brief descriptions
of the main types of beating machinery and an account of several theories
which have been advanced to explain the action of this equipment, which is
of great technical importance. One theory — a composite of several theo-
ries- -will be applied to interpret the effects of beating and allied phenomena
on the fibers, pulps, and papers. Brief mention will also be made of the
influence of the chemical composition of the pulps on their papermaking
performance.
Those who are interested primarily in cellulose reactions should not
dismiss lightly this chapter on the qualities of pulps which impart different
properties, such as strength, to paper. The forces that hold two fibers to-
gether in paper are essentially the same as those that hold the fibrils to-
gether in an individual fiber. Conditions of drying and the cohesion of
the elements that lead to a strong sheet of paper also make a fiber more
difficult to swell and less reactive.
A. PHYSICAL PROPERTIES OF PAPER PULP
Paper is made from a water suspension of less than 0.5% of suitably
prepared fibers, by continuously flowing it onto, then draining the water
621
622 CELLULOSE
away through, a traveling, endless, fine-meshed screen called the "wire."
Usually modifying materials, such as size, papertnaker's alum (AkCSO^s),
color, and mineral filler (often china clay) are added to the fiber mixture
before dilution. After draining, a matted or felted web of wet fibers re-
mains on the wire from which additional water is removed by suction.
The web is then carried by an endless woolen "felt" — actually a woven
blanket — through two or three pairs of press rolls to squeeze out as much
more water as is possible. The remaining water, now present to the extent
of about twice the weight of the fiber, is evaporated as the web is pressed
against the successive smooth surfaces of revolving, steam-heated cylinders
by an endless, porous "dryer felt." The dried web, usually containing
from about 3 to 8% of moisture, passes between calender rolls to smooth it,
after which it is wound into rolls (reeled up). One or both sides of the web
may be coated either on the paper machine or afterwards, to provide a
special surface for printing or other purposes.
The characteristics of the finished paper are influenced by each step
in the process, but the main factors are the type of fibers used and especially
the manner of preparing them. So far, the evaluation of the quality of
pulp by the industry has been largely empirical. Besides estimation of the
whiteness (brightness) of the pulp and its cleanliness, the usual method
consists of making the pulp into sheets after one or more degrees of beating
and subjecting the resulting sheets to physical tests normally applied to
paper.1 Measurements made are basis weight (i. e., weight per unit area,
for pulp usually expressed in grams per square meter) , resistance to burst-
ing and tearing, and less frequently folding endurance, tensile strength,
and permeability to air. An arbitrary measure of the ease of draining water
from the prepared pulp suspension is also often made.
As would be expected, desirable qualities of a pulp depend largely on those
of the paper into which it is to be transformed. Most paper made is sub-
sequently printed, and whan under the pressure of the block in common
letterpress printing, it should have a smooth, even, ink-receptive surface.
This calls for a sheet having uniform thickness and density, so that when
viewed against the light, its appearance should approach that of milky
glass and not be "wild" or mottled. Smoothness of surface and ink re-
ceptivity are more readily obtained by using pulp having thin fibers;
also, the shorter the fibers are, the easier it is to secure a good formation
or "look-through." When shorter fibers are used, the strength of the sheet
is decreased. However, unlike papers for wrapping purposes, a high
1 A comprehensive discussion of paper tests may be fotind in J. P. Casey, Pulp and
Paper Chemistry and Technology, Vol. II, Interscience, New York-London, 1952.
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 623
strength is not essential for printing papers, so that finely ground wood,
reworked waste paper, soda pulp from deciduous or hardwood trees like
poplar, gum, and birch, and from esparto grass, all of which have an aver-
age length in the order of 1 mm., constitute the bulk of the ' 'furnish/*
Wrappers and other papers that must have a high strength to withstand
applied forces are commonly made from coniferous or softwood trees such
as spruce, pine, hemlock, and fir, which initially have average fiber lengths
in the order of 3.5 mm. Fibers from some of these species, for example,
most southern pines, are inclined to be not only longer but also relatively
coarser, which makes them less well suited for conversion into fine printings.
Cotton, and especially bast fibers such as linen, can be readily beaten so as
to split lengthwise into fine and relatively long fibrils and thus may be
converted into very fine-textured, strong sheets such as for currency, fine
writing, and cigaret papers. As an example of a special requirement,
pulp for cigaret papers is usually prepared from flax because of the better
smell and taste of its smoke ; also, to make the sheet made from the well-
fibrillated fibers soft and porous, it is "filled'* with about 30% of precipi-
tated chalk.
1. The Strength of Paper
In the pulp and paper industry, "strength" has no definite meaning
except the ability of the paper to withstand one or more kinds of applied
force. In North America, standard paper tests are carried out with the
specimens conditioned in an atmosphere of 50% relative humidity and
73°F.; elsewhere, these tests are run often at the old standard of 65%
relative humidity and a slightly lower temperature.
The least complex strength test, the tensile breaking load, is deter-
mined by applying a gradually increasing pull to a strip of paper, usually
15 mm. wide and 180 mm. long, at such a rate that a break occurs in about
a quarter of a minute. The percentage of elongation of the strip before
fracture — that is, its "stretch" — is often measured simultaneously.
When a pull is applied to a strip of paper, the resisting components are
the structure itself and the individual fibers. The fibers lie mainly in the
plane of the paper and are intertwined, kinked, and twisted. Provided
that mutual contact was made while the fibers were wet and was not
brought about by calendering after the sheet was dried, the joints extend
over an area and cohere to a degree governed by the nature of the fiber,
its preparation, and the way in which the paper was made. When a small
tensile force is applied to the paper strip, an almost perfectly elastic and
624 CELLULOSE
recoverable strain will result, corresponding to the unbending or uncurling
of the component fibers as happens with a woven fabric. A greater pull
will cause more unbending, but soon in one or two places the joints between
two fibers, one of which is being particularly stressed by the applied load,
will break, thus introducing a small irrecoverable strain in the structure.
If the pull is increased to break a number of joints and straighten out r^iore
fibers in its direction and then is removed, many of the displaced fibers will
require a considerable time to assume their relaxed position. The shrink-
age due to this factor constitutes a delayed recoverable, or frictional,
elastic strain. Also, as the pull is progressively increased, any fiber that
has been straightened out and aligned in the direction of the pull will
prevent a further stretching of the structure until it is either pulled loose or
broken. The position of such a heavily stressed fiber is analogous to that
of a very tough earthworm being tugged from its hole by a robin. If the
tugging has proceeded for some time with such meager results that the robin
releases the worm, the part of the worm outside of the hole will snap back
for a distance corresponding to the elastic strain; the part in the hole and,
to a minor extent, the upper part of the worm will contract more slowly for
a distance corresponding to the delayed recoverable strain; the distance
that the lower end of the worm was elevated by the robin, together with the
extent that the worm may have been permanently stretched, constitutes
the nonrecoverable strain.
The changes inside the noncrystalline or amorphous regions of a fiber
submitted to a pull are probably similar to those in the stressed strip of
paper. The molecules and crystallites in these regions correspond to the
fibrils and the individual fibers in the paper. One difference is that in a fiber
the individual structural elements are almost parallel to each other in each
of the various concentric zones. This subject is dealt with more fully in
Chapter I V-B.
After the development and use of a new testing instrument of ingenious
design, Steenberg2 and co workers,3"5 at the Swedish Forest Products
Research Institute, published a comprehensive study of stress-strain-time
relationships of paper which has done much to focus attention on the vis-
coelastic properties of paper. Because of these properties, the strength
of paper determined at normal loading speeds should not be relied upon
to assess its resistance to sudden shock loads nor to prolonged stress below
* B. Steenberg, Svensk Papperstidn., 50, 127 (1947).
8 B. Steenberg, Svensk PappersMn., 50, 346 (1947).
4 B. Ivarssori and B. Steenberg, Svensk Papperstidn., 50, 419 (1947).
5 B. Ivarsson, Svensk Papperstidn., 51, 383 (1948).
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 625
the breaking load. Ranee6 showed that strips of paper having a tensile
strength of 10 kg. under normal testing conditions broke after 14 hrs.
under an applied load of 8 kg. ; after 20 days under a 6-kg. load; and after
220 days under a 4-kg. load. A summary with some other applications of a
study of stress-strain-time relationships has been published.7
2. Common Tests for Paper
The outcome of the common physical tests for paper is dependent on the
basis weight of the sheet, usually expressed as pounds per 500-sheet ream of a
certain size, often 24 by 36 in. The bursting test, the most frequently
used test, is made by clamping an area of the sheet, effectively 1.24 inch
in diameter, against a rubber diaphragm and applying increasing hydraulic
pressure beneath the diaphragm until the paper bursts. The maximum
pressure attained is the value reported. The bursting strength more or
less parallels the tensile strength in the direction in which the paper has the
least stretch, usually the strongest direction of the paper. The ratio of
bursting strength to tensile strength increases with the ability of the paper
to stretch.
The tearing resistance is found by a pendulum type of instrument which
measures the work done to tear several sheets through a fixed distance and
is reported in terms of force (in grams) required to tear a single sheet. The
measured work also includes bending the specimen and rubbing the torn
edges past each other. The result depends mainly on the length of the
fibers, on the bulkiness of the sheet and its stretch, and for weak sheets on
the degree of fiber cohesion. For a sheet of fibers that do not cohere well,
the tear increases as the cohesion between the fibers is increased by any
means, until the cohesion becomes sufficient to dissipate the applied tearing
force over an appreciable area around the point where the tear is progress-
ing. After this degree of cohesion is reached, a further increase serves to
diminish the tearing resistance, usually quite rapidly, because increased
cohesion beyond the optimum for tearing resistance has an effect similar to
increased sheet density; both serve to concentrate the applied tearing force
over a smaller area of the sheet and thus make it more effective.
The folding endurance of a sheet of paper is the number of double folds
that a strip 15 mm. wide and under a tension of about 1 kg. will withstand
over a line across its width until fracture. The result is dependent on the
e H. F. Ranee, Proc. Tech. Sect., Paper Makers' Assoc. Gt. Brit. & Ireland, 29, 449
(1948).
7 B. Steenberg, Pulp & Paper Mag. Can., 50, No. 3, 207 (1949).
626 CELLULOSE
third or fourth power of the applied tension, so that folding endurance
as measured by current methods is unduly sensitive to the normal tensile
strength of the strip. True folding endurance is dependent mainly on the
length and fineness of the fibers, their elasticity, and on the stretch of the
sheet.
The air resistance, the inverse of air permeability, is usually measured
in terms of the number of seconds for 100 cc. of air, under pressure of about
12 cm. of water, to flow through 1 sq. in. of paper. It depends on the dens-
ity of the sheet and on the specific surface (i. e., the exposed surface per
gram) of its components.
It has long been known that when a sheet of newly formed paper has
been dried under tension, as when clamped against a polished plate in-
stead of being allowed to shrink freely while drying, the tensile strength
is increased by about 10 to 50%, and the stretch to the breaking point is
diminished by nearly 50%. The loss of stretch is mainly due to the re-
moval of the nonrecoverable component. There is also a corresponding
drop in bursting strength of 10 to 20%, a decrease in tearing resistance of
up to 50%, and usually a severalfold increase in folding endurance. In
consequence, paper tests made on pulp test sheets give highly arbitrary
values which depend on how the pulp was prepared, how the sheets were
formed, and how they were pressed and dried.
Interpretations of such common tests are not only quite complex, but
at best merely serve to indicate whether the pulp is more or less suitable
for a given paper; they give little information concerning the underlying
causes for one pulp to be better or worse than another in any respect.
Occasionally, to supplement these physical data, measurement of viscosities
and such chemical tests as alpha-, beta-, and gamma-cellulose content,
copper number (ability to reduce Fehling solution), and chlorine demand
are made. However, as will be briefly discussed later, unless the results
are extreme, chemical tests by* themselves do not provide a good measure
of any papermaking quality of a pulp.
3. Pulp Testing
A more fundamental approach to pulp testing is needed, and one8 which
has shown some promise involves consideration of six pulp qualities rather
more basic than those just discussed. However, it should be emphasized
that producers and consumers of pulp are mutually restrained from the
easy abandonment of orthodox methods, so that up to the present, only
• J. d'A. Clark, Pulp & Paper Mag. Can., 49, No. 10, 202 (1948).
VTH. PROPERTIES AND TREATMENT OF PULP FOR PAPER 627
meager progress in this direction can be recorded. The following qualities
were chosen, not because they necessarily represent the best for the purpose,
but because as a group they comprise a fairly complete characterization of
the papermaking qualities of pulp which can be deduced from a practical
system of currently available measurements. The factors involved are
(a) fiber length, (b) dry fiber flexibility, (c) wet fiber flexibility, (d) co-
hesion, made up of (1) the extent of the area in bonded contact and
(2) the intensity of the bonding, (e) intrinsic strength of the fibrous ma-
terial, and (f ) ability of the pulp to respond to the wet mechanical treat-
ment or beating. Of these factors, beating is one of the most important
because of its pronounced effects on most of the other factors. For
example, an appropriate beating treatment, without the addition of supple-
mentary materials to the pulp furnish, can so modify fiber length, flexi-
bility, and area of cohesion that an average pulp may be made into such
extremes as a fairly satisfactory blotting paper or a fairly good transparent
greaseproof wrapping paper. The nature and measurement of the first
five factors involved in this newer concept of fiber quality will be discussed
before beating.
(a) Fiber Length
Even today there is no general agreement as to what may be meant by
the "length" of fibers in a pulp, mainly because there is a choice of express-
ing it in a variety of ways, including the numerical average length of the
fibers (which is much affected by the lower limit of the length of the particles
considered to be "fiber1 ') and the weighted average length by length, by
projected area, by volume, and by true weight. Thus the * 'average"
fiber length of untreated black spruce fibers has been reported to vary
from 0.749 to 3.4810 mm., usually without definition of the type of average
employed. Perhaps because of this uncertainty, as well as the tediousness
of making individual fiber measurements, fiber length has been a test some-
what in disrepute. It is clear that just as the molecular length of cellulose
affects the strength of the structure it composes, so the length of the fibers
affects the strength of paper. Although the number-average length seems
to determine the strength of polymers (see Chapter XI), the weighted
average length by weight (which will be hereinafter termed "weight-
average length") was shown by Clark11 to be the important one in the case
9 J. B. Calkin, Paper Trade J., 91, 44 (Aug. 28, 1930).
10 C. D. Mell, quoted by E. Sutermeister, The Chemistry of Pulp and Papermaking.
2nd ed., Wiley, New York, 1929, p. 57.
11 J. d'A. Clark, Paper Trade J., 115, 36 (Dec. 24, 1942).
628 CELLULOSE
of paper, perhaps because the individual fibers, unlike molecules, vary
in thickness and especially because in a test sheet of paper the fibers are
isotropic, lying equally in all directions. He found if L was the weight-
average length either of the longest fraction of wood pulp fibers separated
out, cut, and reclassified, or of rayon fibers cut to various lengths, the
tensile strength of test sheets varied as L1/a, the burst directly with L,
and the tearing resistance as L8/t. Some confirmation of these ratios is
indicated in a subsequent study12 by the U. S. Forest Products Laboratory.
Fiber length had no measurable effect on the density of the paper nor,
contrary to an erroneous common conception which widely persists, on the
ease with which water drained from the pulp. Also, if the thickness of all
fibers was held constant and if equal weights of uniformly long and short
fibers were mixed, the test results on paper made from the mixture cor-
responded closely with tests on a sheet made from uniform fibers having
the same weight-average length as the mixture and did not correspond at all
with tests on a sheet made from much shorter uniform fibers having the
same number-average length as the mixture.
These findings are at variance with several previously reported results on
the effect of fiber length except one,13 but in all the other previous findings
with that exception, the test specimens were prepared after classifying a
whole pulp into portions. There is a significant correlation between the
length of a natural fiber in a pulp and its thickness,14 so it is believed that
most of the effects of varying fiber length reported were masked by the
important influence of accompanying changes in fiber thickness. The
importance of thickness is evident because the flexibility of a rod of circular
cross section varies inversely as the fourth power of its diameter, and the
number of fibers in a unit area of a sheet of paper of a given basis weight
varies as the square of their diameter.
The weight-average fiber length (see TAPPI method T 223 sm-53) is
conveniently determined from the fiber distribution with a classifier of the
Bauer-McNett type11 or more rapidly with the Clark Four-Screen Classi-
fier.15 The specimen fibers, highly diluted in water, are caused to flow
through a series of compartments parallel to the faces of screens of de-
creasing mesh. If the fibers are shorter than twice the screen opening they
can turn a somersault on the edge of an opening and pass through. Other-
1J Sulfite Pulp for Paper, U. S. Forest Products Laboratory, Mimeographed Report
1596 (1943).
18 R. B. Brown, Paper Trade J., 95, 145 (Sept. 29, 1932).
14 J. H. Graff and R. W. Miller, Paper Trade J., 109, 31 (Aug. 10, 1939).
16 A. E. Reed and J. d'A. Clark, Tappi, 33, 294 (1950).
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 629
wise they remain in the compartment ahead of the screen. However, fi-
bers longer than twice the size of the openings in the screens may pass
through if the fibers are bent or if they are very flexible, so there is always
the likelihood of a small error in the determination of length distribution
curves and weight-average lengths, unless this effect is measured and
allowed for.
(b) Dry Fiber Flexibility
Dry fiber flexibility may be defined as the deflection of a dry fibrous ele-
ment which results from an applied bending moment. It depends
markedly on the thinness of the element and its walls, and to a lesser extent
on humidity. Fibers are usually tubular, and when the fibers are dried,
the sides of most thin- walled fibers collapse to form flat tubes ; they there-
fore bend much more easily, are more flexible, and compact together better
than do thick-walled fibers.
In most trees, especially conifers, the surnmerwood fibers are thick-
walled and stiff; the thin-walled springwood fibers, though of equal
diameter tangentially and of even greater diameter radially, are more
flexible. Paper made from the former is bulky, weak, and brash; paper
made from the latter is relatively dense, strong, and pliable. The thin-
walled fibers, although individually weaker than the thick-walled fibers, are
more numerous in sheets of the same basis weight. Thus they distribute a
force applied to the sheet over a greater area and use their strength to
greater advantage by virtue of their greater flexibility as well as their
greater bonded area.
Since the solid cellulosic material in common papermaking fibers does
not vary greatly in its resistance to bending, an approximate measure of
dry fiber flexibility is that of the total length of a gram of fiber. This can be
ascertained11 in about half an hour by optical projection; cumulative
measurements are made on a fiber suspension of known dilution and volume
which has been immobilized by the addition of a gel.16a Another approxi-
mate measure of dry flexibility is a determination of the average diameter
and wall thickness of the fiber from direct measurements of cross sections
made on the unpulped fibrous material, for example, wood. These are
neither difficult nor tedius to make if a photomicrograph of known enlarge-
ment of a representative cross section is available.
Because dry fiber flexibility is closely related to fiber fineness, a measure
of either property provides a fair index of the relative surface smoothness
1Ba See TAPPI method T 232 sm-53.
630 CELLULOSE
and ink receptivity of paper made from the pulp, if an allowance is made for
the possible splitting of the fibers by subsequent treatment.
(c) Wet Fiber Flexibility
Wet fiber flexibility is influenced not only by the factors which govern
dry fiber flexibility, but also by the extent to which the fibers were bruised
or internally split and fibrillated, or rubbed or externally split, during beat-
ing. These factors are not important in dry flexibility because in most
cases the split fibrous elements rebond as they dry.
The ease with which fibers are flexed when wet is most conveniently
measured by the apparent density (or its reciprocal, the apparent specific
volume) of pulp test sheets after these have been wet-pressed and dried
without pressure on a polished metal disk in accordance with the standard
pulp testing method (TAPPI Standard T 205 m-50). Alternatively,
the apparent density may be measured indirectly by the free shrinkage of
the pressed moist sheet when dried in the air. As water leaves a freshly
deposited moist sheet of paper, adjacent fibrous elements are drawn to-
gether by surface tension. Campbell16 has pointed out that the force in-
volved is independent of fiber thickness and, in consequence, may amount
to thousands of pounds per square inch for the finer fibrous elements, which
accordingly will give a harder, denser, more compact sheet. A disturbing
element in the relationship between sheet density and wet fiber flexibility
may be the curliness of the fibers, but this is a minor one unless the fibers
were curled artificially, as will be discussed later.
It should be emphasized that, contrary to widely held opinion with an
occasional dissenter,13 fiber length per se is not an appreciable factor in
sheet density nor air permeability; such opinions originally were based on
data derived from length-fractionated whole pulps, in which experiments,
as already mentioned, fiber thickness was an unconsidered covariable.
(d) Fiber Cohesion
Fiber cohesion includes both the area of fibrous elements bonded together
and the intensity of the bonding. The former may be measured by optical
means as developed by Parsons,17 when the specific surface of the pulp is
known. A measure of fiber cohesion as a whole is the transverse tensile
strength per unit area of a test sheet. A good method for determining this
is the viscosity-velocity product (WP), measurement developed by the
M W. B. Campbell, Can. Dept. Interior, Forest Service Bull 84 (1933).
17 S. R. Parsons, Paper Trade /., 115, 34 (Dec. 17, 1942).
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 631
Institute of Paper Chemistry.18 It consists in passing a heavy roller over a
drop of highly viscous liquid on a face of the sheet and finding the limiting
surface speed of the roller needed to blister or split the paper. The in-
tensity of bonding may be judged by dividing the calculated force to split
the sheet by the estimated percentage of the fractured area in fiber-to-fiber
contact.
From a theoretical point of view, the measurement of the intensity of
bonding is an interesting one, because it must relate to the chemical and
physical nature of the surfaces in contact and is probably not increased by
beating subsequent to the interiors of the fibers being exposed. In prac-
tice, as has already been intimated, fiber cohesion is of importance with
respect to paper strength, not because of any direct resistance offered to an
applied stress, which would be quite low, but because of the frictional
resistance developed by individual fibers tending to be displaced length-
wise by an imposed stress. When the fibrous elements are long in relation
to their diameter, as is the case with fibrils derived from cotton and bast
fibers, then the magnitude of fiber cohesion becomes correspondingly less
important, particularly as the sheet becomes increasingly dense and the
elements are packed closer together. It must be emphasized that only
relatively small cohesive forces are needed to anchor firmly a fiber embedded
in a normal sheet, particularly if the fiber is long.
If the normal tensile strength of a test sheet is divided by a function,
preferably the square root,11 of the weight-average length of the fibrous
elements and by the apparent density of the sheet, a practical index of
fiber cohesion may be derived. This reduces cohesion to a single number
which is reliable at least to the extent of telling if the pulp being tested is
proficient or deficient in this property.
(e) Intrinsic Strength of Fibers
The intrinsic strength of a fiber may be found directly by attaching the
ends of an individual fiber to the ends of two strips of paper with sealing
wax, and measuring the tensile strength of many fibers this way by an ap-
paratus fashioned after a chainomatic balance, then dividing the breaking
load of each by its cross section measured microscopically.
Because of the delicate procedures needed to handle individual fibers and
their variability which entails many data for accuracy, it is much easier
to work with sheet material and determine this property indirectly by the
18 Anon., Paper Trade /., 123, 24 (Oct. 31, 1946); 123, 24 (Nov. 7, 1946).
632
CELLULOSE
tensile strength of the sheet as the span between jaws approaches zero as
proposed by Hoffman Jacobsen.19
The assumption is made in a standard handmade test sheet that half of
the fibers are oriented in the direction of pull. An improved attachment de-
signed by Clark20 for a tensile tester is shown in Figure 1. The test20ais
TANG FOR UPPER GRIP
-SAMPLE
TANG FOR UWER GRIP
Fig. 1. Zero-span jaw attachment for tensile strength tester. Courtesy
of Thwing-Albert Instrument Co., Philadelphia, Pa.
rapid and reproducible. It can be regarded as the ultimate or maximum
tensile strength of a test sheet made with the fibers being tested, after they
have been subjected to a theoretjcally ideal beating treatment for the
optimum time. The test appears to be affected by the degree of polymeri-
18 P. M. Hoffmann Jacobsen, Paper Trade /., 81, 52 (Nov. 26, 1925).
80 J. d'A. Clark, Paper Trade /., 118, 29 (Jan. 6, 1944).
soa See TAPPI method T 231 sm-53.
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 633
zation of the cellulosic material in the fibers (discussed more fully under
topic E-l in this Chapter VIII) and is sensitive to overcooking or over-
bleaching of the pulp. As proposed by Hoffmann Jacobsen,19 the ratio of
the normal to the zero span tensile test gives a good measure of the over-all
entanglement and cohesion of the fibers.
(/) Pulp Evaluation and Response to Beating
The five properties just discussed are sufficient to characterize the physi-
cal properties of a prepared or beaten pulp. Properly interpreted, they
serve to forecast its ' 'running" qualities on the paper machine (in inverse
proportion to the density of the test sheets) and to predict the strength and
most other qualities of the paper made from it.
Pulps differ considerably in their response to beating, depending upon
the structure of the fibers and the chemical nature of their material. The
measurement of beating qualities will be discussed later; however, a prac-
tical and comprehensive set of tests for evaluating the quality of an un-
known pulp, including its beating characteristics, is as follows: Two
portions of the sample are beaten in a standardized manner, one for a
moderate time and the other for twice as long. One of these times should
preferably, though not essentially, be chosen so that after that period the
laboratory treatment will be approximately equivalent in extent of beating
to the treatment of the pulp in practice. A set of standard pulp test sheets
is made for the two beating periods. At the end of the period correspond-
ing to the degree of beating in practice, the average fiber length, the fine-
ness of the fibers (if unknown), the density of the test sheets, and their
normal and zero-span tensile strength are measured. To serve as a check,
it is desirable to apply also all the usual strength tests, such as bursting
and tearing, to the sheets at that beating period. With the other beating
period, only the normal tensile strength of the test sheets is needed, but
it is well to measure also their density and the average length of the fibers.
Provided that the two chosen degrees of beating lie on the straight part
of the plot of tensile strength versus logarithm of beating, which usually
will be the case, the percentage increase in the normal tensile strength of the
test sheets effected by doubling the time of beating gives a good practical
measure of the beating quality of the pulp.
B. BEATING
WHen pulp is beaten, the outstanding resulting characteristic lies in the
ability of fibers so treated and thereafter dried in contact with one an-
634 CELLULOSB
other, to cohere strongly. The beaten fiber mass also acquires a slimy feel
and the property of holding water longer or more firmly when allowed to
drain. The prepared fiber suspension, called "stuff" or "wholestuff" or
"stock" is then said to be "beaten," "wet," "slow," or "hydrated." Before
beating, the condition of the pulp is sometimes termed "half stuff," but more
often "unbeaten" or "raw"; after only a mild beating treatment the pulp
condition is termed ' 'free/ ' "Wetness, ' ' ' 'slowness, ' ' and ' 'f reeness' ' are de-
rived words. These technical terms refer to the ease or difficulty with
which the pulp, when squeezed in the hand or placed on a screen, either
retains or parts with some of its associated water.
Beating is extremely important in its relation to the quality and the rate
of production of paper. In fact, it has been said with much truth that
paper is "made" in the beater. Furthermore, the interest of modern paper
technologists in beating has been heightened by economic reasons. An
enormous amount of power is consumed — from 200 to nearly 2000 kilowatt-
hours per ton of finished paper for certain specialities although rarely is a
figure of 1000 exceeded. The high cost of this one operation in paper-
making is readily understood from the fact that in the United States and
Canada the current annual production of paper and board is over thirty
million tons.
1. Influence of Moisture on Pulp
Because of economy in transportation and also resistance to decay in a
dried condition, most purchased pulps (excluding groundwood, which is
not easily defibered after drying) are shipped after having been dried on
steam cylinders to a dryness of over 70%, usually about 90%, in which
latter condition it is technically termed "air-dry." The saturation of the
fibers with water and their consequent swelling and softening is the first
step in beating, although scarcely recognized as such. The rate of this
swelling is somewhat dependent, inversely, on the extent to which the
fibers have been dried. When swollen, they become more flexible and less
brittle; because of this, they are better able to withstand cutting during
the violent mechanical action of beating. Pulps which have never been
dried are initially much softer and more swollen than dried fibers that have
been soaked in water and, in consequence, behave differently upon being
made into paper, especially after only a limited beating period. In most
cases, at least one hour of mild beating treatment is required before the
papermaking characteristics of a normally dried pulp revert, in part only, to
those it had before being dried. This swelling of the fibers has been termed
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 635
"imbibition" by some authorities and at one time was considered to be an
integral part of the beating process itself, various vacuum and pressure
devices being proposed and built into the beater to accelerate imbibition.
However, except for the undisputed fact that the rubbing and squeezing
of the pulp during the beating period serve to accelerate the sorption of
water, imbibition is now commonly regarded as more or less incidental to
beating provided that fibers are properly soaked before being subjected to
intensive treatment.
2. Beating Equipment
Before considering the changes of pulps effected by beating and discuss-
ing the various theories explaining them, a brief description will be given of
the principles of four types of commercial modern beating equipment.
Prior to the adoption of this equipment, heavy stampers were used to break
up the mass of tangled rags, commonly used as a source of fibers, into a
uniform pulp.
(a) Beaters, Jordans, and Refiners
The hollander (usually termed the beater), introduced about 1670,
marks the beginning of modern beating. This beater, which has changed
but little in its original design, is still supreme in the fields of almost all
but coarse papers and certain specialities. Figure 2 shows a cross section
of a modern beater. It has the form of an oval-shaped trough A divided
into two channels by a "mid-feather" B. Across the bottom of one chan-
nel, is placed the "bedplate" C which is either single or multiple; this plate
is provided with projecting blades or "bars" of steel or bronze, usually from
3/i« to 8/s in. thick and spaced about the same distance apart. These blades
are set at a slight angle to the plate and project from the wooden spacers
about */2 in. Over the bedplate is mounted a heavy roll D, provided with
blades or "bars" which are set parallel to the axis of the roll, which project
about 2 in., and usually are spaced about 2 in. apart. The bearings sup-
porting the roll are precisely adjustable to vary the clearance or pressure
between the bars in the roll and in the bedplate. When the beater is in
operation, the pulp suspension is circulated round the trough by the paddle
action of the bars in the roll. Some of the fibers in the pockets between the
bars are commonly supposed to be caught on the edges of the bars in the
form of a "fibrage" and banged and rubbed across the bedplate bars and
their edges. The pulp is discharged over the backfall E to flow round
the trough to the front of the roll again. Usually the beater roll is lowered
636
CELLULOSE
as beating proceeds and after the fibers have become sufficiently flexible
to withstand a more drastic pressure without being so readily cut. The
peripheral speed of the roll is approximately 2000 ft./min., the pulp circu-
lates in the trough at the rate of about one revolution in 2 or 3 min., and
the treatment is usually continued from one to a number of hours, depend-
ing on the kind of stock. The consistency of the pulp (that is, the percent-
age of pulp solids in a measured volume of suspension) is usually from 5
to 8%. Milne's two papers21'22 should be consulted for further practical
information on beaters and beating.
Fig. 2. Sectional view of a modern beater and detail of bedplate adjustment.
(A) Beater tub; (B) midfeather; (C) bedplate (multiple shown); (D) roll; (E)
backfall. The curved arrows within the diagram indicate the direction of flow
of the pulp. Courtesy of E. D. Jones & Sons Co., Pittsfield, Mass.
The "Jordan51 has almost entirely supplanted the beater in modern
mills manufacturing newsprint and kraft wrapping papers. This type of
beating equipment consists of a conical drum rotating in a horizontal coni-
cal casing, the rubbing surfaces being provided with blades or "knives"
similar to those described for the beater. The clearance between the knives
is adjusted by sliding the drum in or out of the cone. The pulp, at a con-
sistency between 2 and 5%, enters the casing at the small end and is moved
by both pressure and centrifugal action to the exit at the large end. These
machines are built in various sizes, some having a capacity for treating
considerably over 50 tons of stock per day and consuming as much as 500
horsepower in so doing. In general, the action of a Jordan is more that of
81 S. Milne, Paper Trade J., 84, 54 (June 16, 1927).
« S. Milne, Pulp & Paper Mag. Can., 37, No. 8 442 (1936).
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 637
cutting and especially of a ''clearing'1 of clots and fiber bundles in contrast
to that of a beater. Thus, jordans are now almost universally used on the
stock for this purpose after the beaters and just prior to the paper machine.
The refiner,23 which again is coming into favor, has found its greatest
use in making kraft papers but its use is being extended to other grades.
This equipment consists essentially of two large steel disks in a casing, one
or both of which revolve. In a variation of this design, so as better to bal-
ance the large pressures involved, a center disk is made to revolve between
two stationary ones. Like millstones, the disks have bars or serrations
machined into their two adjacent faces. Pulp is fed through an opening to
the center of one of the disks and passes then between the faces to the out-
let on the periphery of the casing. The distance between the faces of the
disks is adjustable with great precision so as to vary their action on the pulp.
More recently, a new type of refiner called an Idar has shown promise,
especially for treating cotton, linters, and bast fibers. The faces of the
disks of the new machine, one of which is rotated, are serrated with V
grooves about % in. apart to form a myriad of small sharp pyramids.
The pulp is fed through into the center of the disks, where it is dewatered
into clots ' aving a high consistency. As the clots proceed outwards, pre-
sumably i? y are seized by the apices of the pyramids, wrenched apart,
and re-formed to be wrenched apart again. In this way, a maximum of
splitting and rubbing action is secured with a minimum of cutting, espe-
cially with those fibers which readily split along their lengths.
(b) Special Machines
Although the hollander, Jordan, and refiner are the most widely used type
of beating equipment, three other types — the rod mill, the kollergang, and
the Curlator — are in limited use commercially for special purposes.
A rod mill24 consists of a large rotating steel cylinder filled partially with
heavy steel rods. The pulp, which may have a high consistency (up to 20%
or more), is fed in at one end, and is subjected to the rolling and pounding
action of the rods as it proceeds to the outlet at the other end. However,
because of power requirements, wear of both the rods and the lining of the
cylinder, and intense noisiness, very few remain in operation.
The kollergang or edge-runner25 is used in some mills, usually paper mills,
18 T. W. Chambers, Paper Making and Its Machinery, Constable, London, 1920, pp.
50-54.
*< S. D. Wells, Paper Trade /., Ill, 89 (Aug. 29, 1940).
16 W. Brecht, Papier- Fabr., 35, 259, 313 (1937).
638 CELLULOSE
mostly in northern Europe. A typical machine consists of two heavy
granite disks, about 3 ft. in diameter and 10 in. wide, mounted on horizontal
axes to track around a horizontal annular trough at about 15 r.p.m.
Pulp at about 30 per cent consistency is fed into the trough where it is
subjected to a rolling and twisting action by the runners. Because of the
very high consistency of the pulp and the manner of operation, relatively
few fibers are cut. The high power consumption per ton of pulp treated
and the low output per machine are reasons why, despite their desirable
action, few kollergangs are used in North America.
A machine called a Curlator26 has been placed in limited commercial use
in the United States and Canada during the last few years. It consists of
two rough-surfaced steel disks about 4 ft. in diameter, mounted horizon-
tally one above the other. The upper disk, which bears upon the bottom
disk with a pressure of about 15 Ib./sq. in., is given a gyratory motion
on a small radius. The action continuously dewaters and rubs the pulp
fibers or fiber bundles flowing outwardly between the disks and curls the
fibers. In consequence, test sheets made from the treated pulp are bulk-
ier and thus weaker in burst and tensile strength and stronger in tear
resistance and more absorbent than if made from untreated fibers.
Rod mills, kollergangs and Curlators find use mainly in pulpi and paper
mills making some special grades of paper. The machines are well adapted
to defibering shives or slivers of wood which are incompletely softened
by cooking. Their use is being extended because of a recent trent toward
more "semichemical" pulping. This process entails less complete cooking
of the wood chips and their subsequent reduction to fibers by mechanical
treatment, usually while the material is still hot. Refiners, usually in
conjunction with jordans, are also being increasingly used for the re-
duction of the wood chips. Pulp yields of upwards of 60% of the wood
substance are attained.
3. The Action of Beating Equipment
When the bar of a beater, Jordan, or refiner sweeps through the pulp,
it is commonly held that a layer of fibers is caught and draped over the edge,
mostly at right angles, to form a "fibrage"27 or mat having elements lying
mostly at right angles to the bar edge. This fibrage is then banged against
the leading edge of one of the stationary bars, and intense pressure is de-
* H. S. Hill, J. Edwards, and L. R. Beath, Tappi, 33, 36 (1950).
n S. Smith, The Action of the Beater, Tech. Sect., Paper Makers' Assoc. Gt. Brit. &
Ireland, London. 1923.
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 639
veloped on the assembled fibers across the zone of contact. Some of them
are cut; some are split, and the tail of the outer surface of the fibrage
remaining on the moving edge is rubbed across the edge and face of the
stationary bar. During this action some of the fibers are transferred to
the leading edge of the stationary bar and may form a secondary fibrage
there. The outer face of this secondary fibrage is in turn treated by the
edge and face of the following moving bar. Fibers are also rubbed against
parts of the equipment and against each other as the pulp is circulated around
the machine.
However, it is not certain that the action is as simple and as straight-
forward as just described. Because of the induced turbulence, there is
some doubt as to the ability of the leading edge of a flat bar moving flat-
wise through the pulp to gather other than momentary fibrages over short
lengths. If this is the case, then it is unlikely that such fibrages would be
well oriented ; they would be more in the nature of clots.
The predominantly pounding action of the rod mill, the rubbing and
twisting action of the kollergang, and the rubbing and wrenching action of
the Idar have been mentioned. These do not preclude cutting and
splitting the fibers, since, when two or more wet fibers cross each other and
are compressed between two hard surfaces, mutual cutting of the fibers
can and does occur.
On a greatly magnified scale, the beating operation would be analogous to
pouring bushel baskets of very long cigars between two trains passing each
other at high speed with a clearance between them of only a couple of
inches. Under this action, the cigars would be cut, or would cut each other;
they would be split externally, unravelled, and rubbed. Much of the
" wrapper'* would be loosened and removed and the inner filler would be
split, separated, and bruised. Some cigars would be bent, and a large
amount of tobacco dust and small pieces would also result. In all beating
operations, it will be clear that fibers are cut, split, bruised, and, usually
to a minor extent, bent or curled. Fiber debris is always produced, and
new surfaces are opened up and exposed to water which is immediately
adsorbed. Accordingly, the beating action given to fibers by any piece of
beating equipment may be characterized and measured in terms of (1)
cutting (shortening), (2) splitting, which in turn may be subdivided into
(a) surface splitting or roughening — as a result of rubbing, and (b) internal
splitting — as a result of bruising, and (3) deformation or curling of, the
fiber — usually negligible unless the fibers have been treated with a Curlator
or in a commercial kollergang at high consistency. It should be emphasized
that these three basic actions, together with the adsorption of water on
640 CELLULOSE
exposed cellulosic surfaces, are the only things that can possibly happen
during beating. The so-called "hydrating" action of the beater is presumed
to be equivalent to the splitting of portions of the exterior and interior of
fibers and the sorption of water on the newly exposed surfaces.
The over-all and relative extent of the three basic actions on a given
pulp will depend on the design of the equipment, the manner of its opera-
tion, and the consistency of the pulp being treated; temperature has a
small, usually negligible, effect. For a given pulp, cutting of the fibers will
be promoted relative to the other actions by the following: (a) Low pulp
consistency — Thinner fibrages or clots will be formed on the bar edges so
there will be less cushioning at impact and more intensive pressure, (b) Bar
sharpness — There is much erosion and corrosion of the bar edges in beating
equipment. New bars cut more than worn ones, thin ones more than thick,
and, despite contrary statements in the literature, relatively soft noncor-
roding metal bars like bronze or monel often cut more than do steel bars,
especially hard steel bars, which do not wear away nearly as fast as their
leading edges become rounded, (c) Bar angle- —It has been held that the
less parallel are the fixed and moving bars, the greater is the scissoring or
cutting action. However, unless the bars are very sharp it is very doubtful
if this is much of a factor, (d) Bar speed— For a given applied horse-
power, a lower speed means a lesser clearance between the bars and con-
sequently a greater cutting action. On the other hand, fibers, especially
when wet, are capable of extensive plastic deformation, so that, if, during
beating, cutting forces are applied more suddenly, increased bar speed,
especially beyond about 2500 feet per minute, favors relatively more cutting.
However, as pointed out by Ranee,28 there is an analogy between the mech-
anism of oil lubricating a bearing and the pulp suspension acting as a
lubricant between the fixed and moving bars. Thus an increase in the
speed of the moving bars may sweep more fibers into the zone of action
and to that extent reduces the~tendency for cutting to occur, (e) Rigidity
— The firmness and inertia of the bars and bar-holding parts intensify
the imparted shock, which, for the reason just discussed, favors relatively
increased cutting. Accordingly, in general, mill equipment will do con-
siderably more cutting and smashing of fibers to produce debris than will
laboratory equipment, even if other factors are substantially the same,
(f) Bar clearance or pressure — This is the main adjustable variable for
control purposes and governs the power input to the beating machine.
Internal splitting depends on the nature of the fibers, but will be increased
88 H. F. Ranee, Proc. Tech. Sect., Paper Makers' Assoc. Gt. Brit. & Ireland, 32, 360
(1951).
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 641
by high consistency of the pulp, blunt bars, and medium bar pressure.
Surface splitting or rubbing is favored by high consistency, high bar speed,
and low bar pressure. If a ' 'stone" bedplate and a "stone" roll, actually
made of basalt lava, the roll having wide, deep channels cut across its
face like a broad gear wheel, are substituted for the usual beater roll with
metal blades, or if the bedplate is made of stone and the roll of metal,
the rubbing and bruising action of the beater will be intensified. Accord-
ingly, such equipment is especially desirable for greaseproof papers and
some specialties, such as ultrathin electrical condenser tissues, but un-
fortunately the lower output and thus higher power consumption per ton
of stock prepared with stone tackle are considerable disadvantages.
It is almost axiomatic that the more power that can be applied to any
given beating machine working on pulp, the less power per ton is needed to
beat the pulp. This is so because, in general, a substantial portion of the
applied power is largely wasted in circulating the mass of pulp and water,
and the power loss for this does not increase as the total applied power is
increased. However, unless the consistency of the pulp can be raised to
offset the increased bar pressure involved, the quality of the prepared stock
is then likely to suffer to some degree.
4. Measurement of Beating
The effects of beating which have been most used for its measurement
have been (1) reduction in fiber length, (2) difficulty in draining water
from a mass of pulp, (3) increase in the density of test sheets, (4) increase
in the specific surface of the pulp, and, as already discussed, (5) increase in
the tensile strength of test sheets.
The oldest and most straightforward method of judging the degree of
beating is the estimation of the length of the fibers, originally by direct
inspection, more recently with a magnifier or by optical projection. By
modifying the operation of a high-speed fiber classifier,16 the percentage of
fibers longer than a chosen length may be determined in a few minutes;
it is possible, therefore, that this method will again take the leading place.
Since the beginning of this century, scores of tests and instruments be-
came available for measuring the rate at which "free" or "fast" water
drains from pulp. These tests are empirical and are conducted under
specified conditions. Results are usually reported as (a) drainage time,
that is, the time required to form sheets of a certain moisture-free ,basis
weight (TAPPI Standard T 221 m-51), or as (b) the "freeness," that is, the
volume of "fast" water drained from the pulp (TAPPI Standard T 227
m-50).
642 CELLULOSE
Unfortunately, the freeness test is affected by many different factors,
particularly the quantity of fine debris present in the pulp. In fact the
"freeness'1 of a well-beaten pulp can be almost fully restored to its original
value by washing beaten pulp with water on a 200-mesh sieve and so re-
moving the debris without greatly affecting the strength.8 In consequence,
because of the highly questionable, and indeed erroneous, implicit assump-
tion that the quantity of fine debris produced by all kinds and sizes of
beaters and other machines is proportional to their beating action, whereas
it may vary greatly even with a single laboratory beater, the custom of
reporting the strength or other properties of pulps beaten to a given free-
ness is always indefinite and often misleading.
Clark8'29 has found that when the apparent density of the test sheets, or
better still its inverse, the apparent specific volume, is plotted vs. loga-
rithm of either the amount of energy supplied or the beating time, if the
energy supplied per pound of pulp is constant, the resulting plot is a
straight line. This may except the initial point for unbeaten pulp which is
sometimes disturbed in position by the presence and quick removal of the
primary wall, and will except the final points if the pulp is beaten consider-
ably beyond the normal commercial limit. It was also found8 that the
slopes of the lines so plotted for a wide variety of pulps treated with the
same beater and under the same conditions are surprisingly parallel.
Thus, the plot for a straw pulp gives a slope similar to a wood pulp or a
rag pulp but commences at a lower specific volume — about 1.5 instead of
about 2 and 2.5 cc./grarn, respectively. These findings make the specific
volume of the test sheets (or their density) a most useful basis for measuring
the degree of beating. A minor drawback to using either the density or
specific volume is their comparative insensitivity and consequently the
time and care needed to make a measurement with sufficient precision.
The longitudinal shrinkage of the test sheets before and after drying is an
analogous test which is more Sensitive and may serve for control purposes
when suitable equipment for determining density, such as that described
in TAPPI Standard T 205 m-50, or adequate time, is not available.
The specific surface of a pulp (see TAPPI method T 226 sm-52) ap-
pears to increase linearly with the logarithm of the amount of beating,
at least during the initial and intermediate stages and at a different rate
with different pulps. It would make an excellent measure of beating if it
were not so difficult to determine reliably.
When the tensile strength or its function, breaking length, is plotted
against the logarithm of the amount of beating, the result is linear over
» J. d'A. Clark, Paper Trade /., 116, 31 (Jan. 7, 1943).
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 643
most of the practical range of beating and the slope varies with different
kinds of pulps. When the slope of this line is determined with standardized
reproducible laboratory beating equipment as discussed in Section A-3-f
of this Chapter VIII, it constitutes possibly the best current practical
measure of the beating quality of a pulp.
C. THEORIES OF BEATING
Because of its importance, beating has been the subject of many pub-
lished articles and discussions, but even yet there is not complete unanim-
ity concerning the cellulose-water relationship in beating.
In order to explain the various known effects of beating, hypotheses have
been offered which may be divided into two main groups: chemical and
physical, with the colloidal aspects of the latter receiving increasing sup-
port. Associated with these differing views, respectively, are the names of
Carl Schwalbe,30 James Strachan,31-33 and W. Boyd Campbell,16'84 whose
main theses may be found in the references noted. Also among the more
comprehensive papers on the subject are those of Bell,35-36 Cottrall, 37'w
Harrison,39 Kress and Bialkowsky,40 and Jayme41 and extended treatments
of special phases of the problem by Katz42 and especially Stamm,48 and the
reviews of the subject by Simmonds,44 Rowland,45 Clark,46 and Suter-
meister.47
*> C. G. Schwalbe, Paper Trade J., 72, 58 (Mar. 3, 1921).
81 J. Strachan, Proc. Tech. Sect., Paper Makers' Assoc. Gt. Brit. & Ireland, 6, 139
(1926).
82 J. Strachan, ibid.. 13, 61 (1932); 14, 447 (1933).
88 J. Strachan, ibid., 19, 171 (1938).
84 W. B. Campbell, Paper Trade J., 100, 35 (Feb. 14, 1935).
85 J. H. B. Bell, /. Soc. Chem. Ind.t 52, 109T, 119T (1933).
88 J. H. B. Bell, Proc. Tech. Sect., Paper Makers' Assoc. GL Brit. & Ireland, 15, 401
(1934).
87 L. G. Cottrall, ibid., 14, 241 (1933).
38 L. G. Cottrall, Tappi, 33, 471 (1950). See especially L. G. Cottrall, Introduction to
Stuff Preparation for Papermaking, Griffin, London, 1952.
89 H. A. Harrison, Proc. Tech. Sect., Paper Makers' Assoc. Gt. Brit. & Ireland, 11, 303
(1930).
40 O. Kress and H. Bialkowsky, Paper Trade J., 93, 35 (Nov. 12, 1931).
41 G. Jayme, Australian Pulp & Paper Ind., Tech. Assoc. Proc., 3, 432 (1949).
48 J. R. Katz, Cellulosechemie, 11, 17 (1930).
48 A. J. Stamm, U. S. Dept. Agr., Misc. Pub. 240 (1936).
44 F. A. Simmonds, Paper Trade /., 101, 35 (July 18, 1935).
« B. W. Rowland, Paper Trade J., 101, 98 (Sept. 26, 1935).
46 J. d'A. Clark, Paper Trade J., 97, 25 (Dec. 28, 1933).
47 E. Sutermeister, Paper Ind. and Paper World, 29, 361, 542 (1947).
644 CELLULOSE
1. Chemical Theory of Beating
The chemical theory of beating, which was held by such prominent in-
vestigators as Cross and Bevan,48 Sindall,49 and later championed especially
by Schwalbe,30 suggested that during beating, the fibers were cut and fibril-
lated and at the same time a cellulose hydrate or slime, which was presumed
to be a chemical compound of cellulose and water, was formed round the
particles and provided a strong adhesive which, when dried, cemented the
structure together. Cross and Bevan were somewhat noncommittal about
whether or not a true cellulose hydrate was formed, and chose to term it
"physicochemical" combination, but both Schwalbe and Sindall found
experimentally that beaten pulp was slightly more hygroscopic than un-
beaten pulp and, accordingly, believed that some hydrate was formed.
Subsequently, very precise measurements16'50 of the equilibrium moisture
content of the same pulps, both beaten and unbeaten, have disclosed that
the former usually, but not always, do contain more moisture, that is,
up to 4% of the total moisture present. However, because of no definite
experimental evidence as to the presence of a hydrate, the slight excess of
moisture in the beaten pulps may be very satisfactorily explained on the
basis that intensive beating has opened up the internal structure of the
fibers and provided fresh surfaces to adsorb additional moisture.
Some investigators have suggested, on purely hypothetical grounds,
that a liquid mucilage of carbohydrate material is formed in greater or
lesser quantities when water permeates a fiber and this is squeezed to the
surface when the fiber is subjected to mechanical treatment. Wurz,51
for example, has postulated that pectin-like substances, uronic acids, are a
necessary content of pulps from which well-bound papers like greaseproof
can be satisfactorily made.
The chemical theory of beating was a very comfortable one to the paper-
maker as it could be made to explain satisfactorily almost, if not quite all,
the more practical beating phenomena which have been referred to above.
In particular, by the apparently reasonable postulation that the hydrate
was glue-like in character and increased in quantity as beating proceeded,
a good explanation was provided for the characteristic way in which the
48 C. F. Cross and E. J. Bevan, A Text Book of Paper making, 5th ed., Spon, London,
1920, especially Chapter VII.
49 R. W. Sindall, The Manufacture of Paper, Constable, London, 1908, especially
Chapter IX.
60 S. E. Sheppard and P. T. Newsome, Ind. Eng. Chem., 26, 285 (1934).
11 O. Wurz, Papier-Fabr., 35, Tech.-wiss. XL, 54, 57 (1937); 38, Tech.-wiss. TL, 87
(1940).
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 645
strength of sheets made from cellulose fibers increased with beating.
Indeed, Schwalbe52 not only proposed the use of chemical agents, such as
acids and alkalies, to expedite beating by the formation of a hydrate, but
he even carried out one or two apparently satisfactory commercial demon-
strations.
2. Physical Theories of Beating
The main difficulty in securing a general acceptance of any one physical
theory, on the other hand, has been the absence of an equally satisfactory
and simple explanation of the mechanism of bonding as was provided in the
chemical theory or by the mucilage or glue hypothesis; it is with this point
especially that many investigations and differences of opinion have been
concerned.
The chemical theory was severely criticized by Strachan in 1926.81 He
insisted that the taking up of water by pulp in beating should be termed
"imbibition" and not hydration. He described the water content of paper
stock as follows: (1) water of suspension, in which the fibers float; (2)
capillary water held between the fibers and in their canals and pores;
(3) colloidal water composed of (a) water of "imbibition" adsorbed by the
unbeaten fiber and (b) water of "hydration" or an increase of water of
imbibition due to beating. The colloidal water was stated to be held partly
by adsorption and partly by occlusion.
Strachan described a number of experiments which he conducted on
artificially made fibers having a density of 1.2, prepared from rag paper
moistened with zinc chloride and compressed together followed by thorough
washing ("vulcanized" fiber), and demonstrated quantitatively the effects
of vacuum, pressure, temperature, papermaker's alum, acids, and alkalies
on the rate of imbibition of his material, which was presumed to be anal-
ogous to the fiber of a pulp.
(a) Fibrillar Entanglement
After describing his concept of the internal structure of the cellulose
fiber, confirmed later by others,53-64 Strachan postulated that the layers
of compacted fibrils, being porous, allow water to penetrate and thus cause
the fiber to swell. When the fiber is then subjected to the beating action,
the outer layers are loosened, and the surface becomes fibrillated, there is
easier access df the water to the inner layers, and the rate of water pene-
62 C. G. Schwalbe, Paper Trade J., 80, 46 (Jan. 22, 1925).
*3 G. J. Ritter and G. H. Chidester, Paper Trade J.t 87, 131 (Oct. 25, 1928).
64 G. W. Scarth and J. I). Spier, Trans. Roy, Soc. Can., 23, 281 (1929).
646 CELLULOSE
tration is thus increased. He considered that beating increases the water
content of the fibers by the dual effects of fibrillation, which enables the
fibers to hold more water, and by the pressure to which the fiber is subjected.
For example, increasing the consistency of pulp in a beater results in a
greater effective pressure being applied and, consequently, faster ' 'hydra-
tion" or increase in imbibition. He pointed out that only the surface of
the fibers is affected by prolonged light beating, and under these condi-
tions a strong opaque papqr, such as currency paper, is obtained; whereas,
if the fibers are subjected to intensive beating, the whole mass is bruised
and softened and fibrillation takes place throughout the fibers, giving a
more transparent, greaseproof type of paper. When a sheet of paper is
made, the fibers of beaten stock are soft and fibrillated and lie and adhere
together more closely than unbeaten fibers on the forming wire of the
paper machine. Between the press rolls, the fibers are squeezed and are
drawn in intimate contact according to their degree of plasticity and degree
of fibrillation. As the fibers are dried, the water of imbibition is forced out
and they shrink. The outer layers of the fibrils on the fibers also shrink
together, so that the dried fibers become coated with a layer of denser
structure than was previously the case. The fibers are no longer as porous
as they were before beating, because the air spaces have been largely closed,
producing hardness and transparency in the mass. The compacting of
,and cohesion between the fibers, finally, is somewhat increased by the
subsequent calendering. During the process of beating, some of the fibrils
become detached from the fiber ; Strachan held such material to be useless
from the papermaking viewpoint. He assumed cohesion between the
fibers to be caused primarily by the mutual entanglement of the fibrils
produced by beating, the detailed mechanism involved being described
merely by a statement that it is in accordance with ' 'well-known physical
laws."
Stracban's explanation of the beating act on was held by several critics66
to be incomplete in a number of particulars. Some exception was taken
to his denial of the existence of a true cellulose hydrate, such as has already
been discussed, but more important was the fact that his theory did not
explain the marked increase in strength obtained in paper made from pulp
which was given only a mechanical stirring treatment in water, even though
previously the pulp was thoroughly soaked. Strachan, however, claimed
that even with such a slight treatment, an exceedingly fine degree of fibril-
lation, which was sufficient to account for the increased strength, actually
occurred.
» Discussion, Proc. Tech. Sect., Paper Makers9 Assoc. Gt. Brit. & Ireland, 13, 1 (1932).
V11I. FKUJftCKllKb AJNU IKKATMKJMT OF FUJUP FUK FAFKK
The work of Kress and Bialkowsky40 confirmed the opinion that no ap-
preciable chemical changes occurred with beating except an increase in
the sensitiveness of the pulp to hydrolysis and a slight decrease in the
cuprammonium viscosity, effects which may be ascribed to mechanical
changes. They noted that a beaten pulp sorbs no more methylene blue
dye from an aqueous solution than does an unbeaten pulp. Although
methylene blue is positively charged in solution, it appears to be able to
penetrate the intermicellar structure of the cellulose quite easily.
This finding was later confirmed by Strachan33 and was held by him to
prove that the total (i.e., including the internal) specific surface of the cellu-
lose was not increased by beating. Strachan32 also showed that the re-
tention of positively charged sols like silver oxide, especially when pre-
cipitated in the presence of pulp, increased markedly as beating progressed,
which indicated an increase in the external specific surface.
Kress and Bialkowsky40 furthermore reported the results of a number
of experiments with pulp and various organic liquids. They measured
the swelling action of the different liquids on cellulose and found that the
swelling power corresponded with the degree of beating obtained and with
the strength developed in the resulting paper when cellulose was beaten in
that liquid. No strength or swelling was developed in oil, very little in
alcohol, considerable in ethyleneglycolandin water, and still slightly more
in formamide. No visible fibrillation was developed in oil or alcohol even
after intensive beating. They conclude that " these physical changes are
mainly occasioned by the swelling of the fibers by the liquid in which the
stock is beaten, with the result that there is an increase in the volume of
the fibers. In the wet swollen condition, the fibers are low in strength
but are highly plastic and ductile, and any mechanical action will tend to
bruise and fibrillate the fibers lather than to produce a sharp cutting action.
The hard close nature of the paper made from so-called hydrated stock
seems to be due entirely to the fibrillation of the fibers and the shrinkage
of the fibers when the liquid producing the swelling and shrinking is re-
moved by drying. " It will be noted that this conclusion is in accord with
that previously reached by Strachan.
(b} Partial Solubility of Cellulose
A different view of the mechanism of bonding was presented by Camp-
bell56 in 1932, and most of the subject matter in this and some later articles
on the subject were published in 1933 in pamphlet form.16
* W. B. Campbell, Paper Trade /., 95, 29 (Oct. 25, 1932).
648 CELLULOSE
A few years previously, Urquhart57 had advanced the hypothesis that,
during the formation of cellulose in nature, a precipitate was formed in the
presence of water and, hence, the hydroxyl groups would have molecules
of water attached. As the fiber dried, the hydroxyl groups would be freed
from water and their residual valences would be extensively satisfied by
those of adjacent cellulose molecules. When the fibers were again wetted
with water, some of these bonds would be broken and some of the hydroxyl
groups would reattach themselves to water, giving a looser form of struc-
ture. If this water is driven off again, the groups once more mutually
bind together. This concept was accepted and extended by Campbell
who suggested that, because of their "partial solubility," the molecules,
especially the short-chain molecules, when exposed on the surface of a fiber
by beating, were on the verge of solution. They were thus endowed with a
freedom that enabled the molecules of adjacent fibers to so orient them-
selves that, upon drying, many of their hydroxyl groups could bind together
by means of secondary valence forces.
In connection with "partial solubility," Wislicenus and Gierisch58 found
that, after breaking down pure cellulose fibers by very fine dry grinding,
up to 0.4% of the resulting powder became soluble in water and, because
the amount of ash in the dissolved portion was little more than in the
original paper, it was clear that part of the cellulose itself had dissolved;
chemical tests indicated that the dissolved material had undergone con-
siderable degradation. Strachan published data to show that appreciable
(though very small) quantities of material, from 13 to 21 parts per million,
were dissolved by cold water from a carefully purified cotton cellulose, even
after as many as fifty extractions. Upon evaporation, the hydrolyzed
residue had a reducing action equivalent to about 15% of its weight of glu-
cose. Also, in a discussion of Strachan's paper, Turner69 described experi-
ments he had carried out by dipping a highly purified cotton fabric in con-
ductivity water. The brown line formed where the water evaporated
consisted of material which was soluble in alcohol and which possessed con-
siderable reducing power. The line could be re-formed at lower and lower
levels on the same piece of cloth with undiminished intensity, which indi-
cated that, by some chemical change, the cellulose was being transformed
into a water-soluble material.
There thus exists some indirect experimental support fortheideaof "partial
67 A. R. Urquhart, /. Textile Inst., 20, T125 (1929).
68 H. Wislicenus and W. Gierisch, Kolloid-Z., 34, 169 (1924).
69 H. A. Turner, Proc. Tech. Sect., Paper Makers' Assoc. Gt. Brit. & Ireland, 19, 182
(1938).
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 649
solubility" of the cellulose surfaces in water as expressed by Campbell
and for the concept that some of the molecules on the surface are on the
verge of, or are even completely in, solution, especially those molecules of
cellulose and allied materials having a relatively short chain length. Fur-
thermore, it has been shown that a completely acetylated pulp has little
strength if made in water but has considerable strength if formed in alcohol
or acetone in which it is partially soluble.50 Also, partial methylation61
or partial acetylation60-62 under carefully controlled conditions appreciably
increases the strength of paper formed from the pulp.
Clark63 has ventured an opinion that essentially the primary hydroxyl
groups are involved in the linkages holding the cellulose chains together.
This structure cannot readily be broken by water, but if these linkages
are disturbed by the primary hydroxyl groups becoming either loosely
engaged with cuprammonium complexes or xanthates or by being converted
to carbonyl, ester, or ether groups, the material becomes much more
hygroscopic and even soluble in water, provided that most of the secondary
hydroxyl groups are left substantially intact and are not transformed to
hydrophobic radicals. This, statistically at least, appears probable during
the early stages of conversion of the cellulose to a derivative.
(r) The Mechanism of Bonding
Campbell16 also showed by calculations that, especially with the finer
fibrils, as the water is removed the normal surface tension forces, which
with pure water amount to about 2 X 70, or 140 dynes per centimeter of
filament irrespective of diameter, give rise to enormous compacting pres-
sures, which further increase the surface areas in contact. This explains
why well-beaten stock compacts so easily on drying, giving a hard, dense,
strong paper. That this is the mechanism responsible appears to have
been confirmed by Van den Akker.61 He showed that if an undried pulp
test sheet was frozen and the water removed by sublimation in a freezer,
the resulting dry sheet was bulky, opaque, fluffy, and weak.
Campbell concluded that, when fibers have access to moisture, a layer of
water is adsorbed on every exposed crystal surface and a layer of hydrated
cellulose (in a colloidal sense) is thus produced. The association of the
60 J. C. Bletzinger, Ind. Eng. Chcm., 35, 474 (1943).
01 G. Jayme and D. Froundjian, Cellulosechemic, 18, 9 (1940).
fi2 W. H. Aiken, Ind. Eng. Chem., 35, 1206 (1943).
83 J d'A. Clark, Proc. Tech. Sect., Paper Makers' Assoc. Gt. Brit. & Ireland, 24, 30
(1943); Paper Ind. and Paper World, 25, 382, 507 (1943).
6< J. A. Van den Akker, Tappi, 35, 13 (1952).
650 CELLULOSE
hydroxyl groups with water lowers the attraction of the crystallites for
each other along the natural crevices and cleavage planes in the fiber, the
interior of which, it will be recognized, is normally under tension from
having been dried, so that small passages are split open and additional
water enters, resulting in further adsorption and swelling. Beating bends
and stresses the swollen and softened structure, and the fiber becomes still
more flexible. Fibrillation also occurs, greatly increasing the external
surface of the fibers. When the fibers are made into paper and the water is
removed, the extent of bonding depends on the flexibility of the fibers and
the amount of surface exposed, both being clearly dependent upon the
degree of beating. The degree of orientation of the crystallites in contact
with each other, which also depends somewhat on flexibility, is an important
factor in the degree of bonding; thus, if the crystallites happen to be
parallel and point the same way when they come together, the degree of
bonding will be at its maximum.
In practice, there is a range of bonding from the slight bonding shown
in blotting paper (which is both weak and opaque) to the extreme case of
glassine paper where the transparency, degree of bonding, and water in-
solubility of the bonds are at a maximum for paper.
This explanation of the process of bonding, akin to the phenomenon of
crystallization, was at variance with the conclusions of Strachan,32 who
took exception to it on the ground that the union of crystal surfaces de-
manded that the surface must be either malleable or liquid (e. g., dissolved),
so that the crystal surfaces could be oriented similarly. He stated33 that
the first condition was obtained to a limited extent in the parchmentizing
process (of vegetable parchment) and the second in the case of making
cellophane, but neither condition was fulfilled with cellulose under normal
papermaking treatments. He re-emphasized his opinion that * 'during
beating, the surfaces of the fibers are rendered plastic by fibrillation. The
fibrils of the two beaten surfaces become entwined and bind together.
We have not only intertwining of the fibers themselves, but also inter-
twining of minute fibrils. ... As the external surface of the fibers increases
with beating, so the cohesion between such surfaces increases in the presence
of water which brings them into contact."
However, Strachan 's grounds for insisting upon that mechanism of
bonding exclusively, disappeared when Clark46 published his findings of the
surprising shear strength developed by the bond formed when two sheets
of normal uncoated cellophane, which certainly were not fibrillated, are
wetted, cleaned, then pressed and dried together. He also showed that mo-
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 651
lecular orientation, as Campbell had suggested, played an important part
in their cohesion.
Harrison, in the course of several contributions as a result of many
careful experiments on beating,39 published65 some photomicrographs
showing mildly beaten fibers apparently devoid of fibrillation which
nevertheless formed quite strong sheets of paper and concluded accordingly
that fibrillation was not alone responsible for strength. Indeed, he pointed
out that when fibrillation became pronounced after much beating, the rate
of strength development was then barely increasing. Cottrall38 published
other photomicrographs later which appear to support the same viewpoint.
In his recent book,38 an interesting theoretical arid practical review of
beating and refining, Cottrall continues to minimize the importance of ex-
ternal fibrillation and instead emphasizes wet fiber pliability enhanced by
the presence of hemicelluloses and internal fibrillation brought about by
beating
Edge*5 showed that, when coarse fibrillation was developed by beating a
pulp in toluene, if the toluene was then displaced by water and sheets
were formed, only a small fraction of the normal strength was obtained as
compared with both beating the pulp and forming the sheets in water.
Bell35 gave an account of an extensive series of experiments dealing with
the drainage, shrinkage, and properties of masses or cakes of beaten and
unbeaten pulps and several liquids, together with a discussion of beating
on the basis of chemical, colloidal, and physical properties of the fibers.
In this and in a later paper36 where the question of cohesion is further dis-
cussed, he concluded: ' 'Sheet strength must be due to a loose chemical
bonding between the fibers. . .and is probably effected through the free
hydroxyl linkages in the outer transverse surfaces of the cellulose micelles . . .
Beating exposes . . a greater external surface with its quota of free hydroxyl
groups. If, however, we can obtain sheet strength without much fibrilla-
tion, there must be some other way of liberating these hydroxyl linkages
during beating." He then suggested that the actual existence of the
Ltidtke86 noncellulosic membrane system, enclosing the fiber elements, and
its rupture, would account for this, and was inclined to think that fibrilla-
tion played a subordinate part.
(d) The Surface of Fibers
It will be noticed that, with the possible exception of the mucilage-
66 S. R. H. Edge, Proc. Tech. Sect., Paper Makers' Assoc. Gt. Brit. & Ireland, 16, 273
(1935).
66 M. Liidtke, Cellulose >chemie, 13, 169, 191 (1932); 14, 1 (1933).
(552 CELLULOSE
formation hypothesis which so far has no valid experimental basis, none of
these physical theories accounts satisfactorily for the considerable rate of
increase of strength always noticed during the preliminary stages of beating.
To investigate the problem further, Clark67 devised a method for de-
positing a fine film of metallic silver on wet cellulose fibers. This permitted
measurements of the specific surface of beaten and unbeaten pulps to be
made with a reasonable degree of accuracy (TAPPI method T 226 sm-52).
In addition, the finer structure of beaten fibers, including the ultia-fine
fibrils described by Strachan, could be then observed with ease. The
silvering step may be recommended as an excellent morphological stain
for the microscopy of cellulose, since many normally invisible details thereby
become resolved. The colors vary from pale yellow through a range to
dark brown and black.
Figure 3 shows an interesting microphotograph of two bleached sulfite
fibers from a sample of pulp which was subjected to rapid stirring only,
at 3% consistency for ab'mt 100 minutes, corresponding to a very mild
beating treatment. A thin film of silver was then deposited on the fibers,
just thick enough to cover the surfaces and make the very fine fibrils visible;
this also revealed the structural details of the loosened primary wall
material which, because of its thinness, is very difficult to see under usual
conditions. This sheath, which apparently has less reducing power than
the body of the fiber at this initial stage of silvering, is colored a golden
yellow by transmitted light, whereas any fibrils on the secondary walls
are almost opaque. Impressions of the bordered pits may be seen on the
loosened sheath on one of the fibers in Figure 3. Where the sheath is still
on the fiber or where it is folded over itself, there is no trace of fibrils or
fuzz on its surface; on the other fiber, the skin of which apparently has
been rubbed off, there is a uniform covering of very fine fuzz. Without the
silvering operation, which requires a high degree of skill, this fuzz is ex-
ceedingly difficult to resolve and requires a microscope with a high numeri-
cal aperture and critical illumination. The presence of the fine fuzz on
the body of the fiber and its absence on the outer sheath have since been
confirmed with the electron microscope. It is this fuzz which Strachan
maintained was present on lightly beaten fibers. In this connection neither
he nor others had previously described the functions of the primary
wall or outer sheath before or after beating, the freedom of the surface of
the fibers from fuzz or fibrils prior to its removal, and the subsequent auto-
matic appearance of the very fine fibrils on the secondary wall beneath.
67 J. d'A. Clark, Paper Trade /., 115, 32 (July 2, 1942).
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER
053
There has been some confusion in the literature between the identities
of the primary wall and the outer layer of the secondary wall. Wardrop
and Dadswell68 have recently given a clear summary and account of the
Fig. 3. Portions of two bleached suliite fibers from western hemlock after
long stirring and after being lightly silvered (Clark67).
matter with photomicrographs of the structure. It may be mentioned
that the constrictive spiral windings observed on solvent-swollen wood
pulp fibers that have not been overcooked or overbeaten are derived not
68 A. B. Wardrop and H. E. Dadswell, Australian Pulp & Paper Ind., Tech. Assoc.
Proc., 4, 198 (1950); 5, 204 (1951); Holzforschun^ 7, 83 (1953) (in KnKli-h*.
654 CELLULOSE
solely from the primary wall but from the outer layer of the secondary
wall, which is a well-oriented, strong, crystalline, fibrous cellulose structure.
On the other hand, the primary wall is usually a weak, somewhat brittle
membrane, which, as Wise and others69 have recently indicated, appears to
consist largely of mannans. It is probably incapable of causing the ob-
served constrictions and is incapable of fibrillating although, of course, it
may be shredded.
By relating the action of a known weight of silvered fiber in the cat-
alytic decomposition of a standardized dilute solution of hydrogen per-
oxide with the action of a measured area of small shreds of cellophane sheet-
ing also silvered, it was found, as mentioned, that the external specific
surface of the pulp increased in proportion to the logarithm of the amount
of beating, that is, very rapidly in the initial stages of beating, conespond-
ing closely with the rapid increase in the strength of the resulting papers in
those early stages. Notwithstanding the mildness of the treatment to the
fibers shown in Figure 3, their specific surface was 36% higher, and the
tensile strength of sheets made from them 270% higher, than the pulp
before treatment.
3. Composite Theory of Beating
Considering these later observations with others, especially those of
Strachan, Campbell, and Bell, the following modified theory of beating
was formulated by Clark.63
During the chemical and mechanical treatments involved in the prepara-
tion of pulp, the primary wall of the fiber, which is permeable to but is not
swollen by water, is partially cracked, rubbed loose, or removed to expose
some of the underlying surface of the fiber. In the case of wood pulps,
this underlying surface is the spirally wound outer layer of the secondary
wall.
When water enters the interior of the secondary wall (body of the fiber)
the fiber commences to swell to almost the original size it had in the living
plant because the water penetrates the voids between the micelles and
breaks a number of bonds holding the structure tightly together, possibly
those between the secondary hydroxyl groups by combination and those
involving the primary hydroxyl groups by the splitting action of the
elements as the fibers swell.
The rubbing and the partial solvent action of the water almost immedi-
ately form a kind of two-dimensional colloidal suspension of the cellulose
69 L. E. Wise, J. W. Green, and R. C. Rittenhouse, Tappi, 32, 335 (1949).
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 655
and other carbohydrate material on the exposed and interior wetted sur-
faces in the amorphous regions of the secondary wall. The "concentration"
of the "suspension" is believed to be dependent inversely on the degree of
polymerization of the cellulose and allied material that is exposed. This
state is regarded as a colloidal system in which the particles are fixed or
anchored at one end but otherwise subject to all the usual laws of colloidal
behavior. The surfaces of these particles, bearing hydroxyl groups,
are strongly hydrophilic and attract and bind a number of consecutive
layers of water molecules which, especially near the solid surfaces, are
well oriented and closely packed.70 It is this water envelope that is
the so-called "water of hydra tion" and that gives rise to the various
phenomena associated with it. The molecules of water adjacent to the
cellulose surfaces are probably held by hydrogen bonding to the hydroxyl
groups; if a hydrogen bond is regarded as a chemical bond, then this
portion of the water may be regarded as being chemically bound. The
primary wall is not in any way water soluble, and in fact Brauns and
Lewis71 found that the material from the outer surface of wood pulp fibers,
presumably the primary wall, was not easily soluble even in cuprammoriiuin
solution. Also, unlike both the outer layer and inner body of the secondary
wall, the primary wall is largely membranous arid docs not fibrillate, at
least on its exterior surface, so that whatever portion of it remains, forms a
smooth, protective layer on the outside of the fiber. When the fibers are
almost wholly covered with this layer, as is usual in the unbeaten state,
adjacent fibers in a wet web of paper are prevented from appreciably
adhering together on drying, not only because of the absence of a surface
"suspension" on the sheathed part of their exteriors but also because the
diameters of the fibers are large compared with the diameter of fibrils
subsequently produced by beating. Stiff elements result from large diam-
eters; however, since the surface tension per unit length of element is inde-
pendent of its diameter, only a relatively small compressive effect on the
sheet is noticed with unbeaten stock when the water is removed on drying.
To some extent the greater rigidity of the wet fibrous structure also may be
the result of the presence of part of the primary wall, and this factor too
would prevent the fibers from coming into intimate contact when the sheet
is dried.
As beating proceeds, accompanied by the rubbing off of the brittle pri-
mary wall and the further wetting and swelling of the fibers, the underlying
material becomes more coarsely fibrillated ; this not only permits stronger
70 O. Maass and W. B. Campbell, Pulp & Paper Mag. Can., 40, No. 2, 108 (1939).
71 F. E. Brauns and H. F. Lewis, Paper Trade J., 105, 35 (Sept. 2, 1937).
656 CELLULOSE
surface tension effects to compact the sheet better, but also results in a
greater extent of bonding surface. The rubbing probably also increases
the "concentration" of the surface "suspensions," and any cellulosic ma-
terial that is rubbed off entirely serves as an adhesive filler between the
interstices of the larger fibers. Some of the very finest material may
possibly also function as a "protective colloid," forming a more stable sur-
face "suspension" by preventing long molecules or fibrils which have been
raised from the surface of the fibers from re-embedding themselves.
When beating exceeds a certain point, a state is reached where the in-
crease in bonding material and facilities for bonding are offset by decreased
fiber length and a weakening of the fibers themselves by mashing and reticu-
1 ition . At this point the pulp reaches the maximum strength.
This composite theory appears to account for all the observed facts
known at present about the beating action. The picture includes the fol-
lowing operations to the water-immersed fibers : ( 1 ) removing the primary
walls, at least in part; (2) loosening underlying elements of the fibers and
thus enabling subsequent interlocking of these elements; (3) softening and
swelling of the fibers and their fibrils so that they will better fill the space al-
lowed by their neighbors and thus increase the area of contact between
them ; and (4) creating smaller bodies, either attached to or separated from
the fibers, which can nest in the crannies between adjacent elements, thus
further increasing contact. The ultimate mechanism of cohesion of the
changed structure may be considered a chemical one, if the nature of hy-
drogen bonding is defined as chemical. Because of the relative spatial
positions of the adjoining cellulose chains, presumably the primary hy-
droxyls play the decisive role in the bonding between them. If the hy-
droxyls of these chains are hydra ted, they do not have much residual at-
traction for one another. If, however, hydrated chains in close contact
with one another lose their water of hydra tion, they will seek to establish
hydrogen bonding with the hydroxyl groups of their immediate neighbors,
whether these neighbors belong to the same or to another fiber, thus bond-
ing the fibers and also the fibrous structure together. The great impor-
tance of water for this process is evident.
When the bonding is between the more flexible fibrils and smaller ele-
ments, the sheet on deformation can adjust itself so that some stress is
thereby placed on many bonds before overloading the most stressed. Even
if unfibrillated whole fiber surfaces could be made to adhere as strongly to
one another, the slightest movement of adjacent fibers after drying would
put large strains on the joints because of the rigidity of the thick fibers.
In consequence there would be a correspondingly low resistance to a break-
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 657
ing of the structure of a paper which depended for its strength solely on
adhesion between surfaces of the fibers rather than also upon interpenetra-
tion and cohesion of their smaller and more flexible fibrous elements.
D. RESULTS OF BEATING
The phenomena and changes which take place when pulp is beaten and
when beaten pulp is made into paper are many. The more important re-
sults of beating can be grouped, for purposes of discusssion, according to
the effects on fibers, pulps, and paper. These and some miscellaneous fac-
tors will be discussed in the light of what has been said regarding the action
of the beater and the foregoing composite theory of beating.
Artificial cellulose fibers do not respond at all to beating because of their
solid structure, which cannot fibrillate. Mechanical or groundwood pulp
and most "semichemical" pulps (that is, pulps which have been prepared by
giving the raw material a mild chemical cook with a resulting yield of over
70%, followed by mechanical reduction to fibrous elements) do not have
their structures sufficiently delignified or loosened by the cooking to
fibrillate easily. In consequence, these do not respond well to beating;
they are merely comminuted. The following discussion applies to natu-
rally occurring cellulose fibers from which the lignin has been substantially
removed by one of the conventional pulping processes.
1. Effects of Beating on the Fibers
(1) Swelling. An appreciable swelling of the fibers occurs, until the
voids and accessible amorphous regions are thoroughly saturated with
water. This swelling, of the order of 20 to 30%, does not noticeably in-
crease with beating until the internal fiber structure is loosened, when the
fibers may swell rapidly to twice their original diameter.28
(2) Rubbing. At least a portion of the primary wall, the relatively brittle
sheath that surrounds the fiber, is removed.63 Indeed, usually quite no-
ticeable quantities are removed by the pumping, agitating, and other me-
chanical actions to which the fibers are subjected after pulping. This ex-
poses an exceedingly fine, almost invisible microscopic fibrillation or fuzz
on the underlying secondary wall as shown in Figure 3.
Observations indicate that the primary wal1 on acid-cooked sulfite pulp
is more brittle and comes off more readily than that on alkali-cooked kraft
or soda pulps. This may well be a contributing factor in the slower rate of
beating which is a characteristic of alkaline pulps in general.
(3) Cutting. Wet fibers are not very strong in shear or in tension and may
658 CELLULOSE
be cut by the intense pressure developed across clots or fibrages caught by
the bars, especially when they are thin or sharp. When the fibers are
well rubbed, bruised, and swollen by gentle preliminary beating, they
become softened, more slippery, and will stretch more. They do not then
cut so easily as do those from a dried pulp that is insufficiently swollen
before being subjected to intense bar action.
(4) Splitting. The fibrous structure is ultimately unraveled or split, with
the formation of fibrils which are visible under relatively low magnifica-
tion. The manner in which fibers split and fibrillate depends upon their
structure. Bast fibers like linen split readily into long, fine fibrils; wood
pulps, which have their fibrils lying more or less circumferentially, espe-
cially on the outer layer of their secondary wall, tend to produce coarser and
shorter fibrils. Many of these break off and, together with the particles
from the primary wall and short cells, form a debris which, being mobile,
tends to plug the pores in a mass of pulp very effectively when water at-
tempts to flow from or through the mass but to a lesser extent when the
mass is vibrated as on a paper machine wire. The newly exposed surfaces
immediately adsorb a layer of water.
(5) Bruising. Internal splitting (i. e., bruising) results in the penetration
of water, which thereupon becomes adsorbed on the newly opened sur-
faces and keeps them from recrystallizing together again. In consequence,
the fibers become more limp and flexible in their wet state. When the
fibers are dried, the sorbed water is removed, and most of the split fiber sur-
faces and some of the previously existing voids or canals may "heal" to-
gether again so that the fiber becomes as stiff as or stiffer than before.
(6) Deformation. When the consistency of the pulp is high and with
equipment that produces a rolling or twisting action, many of the fibers
become more or less permanently curled or bent (i. e., deformed) because
of their appreciable plasticity when moist.
2. Effecbf of Beating on the Pulp
(7) Surface. According to measurements made by the silvering technique,
there is a rapid increase in the specific surface of the pulp, especially
during the early stages of beating. This rapid increase appears to be
due to the removal of the primary wall; obviously its complete removal
would triple the exposed surface. When using the liquid permeability
method of measuring the specific surface, Robertson and Mason72 found
that the surface increase is slow at first and rises more and more quickly as
78 A. A. Robertson and S. G. Mason, Pulp & Paper Mag. Can., 50, No. 13, 103 (1949).
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 659
beating progresses. These findings, based on the Kozeny-Carman73 equa-
tion for flow, seem incompatible with microscopic observation, especially if
the silver staining technique or the electron microscope is used, and incon-
sistent with the corresponding very rapid increase in strength during the
early stages of beating.
(8) Greasiness. The pulp develops a slimy or greasy feel; it also flows
more readily around the tub of an ordinary hollander, and its surface looks
smoother. These effects are due to the fibrillation and bruising of the
fibers, with the subsequent sorption of water on the freshly made surfaces
and crevices of the fibers and their consequent swelling, and also to the re-
duction of length of the fibers.
(9) Wetness. There is an increase in the time taken for water to drain
away from a mass of wet pulp placed on a screen and an increased resistance
to the passage of further water through the wet mat. The amount of
water retained by a given dry weight of pulp under a given pressure has
been reported31 to be greater for beaten than for unbeaten pulp. Several
observers,35'74 however, have found that the quantity of water retained by
either pulp under the same pressure is the same, provided that enough time
elapses for equilibrium to be reached.
(10) Flocculation. When stirred in a dilute suspension of water, well-
beaten pulps usually, but not always, disperse more readily than unbeaten
pulps; after standing, these pulps do not floe as readily. Fibrillation de-
veloped during beating promotes clotting of the fibers in suspension because
of the resulting mechanical entanglement upon chance contact. However,
this effect is usually more than offset by the concurrent reduction of fiber
length during beating which enables the fibers to move and turn in suspen-
sion more freely without coming into mutual contact.
(11) Sorption. The sorption of most basic dyestuffs, such as methylene
blue, in aqueous solution is scarcely affected,32'40 but with certain direct
dyes (such as Purpurin 4B) and also alumina sol and other positive col-
loids, some increase in sorption does occur, probably, in part at least, be-
cause of the increased external surface.
(12) Equilibrium Moisture. Beating increases the equilibrium moisture
content of the pulp only very slightly. 16'50|75~~77 It may be mentioned that
most cellulosic materials contain approximately one-eighth additional
78 P. C. Carman, Trans. Inst. Chem. Engrs. (London), 15, 150 (1937).
74 A. Sedoff, C. V. Holmberg, and E. C. Jahn, Paper Trade J., 109, 42 (Dec. 28, 1939).
76 W. B. Campbell and L. M. Pidgeon, Pulp 6f Paper Mag. Can., 20, No. 6, 195 (1930).
7« C. O. Seborg and A. J. Stamm, Ind. Eng. Chem., 23, 1271 (1931).
77 J. K. Russell, O. Maass, and W. B. Campbell, Can. J. Research, 15B, 13 (1937).
660 CELLULOSE
moisture at a given atmospheric humidity and temperature if approached
from a wetter instead of from a drier state. As a rule also, the purer the
cellulose is, the less is the moisture content under the same conditions.
Some increase in the sorbed moisture would be expected with the greater
surface exposed by the splitting of the fibers.
(13) Copper Number and Viscosity. The effect of beating on the pulp
properties measured by cuprammonium viscosity (degree of polymeriza-
tion) and copper number (end groups) is open to question. Some observers
report that intense beating slightly reduces the viscosity of a solution of
the pulp in cuprammonium hydroxide, increasing at the same time the cop-
per number.78 Negative results have been obtained by other observers.
It may be concluded, therefore, that the effect of beating on the chemical
characteristics of pulp under normal conditions is not appreciable. How-
ever, if the mechanical action is sufficient to disrupt some of the molecules,
as happens with dry grinding of pulp, a slight increase in both of the above
properties would be expected, as has been found.68
(14) X-ray Pattern. The x-ray diffraction pattern of the cellulose is un-
changed. It has been shown that a very drastic treatment such as mer-
cerization is required to change the pattern,79 so that beating could hardly
be expected to have an effect.
(15) Concentration of Inert Substances. The concentration of inert sub-
stances, such as sugar, dissolved in water in which the fibers are placed is
not appreciably changed by beating.36 This, incidentally, is not a very
sensitive method for the detection of combined water in the presence of ad-
sorbed water. Beating may alter the pH of the suspension appreciably, es-
pecially of an alkaline pulp, by squeezing out sorbed chemicals from the in-
terior of the fibers.
( 1 6) Zeta Potential. The electrokinetic potential of the pulp is stated to
increase.80 The difficulty of making such measurements with precision
has been discussed by Mason.81 In addition, the exposed surfaces of the
fibers increase greatly, and the possible influence of the removal of the sur-
rounding primary wall is uncertain. For these reasons, the reported results
would appear to require confirmation.
(17) Heat of Wetting. The heat of wetting of dried cellulose when placed
in water is not appreciably affected by beating and may be decreased, but
crushing cellulose while dry increases the heat of wetting by about 25%. ^
78 C. E. Curran, F. A. Simmonds, and H. M. Chang, Ind. Eng. Chem., 23, 104 (1931).
79 C. Trogus, Verein Zellstoff- Papier- Chemiker u. -Ingenieure, Jahresber., 1928, 140.
80 K. Kanamaru, /. Soc. Chem. Ind., Japan, 34, Suppl. binding, 39 (1931).
81 S. G. Mason, Tappi, 33, 413 (1950).
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 661
This would indicate that fresh internal surfaces opened up by beating are
mostly closed again on drying and that the reformed bonds between adja-
cent cellulose micelles are not capable of being opened again by wetting. In
consequence, if pulp is beaten before drying, it is considerably less reactive.
3. Effects of Beating on the Paper
(18) Density. The fibrous structure of paper made from beaten stock
will shrink more on drying and give a harder, more dense, and less perme-
able sheet. This phenomenon, which increases in proportion to the
logarithm of the length of the beating time,8'29 is due to the production of
the finer fibrils, the softening of all the elements, and the drawing together
of these elements during drying by surface tension. Whether the sheet is
permitted to shrink freely or is dried under tension does not appear82 to in-
fluence the apparent density of the structure if the thickness of the sheet is
measured with a micrometer having a ballpoint pressure foot.
(19) Opacity. The opacity of the paper decreases. Very well beaten
pulp forms a paper quite translucent and, incidentally, one less pervious to
oil (e. g., greaseproof paper or imitation parchment). Additional beating
ind heavy calendering form a transparent paper (e.g., glassine). The opac-
ity of a paper is dependent, among other things, on the extent of the solid-
itr interfaces present in its structure. These in turn depend on the total
specific area of its components less the area in solid contact. With beating,
the latter increases more rapidly than does the former.
'20) Color. A given quantity of dyestuff will give an appreciably greater
ntensity of color to finished paper made from beaten stock, and a white
»heet will appear to be darker than one from the same stock lightly beaten.
This follows from the changes noted in the preceding paragraph.
'21) Sizing. The degree of sizing of paper (i. e., resistance to aqueous
luids) with a given quantity of sizing materials, such as rosin size and paper-
naker's alum, improves markedly to begin with because of the greater
compactness of the wet structure. However, a very well-beaten stock, such
is that prepared for glassine, becomes difficult to size, probably because of
he very low retention of any added materials other than sugars or poly-
iaccharides, especially if the added materials like sizing agents are hydro-
>hobic. Even starch is then poorly retained.
22) Strength. Paper made from beaten pulp will have a higher tensile
ind bursting strength and will stretch more before rupture, for reasons dis-
82 G. F. Glover, P. F. Ray, and E. J. Pritchard, World's Paper Trade Rev., 135, 51
1951).
662 CELLULOSE
cussed earlier in this chapter. Excessive beating under certain conditioi
will reduce the bursting strength from its maximum value but seldom tl
tensile strength. Except for a rise in the resistance of the paper to tear, £
is frequently observed during the early stages of beating in practice, th
tearing strength progressively diminishes, mainly because of shortened f
bers and increased sheet density. As beating increases, the ability of th
paper to withstand repeated folding rapidly rises to a maximum, then ust
ally diminishes after the paper becomes brittle.
4. Factors Affecting the Rate of Beating
(23) Chemical Composition. Fibers that contain considerable quantitie
of hemicelluloses beat more rapidly than purified fibers. Pulp hydrolyze<
with acids also beats faster. These and other chemical matters are dis
cussed later in this section.
(24) Temperature. An increase in the temperature of the pulp decrease
the rate of beating; conversely, low temperatures promote more rapi(
beating.83'84 Also, when beaten stock is heated especially to a high tern
perature, its wetness is considerably reduced.85 These effects which hav<
been well known in practical papermaking for a long time, are both prob
ably concerned with the fact that, as with most other adsorption phenom
ena, a decrease in temperature enhances the adsorption of water on ex
posed cellulose surfaces and thus promotes greater swelling and fibrillation
especially microfibrillation.
(25) Effect of Predrying. Pulp that has been dried is much more difficuli
to beat and, with treatment identical to an undried pulp, results in a weaker
softer, bulkier, and more opaque sheet. When pulp is derived from papei
or "broke" by repulping, these effects are accentuated. It seems likelj
that when cohered dried fibers are forcibly parted, especially if well beater
and if not well soaked in^water, the most efficient fibrils and microfibrilj
(that is, the ones that took part in the previous bonding) are stripped ofl
their surfaces.68 Furthermore, the drying of beaten fibers also causes theii
internal splits and voids to recrystallize or heal, giving a more compact and
stiffer structure. Considerable mechanical treatment is required to break
a substantial number of these internal bonds, to permit the penetration oi
water again, and to refibrillate the surfaces. During the course of this
83 A. Noll, Papi.er-Fabr.t 35, Tech.-wiss. TL, 393, 401 (1937).
*4 O. K. Ronney and C. E. Libby, Tappi, 34, 223 (1951).
86 T. R. Le Compte. Paper Trade /., 93, 42 (Oct. 1, 1931).
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 663
treatment, the fibers are further cut and shortened so that the drying of
moist pulp constitutes practically an irreversible process.
(26) Beating in Other Liquids. The progressive decrease in the strength
of the resulting paper when pulp is beaten in liquids progressively less polar
than water, described by Strachan,86 Kress and Bialkowsky,40 and Edge,85
has already been mentioned. The beneficial effect of formamide on beat-
ing has been confirmed by Musser and Engel.87 During beating, the degree
of "peptization" or dispersion of the elements on the surfaces of the cellulose
will depend on, among other things, the attraction of the liquid to the sur-
faces of the hydrophilic cellulose. Oils and benzene are attracted not at
all, whereas formamide is even more strongly attracted than water.
5. Phenomena Relating to Beaten Pulp
(27) Effect of Temperature. Clark46 has found, contrary to the findings
of Nakano as discussed by Le Compte,86 that the strength and other quali-
ties of the resulting paper may not be impaired by heating a beaten stock.
Also, the wetness of the stock is not recovered by later soaking it for several
days, but it may be recovered, at least in part, by vigorous stirring of the
stock at normal temperatures.
Jayme41 found that the degree of swelling of pulp was decreased by boil-
ing for an hour and that the strength of the sheets was somewhat dimin-
ished. The wetness of the pulp was not greatly improved by soaking the
boiled pulp in cold water for a day. Lyne and Gallay88 also found that
heating a beaten pulp as a suspension and again also heating the test sheets
after forming and especially after pressing them, resulted in a decrease in
tensile strength and an increase in the bulk and the tear. It is evident that
wetness of pulp will be decreased by boiling, because the surface * 'suspen-
sion" would be dehydrated and collapse and, if it is conceded that the sur-
face tension effect of the water as it is removed from between the fibers on
drying is sufficient to raise most of the fibrils and smaller components of the
"suspension," this may be the reason why boiling does not appreciably af-
fect the strength of the resulting paper in some instances. If the fibers in
a moist sheet are heated and are left with insufficient water between some
elements to re-form the suspension, and are then dried, it is clear that the
density of the paper will be less, and consequently the tensile strength will
be decreased and the tear increased.
86 J. Strachan, Proc. Tech. Sect.. Paper Makers9 Assoc. Gt. Brit. & Ireland, 6, 181
(1925).
w D. M. Musser and H. C. Engel, Paper Trade J., 115, 33 (Aug. 20, 1942).
" L. ML Lyne and W. Gallay, Tappi, 33, 429 (1950).
664 CELLULOSE
(28) Addition of Inert Materials. The incorporation in pulp of any filler,
such as china clay, calcium sulfate, or chalk, reduces the strength of paper
made from it. The addition of oils or waxes has the same effect, but to a
greater degree. Resin size weakens the sheet except sometimes with an
unusually weak, lightly beaten pulp when the adhesion provided by the
resin exceeds the natural cohesion of the fibers. Sizes made from bitumi-
nous emulsions have a similar weakening effect on strong, well-beaten pulp
but similarly may improve the strength of weak pulps such as from old pa-
per often used for making some kinds of boxboards, after scarcely any
beating. On the other hand, the addition of certain hydrophilic materials,
such as locust bean and guar gums,89 methyl90 and carboxymethyl9 1 celluloses,
soluble resins,92 hemicelluloses,93 and cooked starches,94 improves the
strength of the resulting sheet unless the pulp is very highly beaten.
These effects follow because the natural bonding of well-beaten fibers by
fibrillar entanglement and recrystallization is a very efficient process. The
addition of suitable adhesive material will supplement the natural bonding
if the pulp is not well prepared, but otherwise it will be a hindrance to the
natural bonding.
(29) Addition of Electrolytes. Cohen,95 in a recent careful study, found
that monovalent cations added to the pulp cause a slight increase in the
strength of the sheet, divalent cations have little effect, and tri- and tetra-
valent cations cause a substantial reduction in strength. Water flows
through pulp more readily in the presence of electrolytes, the effect increas-
ing with valency of the cation. The adverse effect of papennaker's alum
on sheet strength and on wetness has been known for a long time. Also, it
may be observed that when alum is added to the tub of a beater, the level
of the stock in front of the roll is lowered appreciably because the mass be-
comes less fluid.
These results could be expected from a knowledge of the flocculating ef-
fects of polyvalent cations oji hydrophilic colloid suspensions, if the same
principles are applied to the two-dimensional "suspension" on the sur-
faces of the fibers. The weakening effect of the trivalent cations is prob-
89 B. W. Rowland, Paper Ind. and Paper World, 27, 1398 (1945).
« D. M. Musser and H. C. Engel, Paper Trade J., 115, 85 (Aug. 20, 1942).
*l S. R. H. Edge, Proc. Tech. Sect., Paper Makers' Assoc. Gt. Brit. & Ireland, 27, 189
(1946).
M C. G. Weber, M. B. Shaw, M. J. O'Leary, and J. K. Missimer, Paper Ind. and Paper
World, 30, 83 (1948).
" L. E. Wise, Paper Ind. and Paper World, 29, 825 (1947).
»4 J. P. Casey, Paper Ind. and Paper World, 26, 1277 (1945).
w W. E. Cohen, Paper Trade /., 132, 19 (June 22, 1951).
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 665
ably due to their causing a precollapse (analogous to precipitation) of the
microfibrils on the fibers. Thus, when the microfibrils are collapsed and
the fibers are placed and dried together, even the surface tension effect of
the departing water cannot re-elevate the normal quota of microfibrils to
form the "suspension11 to bond with that on neighboring surfaces. The
slight strengthening effect observed with monovalent cations may be due to
their displacing the divalent cations (calcium) normally present, and thus
facilitating the "suspension" of the microfibrils.
For the same reasons, materials which have strong flocculating and dis-
persion actions on hydrophilic colloids also profoundly affect the strength of
paper. It is known that the addition of tannin96 has a marked depressant
effect on the strength of paper and on wetness of the pulp. On the other
hand, Cohen and others in his laboratory 97 found that sodium hexameta-
phosphate in concentrations of about 0.0015 N improved the strength of
several pulps in the order of 15% and the air imperviousness of resulting
paper as much as 70%.
(30) Bacterial Action. Keeping moist beaten stock sterile and wet for
months at normal temperatures does not appear to alter its character.98
However, keeping beaten pulp for a day or two in a stock chest (i. e., a
large vat), where it is subject to bacterial action, usually reduces its wet-
ness and makes it drain faster on the paper machine; keeping unbeaten
stock in a chest under the same conditions usually increases the ease with
which it can be beaten.
It appears that when cellulose-consuming bacteria are present, they
would first consume the microfibrils from the surfaces of the beaten pulp
and so reduce its wetness and the strength of the paper. In the case of an
unbeaten stock, it is conceivable that under certain conditions the bacteria
would weaken or even cleave the internal structure of the fibers and permit
easier separation of its components and consequently faster beating.
E. PULP CHEMISTRY AND PAPERMAKING PROPERTIES
In the cooking and bleaching of pulps and in certain chemical tests where
complete solution is not attained, the chemical operations are progressive
in depth (topochemical), and the results depend to a degree on the chemical
" H. P. Dixon, Jr.. Paper Trade /., Ill, 29 (July 18, 1940).
07 W. E. Cohen, Gwenneth Farrant, and A. J. Watson, Paper Trade /., 133, 16 (July
27, 1951).
M Second Report of Pulp Evaluation Committee to the Technical Section, Paper Makers'
Assoc. Gt. Brit. & Ireland, London, 1936, especially p. 85.
666 CELLULOSE
and physical nature of the successive layers accessible to the chemical re-
agents. Other important factors are the heterogeneity of the chemical con-
stituents of most pulps, varying from highly active to inert, and the dis-
proportionate ease with which the amorphous zones are chemically at-
tacked or dissolved as compared with the crystalline zones. Another
complication is the empirical nature of the tests themselves, the results of
which are appreciably altered by differences in the chosen times and tem-
peratures of reaction and the concentrations of the reactants. Papermak-
ing properties of individual pulps reside mainly in the exposed surfaces of
the fibers. This is especially so under modern conditions where most of
the exterior of the fibers are not extensively split open, so that it is difficult
to deduce papermaking properties from the chemistry of the entire fibrous
bodies which, on the contrary, derives mainly from the interior compo-
nents. Moreover, because of the nature of the paper structure, a number,
perhaps not less than five, of distinctly different properties of a pulp (de-
scribed earlier in this chapter under "Pulp Testing'*) must be determined
and considered together before an unknown sample can be characterized for
its papermaking qualities with any degree of certainty. Accordingly, it
would seem reasonable to suggest that mere chance plays quite an impor-
tant role in many and varied correlations which have been found to exist
between one or two individual chemical tests and the papermaking proper-
ties of the pulps, especially since exceptions are so frequently found.
As far as papermaking is concerned, the main function of applying
chemistry to pulps is to provide an understanding of the reasons underlying
differences in their physical properties or to furnish a basis for process con-
trol rather than to provide for an actual evaluation of the pulps. This
limitation should be kept in mind during the following discussion of the
three main groups of chemical tests which have proved the most interesting.
1. Degree of Polymerization
The tests related to degree of polymerization (D.P.) include viscosity,
chain-length distribution, alpha-cellulose content, and copper number.
It may be said that, in general and in accordance with Schur and Lewis,"
sulfite pulps with a high D.P., that is, around 1500 (which corresponds to a
TAPPI viscosity of a 1% solution in cuprammonium of about 70 centi-
poises), resist cutting during beating and can be well fibrillated before be-
coming too short. The pulp probably would be especially suitable for cur-
rency and high-class papers such as for records of deeds where permanency
•• M. O. Schur and H. F. Lewis, Tappi, 33, 392 (1950).
VIII. PROPERTIES AND TREATMENT OP PULP POR PAPER 667
of the paper is also an important factor. Pulp with a low D.P., say 600
(10 centipoises, TAPPI) or below, will not be resistant to cutting and will
beat much faster. The resulting paper will be relatively soft, bulky, weak,
absorbent, and opaque and suitable for featherweight papers often used for
novels. This is so to some extent even if the low viscosity is achieved by
degrading the pulp with a mineral acid,100 but it should be emphasized that
if the D.P. of a pulp has been reduced by a greater degree of cooking and in
another case "artificially" reduced by acid hydrolysis, their characters will
differ considerably. The entire cellulose structure, including existing or
potential fibrils, will be weakened uniformly throughout by treatment
with the acid. A pulp with an intermediate D.P. is desirable for papers
having a balance of properties, good strength, and good formation (which
in practice requires that the pulp be shortened somewhat during beating),
for example, fine writing papers.
Corresponding figures for the D.P. of sulfate pulps, or of alkaline-pre-
pared pulps having similar characteristics as far as beating and strength are
concerned, are much lower, perhaps only half as much.101 Jayme41 holds
the view that the surfaces with the highest D.P. form the strongest bonds,
and that well-beaten sulfate pulp is much stronger than well-beaten sulfite
pulp of the same D.P. because the D.P. of surface material on the sulfite is
much lower. However, this view does not appear to account for the fact
that if the two pulps were given only a small degree of beating, under ordi-
nary circumstances paper made from the sulfite pulp would be much the
stronger, as would be expected from the composite theory of beating previ-
ously discussed, since the sulfite pulp, having the lower D.P., would ac-
cordingly provide surfaces with greater cohesiveness.
A highly purified wood pulp, that is, one high in alpha-cellulose content,
is exceedingly difficult to beat.102 This too would be expected from the
composite theory because of the relative difficulty of forming a substantial
surface suspension with only long-chain molecules, the short ones having
been removed by the purification process. On the other hand, rag pulps
which have the same or a higher alpha-cellulose content, especially if from
100 D. M. Musser and H. C. Engel, Paper Trade J., 113, 31 (July 10, 1941); 114, 29
[Apr. 9, 1942).
101 If chips from the same wood are pulped to the same degree of lignin removal, de-
spite the fact that the viscosity of the sulfate pulp will be substantially less than that of
the sulfite pulp, the ^ero-span tensile strength of the former will be significantly higher.
This evidence suggests that the factor used for converting the viscosity of differently
prepared pulps to D.P.'s may not be independent of their preparation as is generally as-
sumed.
101 G. A. Richter, Ind. Eng. Chem.. 23, 131, 266 (1931).
668 CELLULOSE
well-worn materials, are not so difficult to beat. This is due to the dif-
ferent manner in which the various fibers disintegrate under the action of
the beater. Cotton and especially bast fibers, such as linen, are frayed out
and unraveled into a mass of very fine, long fibrils by beating, whereas wood
breaks up into relatively coarse, short particles or, at best, into coarse
fibrils.
Degraded cotton rags, a bleached coniferous sulfite, or a hardwood kraft
may each have the same measured viscosity and yet have widely different
beating qualities and yield quite different papers. Thus, except possibly
as a control test applied to pulps from the same source and prepared under
the same conditions, the viscosity test is of little or no value by itself in
predicting their papermaking qualities. However, it can be stated that a
pulp with both a high D.P. and a high hemicellulose content can usually
be made to yield strong papers.
The distribution pattern of the molecular chain lengths in cellulose has
also failed to show promise for evaluating papermaking qualities. This is
not surprising because, by analogy, Clark,11 as already mentioned, has
shown that paper made from a mixture of long and short fibers had very
similar properties in all respects to paper made from fibers of uniform length
equal to the weight-average fiber length of the mixture.
2. Hemicellulose Content
Tests related to hemicellulose content include beta- and gamma-cellulose
contents as determined by the Cross and Bevan procedure (see Chapter
XII), solubility of the pulp in alkalies of various strengths, and pentosan
and polyuronide contents.
It is well known that impure fibers, for example, unbleached pulps and
especially ordinary straw pulps with their high hemicellulose content, beat
more rapidly than purer fibers. The pentosan content of the pulps appears
to have a marked influence on beating,103 but Klingstedt104 and March106
point out that the factor is the proportion of alkali-soluble material in the
pulp rather than the pentosan content. Bleaching removes some of the
hemicellulosic materials, especially if the pulps have been subjected to a
caustic extraction during the bleaching process. This usually results in
slower beating of the pulp. However, if the original pulp was somewhat
raw to begin with, then removal of lignin by bleaching may increase the
1WO. H. Young and B. W. Rowland, Paper Trade /., 97, 44 (Oct. 12, 1933).
104 F. W. Klingstedt, Svensk Papperstidn., 40, 412 (1937).
108 R. E. March, Paper Trade /., 127, 51 (Oct. 21, 1948).
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 669
rate of beating. Alkali-cooked pulps from coniferous woods have more
hemicellulose removed than sulfite pulps, and accordingly they beat more
slowly. On the other hand, if hardwood pulp is prepared by the sulfite
process (which is seldom) the resulting pulp is both weak and difficult to
beat. With deciduous woods, it may be that the material is so degraded
by the acid cooking liquor that it becomes soluble, and therefore nearly all
the hemicellulose on the surfaces is removed.
Cottrall88 has recently presented a comprehensive review of the effect of
hemicelluloses on the papermaking qualities of wood pulp, and additional
data showing that pulp becomes less responsive to beating and yields a
weaker sheet as hemicellulose is removed. He also shows that the gamma-
cellulose content of a number of different kinds of wood pulp correlates
quite well with the resulting papers. He agrees with Jayme and Loch-
mfiller-Kerler106 that there is an optimum hemicellulose content for pulps
above and below which the strength of the resulting sheets is less. March106
also found this to be the case. Cottrall refers to the linear relationship
between the pentosan content of softwoods and their swelling capacity
found by Young and Rowland103 and subscribes to an opinion that the
main role of hemicellulose is its action as a plasticizing agent for the fibers.
This view is in accord with the findings107 that the addition of hemicelluloses
to pulps deficient only in this respect has a relatively small, but nevertheless
significant, beneficial effect on the rate of beating and the resulting strength
of the sheet.
With respect to the three Cross and Bevan celluloses, Giertz108 has pre-
sented evidence to show that the alpha-cellulose content of wood pulp pro-
duced by a wide range of cooking conditions is remarkably constant — at
43% for spruce. It would appear to represent the highly crystalline and
resistant cellulose composing the "strings" or micelles in the secondary wall
of a fiber, which may be seen in an electron microscope after disrupting the
fiber by ultrasonic treatment, as uniform bodies about 75 A. thick.109 In
the fiber they lie together as bundles, and between these crystalline ele-
ments are amorphous regions containing the hemicellulose or the gamma
fraction, which Giertz showed corresponds closely with the quantity of
easily hydrolyzable material in a number of pulps, as determined by Nicker-
son's method. The beta-cellulose fraction becomes appreciable only after
106 G. Jayme and E. Lochmiiller-Kerler, Papierfabr. Wochbl. Papierfabr., 1944, 223.
107 H. E. Obermans, Paper Trade J.t 103, 83 (Aug. 13, 1936).
108 H. W. Giertz, Proc. Tech. Sect., British Paper & Board Maker's Assoc., 33, 487
(1952) Part 3; World's Paper Trade Rev., 136, 1451 (1951).
*• B. G. Rlnby, Tappi, 35, 53 (1952).
670 CELLULOSE
the pulp has been subjected to degradation as with acids or prolonged
cooking. Consequently, this fraction would appear to represent any short-
chained fragments derived from the crystalline cellulose by cleavage.
R&nby110 has confirmed this with further electron microscope and x-ray
studies of the various fractions.
Giertz111 has also pointed out that since the pentosan and noncellulosic
material is contained mainly in the noncrystalline parts of the fibers, it
may not itself influence beating but rather be a proportionate indication of
the presence of amorphous cellulose which can swell and be beaten much
faster than the crystalline portions. He questions Wurz's conclusion61
that the presence of pectins and polyuronides is of primary importance in a
pulp suitable for making a good greaseproof paper (a hard transparent
sheet) because of the almost insignificant percentage — about 2% — required.
Giertz suggests that because these constituents are so easily hydrolyzed by
adverse cooking conditions, their presence is merely a sign that the sulfite
cooking condition was mild enough that a large proportion of hemicellu-
lose, and thus the original amorphous cellulose, is left in the fibers. The
greater amorphous cellulose content will permit the pulp to swell and be
beaten more easily, and thus, when the sheet is dried, the fibers will col-
lapse to give both the fibers and the sheet a dense, translucent structure.
This view appears to be in good accord with observed phenomena and
may explain anomalies in the relation between the chemical analysis of
many pulps and the strength of the resulting papers.
3. Lignin Content
The usual tests for lignin content comprise the chlorine demand of the
pulp, its bleachability, and its permanganate number.
For a long time it was thought that the presence of lignin in "strong."
i.e., relatively undercooked, pulps was responsible for their generally higher
strength, and it was common to grade pulps on that basis. However, it has
become clear that if the lignin was progressively removed with chemicals
like chlorine dioxide112 and sodium chlorite113 that did not appreciably im-
pair the cellulose and hemicellulose content, the strength of the pulps was in
fact improved.
Raw pulps, such as are prepared by "semichemical" processes and which
uo B. G. Ranby, Svensk Papperstidn., 55, 115 (1952) (in English).
111 H. W. Giertz, Cellulosa och Papper SPCI 40th Anniversary No. 1908-1948, 417
(1948).
11S E. C. Jahn and C. V. Holmberg. Paper Trade /., 114, 203 (Apr. 23, 1942).
111 G. Jayme, Papier- Fabr., 40, 137, 145 (1942).
VIII. PROPERTIES AND TREATMENT OF PULP FOR PAPER 671
contain a large proportion of their original lignin content, do not beat quickly
and do not yield very strong papers. It would appear that if present in
quantity, for example in a wood pulp having a yield of upwards of 70% of
the original wood substance, the lignin continues to bind the structure of
the individual fibers together so tightly that they cannot be fibrillated
easily. Thus, as with artificial cellulose fibers, attempts to beat them merely
cause a reduction in their size. In either case, even with the addition of
mucilage or other binding agent, with the complete absence of fibrillation
on viscose or only the sparse fibrillation developed on highly lignified
pulp fibers, no tough cementing action between the fibers can be achieved.
4. Cooking Reactions
When cellulose-containing material is cooked to make pulp, the compo-
nents of the fibers and of the encrusting material are attacked at different
rates. The proportionate rate of attack on different components of the
fibers also depends on the nature of the cooking liquor and on the time and
temperature schedule employed. It should be pointed out that although
the alkaline processes avoid degradation by acids, degradation due to heat
is greater because of the higher cooking temperatures used.
Varying the acid concentration and calcium content in a sulfite cook or
the alkali concentration and degree of sulfidity (which has a protective
influence on the fibers) in a sulf ate cook yields pulps of various compositions
as well as kinds. The attack is mainly on the amorphous or less crystalline
parts of the fibers so that, as Giertz points out, the quantity of amorphous
cellulose remaining is an important determining factor in the papermaking
quality of the pulp.
The removal of lignin from the fibers by the cooking liquor parallels, to
some extejit, the attack on the different cellulose components. In conse-
quence, by varying the cooking conditions, pulps can be produced having
the same chlorine demand or lignin content and perhaps the same hemicellu-
lose content, but having quite different papermaking characteristics.
A coniferous pulp prepared by the sulfite process (acid) with a moderate
degree of cooking, as determined by its lignin content, beats more rapidly
than a pulp prepared by an alkaline cook (e.g., soda or sulf ate process)
and having the same lignin content. Mitscherlich pulp (pulp prepared by
a relatively mild slow sulfite cook) cooked to the same degree beats still
more rapidly.
In accordance with the composite theory of beating, when pulp is pro-
duced by any mild process giving a high yield, or by an acid process which
672 CELLULOSE
tends to hydrolyze the material, there is more short-chain material exposed
or created on the outer surfaces. This will form a more ' 'concentrated'*
and, hence, a more cohesive surface ' 'suspension"; conversely, a caustic
treatment, which normally removes the short-chain material and which,
unlike anacid,does not effect hydrolysis of the high D.P. material, gives pulp
with much less cohesiveness because of the longer chain lengths and thus a
lower "concentration" on the exposed surfaces. (In the case of the sel-
dom-made hardwood sulfite pulps, as already mentioned, the short-chain
material may be so hydrolyzed that it is dissolved away.) At the same
time, consideration must also be given to the effect of the pulping process
on the intrinsic strength of the fibers and fibrils. It follows, therefore, that
pulp prepared from hardwoods or softwoods by the mild neutral sodium sul-
fite process which gives a relatively high yield, is exceptionally strong as re-
gards fiber structure and its possession of considerable short-chain adhesive
material and amorphous cellulose. Pulp made with the normal (acid) sul-
fite process is weakened in fiber structure, but, if a softwood, it too beats
rapidly, presumably because of hydrolysis or shortening of the material re-
maining on the surface of the fibers. Soda and especially sulfate pulp re-
tain their fibrous structural strength to a great extent, but, because of the
dearth of short-chain material and amorphous cellulose on the outer sur-
faces of the fibers, beat relatively slowly. If both the sulfite and the sulfate
processes are applied to the same softwood chips, then a carefully produced
sulfate pulp, because of its greater D.P. and thus greater intrinsic strength,
after extended beating to develop fibrillation and cohesion, can be made
ultimately into a considerably stronger paper.
Chapter IX
DERIVATIVES OF CELLULOSE
A. REACTIVITY AND REACTIONS OF CELLULOSE
HAROLD M. SPURLIN
The usefulness of cellulose depends not only on the excellent physical
properties of cellulose itself, but also on its ready transformation into
derivatives. These derivatives are useful because of their solubility charac-
teristics not possessed by cellulose itself, their ease of forming at high tem-
perature, or their greater softness and flexibility.
Many different features of cellulose behavior need to be considered if
cellulose reactions are to be understood. The chemical nature of cellulose
(Chapter III) and the structure of cellulose fibers (Chapter IV) are ob-
viously important. Many peculiarities associated with individual reaction
conditions will be treated at length in later Sections of this Chapter IX.
Some ideas about the nature of forces holding the fiber structure together
can even be gathered from the nature of fiber-fiber interactions in paper
(Chapter VIII). In spite of the fact that some repetition is involved, it
appears wise to devote this Section A to a unified treatment in an effort to
reconcile the many apparently inconsistent experiments and interpretations
thereof to be found in the literature. The viewpoint adopted here owes
much to the influence of Staudinger. However, it is believed that a more
realistic attitude is expressed, especially in questions that deal with the
uniformity of reaction of fibrous cellulose.
The topics considered are: (1) uniformity of reaction, with special
emphasis on the nature of derivatives prepared in homogeneous solution;
(2) the influence of fiber structure on uniformity; and (3) chemical factors
which limit the attainment of uniform products.
Almost all of the commercially important cellulose derivatives are either
esters or ethers, prepared by reactions typical of compounds containing
hydroxyl groups. Even with those derivatives that are not esters or
ethers, such as N2O4-oxidized cellulose or derivatives in which the hydroxyl
673
674 CELLULOSE
group has been replaced by halogen, amino, or other groups, the methods
of preparation are exactly analogous to those of similar compounds of low
molecular weight. The peculiarities of cellulose reactions are thus not due
to any characteristic difference between the. innate reactivities of the hy-
droxyl groups of cellulose and the hydroxyl groups of other types of com-
pounds. Rather, the problems encountered are due to two main factors:
(1) The cellulose hydroxyls may not be available for reaction because the
crystallinity or insolubility of the cellulose hinders access of the reagent
to the hydroxyl groups; (2) excessive amounts of degradative side reactions
must be avoided because cleavage of the cellulose chain would result, giving
products with unsatisfactory properties. Fortunately, the degradation
reactions may be held within acceptable bounds in the important cases.
Most of the technical problems of cellulose reactions thus center about the
question of availability of the hydroxyl groups for reaction. Sometimes,
as in the case of direct esterification with acetic acid, the lack of availability
simply prevents the reaction from going in a satisfactory manner. More
often, difficulty arises because the differences in accessibility of different
portions of the sample result in very nonuniform products. Much of the
science of cellulose derivatives is concerned with methods designed to hold
this nonuniformity within acceptable bounds.
1. Uniformity of Substitution
Care is necessary in defining the uniformity of a cellulose reaction. The
technologist usually has a clear idea of what is meant by a uniform product.
It will be completely soluble in a variety of solvents. Solutions, films, or
molded articles will be clear and sparkling, which means that they will be
free of suspended matter and will have little tendency to separate into two
phases. A more exact consideration will soon demonstrate that even
though the above conditions jxe fulfilled, no cellulose derivative can be
really uniform. In fact, statistical considerations show that in no sample
of a partially substituted cellulose derivative will there be two identical
molecules of long chain length.1 This variation in arrangement of sub-
stituents will be superimposed on a distribution of chain lengths and on dif-
ferences in degree of substitution (D.S.) between different cellulose chains.2
The important question of evaluation of uniformity will depend on a
1 H. M. Spurlin, Trans. Electrochem. Soc.t 73, 95 (1938).
1 A. J. Rosenthal atad B. B. White, Ind. Eng. Chem.. 44, 2693 (1952).
DC. DERIVATIVES OF CELLULOSE 675
definition of a "perfectly" uniform material and then on some criteria for
the degree of deviation from this norm. Uniformity of chain length will
be treated in Chapter X-D. The discussion of uniformity of substitution
in this Section A of Chapter IX will be based on the following definition of
uniformity of a cellulose derivative: The highest possible degree of uni-
formity is that resulting when every anhydroglucose unit has had an equal
amount of exposure to the reagents involved. According to this definition,
products of the same degree of substitution and the same molecular weight
might be different in properties because of a difference in the manner of
arrangement of substituents along the cellulose chain, and yet each product
would be considered uniform. This state of affairs is quite possible; uni-
form cellulose acetates of the same degree of substitution but prepared in
different manners actually differ appreciably in properties.8 This differ-
ence is due to varying extents of reaction of the primary and the secondary
hydroxyls. (See also Section C of this Chapter IX.)
The definition of uniformity of reaction adopted here, namely, equality
of ease of access of the reagents to the individual anhydroglucose units,
naturally brings up the question of whether this equality is attainable.
The answer appears to be definitely yes. Modern polymer research is en-
tirely in harmony with the concept that individual segments of a dissolved
polymer chain will have the same reactivity as a molecule of low molecular
weight and similar structure. Furthermore, the influence of other portions
of the molecule on the reactivity of a given group will decline very rapidly
with increasing distance from that group. An example may be taken from
the field of polymerization : The rate of interaction of monomer with the
end of a growing polymer chain during copolymerization depends almost
entirely on the nature of the terminal monomer group, and not on the rest
of the chain.4 The hypothesis that the individual anhydroglucose units
of cellulose or one of its derivatives in solution are equally available for
reaction may therefore be accepted with confidence.
It is equally certain that this uniformity of access will not be possible if
fibrous or crystalline cellulose reacts in a heterogeneous manner. Even
in this case a product meeting the above definition of uniformity is possible
in the case of an equilibrium reaction or in the case of complete reaction.
If the reaction equilibrium is at some point short of complete substitution,
1 C. J. Malm, L. J. Tanghe, B. C. Laird, and G. D. Smith, /. Am. Chem. Soc.t 75, 80
(1953).
4 T. Alfrey, Jr., J. J. Bohrer, and H. Mark, Copolymerization (High Polymers, Vol.
VIII), Interscience, New York-London, 1952,
676 CELLULOSE
ease of access may not have been equal at the beginning; however, if suf-
ficient time is allowed, complete access to all hydroxyls may be obtained.
The quantitative evaluation of degree of uniformity of a product depends
on the development of suitable experimental methods and on correct
mathematical interpretation of the results. Since many misleading state-
ments in the literature are based on the wrong mathematics, it is desirable
to discuss in detail the nature of the distributions of substituents that can
be expected.
(a) CALCULATION OF THEORETICAL ARRANGEMENT OF SUBSTITUENTS
If the principle of equal availability of all anhydroglucose units is ac-
cepted, some important conclusions can be reached about the arrangement
of substituents along the chain. The nature of the substitution on an
anhydroglucose unit will be governed by probability considerations. The
commercially important partially substituted derivatives will be expected
to have higher local concentrations of hydroxyl groups along some portions
of the chains than along others even if they are uniform in the sense used
here. These fluctuations of composition undoubtedly have much influence
on the properties of the products. For instance, the possibility of micro-
bial attack appears to be associated with the presence of unsubstituted
anhydroglucose units along the chain (see Chapter III-C-5). Also, the
outstanding physical properties of cellulosic plastics are possibly due to
these local high concentrations of hydroxyl groups. There has therefore
been a great amount of effort expended on the study of the distribution of
substituents along the cellulose chain.
The nature of this distribution is easily calculated on the basis of the as-
sumptions that (1) availability of all anhydroglucose units is equal, (2)
the influence of the state of reaction of one hydroxyl on the reactivity of
another declines very rapidly with increasing distance between the two,
(3) the ratios of the reaction rate constants to one another remain constant
throughout the reaction, and (4) end-group effects are negligible. If as-
sumption (3) is fulfilled and the back reaction is negligible, the distribution
of substituents at a given degree of substitution will be the same as that ob-
tained in first-order reactions of the hydroxyl groups with nonvarying re-
action constants.
All of the available data on distributions controlled by the rates of reac-
tion of the three sorts of hydroxyls may be correlated if assumption (2)
is modified to read: The only interference between hydroxyls occurs be-
tween positions 2 and 3 of the same anhydroglucose unit. With this pro-
viso, the distribution in the case when only one type of substituent is in-
IX. DERIVATIVES OF CELLULOSE 677
volved may be described by five reaction rate constants : k%, k$, and k* for
initial reaction at positions 2, 3, and 6 of the anhydroglucose ring; ka for
reaction at position 3 if position 2 is substituted; and &6 for reaction at
position 2 if position 3 is substituted. The total substitution, in substit-
uents per anhydroglucose unit, will be called S. The fraction of unsub-
stituted anhydroglucose units will be designated 5n; the fraction monosub-
stituted at position 2, 3, or 6 will be designated sz, s3, or s* ; the fraction di-
substituted at positions 2 and 3, 2 and 6, or 3 and 6 will be designated
$2,3, s2,e, or £3,6,' and the fraction trisubstituted will be designated $2,3,6.
The following symbols are also useful in simplifying the writing of equa-
tions :
(1)
.
#2 T *3 — Ka
N = - - - (2)
£2 + £3 — kt
The equations for the fractions of anhydroglucose units that are substituted
in various manners are :
52 = Me~ (k° + k'}t - Me~ <kl + *• + *8)< (4)
53 = Ne~(kb + kt)t - Ne~(k* + k' + Mt (5)
S6 = e- <*• + *«)« _ e- (*, + *,+*,)/ (6)
,M = e-ktt - Me(k" + k>)t - Ne-<k" + Mt +
(M + N - l)e-(k> + kt + ftl)' (7)
52,6 = Me-k°* - Me-(k° + *')( - Me~(k' + *'>' + Me~(kl + kl+ kf>t (8)
536 = Ne~kbt - Ne-(kb + *')( - Ne-(k' + *'" + Ne~(k* + *' + *'}' (9)
52.3.6 = [1 - e-**} fl - Me-"'1 - Ne-kbt +
(M + N - l)e-(kt + k>}t] (10)
S = 3 - Me~kai - Ne~ktt - e~k>t + (M + N - 2)e-(kl + Mt (11)
It is frequently convenient to consider the case where all reactivity con-
stants are equal. In this case, a notable simplification of the equations
results, and they may be solved in terms of S.6 These simplified equations
follow:
1 T. E. Timell, Studies on Cellulose Reactions, Esselte A/B, Stockholm, 1950; also,
Ing. Vetenskaps Akad., Handl., No. 205, Stockholm, 1950.
678 CELLULOSE
If &2 » Jfej •» #6 = #a = ^ft
(13)
*.., -
If no mutual interference with reactivity at positions 2 and 3 is encoun-
tered, kz = kj>, kz = ka, and M = N = 1. If the interference between re-
action at positions 2 and 3 is so complete that there is never any substitu-
tion at both of these positions in one glucose unit, ka = kb = 0 and M + N
= 1.
In using these equations, it should be remembered that / is not propor-
tional to time unless all conditions remain constant during the reaction.
Rather, / is to be considered as a parameter which includes a time factor,
by means of which corresponding values of S and the individual types of
substituted glucose units may be calculated.
Similar equations may be calculated for the case where the extent of
reaction is governed by equilibrium constants rather than rate constants.
These equations have been published6 for the important case where there is
no mutual interference between substituents in positions 2 and 3, and will
not be repeated here. The distribution curves are identical for rate-
controlled and equilibrium-controlled distributions if all three hydroxyls
have equal reactivity. If the reactivities are different, the distributions
are similar for the two cases* It should be noted, however, that with
equilibrium control the maximum of any monosubstituted species ($2, $s,
or $e) will occur at an average substitution 5=1, and of a disubstituted
species at 5 = 2. This is not true for rate-controlled reactions unless
either the reaction rate constants are all equal or ka = £& = 0.
In order to visualize the implications of the distribution equations, it is
helpful to compare calculated plots for several assumed ratios of reactivi-
ties. In Figure 1, the fractions of the anhydroglucose units present in the
unsubstituted, monosubstituted, disubstituted, and trisubstituted forms
• H. M. Spurlin, /. Am. Chem. Soc.. 61, 2222 (1939)
IX. DERIVATIVES OF CBLLULOSK
679
are plotted against the total substitution, S, for several assumed ratios of
equ&brium and reaction rate constants and with the further assumption
of no mutual interference between positions 2 and 3. One important point
is immediately apparent from Figure 1. There is not a great deal of dif-
ference among the curves for a given value of c, in spite of fairly wide dif-
ferences of reactivity that were assumed. It is hardly likely that analytical
o.i -
O.o
Fig. 1. Theoretical over-all distribution of substituent groups in cellulose (Spurlin6).
Abscissa: Average number of substituents per anhydroglucose unit.
Ordinate: Fraction of total number of anhydroglucose units substituted as indicated
Co, unsubstituted; SQ
it monosubstituted ; 52 + $3 -h so
Ratio of equilibrium constants :
1:1:1
1:4:16
2t disubstituted ; $2,3 + ^2,5
d, trisubstituted ; s9
Ratio of rate constants:
1:1:1
— O— 1:1:10
•-- 1:1:10
methods for the quantities plotted in Figure 1 will ever be good enough to
allow valid conclusions to be drawn about the relative reactivities of the
three sorts of hydroxyls. It is even less to be expected that such deter-
minations will allow the accurate comparison of the uniformity of two dif-
ferent samples of the same degree of substitution. As a matter of fact,
most of the data obtained by investigators who determined only the amount
of mono-, di-, and tri-substituted anhydroglucose units present in uniform
680
CELLULOSE
derivatives can be correlated within experimental error by the assumption
that all three sorts of hydroxyls have the same reactivity6 with no inter-
ference between positions 2 and 3. Such interference actually occurs in
many cases, as will be discussed below.
0.6
0.5
0.4
0.3
0.2
0.1
/
** —
•"s
/
\
/
\
1
\
/
\
/
\
^
\
/
/
\
\
1
tfi
*/
\
\
/
ipj
?
\
\
/
/
'
\
\
/
&^
*S
s^T
\,
\
0123
Fig. 2. Proportions of the six different mono- or di-substituted glucoses, if the
nature of substitution is determined by the rates of reaction of the three hydroxyls
in the ratio 1 : 1 : 10 (Spurlin6).
Abscissa: Average number of substituents per anhydroglucose unit.
Ordinate: Fraction of total number of anhydroglucose units substituted as
indicated.
52>a, $2,6, 5s,6 =* fraction of anhydroglucose units with two substituents, in 2,3-,
2,6-, or 3,6- positions.
5j, Sj, $e «• fraction of anhydroglucose units with one substituent, in 2-, 3-, or 6-
position.
The relative reactivities of the three sorts of hydroxyls may be deter-
mined with reasonable precision if all eight possible manners of substitution
of the anhydroglucose units are measured separately. That this is so is
indicated in Figure 2. As is evident, a tenfold greater reactivity of the
primary hydroxyl than either of the two secondary hydroxyls is readily de-
tected by comparing the ratio of SB to $2 or 53 at a substitution of 0.5-1.0,
or of 52,6 or $st« to $2,3 at a substitution of 1.5-2.0. The determination of
reactivity ratios under conditions of uniform reaction, as in solution, is
at present in a satisfactory condition, and will be reviewed below.
It would be desirable to be able to use the available information about
IX. DERIVATIVES OF CELLULOSE 681
reactivity ratios to evaluate the uniformity of cellulose derivatives by
analysis for the different types of substitution. By comparison of these
results with the theoretical distribution, a quantitative evaluation of uni-
formity could be obtained.
Unfortunately, it is still not possible to specify methods for the exact
quantitative evaluation of uniformity by using these principles. The dif-
ficulty is twofold. To begin with, it is apparent from Figure 2 that the
curves have rather flat shapes in the regions where content of a given species
is appreciable. Accordingly, a mixture of material of, say, D.S. 0.8 and
1.2 would be hardly distinguishable from a uniform material of D.S. 1 if
only the six quantities in Figure 2 were determined.
It turns out that the detection of nonuniformity depends on exact analy-
sis for species that would be present in small quantity in a uniform product.
For example, the amount of unsubstituted glucose in a product of D.S.
2.5 is a good criterion of the uniformity.
The other important difficulty in a statistical study of the evaluation of
uniformity of the substituted glucose content arises from the break-
down of the assumptions involved in the calculations. It is becoming
increasingly apparent that interference between positions 2 and 3 is an
important factor in rate-controlled reactions.7 This necessitates the use
of the complete mathematics of equations 3-1 1 which has never been done
over a wide enough range of D.S. to allow the evaluation of ka and k*.
Another complication arises if the ratio of reactivities changes during re-
action. For example, in etherification to a high D.S., there is a profound
change in the nature of the medium surrounding the individual hydroxyl
groups from the initially hydrophilic alkali cellulose to the hydrophobic
ether. It would hardly be surprising if there was a change in the reactivity
ratios as a consequence of the change of medium. The mathematical
difficulties of handling such a situation would not be insuperable, but there
appears to be little chance that the necessary precise data will be obtain-
able.
As a consequence of the above difficulties, it is scarcely surprising that,
in industry, the evaluation of uniformity is 'on a purely empirical physical
basis, depending on strict specifications of turbidity in solution, solubility
range, viscosity-concentration behavior, and the like. It must not be
supposed, however, that the effort that has been expended in the study of
the statistics of substitution reactions has been wasted. Many erroneous
notions have been dispelled. Above all, a very good idea of that which is
practically attainable has been reached during the last twelve years.
7 T. E. Timell and H. M. Spurlin, Svensk Papperstidn., 55, 700 (1952)
682 CELLULOSE
(b) EXPERIMENTAL EVALUATION OF DISTRIBUTION THEORY
The most important use of the calculated distribution curves is to
serve as a logical framework for the correlation of experimental results on
the nature of partially substituted derivatives. The experimental work
usually has purposes other than mere confirmation of the equations. The
most frequent objective has been to study the availability of fibrous cellu-
lose for reaction. There has also been important work on the directing
influence of specific reaction conditions, such as etherification of the as-
sociation complex of cellulose with sodium and cupric hydroxides (see
Table 15, Section E of this Chapter IX). There has, however, been enough
work to furnish experimental confirmation for the statistical theory of the
arrangement of substituents in cellulose derivatives.
A complete confirmation of the theory would require a study of the in-
fluence of the state of substitution of one anhydroglucose unit on that of
its neighbors. Because of the experimental difficulty of such an approach,
attention has been confined up to the present to the following alternative
methods.
In the first method, a series of derivatives of increasing substitution is
prepared, and the amount of some specific type of substitution is deter-
mined. For example, the amount of primary ^substitution may be esti-
mated by the techniques of tritylation,8 tosylation-iodination,9'10 rate of
tosylation.11 In addition, the amount of unsubstituted glycol groups in
the 2,3- positions (i.e., the glycol number) can be estimated by cleavage
by periodate12 or lead tetraacetate.12 Because of the limited amount of
information obtainable by these methods, this type of approach cannot
be expected to give a completely satisfactory confirmation of the theory.
However, it is satisfying to know that results so far obtained by these
methods on uniform products are in agreement with the theory.
In the second method of approach, an attempt is made to determine
the number of anhydroglucose units substituted in each of the eight pos-
sible manners for a series of derivatives of increasing total substitution.
Unfortunately, this method is applicable only to cellulose ethers, since the
available methods of analysis depend on hydrolysis to the monomeric sub-
• I. Sakurada and T. Kitabatake, /. Soc. Chem. Ind.> Japan, 37, Suppl. binding, 604
(1934).
• C. J. Malm, L. J. Tanghe, and B. C. Laird, /. Am. Chem. Soc.t 70, 2740 (1948).
10 F. B. Cramer and C. B. Purves, J. Am. Chem. Sac., 61, 3458 (1939).
u T. S. Gardner and C. B. Purves, /. Am. Chem. Soc.t 64, 1539 (1942).
11 J. F. Mahoney and C. B. Purves, /. Am. Chem. Soc., 64, 9 (1942).
11 H. H. Brownell, Thesis, McGill Univ., Montreal, 1953.
IX. DERIVATIVES OF CELLULOSE 683
stituted glucose stage without the loss of substituents which occurs on hy-
drolysis of other derivatives. Of course, analytical information on the
hydrolyzed products can be supplemented with information of the first
kind, such as the glycol number of the unhydrolyzed material.
It would be especially desirable to isolate, characterize, and weigh the
eight types of substituted glucose, rather than to depend on frequently
rather indirect analytical methods. With the present highly developed
status of paper chromatography, countercurrent distribution between
solvents, and high-vacuum distillation methods, such an approach should
not be too difficult. It is encouraging to note that even in the complicated
case of hydrolyzed hydroxyethyl cellulose, it was possible by the use of
paper chromatography to resolve all the glucose substitution types except
those substituted in the 2- position from those substituted in the 3- posi-
tion.18 The early attempts by Traube and his coworkers14*15 to separate the
methylated glucoses by distillation were not so successful, though resolu-
tion of unsubstituted, mono-, di-, and tri-substituted glucoses was reason-
ably complete even without the use of a distillation column.
By selecting from the methods outlined above and when necessary de-
veloping new methods, Timell5'16~"19 succeeded in establishing that the
statistical theory of distribution of substituents does indeed hold for reason-
ably uniform ethers (see Table 17, Section E of this Chapter IX). In the
case of methyl cellulose, the reactivities of the three sorts of hydroxyls did
not appear to differ by factors of more than 2 : 1 when either solutions of
cellulose in quaternary bases or fibrous alkali cellulose were reacted with
methyl sulfate or methyl chloride. On the other hand, in the presence of
excess cupric hydroxide, alkylation in the primary position is suppressed
(Table 15, IX-E).
As had been expected by earlier investigators, the presence of ionized
substituents in either secondary position greatly reduces the ease of intro-
duction of a second ionized substituent in the adjoining secondary position.7
In fact, with carboxymethyl celluloses of D.S. 1 or less, there was no evi-
dence for substitution in both secondary positions at the same time, and
the mathematics of equations 3-11 with ka = &» = 0 appeared to apply.
This conclusion cannot be strictly true, since by the use of drastic reaction
14 A. Funk, Dissertation, Berlin, 1935; H. J. Schenck, Dissertation, Berlin, 1936.
l« W. Traube, R. Piwonka, and A. Funk, Ber., 69B, 1483 (1936).
16 T. E. Timell, Svensk Kern. Tid., 62, 49, 129 (1950).
17 T. E. Timell, Svensk Paperstidn.t 55, 649 (1952).
18 T. E. Timell, Svensk Paperstidn., 56, 311 (1953).
" T. E. Timell, Svensk Paperstidn., 56, 483 (1953).
684 CELLULOSE
conditions it is possible to prepare carboxymethyl cellulose of substitution
greater than 2.
As a result of TimelTs thorough investigations, there can remain no doubt
about the usefulness of statistical considerations as a guide to the study
of the manner of arrangement of ether substituents along the cellulose
chain. Furthermore, it is interesting to note that under usual technical
conditions of etherification, the initial reactivities of the three sorts of hy-
droxyl groups are rather close to one another. This is particularly true of
the 2- and 6- positions, the reactivity in the 3- position being uniformly
somewhat lower.5
It is also becoming increasingly apparent that the original supposition6
is correct that there is frequently marked interference between reaction at
positions 2 and 3. Although so far firmly established only in the case of
the carboxymethyl group, similar effects may be expected in all cases in
which ionized or bulky substituents are involved. It may be anticipated
that tosyl, trityl, and benzyl groups will give large mutual interference.
The presence of even a relatively small adjacent group, such as ethyl or
acetyl, would be expected to reduce the rate of introduction of tosyl or
trityl groups into the adjoining positions.
This mutual interference does not in any way invalidate the statistical
treatment. It simply necessitates the determination of more reaction rate
constants in order to specify the nature of the system. For example, if the
rate of tosylation is to yield information about the amount of substitution
of each sort of hydroxyl in cellulose acetate, it will be desirable to consider
the following seven rate constants: one constant for the primary group;
three constants for position 2, depending on whether position 3 is unsub-
stituted, carries an acetyl group, or carries a tosyl group; and three similar
:onstants for position 3.
(c) UNIFORMITY OF METHYL CELLULOSE
The application of the concept of statistical distribution of substituents
to the evaluation of uniformity may be illustrated by the comparison of the
nature of methyl celluloses prepared in different manners. The question is
important because some of the conclusions that are drawn, even in the cur-
rent literature,20 are not reconcilable with the principles adopted in this
book.
The principal point at issue is the nature of cellulose xanthate in solution
(Section F of this Chapter IX and Chapter X-C). By reaction of this
M T. Lieser, Kurzes Lehrbuch for CMulosechemie, Borntrager, Berlin, 1963.
IX. DERIVATIVES OF CELLULOSE 685
solution with diazomethane generated in situ> Lieser21 obtained methyl
celluloses of D.S. about 0.5. It is not very probable that the methyl groups
were introduced solely in the positions originally occupied by xanthate
groups.22 The following arguments apply only to the uniformity of the
methyl celluloses. These, when subjected to acetolysis at low temperature
with a mixture of acetic anhydride, acetic acid, and sulfuric acid, gave
cellobiose octaacetate in yields of 11-14 g. per 100 g. of methyl cellulose
used. Similar experiments with unmethylated cellulose gave yields of 46%.
These experiments prove that the methyl cellulose samples contained
many of their anhydroglucose units unsubstituted and adjacent in pairs.
Further conclusions that can be drawn from these results depend on what
assumptions are made. Lieser assumed that a methyl cellulose of D.S.
0.5 would not contain more than one substituent on any individual anhy-
droglucose unit, and that the yields from methyl cellulose should be cor-
rected by a factor corresponding to the yield from cellulose (i.e., 46% for
these experiments). He further assumed, without carrying out the neces-
sary calculations, that the observed yields could not be explained on the
basis of random arrangement of the substituted anhydroglucose units.
Lieser therefore concluded that the cellulose in viscose is only 50% available
for reaction, and that the samples consisted of mixtures of monomethyl
cellulose and unaltered cellulose. This conclusion is unjustified, as was
pointed out in the first edition of this book as well as by Staudinger and
Zapf 23 and by Timell.6 Obviously, more than half of the anhydroglucose
units of a methyl cellulose of D.S. 0.5 would be unsubstituted. As is
evident from Figure 1, the content of unsubstituted anhydroglucose units
for D.S. 0.5 can go as high as 57.8% if all hydroxyls have equal avail-
ability. There would be a better than 25% chance that any cellobiose
unit that was formed by hydrolysis would be unsubstituted. The ob-
served yields of cellobiose octaacetate (about 25% of that from cellulose)
are thus in excellent agreement with those to be expected on the basis of a
perfectly uniform methyl cellulose.
There is a further point that needs attention in this connection. The
assumption is implicit in the above reasoning that the rate of hydrolysis of
the 1,4-glucosidic bonds in the chain structure will not depend on the state
of substitution of the adjoining anhydroglucose units. This assumption
21 T. Lieser, Ann., 483, 132 (1930).
82 Chian-Yuang Chen, Ralph E. Montonna, and C. S. Grove, Jr., Tappi, 34, 420
(1951).
» H. Staudinger and F. Zapf, /. prakt. Chem.t 156, 261 (1940).
686
CELLULOSE
is far from being justified. Lieser21'24 recognized that methyl celluloses
undergo the acetolysis reaction more rapidly than does cellulose, and that
incompletely reacted products isolated from the reaction mixture were
lower in methoxyl content than was the starting material. During the
acetolysis reaction, a mixed acetate-methyl ether is being degraded. It is
now known (see Section G of this Chapter IX) that cellulose ethers undergo
TABLE 1
Acetolysis of Cellulose and Methyl Cellulose
(From data of Lieser and Jaks81*84)
Cellobiose
Cellobiose
octaacetate
Cellobiose
octaacetate
yield,
content
isolated,
calculated,
calculated,
g./lOO g.
g,/100 g.
mole
of methyl
of methyl
Type of product
D.S.
per cent
cellulose
cellulose*
Fibrous cellulose
0
100
45
—
Fibrous methyl cellulose*
0.28
54
37
23
0.46
36
23
15
0.65
22
22
10
0.73
17
26
8
0.74
17
19
8
1.26
3.5
14
1.5
Technical methyl cellulose6
>1.5
<0.2
0
<0.1
Methyl cellulose from xanthated
0.50
32
11
14
0.46
36
14
15
* Based on a 45% yield from the cellobiose content for fibrous products,24 46% for
xanthate.11
6 Prepared by reaction of cellulose with methyl sulfate in the presence of 20% NaOH.
c "Tylose," presumably prepared with methyl chloride and 35% NaOH.
* Prepared by reaction of diazomethane with viscose solution.
acid degradation more rapidly than does cellulose acetate. It may there-
fore *be expected that the methyl-substituted portions of a mixed acetate-
methyl ether will be hydrofyzed preferentially. This will lead to an in-
creased chance that any remaining dimer unit will be substituted only with
acetyl groups, and will enhance the yield of cellobiose octaacetate above
that to be expected on comparison with cellulose itself, for which the theo-
retical yield is only 67 mole per cent,26 even if all cellobiose units, once they
are formed, are protected from further degradation.
In spite of the uncertainty arising from the above considerations, ace-
tolysis is a potentially valuable tool for the evaluation of the uniformity of
«4 T. Lieser and R. Jaks, Ann.. 548, 204 (1941 ).
» W. Kuhn, Ber., 63B, 1503 (1930).
IX. DERIVATIVES OF CELLULOSE 687
cellulose ethers. As was pointed out above, it is desirable for this purpose
to use a method for the determination of some constituent that would be
present in very small quantity in a uniform product. Above a D.S. of 1,
the content of unsubstituted cellobiose units fulfills this requirement ad-
mirably. It is therefore interesting to compare with the theory for a uni-
form product some yields of cellobiose octaacetate obtained by Lieser and
Jaks24 from methyl celluloses obtained by reaction of fibrous cellulose with
methyl sulfate in the presence of 20% NaOH. These data are given in
Table 1, together with comparable data for cellulose, a technical water-
soluble methyl cellulose, and the previously mentioned methyl cellulose
samples prepared from cellulose xanthate. It is immediately apparent
from these results that the yields of cellobiose octaacetate from the fibrous
products prepared with 20% NaOH were higher than could have been
expected from a uniform product. The technical product and the ethers
prepared from xanthate gave yields in very good accordance with the theory
for a uniform product. These differences are in agreement with general
experience. The technical products, prepared with strong NaOH and
methyl chloride, are much more uniform than are ethers prepared with
methyl sulfate and weak NaOH (see Section E of this Chapter IX). The
product prepared in solution from the xanthate would be expected to be
uniform.
On the basis of these results, the acetolysis method would be expected
to be especially suited to the evaluation of uniformity of technical products
in the substitution range 0.7-1.2.
(d) COMBINED RATE AND EQUILIBRIUM CONTROL
Until recently, very little progress had been made in the study of the
distribution of substituents in cellulose esters. This is, of course, a diffi-
cult field because of the ease of removal and migration of ester groups.
To be sure, previous investigators8'9'11 had shown that the secondary
cellulose acetates of commerce had roughly equal amounts of hydroxyl
content in the primary and secondary positions. The situation has now
been largely clarified by the efforts of Hiller26 and Malm and co-
workers.8'27""29 These workers found that the normal secondary cellulose
acetate of commerce (D.S. 2.3-2.4), when dissolved in acetic acid contain-
88 L. A. Hiller, Jr., J. Polymer Sci., 10, 385 (1953).
87 C. J. Malm, L. J. Tanghe, and B. C. Laird, /. Am. Chem. Soc.t 70, 2740 (1948).
88 C. J. Malm, L. J. Tanghe, and B. C. Laird, /. Am. Chem. Soc., 72, 2674 (1950).
20 C. J. Malm, L. J. Tanghe, B. C. Laird, and G. D. Smith, /. Am. Chem. Soc., 74,
4105 (1952).
688
CELLULOSE
ing 2-3% water, would initially increase slightly in substitution and then
gradually decrease. With products of lower substitution (D.S. 1.7), the
initial rise in substitution was much more pronounced. This behavior is
illustrated in Figure 3 for three samples of cellulose acetate. The sulfuric
acid used as a catalyst in this case did not change the general nature of the
results, though the higher temperatures necessary to secure acceptable
reaction rates in the absence of catalyst decreased the initial rise. It was
48 72 96
TIME(hourt)
Fig. 3. Acid-catalyzed reaction of secondary cellulose
acetates with 97.2% acetic acid (Hiller26). Concen-
tration of H2SO4, 0.1%; temperature, 65.7 °C.
clear to both sets of workers that the explanation of this behavior was
simple. The free primary hydroxyl content of the sample was initially
greater than that corresponding to equilibrium with the acetic acid-water
mixture used, and these primary hydroxyls were acetylated fairly rapidly,
resulting in a tendency for the substitution to rise. At the same time, the
content of free secondary hydroxyl was initially lower than the equilibrium
value, so that deacetylation occurred in these positions, but at a lower rate
than the initial acetylation in the primary position. As the more rapid
reaction in the primary position approached equilibrium the continuing
hydrolysis in the secondary positions became dominant. The validity
of these conclusions was checked by tritylation experiments. It was also
found by Malm29 that the rate of the acid-catalyzed acetylation by acetic
anhydride was more rapid in the primary position than in the secondary
positions. Furthermore, there was a good correlation of optical rotation
DC. DERIVATIVES OF CELLULOSE 689
with the relative amount of free primary and secondary hydroxyls, allow-
ing the redistribution reaction to be followed readily.
Hiller26 undertook a very elaborate mathematical analysis of his rate
data, with the objective of calculating the pertinent rate and equilibrium
constants. He was unable to segregate the effects of reaction at the two
secondary positions, and therefore based his calculations on the question-
able assumption that the rates in the secondary positions could be charac-
terized by a single constant. This is a fairly serious drawback of those
portions of his calculations based on observations after long elapsed times,
and particularly of his estimates of the equilibrium constants for reaction
in the secondary positions. However, his results should prove of great
value in the interpretation of the commercial process of hydrolysis of cellu-
lose triacetate to secure uniform products of lower substitution. Killer's
values of rate constants are given in Table 2 and the equilibrium constants
TABLE 2
Rate Constants for Acetylation of Cellulose (Hiller26)
Rate constant X 10s, mole~J hr. ~l
Primary groups
Secondary groups
Temperature, ° C.
Acetylation Deacetylation Acetylation
Deacetylation
Uncatalyzed reaction
84.4
2.6 ± 1.8
1.9
0.74
0.057
94.2
6.5 =fc 3.0
3.6
1.7
0.15
104.4
9.2 d= 3.0
4.9
3.6
0.40
115.1
14.8 =b 6.0
7.4
7.6
0.94
Acid-catalyzed reaction
(0.1%
H2SO4)
46 0
1.5 ± 0.2
2.5
0.92
0.035
54.8
3.8 db 1.8
4.7
1.3
0.060
65.7
6.3 ifc 1.0
6.3
2.0
0.12
are shown in Figure 4 as a function of temperature. The net result of the
interaction of the rate and equilibrium constants is that in a given acetic
acid-water mixture, the equilibrium extent of acetylation of the primary
hydroxyls tends to be high, and the rate of attainment of this equilibrium
is relatively high. Conversely, with the secondary hydroxyls the equilib-
rium substitution is low, as is the rate of attainment.
In commercial practice, the hydrolysis of primary cellulose acetate is
carried out in the presence of so much water that the equilibrium substitu-
tion would be very low for both primary and secondary hydroxyls. As a
690
CELLULOSE
consequence, the reaction is largely rate controlled, and the primary hy-
droxyl content is considerably higher than the equilibrium value. This
has important consequences, since a high ratio of free primary to secondary
100
10
I
o
u
2
flC
CD
O
lit
1.0
O.I,
I
I
2.7
2.9 3.1
; 10* (degree"1)
3.3
3.5
Fig. 4. Temperature dependence of equilibrium constants for hydrolysis of
cellulose acetate (Hiller26). K\ is equilibrium constant for hydrolysis of primary
hydroxyls; K*t for secondary hydroxyls; K, for over-all reaction.
hydroxyl in cellulose acetate results in a wider range of solubility than the
converse situation. A similar situation is encountered with ethyl cellulose
of low D.S.; the product substituted only in the secondary positions, pre-
pared in the NaOH~Cu(OH)2 system, is more soluble in water than the
usual product prepared in the absence of copper.15
IX. DERIVATIVES OF CELLULOSE 691
In addition to his studies of rates and equilibrium of acetylation, Hiller26
also determined the rate of degradation of his samples. This facet of his
work will be treated in more detail in Section G of this Chapter IX. The
result of paramount importance in connection with this discussion is the
fact that even under the most favorable conditions of low temperature and
high catalyst concentration, the rate of degradation is so high that equilib-
rium cannot be approached closely without excessive degradation. It
is therefore clear that a uniform cellulose acetate of D.S. 2.0-2.5 and of
acceptably high degree of polymerization (D.P.) will never be secured by
reaction of a nonuniform product with acetic acid-water mixtures until the
equilibrium is closely approached. This conclusion is especially true for
the direct acid-catalyzed esterification of fibrous cellulose. For this reason
the commercial process involves acetylation nearly to the triacetate, which
must then be uniform. This product can then be hydrolyzed in solution
to give a uniform secondary acetate.
2. Reaction of Cellulose Fibers
The raw material for the preparation of cellulose derivatives is nearly
always fibrous cellulose derived from plant sources. However, a large
fraction of the cellulosic material in these native fibers is crystalline in
nature (see discussion of fiber structure in Chapter IV-B), and the interior
of the crystalline regions is inaccessible to chemical reagents as long as
this native crystalline structure is maintained. All reactions of the crystal-
line region have to start at the surface and proceed gradually inward. It
has long been apparent that this inaccessibility of a portion of the cellulose
hinders a uniform reaction. Means for overcoming this tendency toward
nonuniformity constitute a large portion of the technology of cellulose
reactions. The situation may be much more serious than mere nonuni-
formity of the product; cellulose may not react at all or only with extreme
difficulty in a reaction that proceeds nicely with low molecular weight
compounds. Therefore, the rate of reaction of cellulose as well as the uni-
formity of reaction must be considered. In addition, the possibility of con-
ditions promoting serious degradation always exists.
The difficulties of rate and uniformity of reaction of fibrous cellulose are
due, beyond a shadow of a doubt, to the difficulty of getting the reagents
to the cellulose hydroxyls. In order to point out the nature of the stages
of difficulty that are encountered, a simplified r£sum6 of some of the de-
tails of fiber structure discussed in Chapter IV will be presented.
692 CELLULOSE
(a) FIBRILLAR STRUCTURE OF NATIVE CELLULOSE
The basic unit of organization of native cellulose, whether fiber, mem-
brane (Valonia), or unorganized products such as bacterial cellulose, is a
fibril about 100 A. in diameter and of great but indefinite length.
As was mentioned in Chapter IV-B, the current theories of cellulose
fiber reactivity do not take the fibrillar structure into account. Some of
them are even incompatible with the idea of discrete fibrils. In the follow-
ing discussion, an attempt will be made to show that the peculiarities of
cellulose reactions are easily explained on the basis of behavior of and inter-
actions of fibrils.
The literature of cellulose reactions contains many designations founded
on preconceived ideas of mechanism or else designed to conceal ignorance.
Among these terms are microheterogeneous and macroheterogeneous re-
action, micellar surface reaction, permutoid or quasihomogeneous reaction.
Since cellulose reactions are not as clean-cut as these terms would indicate,
and in any case terminology does not assist in understanding, no further
use of these designations will be made in this Section.
Each fibril behaves as if it were largely crystalline in its core, with the
degree and perfection of organization decreasing as the surface is ap-
proached. There is no evidence at all for a separate, amorphous phase or
cementing material holding the fibrils together in fibers of purified native
cellulose. The bonding between fibrils in fibers appears to be due to the
same factors that operate in bonding fibers to make paper (Chapter VIII).
The bonding in paper seems to be due to several factors. The surface
area of the fiber is increased by tearing and splintering. The newly
created surfaces are swollen by water and made easily deformable. When
these surfaces are dried in contact with each other, strong bonds allied in
character to the formation of interpenetrating crystalline regions are formed.
It is especially characteristic that these bonds between fibers are preserved
when the paper is nitrated or acetylated. In the same manner, a fiber
can preserve its identity when transformed into a derivative. This anal-
ogy furnishes good evidence that cellulosic surfaces created in the pres-
ence of water will adhere strongly when dried in contact.
In the water-swollen state during growth of the fibers, crystallization is
not complete although the cellulose molecules must be largely oriented in
the direction of the axis of the fibril.30 In this condition the fibrils will be
easily deformable and will therefore pack fairly closely in the fiber. As
water is withdrawn by growth processes or drying, crystallization in the
fibril will occur. It is plausible to postulate that occasionally this crystal
» E. E. Berkley and T. Kerr, Ind. Eng. Chem.. 38, 304 (1946).
IX. DERIVATIVES OF CELLULOSE 693
growth process will occur in such a fashion that it can cross over into an
adjacent fibril, thus forming a very strong bond between fibrils. More often,
the structure will remain disorganized at the interface between fibrils and
there will remain a surface of weakness. The degree of interfibrillar crys-
tallization may well be an important factor in the phenomenon of "horni-
fication" (very poor reactivity) referred to later. For the present dis-
cussion, it will be assumed that a normally reactive cellulose sample is
being co usidered, and that the bonding between fibrils is loose.
The fibrillar structure of the fiber enables distinguishing three stages of
penetration of reagents. (1) The fiber with a diameter of about 100,000
A., may react only on the surface. (2) The surface of the fibril, with a
diameter of 100 A., may be available for reaction. (3) The cellulose mole-
cule, with a diameter of less than 10 A., may be freely accessible to the re-
agents. On the basis of these considerations alone, it is evident how much
can be gained if reagents can penetrate the fiber. A thousandfold decrease
in the time required for reaction would be anticipated if the surface of the
crystalline regions could be made freely available, as compared to the time
required if a reaction has to work its way inwards from the outer fiber
surface. Of course, a further increase in rate would be expected if, by the
destruction of the crystalline structure, the individual cellulose molecules
could be made available for the reaction from the beginning.
The preceding discussion allows some insight into one of the most im-
portant facts of cellulose behavior, that is, that reactivity improves as the
degree of swelling in the reaction medium increases, other factors being con-
stant. Since, however, the driving force for the reaction is a function of
reagent concentration in the reaction medium, two factors must always be
kept in mind. The medium should allow swelling to occur, yet conditions
favorable for the chemical reaction in question must be maintained.
That swelling is sufficient to ensure access of reagents to the inside of the
fiber is not immediately evident. It might be thought that it would be
necessary to have actual capillary channels into the fiber. In capillary
channels of a size that would be consistent with other evidence, however,
the viscous resistance to actual flow would be enormous and all transfer of
reagents would have to be by diffusion. Consequently, all that is involved
is the variation of rate of diffusion with the degree of swelling. It has long
been known that the rate of diffusion of substances of low molecular weight
in dilute solutions or gels of high polymers is very little less than in
water,81'82 in spite of the macroscopic high viscosity of the solution. It is
" R. Taft and L. E. Malm, J. Phys. Chem., 43, 499 (1939).
32 J. J. Bikerman, /. Phys. Chem., 46, 724 (1942).
694 CELLULOSE
now realized that the local viscosity in a swollen polymer controls the rate
of diffusion as well as the flexibility and extensibility of the structure.
If swelling has caused the sample to be limp and easily deformable, the rate
of diffusion will be high. It is a matter of everyday knowledge that cellulose
fibers and films a^e greatly softened by the uptake of a few per cent of water.
The effect of this water of swelling on the rate of diffusion is very pro-
nounced, as may be seen in Table 3.33-34 A vast amount of experience on
TABLE 3
Influence of Swelling on Diffusion of Water Vapor through Cellophane
(From data of Hauser and McLaren,38 Doty, Aiken, and Mark84)
Permeability,
moles/sq. cm. /sec. /cm.
mercury pressure difference
Relative humidity, % for 1-cm. thickness
100 2700 X 10 -»
50 300
0 1.5
the influence of the degree of swelling on the rate of diffusion in polymers
indicates that the above behavior may be taken as typical of other swelling
agents and other diffusible substances.
Some speculation is in order about the details of the swelling mechanism
of native fibers, particularly with agents such as water and pyridine, which
open up the interior of the fiber without noticeable attack on the crystal-
line regions. It is probable that the outer wrapping layers of the fibers,
which interfere very markedly with the action of strong swelling agents
(Chapter IV-B), are not important in the case of the limited swelling pos-
sible with water, pyridine, or acetic acid. After all, the fiber was more
highly swollen when originally laid down, and the wrapping layers had to
be able to accommodate the corresponding degree of swelling. It is thus
probable that the outer portion of each fibril throughout the fiber swells
slightly, causing the fiber as a whole to expand; at the same time the fiber
loses some of its flattened shape and becomes more cylindrical. There will
then be an intercommunicating network of swollen material, allowing ready
diffusion of reagents into the fiber.
Inspection of electron micrographs of clumps of fibrils in cotton fiber dis-
integrated under water gives the impression that the fibrils have enough
elasticity to spring apart under swelling conditions and thus create actual
38 P. M. Hauser and A. D. McLaren, Ind. Eng. Chem., 40, 112 (1948).
34 P. M. Doty, W. H. Aiken, and H. Mark, Ind. Eng. Chem., Anal. Ed.. 16, 686 (1944).
DC. DERIVATIVES OF CELLULOSE 695
voids in the structure (Chapter IV-C). Such an interpretation must be
viewed with caution, since these observations were made on samples re-
dried without the constraints present in the intact fiber. However, the
ease of splitting along the fibrillar interfaces demonstrated by the electron
micrographs certainly proves that lateral bonding between fibrils is weak
in the presence of swelling agents. That actual channels are opened up
by the elastic forces between the somewhat disordered fibrils is thus quite
plausible. It must be emphasized, however, that these channels are prob-
ably not intercommunicating and certainly do not occupy a large fraction
of the increase of volume of the fiber that is observed on swelling. Also,
fibers such as ramie show no evidence of channels between fibrils, yet ramie
with its beautifully parallel fibrils is nearly as reactive as cotton or wood
pulp with a more irregular structure (see Fig. 49, Chapter IV-C). It is
therefore very probable that most of the transfer of reagents in swollen
fibers is through a zone of swollen material surrounding the fibrils, rather
than primarily through empty channels.
(b) VARIATION OF REACTIVITY OF CELLULOSE
The above concept of interfibrillar swelling has been discussed in detail
to help explain the very wide variation in the reactivity of different cellulose
samples. This variation is particularly evident in reactions such as acetyla-
tion, where the swelling power of the medium is necessarily limited by the
lack of technically suitable chemically inert swelling agents. The reactiv-
ity of different samples of cellulose in such cases can range from very
good to very poor. In cases of poorly reactive cellulose samples, reaction
is always observed to be largely confined to the surface of the fiber. If
drastic conditions or very long times are used, the whole fiber will, of course,
eventually react. The difficulty here is that those portions of the fiber
that react first are subjected to the full degrading action of the medium for
a long time, and the final product will be too low in D.P. to be useful. It
is therefore desirable to use cellulose of high reactivity, even in cases where
a long reaction time would be economically feasible.
From experimental evidence, much of the observed difficulty with poor
reactivity can be traced to a decrease in the reactivity of initially reactive
cellulose, that is, the hornification referred to previously. If, for example,
a water-wet sample of reactive cellulose is dried slowly at 100°C., there will
be a pronounced drop of reactivity as measured under normal acetylating
conditions. The effect is much more pronounced if the cellulose is first
swollen in aqueous NaOH, water-washed, and then dried hot.
The opinion seems to be gaining ground that this loss of reactivity is due
696 CELLULOSE
to an effective cross-linking of the cellulose, which reduces or even com-
pletely inhibits the ability of the fiber to swell in usual acetylation baths.
There have been numerous proposals of actual chemical cross-linking, but
there seems to be no necessity to invoke such action. The possibility of
crystal growth between fibrils appears to be an adequate explanation
The conditions of heat and moisture that lead to hornification are also those
that promote crystallization of cellulose. Another indication that partial
cross-linking of fibrils by crystallization is the correct interpretation is the
fact that the effect of hornification is not nearly so pronounced with re-
action media, such as NaOH or nitrating baths of high HNOa content,
which have the ability to penetrate and modify the crystal structure of
cellulose.
(c) ENHANCEMENT OF REACTIVITY
The reactivity of fibrous cellulose that has not been subjected to adverse
conditions, such as high-temperature drying, is adequate for most purposes.
However, much higher reactivity in media of even very poor swelling ability
can be obtained if desired. The basis of all pretreatments to enhance re-
activity is the great hysteresis of deswelling and crystallization character-
istic of cellulose. An example of this hysteresis is the greater moisture
content of fibers conditioned to constant humidity from a higher humidity,
as compared to conditioning to the same humidity from a dry state (Chap-
ter IV-D). This hysteresis is much more pronounced if the dehydration is
carried out from an initially higher degree of swelling than is possible with
water alone. The enhancement of reactivity is most effective if dehydra-
tion is accomplished by displacement of water by organic solvents. For
example, if cellulose is first dissolved in cuprammonium, precipitated in a
nearly amorphous state, water-washed, and the water then displaced with
organic solvents, a very reactive product is obtained.35 Acetylation of
such an expanded cellulose "with pyridine-acetic anhydride mixtures pro-
ceeds easily to give a soluble product, whereas native fibers under the same
conditions acetylate with difficulty to give insoluble products.86'87
One of the most complete studies of the comparative ease of acetylation
of cellulose subjected to swelling and then to either vacuum drying or sol-
vent displacement of water from the swollen condition has been carried
36 P, Karrer, Einfuhrung in die Chemie der Polymer en Kohlenhydrate, Akadem. Ver-
lagsgesellschaft, Leipzig, 1925, p. 176.
M H. Staudinger and B. Ritzenthaler, Ber., 68B, 1225 (1935).
» H. Staudinger and G. Daumiller, Ann., 529, 219 (1937).
IX. DERIVATIVES 'OF CELLULOSE 697
out by Staudinger and coworkers.88 These results Will be cited in de-
tail in order to give a complete picture of the observed effects. The swell-
ing was brought about by water or by 20% NaOH, as well as by reprecipita-
tion from cuprammonium solution. After thorough water- washing, the
swollen samples were either vacuum dried at 40°C. or solvent displaced
with acetic acid, pyridine, or methanol followed by benzene. With acetic
anhydride-pyridine and cotton at 60° C. the order of decreasing reactivity
and, in parentheses, the corresponding acetyl content after 24 hrs., were:
(1) reprecipitated, pyridine displaced (27%); (2) mercerized, pyridine
displaced (14%); (3) water-swollen, pyridine displaced (8%) ; (4) water-
swollen, dried (5%); (5) mercerized, dried (0.7%). Ramie, linen, and
hemp reacted at practically the same rate as cotton. The degree of poly-
merization also had little influence. All of the above products were nearly
completely insoluble in chloroform, showing the absence of a triacetate
layer on the surface. The solvent-exchanged samples appeared uniform
under microscopic examination.
Similar comparisons were made in another fibrous acetylation process,
with H2SO4 as catalyst and with benzene in the reaction medium to prevent
solution of the product. The mercerized, benzene-displaced sample had
an acetyl content of 25% after 1 hr. and 44% (the triacetate) after 24 hrs. ;
the mercerized, acetic acid-displaced sample, 21% and 44%; the water-
swollen, acetic acid-displaced sample, 17% and 42%; the water-swollen,
dried sample, 2% for high D.P. and 5% for low D.P. after 24 hrs.; the
mercerized and dried sample, 1% after 24 hrs. In most cases, again, the
D.P. and fiber type (i.e., ramie, hemp) had little influence. It was further
observed that the benzene-displaced samples retained their high reactivity
after vacuum drying, and retained 4-8% benzene that could not be removed
by vacuum drying alone. The degradation after 24 hrs. acetylation time
decreased as the reactivity increased. The superiority of the more re-
active samples could doubtless have been much more marked if the compari-
son of degradation had been made at times required for a given degree of
substitution to be reached. It was found that the dried celluloses gave
partially substituted products from which most of the acetyl content could
be extracted with chloroform as cellulose triacetate. Only a small fraction
of the acetyl content of the acetic acid-exchanged products could be re-
moved with chloroform, up to an acetyl content of more than 35%. Be-
yond this point, the solubility increased rapidly.
Staudinger interpreted his results to mean that all of the cellulose mole-
88 H. Staudinger, K.-H. In den Birken, and M. Staudinger, Makromol. Chem., 0, 148
(1963).
698 CELLULOSE
cules in the solvent-exchanged fibrous products were available for reaction,
even in the crystalline regions. So extreme a viewpoint is hardly war-
ranted. In cases discussed below where it is definitely possible for reaction
to occur in the crystalline regions, completely soluble products are obtained
at intermediate degrees of reaction. This is never possible with cellulose
acetate prepared directly from fibrous cellulose. On a submicroscopic
scale, there must be discontinuities in reaction during this process.
(d) AN INTERPRETATION OF REACTIONS OF NATIVE CELLULOSE FIBERS
A satisfactory interpretation of the observed facts can be built up along
the following lines. When fibers that do not swell are reacted, the reaction
zone is observed to start in isolated spots at or near the surface of the fiber,
and then to spread gradually, leaving cellulose triacetate behind the reac-
tion front.88-89 All that is necessary is to assume that under conditions of
good swelling, everything is displaced in dimensions by a factor of a thous-
and. The fibril is now exposed to the reagents, and each fibril will begin
reacting along its length at isolated spots that either have greater dis-
order or are more exposed to a direct diffusion path for reagents. In those
cases, for example, fibrous acetylation or nitration, where the product crys-
tallizes as its formation is complete, the reaction zone may be pictured as in
Figure 5. At an intermediate state of reaction, each fibril may be pictured
as having many alternate zones of completely reacted and unreacted cellu-
lose, with cellulose molecules passing between them. Between these zones
will be regions of incomplete reaction, which will probably be more swollen
than indicated in Figure 5. The swelling of these partially reacted regions
may be expected to be much higher than either the reacted or unreacted
zones, for the reasons indicated in the discussion of influence of degree of
substitution on solubility in Chapter X-A. This swelling will facilitate
easy penetration of reagents into the reaction zone. The structure as a
whole will be insoluble until^the zones of complete reaction have coalesced.
If the medium is a solvent for the product, as in commercial cellulose
acetate preparation in contrast to fibrous acetylation, this picture needs to
be modified only slightly. In this case, the reacted portions will swell as far
as is allowed by the constraints imposed by cellulose molecules pene-
trating into the unreacted zones. The fiber as a whole will appear under
the microscope to swell uniformly as reaction proceeds, but the continuing
existence of the fibrils will hold the structure together up to a high extent
of reaction. This is exactly what has been observed in many microscopic
studies of cellulose acetylation.
** K. Kanamaru, Hek. Chim. A eta, 17, 1436 (Id34).
DC. DERIVATIVES OF CELLULOSE
699
There are many observations which support the picture of localized
attack on the cellulose fibril, with the reaction zone then spreading until
the whole structure has reacted. When fibers are degraded with acids un-
der conditions of only slight swelling, they split up into fragments a few
hundred Angstroms long (Chapter IV, Sections B and C). The same type
of fragment can be isolated from esterification reactions, if the conditions
OH
Fig. 5. Reaction zone progressing along crystalline region with retention of fiber
structure (Spurlin1).
are such that the reacted portion of the material is rapidly degraded, thus
cutting the interpenetrating molecules which normally hold the unreacted
zones together.40 If, on the contrary, degradation conditions are not severe
the few per cent or so of insoluble material remaining near the end of a
reaction will consist of gel particles (crystalline fragments surrounded by
a swollen mass of reacted material). The extreme behavior is exhibited
by fibrous cellulose triacetate of high D.P. prepared under nondegrading
reaction conditions. This product is insoluble in the usual triacetate
solvents even at very nearly complete reaction. It has been shown by
<° R. Signer, A. Aeby, F. Opderbeck, and H. Studer, Monatsh., 81, 232 (1960).
700 CELLULOSE
Centola41 that such products still show the x-ray diffraction pattern of
cellulose, and apparently even 1% of unreacted cellulose can cause insolu-
bility in such cases.
It may be concluded from the discussion above that in many cellulose
reactions the surface of the crystalline region is readily available for re-
action, whereas in other cases reaction is primarily from the fiber surface
inwards. This is not a matter of mutually exclusive alternatives; rather,
all intermediate stages of availability can t>e expected.
Another point to remember is that the progress of the reaction on the
surface of the available regions may impede further reaction. This seldom
happens in reactions in organic media, since the reaction products are
normally more highly swollen and more compatible with the reagents
than is the initial cellulose. However, in the benzylation reaction the
initial product is hydrophobic and is not swollen by the alkali which is
essential for the further progress of the reaction;42 the reaction thus be-
comes more difficult as it proceeds. Another possibility arises when the
reaction cross-links the cellulose. For example, the difficulty of reaction
of cellulose with more than a few per cent of formaldehyde43'44 may be due
to the progressive cross-linking of the structure, which must impede swell-
ing and diffusion.
There is no evidence for a difference between the inherent reactivity
of a crystalline zone of any of the crystalline modifications of cellulose.
The differences of reactivity of native fibers, mercerized fibers, and regener-
ated cellulose can all be explained on the basis of differences in the amount,
size, and degree of perfection of the crystalline regions, and on the greater
ease of hornification of the more expanded structures when subjected to
drastic drying conditions.
(e) REACTIONS OF FIBROUS ADDITION COMPOUNDS OF CELLULOSE
Crystalline addition compounds of cellulose, such as alkali cellulose
(Chapters IV-B and IX-D), are uniformly much more reactive than would
be expected by comparison with native cellulose. This difference seems to
be due to three factors. The media in which the addition compounds are
prepared are good swelling agents for cellulose. The crystalline regions
are smaller. Most distinctive is the good evidence that reagents can pene-
«
41 G. Centola, Atti X° congr. intern, chin., 4, 129 (1939); Chem. Abstracts, 34, 2169
(1940).
42 E. J. Lorand and E. A, Georgi, J. Am. Chem. Soc.t 59, 1166 (1937).
48 R. E. Wagner and E. Pacsu, Textile Research /., 22, 12 (1952).
44 C. F. Goldthwait, Textile Research J., 21, 55 (1951).
IX. DERIVATIVES OF CELLULOSE 701
trate the crystal lattice of the addition compounds as distinguished from
the native cellulose lattice.
This last factor has appeared plausible for many years, in order to ex-
plain the fact that soluble ethers of low D.S. could be obtained by the re-
action of alkali cellulose with etherifying reagents of low molecular weight.6
The most compelling evidence, however, has come from studies of the re-
action of alkali cellulose with CS^ to form cellulose xanthate.45'46 (See also
Chapter IX-E.) In this reaction, an expansion of the crystal lattice be-
gins as soon as €82 absorption starts. The kinetic results support the idea
of a uniform reaction throughout the fiber. If the NaOH concentration is
below that necessary for mercerization, the course of the reaction is entirely
different. An initially rapid reaction slows down very rapidly, so that a
final CS2 uptake of only 30% is found. With NaOH of concentration high
enough to transform the crystal lattice of cellulose to alkali cellulose I,
the final €82 uptake is twice as high. It is probable that with the low con-
centration of NaOH, reaction was largely confined, under the conditions
used, to the more amorphous portions of the cellulose. The above evi-
dence, as well as the evidence from solution properties described in Chapter
X-C, leaves no doubt that reaction has occurred, at least to a considerable
extent, throughout the fiber if the cellulose has been converted throughout
to alkali cellulose.
In general, reactions of alkali cellulose give the most uniform products
with the most water-soluble reagents.6 These would be expected to diffuse
rapidly in the hydrophilic alkali cellulose. Thus CS2 (which possibly
forms a water-soluble complex with NaOH46), sodium chloroacetate,
ethylene oxide, methyl chloride, and methyl sulfate give fairly uniform
products in the fibrous reaction. The ethyl celluloses of low substitution
are relatively nonuniform. Fibrous products of low substitution obtained
with higher alkyl halides are water-insoluble. If, however, the isopropyl
ether is prepared in homogeneous medium, it is soluble.
This trend illustrates the fact that despite the availability of the alkali
cellulose for reaction, no fibrous reaction product of D.S. less than 3 can be
truly uniform unless prepared by an equilibrium reaction. In all fibrous
reactions, some of the material must have a different reaction velocity
than the rest because of either a difference of diffusion path for reagents
or the remaining constraints imposed by the fiber structure. The impor-
tance of such constraints is illustrated by the fact that the mercerization
reaction itself is hindered by tension or other forces applied to the fiber.
46 K. Hess, H. Kiessig, and W. Koblitz, Z. Elektrochem., 55, 697 (1951).
46 H. Grotjahn, Z. Elektrochem., 57, 305 (1953).
702 CELLULOSE
It can easily happen that there are local stresses in the fiber that cause a
different alkali concentration to be attained in different parts of the fiber.
In any case, the crystal lattice forces themselves must cause the available
alkali concentration inside a crystalline region to be different from that in
more amorphous regions.
When all of these factors are considered, it is surprising that the high-
substitution cellulose ethers of commerce are as uniform as they are.5
Several factors are probably involved in this. One important considera-
tion is that the derivatives pass into solution as the reaction proceeds, so
that at least the latter portions of the reaction are under uniform condi-
tions. Another point is the low affinity of partially reacted portions of the
product for NaOH; this low affinity will lower the reactivity of these re-
gions in comparison with portions of lower extent of reaction. Finally,
it appears to be necessary to invoke the hypothesis of large mutual inter-
ference between positions 2 and 3, which will cause the reaction to slow
down when a D.S. of 2 is reached in the most reacted portion of the product
and allow the remainder to catch up.
3. Chemical Factors Influencing Reactivity
The preceding discussion of reactivity has been from the standpoint of
availability of the cellulose hydroxyls for reaction. It has been shown that
if the cellulose hydroxyls can be made equally available for reaction by
operating in solution or by allowing the reaction to come to an equilibrium,
products of an acceptable and easily definable degree of uniformity can be
obtained. There still remains the question of the ease of combining the
requirements for availability of the cellulose* with the chemical require-
ments necessary for a reaction to occur and for the degradation to be held
within acceptable bounds. These questions will now be given a short
discussion for each of the common types of reaction of cellulose. For fur-
ther details of individual substitution and degradation reactions, the sub-
sequent sections of this Chapter IX must be consulted.
(a) ESTERIFICATION
(1) Direct Equilibrium Esterification with Strong Acids
A number of strong acids will react directly with cellulose with a speed
such that degradation is held within acceptable bounds. Nitric acid is the
foremost example, but sulfuric acid will also work well. Formic acid is a
border-lipe example. The outstanding characteristics of the acids that work
well are that relatively concentrated solutions in water will dissolve cellu-
DC. DERIVATIVES OF CELLULOSE 703
lose, crystalline addition complexes with cellulose may be obtained, the
equilibrium degree of esterification is high in the presence of a small
amount of water, and the rate of degradation of the product in acid medium
is lower than that of cellulose. Another characteristic of these acids is
that their esters do not hydrolyze in a normal manner in alkali. Nitrate
esters suffer profound degradation in alkaline medium, with little or no
production of nitrate ion. Sulfuric half-esters (which alone are produced
when sulfuric acid acts on cellulose) act like alkylating agents on alkaline
hydrolysis. If at all possible, they tend to form ether linkages with other
hydroxyls rather than to form free hydroxyl groups. Apparently, the
cellulose oxygen holding the sulfate group may be lost on alkaline hydroly-
sis.
Most other strong acids do not esterify cellulose because the hydrolysis
constant of the product is unfavorable. There is some evidence that an-
hydrous phosphoric acid esterifies cellulose slightly.47 Perchloric acid can-
not form normal esters. Halogen acids are abnormal in that the cellulosic
hydroxyls are split off during reactions. These acids also degrade cellulose
very rapidly. Anhydrous halogen acids are also not good swelling agents.
The tendency of nitric acid and sulfuric acid to dissolve the reaction prod-
uct can be modified by the addition of other substances while still maintain-
ing enough swelling to ensure equilibrium nitration or sulfation. The nitrate
and, to a lesser extent, the sulfuric half -ester are the only well-known ex-
amples of cases where a uniform ester of intermediate substitution can be
obtained by a reaction with acids with retention of fiber structure.
(2) Organic Esters Prepared under Acidic Conditions
As explained previously, the direct esterification of cellulose with organic
acids is not a satisfactory reaction for two reasons. The equilibrium con-
stant is unfavorable. This necessitates the use of media high in organic
acid content and low in water content. Such media are very poor swelling
agents for cellulose, so that the reaction is nonhomogeneous and slow.
The second reason is that the rate of degradation is comparable with the
rate of esterification with organic acids whether the reaction is catalyzed
or uncatalyzed.
The degradation has less relative effect if the driving force for acetyla-
tion is increased above that possible in an equilibrium reaction. This is
done in practice by the use of acetic anhydride. With this reagent, the
47 E. Heuser, W. Shockley, A. Adams, and E. A. Grunwald, Ind. Eng. Chem., 40, 1600
(1948).
704 . CELLULOSE
equilibrium substitution product is the triacetate. It would still be pos-
sible to secure a uniform partially substituted cellulose acetate by direct
reaction if a medium were available in which the reaction could be carried
out in solution from the beginning. At this point, however, chemical in-
compatibilities become important, since there are very few anhydrous
solvents for cellulose which would not destroy the acetic anhydride or
react with the cellulose. Anhydrous sulfuric and phosphoric acids are
obvious possibilities. Phosphoric acid solutions were tried by Heuser,47
with fairly satisfactory results. Products of 20-25% acetyl content could
be obtained by the use of equal parts of acetic acid and 100% phosphoric
acid in what appeared to be a homogeneous reaction. By using acetic
anhydride in place of acetic acid, any level of substitution up to the tri-
acetate could be obtained. The products of D.S. 2.0-2.5 were incompletely
soluble in acetone, however. The low solubility may have been due to the
fact that these products were nearly completely substituted on the pri-
mary hydroxyl group. There is also some probability that solution of
cellulose in concentrated phosphoric acid is not complete (Chapter X-A)
and that therefore a nonuniform product was obtained. No one appears
to have tried the homogeneous reaction of acetic anhydride with cellulose
in solution in 100% sulfuric or trifluoroacetic acids, both of which are re-
ported to be solvents for cellulose. It has been found,48 however, that
activated cellulose reacted with a mixture of 60 parts of acetic acid, 40
parts of acetic anhydride, and 2 parts of trifluoroacetic acid gave a fiber-
free reaction mixture at a D.S. of 2.23. This product was also insoluble
in acetone. There are some processes for the direct preparation of fibrous
cellulose acetate of D.S. 2.0-2.5 that involve the use of large quantities of
H2SO4. In this case, the initial product is a mixed acetate sulfate.
The above results have been cited at length because they indicate that
even if the long-sought goal of uniform direct acetylation of cellulose to a
D.S. of 2.5 could be reached, the product might be unsatisfactory for com-
mercial uses because of its low content of free primary hydroxyl groups.
The situation on the acid-catalyzed esterification of cellulose by organic
acids and anhydrides may be summarized by the statement that it is
almost but not quite impossible to combine the chemical requirements and
the swelling requirements so that perfectly uniform reaction conditions can
be obtained. The present commercial process, based on enough swelling to
get rapid but not uniform reaction in the esterification cycle, followed by
uniform hydrolysis to the desired D.S., appears to be the most satisfactory
solution to the problem.
« P. W. Morgan, Ind. Eng. Chem., 43, 2575 (1951).
IX. DERIVATIVES OF CELLULOSE 705
(5) Reactions Requiring Media with No Active Hydrogens
There are a variety of other reagents that will react with hydroxyl groups
under suitable conditions to give esters. Among these are ketene and
isocyanates. These reagents give very poor results with cellulose because
they react vigorously with nearly all swelling agents and with all solvents
for cellulose. The best that can be done is to use a highly activated cellu-
lose and a tertiary amine as a catalyst and partial swelling agent, and to
carry the reaction to completion. Even under such conditions ketene re-
acts unsatisfactorily because of its tendency to polymerize. Ketene can
be used in an acetic acid acetylation, but under this condition it is first
converted to acetic anhydride.
(4) Esters Prepared under Basic Conditions
Cellulose esters may be prepared from acid anhydrides and chlorides
with basic rather than acid catalysts. (In spite of the fact that the base
may be consumed in the reaction and therefore is a reagent, these reactions
are definitely base-catalyzed. For example, sodium acetate is a catalyst for
the reaction of acetic anhydride with an alcohol.) If the reaction is run
under anhydrous conditions, with a tertiary amine as base, the reaction
proceeds fairly rapidly but in a very nonuniform manner. There is also
difficulty in the case of cellulose esters of sulfonic acids prepared from the
acid chlorides because of the formation of quaternary salt derivatives of
cellulose with the organic base and the replacement of the sulfonyl group
by halogen. Again, this is a case where the swelling and the chemical
requirements are incompatible.
Acid halides may also react with alkali cellulose. This is a surprisingly
satisfactory reaction, not nearly as much reagent being wasted in side
reactions as might be expected. The uniformity considerations are en-
tirely analogous to those in etherification.
Some attempts have been made to prepare cellulose halides by the use
of such reagents as thionyl chloride uTpyridine. It has been impossible
to secure uniform, soluble products, and degradation appears to be severe.
This degradation appears to be a necessary consequence of any reaction
which removes the hydroxyl groups from cellulose.
(b) ADDITION COMPOUNDS AND SOLUTIONS OF CELLULOSE
From the standpoint of reaction rate, the use of addition compounds
of cellulose as intermediates for the preparation of their derivatives is of
great interest. Those stable in the presence of water are formed very
706 * CELLULOSE
rapidly. The crystalline structure of cellulose is greatly disorganized in the
addition compounds, and this disorder is partially maintained on regenera-
tion. This use of addition compounds enables the ready activation of cellu-
lose for other reactions. The addition complexes of cellulose with acids,
bases, salts, and especially copper are also involved in all of the so-called
solvents for cellulose. It is not cellulose itself that dissolves, but a com-
pound (Chapter X-A).
Since the addition compounds are formed and decomposed very rapidly,
there is little interest in the kinetics of their reactions. The equilibria
involved are, however, important. There has been much confusion in this
field because of the neglect of some of the principles of the phase rule and of
the statistical principles of cellulose reaction. For example, attempts have
been made to determine the composition of cellulose addition compounds
by precipitating the addition compound from solution. The assumption
is implied that the ratio of complexing reagent to cellulose is the same in the
precipitate as in solution. With reactions that proceed as rapidly as the
formation and decomposition of cellulose addition compounds, this assump-
tion is entirely unjustified. The cases where this mistake has been made
in the literature will not be listed. Rather, it will be pointed out that there
are entirely valid methods that can be applied to the solutions themselves.
The examination of the variation of optical rotation, of pH, or of light
absorption as the ratio of cellulose to complexing reagent is varied is an
example.49
Even in the case of the solid, crystalline complexes the determination of
the combining ratio is not simple. The lattices of the addition compound
are maintained with only slight, continuous change over a wide variation of
composition of the compound. The case is thus somewj^t analogous to
the formation of a continuous series of mixed crystals. This question is
discussed in more detail in Section D of this Chapter IX and in Chapter
IV-B.
(c) CELLULOSE ETHERS
The chemistry of the formation of cellulose ethers is usually based on
(1) the Williamson reaction of an alkyl halide, sulfate, or sulfonate on a
derivative of cellulose with a strong base, or (2) the addition of an active
reagent such as ethylene oxide, acrylonitrile, formaldehyde, or acetylene
to the hydroxyl group. (The reaction of diazomethane discussed in Section
E of this Chapter IX is an exception.) The technical problems are as-
49 P. Job, Ann. chim.. [10], 9, 113 (1928); W. C. Vosburgh artf R. G. Cooper, /. Am.
Chew. Soc.t 63, 437 (1941).
DC. DERIVATIVES OF CELLULOSE 707
sociated with the difficulty of securing uniformity of reaction and the wast-
age of reagents by reaction with water. There is no difficulty with deg-
radation in any of the commercial processes, since cellulose does not de-
grade rapidly in aqueous alkaline media in the absence of oxygen, and the
products are even more stable under .these conditions.
(1) Aqueous Alkaline Reaction Media
With most of the ethers, there is no possibility of securing a uniform prod-
uct by either reacting to an equilibrium condition or removing substituents
from a trisubstituted product. The control of uniformity depends there-
fore on maintaining uniform availability of the cellulose. As indicated
previously, fibrous alkali cellulose gives sufficient uniformity of products for
most purposes. In this case, the chief problem becomes that of side re-
actions leading to wastage of reagent. As the amount of water in the sys-
tem is reduced, this wastage becomes in general less but the uniformity may
also suffer.
All of the usual ethers may be prepared from aqueous alkaline solutions
of cellulose or from solutions of cellulose esters and may therefore be secured
in the uniformly substituted condition. This method is too expensive for
commercial utilization. The large amount of water present and the low
concentration of base both lead to excessive wastage of reagent in side
reactions. As etherification media, the quaternary base solvents for cellu-
lose are expensive and difficult to recover. The pretreatments necessary
to secure solubility of cellulose in 10% NaOH are also expensive. However,
there has not been enough use of homogeneous etherification for scientific
purposes, in order to secure suitable materials for the determination of
D.P.-viscosity relationships and the like.
(2) Anhydrous Etherification by Use of Metal Derivatives
Attempts to secure greater economy of reagents in etherification by the
use of anhydrous metal derivatives of cellulose have failed, as is pointed out
in Section E of this Chapter IX. Even when alkali cellulose is dried to the
point where the residual NaOH solidifies, reaction ceases.60 These ob-
servations indicate the importance of swelling.
A point that is frequently neglected in the consideration of anhydrous
etherification conditions is the relatively rapid degradation of the anhy-
drous metal derivatives.
M J. Ch£din and A. Tribot, Him. services chim. Hat, 33, 169 (1947).
708 CELLULOSE
(3) Equilibrium Etherification
The normal alkylating agents, such as alkyl halides, cannot give equilib-
rium etherification to a point short of completion under alkaline condi-
tions. This statement is made advisedly, in spite of statements to the
contrary in the literature. Just as in the case of the use of acid anhydrides
in esterification, the free energy change or driving force is enormous in the
etherification reaction and the reaction must go to completion if given
sufficient time.
The base-catalyzed addition reactions of olefins bearing activating groups,
such as acrylonitrile, do not belong in the same category as alkyl halides or
sulfates. The reaction of formation of cyanoethyl cellulose is reversible,
and an equilibrium degree of substitution governed by the composition of
the reaction medium is theoretically possible. The practical difficulty
remains that consumption of reagent in side reactions and hydrolysis of
cyanoethyl cellulose to carboxyethyl cellulose will proceed continuously.
In order to secure an equilibrium reaction, it would be necessary to feed
acrylonitrile and NaOH, while removing sodium acrylate (by crystalliza-
tion, for example) until a stationary state had been established. The
product would be a mixed cyanoethyl carboxyethyl cellulose. No at-
tempt seems to have been made to realize such an equilibrium etherifica-
tion in practice.
(4) Acid-Catalyzed Etherification
Since most chemists are aware that ethyl ether can be made from ethyl
alcohol by using acid catalysts, there has been no lack of attempts to trans-
late this type of reaction to cellulose. It was soon realized that the rate of
degradation of the hemiacetal bond in cellulose under any condition of
acidity was enormously greater than the rate of ether formation, so that
this line of attack is hopeless. ^ A more sophisticated approach is to try
the reaction of an olefin with cellulose. Here, the driving force is greater,
as in the case of esterification with anhydrides. Aside from unconfirmed
patent claims, there is still no evidence that even this reaction may be
made to go faster than the degradation reaction.
Like most generalizations, there is an exception to the rule that ethers of
cellulose cannot be made under acid-catalyzed conditions. This exception
is the reaction of aldehydes with cellulose. The cross-linking reaction of
formaldehyde or glyoxal on cellulose is well known, but the intractable
nature of the product has prevented thorough studies of availability and
uniformity of reaction up to the present. Derivatives containing sub-
IX. DERIVATIVES OF CELLULOSE 709
stituents of the type
RceuOCHOCOR"
i'
can be made very readily by adding an aldehyde or an aldehyde deriva-
tive to a cellulose esterification bath.61 This easy reaction of aldehydes
with cellulose is paralleled by the easy reaction of aldehydes with alcohols
of low molecular weight at moderate temperatures and low acidities.
Cyclic acetals of cellulose, analogous to those produced from glycerol or
pentaerythritol, have not been prepared as yet.
The question of the reactivity of cellulose with urea-formaldehyde,
phenol-formaldehyde, or acetone-formaldehyde resins is allied to the
present subject. There is good reason to believe that these resins do form
ether bonds with cellulose when reacted in the presence of cellulose fibers.
The Zelan process (Section E of this Chapter IX) is another example of the
use of derivatives of formaldehyde in order to secure easy reaction with
cellulose.
(5) Metal Derivatives of Cellulose
The ready preparation of the sodium and other alkali metal derivatives
of cellulose is due to the fact that ammonia and the lower primary amines
are good swelling agents for cellulose and at the same time are chemically
compatible with the alkali metals used in the reaction (see Section D of this
Chapter IX). The nonreactivity of the alkali metal derivatives with
alkylating agents when they are freed of swelling agents is another indica-
tion of the importance of swelling.
(6) Replacement of Hydroxyl Groups
It is possible to write reactions for the replacement of the — OH groups
of cellulose by such groups as — Cl, — NH2, — CN, — SR, and many others.
Alternatively,- the OH group might be oxidized to a keto or acid group.
These reactions, at first sight, appear to be well founded on the basis of
reactions of low molecular weight alcohols. A closer inspection shows that
the conditions are not very favorable. With model compounds such as
propylene glycol it is very difficult to get good yields in these reactions with-
out a host of side reactions, many of which result directly or indirectly in
decomposition of the skeleton of the basic molecule. Such side reactions
as formation of a double bond in the anhydroglucose ring would result in
61 T. F. Murray, Jr., and H. LeB. Gray (to Eastman Kodak Co.), U. S. Patent 1,080,-
145 (Oct. 10, 1933); Chem. Abstracts, 28, 320 (1934).
710 CELLULOSE
very rapid degradation of cellulose. This point will be discussed further
in Section G of this Chapter IX. From the viewpoint of reactions, the
essential point seems to be that any reaction conditions which so loosen a
hydroxyl group that it may be removed will also loosen neighboring bonds
so that the cellulose can degrade as a probable side reaction. An example
would be the introduction of an — NH2 group by preparation of the chloride
from thionyl chloride and reaction of this with ammonia. There would
be two displacement reactions in this case. Probably a sulfite is formed
first, and the sulfite group is then replaced by chloride ion. Finally, the
chloride would be displaced by ammonia. In both of the displacement
reactions, extensive side reactions leading to degradation would be expected.
It appears to be well established that only a few hydroxyl groups of cellu-
lose may be displaced before degradation becomes excessive. The oxidized
products may be prepared with only moderate degradation, but when once
formed, they are very sensitive to alkaline degradation (see Chapter III-
C-3).
(d) CATALYSIS OF CELLULOSE REACTIONS
Up to this point, it has been implied that catalysts for cellulose reactions
act in the same manner as catalysts for the reactions of small molecules.
This is of course true. It is at the same time true, however, that in the
case of cellulose some factors become important that are normally neglected
in other cases. Some of these factors will now be discussed in detail.
(1) Swelling Action of Catalysts
It may be taken as a good rule of thumb that the type of interaction
of a catalyst with a reagent and a hydroxyl group that leads to
more rapid reaction will also leacl to partial solvation of the cellulose and an
increased tendency for the cellulose to swell. In such a reaction as the
acid-catalyzed acetylation of glycerol with acetic anhydride it is not difficult
to secure a homogeneous reaction medium. The rate of this reaction will
then be determined by the acidity, regardless of the type of acid used.
In the case of cellulose, however, the amount of swelling is a complicated
function of the amount of acid, the strength of the acid, and the nature of
the medium. It may be preferable to use as a catalyst a large amount of a
relatively weak acid rather than a small amount of a strong acid, if it is
desirable to secure more uniform reaction conditions.
DC. DERIVATIVES OF CELLULOSE 711
(2) Water Binding by Catalysts
In the case of nitrocellulose, the interesting situation arises that sulfuric
acid, which happens to be a catalyst for the reaction, is added to reduce
the activity of water and to lower the solubility of the product rather than
to speed up the reaction. It is used more as a deswelling agent than a swell-
ing agent, and as a reagent rather than as a catalyst.
(3) Bonding of the Catalyst to Cellulose
If the reaction of glycerol with acetic anhydride is catalyzed with sul-
furic acid, it is a matter of no great moment that the sulfuric acid will
react initially with the glycerol to give some sulfuric half-ester. This
product is still a fairly strong acid and will be able to reach all of the re-
action medium by diffusion. As the reaction proceeds, most of the sul-
furic half -ester groups will be replaced with acetyl groups, and a good yield
will be obtained. In the case of cellulose, two disturbing factors emerge.
Most of the sulfuric acid may be bound initially in the more available por-
tions of the fiber, leaving little available for other portions until that bound
initially has been displaced by acetyl groups. A nonuniformity of catalyst
distribution will therefore be superimposed on the nonuniformity of .avail-
ability. This effect may explain the great differences of acetylation rate
that are observed, depending on when and how the sulfuric acid is added
to the reaction system. Another disagreeable result of the binding of
sulfuric acid is, of course, the fact that great pains must be taken to remove
the bound sulfuric acid from the product.
(4) Degradation by Catalysts
The degradation observed during acid-catalyzed reactions of cellulose is
almost entirely due to hydrolytic scission of the 1,4-glucosidic linkages of
the cellulose chain. The question naturally arises as to whether it is
possible by any means to increase the ratio of the esterification rate to the
degradation rate. It is well established that this ratio is definitely im-
proved as the temperature is lowered.26 The concentration of catalyst
does not appear to be an important factor as long as the temperature can
be controlled. Very rapid acetylation can be obtained by the use of a high
ratio of catalyst to cellulose, without excessive degradation. In fact, it
would be expected that high catalyst concentration would be preferable
from the standpoint of degradation, since the increased swelling would be
expected to lead to more uniform reaction. In order to use high catalyst
712 CELLULOSE
concentration in industrial practice, it is necessary to use a relatively in-
active catalyst such as zinc chloride in order to be able to control the tem-
perature.
It might be expected that the ratio of acetylation rate to degradation
rate might vary from catalyst to catalyst. With some acids, such as HC1,
the chain-splitting reaction can be chemically different from other cases,
since a fragment of the catalyst may be bound in the product as a glucosyl
chloride end group. In such a case, the degradation reaction would be
expected to have a different order and a different rate, compared to the
usual hydrolytic degradation reaction which definitely involves the medium
(water or acetic acid) in its chemistry. This possibility has not been ex-
plored as yet.
It is usually stated that degradation is not a problem in cellulose reac-
tions carried out in basic media. One probable exception should be noted.
In the preparation of tosyl, other sulfonyl, and sulfate esters by use of the
corresponding chlorides in pyridine medium, replacement of the entering
group by chlorine or quaternization with the pyridine is occurring con-
tinuously. As mentioned previously, such reactions are necessarily
accompanied by degradation. It may well be that this type of degradation
is not as pronounced if the reaction is carried out with alkali cellulose rather
than in a pyridine medium.
4. Conclusion
The whole body of evidence supports the view that the reactions of cellu-
lose are exactly analogous to those of compounds of low molecular weight.
The proper analogs must, of course, be chosen. The question of availability
for reaction is not important in much of organic chemistry, but it is en-
countered in such reactions as the esterification of terephthalic acid and in
the corrosion of metals. A reaction of glycerol that gives a 95% yield of
glycerol esters and 5% acrolein might be considered satisfactory. A
cellulose reaction, on the other hand, must leave more than 99% of the
cellulose chain structure unaltered before it can give a technically satis-
factory product.
Investigators in the cellulose field should always pose the question to
themselves: Will the proposed reaction work with methyl glucoside,
cellobiose, or sucrose?
B. INORGANIC ESTERS1
J. BARSHA
, Of the various inorganic esters of cellulose which could be made, the only
one that has achieved large commercial production is nitrocellulose. Con
sequently, a discussion of inorganic esters of cellulose must become pri-
marily a discussion of nitrocellulose, with some mention of other esters of
minor importance. In addition to the interest in nitrocellulose which
stems from its wide use in industry, this inorganic ester is a versatile ma-
terial for studying the chemistry of cellulose. Many advances in under-
standing the structure and properties of cellulose have been derived from
studies of nitrocellulose.
1. Nitrocellulose2"4
Nitrocellulose (more correctly called cellulose nitrate since it is an ester)
is the oldest cellulose derivative. Braconnot in 1832 and Pelouze in 1838
had nitrated various materials including starch, wood fiber, cotton, and
paper with concentrated nitric acid. However, in 1845, Schonbein nitrated
cellulose with a mixture of nitric and sulfuric acids, and he is generally
credited with the discovery of nitrocellulose. The early history of nitro-
cellulose is associated largely with attempts to manufacture it for mili-
tary explosives. It was not until about 1866, when Abel showed that the
stability of nitrocellulose is improved enormously by pulping the nitrated
fibers in a paper beater and then washing out the retained acid, that the
manufacture of nitrocellulose was established on a sound basis.
Nitrocellulose is responsible for many changes in the industrial arts and
1 Editors' note: Because of the importance of cellulose xanthate, this ester is given
separate treatment in Section F of this Chapter IX and Section C of Chapter X.
2 A detailed account of the early history of nitrocellulose and a bibliography of all the
literature up to 1920 are given by E. C. Worden, Technology of Cellulose Esters, Vol. 1,
E. C. Worden, Millburn, N. J., 1921.
8 Patent literature has been systematically compiled by O. Faust, Celluloseverbin-
dungen und besonders wichtige Verwendungsgebiete, J. Springer, Berlin, 1935.
4 The latest review of literature and patents is by K. Fabel, Nitrocellulose; Herstel-
lung und Eigenschaften, Ferdinand Enke Verlag, Stuttgart, 1960.
713
714 CELLULOSE
sciences. The use of the material as a propellant was the first major break
from the traditional use of black powder, which had proceeded without
change for centuries. The next major step in the history of nitrocellulose
was the development of celluloid. Prior to the introduction of this syn-
thetic plastic, some molding had been done with thermoplastic natural
resins such as shellac, but the art of molding and fabricating plastic com-
positions for practical purposes dates from the discovery of celluloid.
Protective surface coatings made little progress and were all essentially oil
or natural resin or oleoresinous compositions until the advent of nitrocellu-
lose lacquer in the years following World War I. Other examples could be
cited.
As a pioneer in opening new fields for exploitation, nitrocellulose has in
some instances met the fate of other pioneers and has been supplanted by
other materials. However, the commercial history of nitrocellulose has
been one of relatively constant growth although of changing markets.
Currently, production is at a high level with lacquer being the most im-
portant industrial use.
The properties of a batch of nitrocellulose which have the greatest in-
fluence on its behavior in actual use are (1) its degree of nitration, and (2)
its solution viscosity, which is a function of the molecular weight or degree
of polymerization (D.P.) of the nitrocellulose. These two properties are
therefore used industrially to characterize every batch of nitrocellulose.
The degree of nitration is most commonly designated by the nitrogen con-
tent expressed as per cent nitrogen or, less frequently, as the number of
cubic centimeters of NO (at 0°C. and 760 mm. pressure) evolved from 1 g.
of nitrocellulose. It is often convenient to designate the degree of nitra-
tion by the "degree of substitution'1 (D.S.) which is the average number of
TABLE 4
General Types of Manufactured Nitrocellulose
Nitrogen
content,
% Field of application Common solvents
10.7-11.2 Plastics, lacquers Ethyl alcohol
11.2-11.7 Lacquers Ether-alcohol; methanol; ethyl, butyl,
and amyl acetates; acetone; methyl
ethyl ketone
11.8-12.3 Lacquers, coated Ether-alcohol; methanol; ethyl, butyl,
fabrics, cements and amyl acetates; acetone; methyl
ethyl ketone
12 . 0-13 . 5 Smokeless powder Acetone
IX. DERIVATIVES OF CELLULOSE 715
hydroxyl groups nitrated per anhydroglucose unit. Although it is possible
to prepare nitrocellulose covering the entire range of theoretical nitrogen
content, the products normally manufactured fall in the broad range of
10-14% N as shown in Table 4.
In industrial practice, the solution viscosity of a batch of nitrocellulose
is referred to simply as the viscosity of the nitrocellulose. In the standard
method for determining nitrocellulose viscosity (see Chapter XII), the
latter is expressed as the number of seconds required for a V ie-inch diameter
steel ball to fall through 10 inches of a nitrocellulose solution (of specified
concentration and solvent composition) at 25 °C. When the nitrocellulose
viscosity is very low, it is usually measured in a capillary viscometer.6
Commercial soluble nitrocelluloses are available in viscosities ranging from
20 centipoises to about 350,000 centipoises in 12.2% solution.
(a) PREPARATION OF NITROCELLULOSE
(1) Cellulose for Nitration
Cotton, the purest form in which cellulose occurs abundantly in nature,
has been used in various forms for the manufacture of nitrocellulose. In
the early days, the cotton was used in the form of skeins, rovings, or waste ;
this was later followed by the use of cotton linters, which are the short
fibers cut from cottonseed at seed-processing mills (see Chapter VI-B).
The advantages of linters over staple cotton in nitrocellulose manufacture
include : (a) lower cost, (b) greater ease of handling and manipulation in
the plant, and (c) less pulping required in purification of the smokeless-
powder grade of nitrocellulose.
Many investigators have studied sources of cellulose other than cotton,
especially in wartime when the demand for nitrocellulose for use in pro-
pellants is very great. Some attention has been given to the preparation
of cellulose from plants other than wood.6""8 However, most of the effort
has been directed toward the development of wood pulps**7-9 suitable for
nitration. It was realized early in this work that, in addition to chemical
purity, the physical properties of the wood pulp are important in obtaining
satisfactory nitration* In 1912, Tedesco10 obtained results which led him
5 Nitrocellulose Handbook, Hercules Powder Co., Wilmington, Del., 1952, 54 pp.
6 M. G. Morin, Mem. poudres, 22, 57 (1926).
7 P. Demougin, Mim. poudres, 23, 268 (1928).
8 E. Afferni and C. Milani, Ann. chim. appticata, 30, 248 (1940).
9 W. Scharrnbeck, Z. ges. Schiess- u. Sprengstofw.t 29, 33, 67, 98, 133, 196, 230, 266
(1934).
10 H. Tedesco, Z. ges. Schiess- u. Sprengstoffw., 7, 474 (1912).
716 CELLULOSE
to believe that nitrocellulose from wood pulp is much less stable than nitro-
cellulose from cotton. However, investigations at about the same time by
Nitzelnadel11 and also by Schwalbe and Schrimpff12 showed that satisfac-
torily stable nitrocellulose for smokeless powder can be made from wood
pulp. During World War I, large quantities of wood pulp (usually in the
form of creped tissue paper) were nitrated in Germany. During the same
period of scarcity of cotton, the nitrocellulose industry in the United States
changed first from mill-run linters to second cut linters and later to hull
fiber. The urgent need in 1918 for extending their supply of raw material
led the Americans to investigate large-scale nitration of crepe paper but
they encountered the following difficulties: (a) The product tended to
gelatinize during nitration, (b) the quantity of acid retained after wringing
was high, (c) the yield of nitrocellulose was only about 1.2 Ib. per pound
of wood pulp as compared with 1.5 Ib. from cotton, and (d) dehydration of
the nitrated pulp with alcohol was slower because of the physical nature
of the product. Mixtures of equal parts of wood pulp and cotton linters
were finally found to give a cellulose with physical form suitable for nitra-
tion, and the nitrocellulose produced from this mixture had satisfactory
stability for the military explosives.18
Considerable progress has been made since the early twenties in the
development of methods for putting wood pulp into suitable physical
form for nitration and for manufacturing well-purified wood pulps suitable
for making high-quality nitrocellulose. The developments in this field
by the Brown Company have been described by Schur and Hoos.14 They
found that if purified wood pulp is shredded in the presence of about its
own weight of water to a form resembling cotton linters and then redried,
the shredded pulp can be nitrated by the same method that is used in the
commercial nitration of linters. They also reported that customary
nitrating equipment and mixed acid gave good results with purified wood
pulp in the form of light-weight paper (tissue). Of greater practical inter-
est was the finding by Schur15 that pulpboard cut into small pieces can be
satisfactorily nitrated in the usual equipment with a mixed acid rich in
11 K. A. Nitzelnadel, Z. ges. Schiess- u. Sprengstoffw., 7, 257, 301, 339, 384, 409 (1912).
11 C. G. Schwalbe and A. Schrirapff, Z. angew. Chem., 27, 662 (1914); A. Schrimpff,
Nitrocellulose aus Baumwolle und Holzzellstoffen, J. F. Lehmann, Munich, 1919.
l» R. G. Woodbridge, Jr., Ind. Eng. Chem.t 12, 380 (1920).
14 M. O. Schur and B. G. Hoos, Ind. Eng. Chem., 29, 26 (1937); this reference lists
the various patents obtained by the Brown Company and also gives a good summary of
other references on the preparation of wood pulp for nitration.
15 M. O. Schur (to Brown Co.), U. S. Patent 1,914,302 (June 13, 1933),
IX. DERIVATIVES OF CELLULOSE 717
nitric acid. Other methods have also been proposed for putting wood pulp
into suitable form for nitration.16-17
The turning point in making the use of wood pulp practical for large-
scale nitration as well as for other chemical treatment came in the mid-
thirties with the invention in the United States of the Stern18 shredder
and process. The shredder produced short pieces of loosely felted wood
cellulose with protruding fibers at the ends and edges. This physical form
was admirably adapted for ready penetration of chemical reagents. The
process developed for nitrating the wood pulp so shredded gave uniformly
esterified nitrocelluloses, was economical and free from the undue difficulties
encountered in nitrating various other physical forms of wood pulp, and
utilized the usual nitrating mixtures and equipment required for the nitra-
tion of cotton linters.
By the beginning of World War II, the use of wood pulp for the manu-
facture of commercial nitrocellulose was well established. In Great Britain,
small tablets cut from wood pulp sheets were nitrated;19'20 in the United
States, the Stern18 process was in full commercial operation. Utilization
of wood pulp for nitrocellulose for military propellants was, however, still
in the experimental stage in the United States.21 This situation was rapidly
remedied so that most of the enormous tonnage of nitrocellulose used by
the armed forces of the United States and its Allies was made from specially
prepared coniferous wood pulps.22
The physical characteristics of the pulps have been shown to have an
effect on nitratability. Coniferous wood pulps were found to be more
suitable for nitration because smaller mechanical losses resulted than when
deciduous wood pulps were used.23 Acid retention by fibers of a variety
of coniferous wood pulps was found by Schur and McMurtrie24 to be higher
16 W. E. Sillick (to Eastman Kodak Co.), U. S. Patent 2,384,853 (Sept. 18, 1945);
Chem. Abstracts. 40, 207 (1946).
17 D. J. Kridel (to Eastman Kodak Co.), U. S. Patent 2,393,783 (Jan. 29, 1946);
Chem. Abstracts, 40, 2305 (1946).
18 R. L. Stern (to Hercules Powder Co.), U. S. Patent 2,028,080 (Jan. 14, 1936).
19 N. Picton and Imperial Chemical Industries Ltd., Brit. Patent 336,235 (May 27,
1929); Chem. Abstracts, 25, 1996 (1931); N. Picton (to E. I. du Pont de Nemours & Co.),
U. S. Patent 1,872,181 (Aug. 16, 1932).
20 N. Picton and N. S. Kelland, Ind. Chemist, 26, 79 (1950).
81 G. A. Richter, OSRD Report No. 71 ; also issued by U. S. Dept. of Commerce,
Washington 25, D. C., OTS, PB Rept. 31203 (1941).
" Anon., Paper Trade J.t 117, 14 (Oct. 7, 1943).
*» L. Zapf, U. S. Dept. of Commerce, Washington 25, D. C., OTS, PB Rept. 4383
(1943).
24 M. O. Schur and D. H. McMurtrie, Paper Trade /., 127, 51 (Sept. 23, 1948);
127, 39 (Sept. 30, 1948).
718 CELLULOSE
in fibers with larger cross section. More complete nitration was obtained
by Brissaud26 with pulps which had received a final alkaline treatment dur-
ing purification as well as with pulp fibers which are long, thick, and elastic
(e.g., pulps from spruce and esparto). Mention should also be made of an
entirely different raw material — that is, regenerated cellulose film (scrap
cellophane) — which has been used in the manufacture of nitrocellulose for
lacquers.28'27
Variations in the conditions under which cellulose loses or absorbs mois-
ture can affect the final nitrogen content reached in nitration with HNOr*
H^SOi-H^O, as well as the solubility and viscosity behavior of the nitrate
produced. Brown and Purves28 obtained nitrates with about 12.2% N
from highly swollen cotton linters, whereas "collapsed1' linters (obtained
by wetting and redrying swollen linters) yielded nitrates with 0.1 to 0.6%
less nitrogen. Tribot and Chddin29 found 'that the addition of water or
pyridine to cellulose increases the rate of nitration by HNOs-H^SOHH^O.
The maximum rate was obtained with cellulose containing about 1 mole
of water or 0.2 mole of pyridine per anhydroglucose unit.
(2) Nitration with HNOv-HtSO*-H*0
Although nitric acid itself in both liquid and vapor form and also mix-
tures of nitric acid with a wide variety of other chemicals have been used
experimentally to nitrate cellulose, the nitrocellulose industry still employs
HNOs-HkSC^-H^O mixtures such as have been used for this purpose for a
century. In other reactions of cellulose which are carried out on a large
scale today, such as the preparation of viscose, cellulose acetate, and ethyl
cellulose, the cellulose dissolves and its original fibrous structure is com-
pletely lost. In contrast to these, the nitration of cellulose is effected with
complete retention of the original fiber structure, and the chief superficial
changes that occur are an increase in the hardness of the fiber and, in the
case of cotton, an untwisting of the convoluted fiber.
Since (as will be discussed later) cellulose nitration is an equilibrium re-
action, the extent of nitration at equilibrium is governed primarily by the
» L. Brissaud, Mlm. poudres, 33, 137 (1951).
* P. B. Cochran, M. V. Hitt, and L. V. Taylor (to E. I. du Pont de Nemours & Co.).
U. S. Patent 1,997,766 (Apr. 16, 1935); U. S. Patent 2,150,205 (Mar. 14, 1939).
87 L. L. Blyler (to E. L du Pont de Nemours & Co.), U. S. Patent 2,399,620 (May 7,
1946); Chem. Abstracts, 40, 4214 (1946).
» R. K. Brown and C. B. Purves, Pulp & Paper Mag. Can., 48, No. 6, 100 (1947);
see also A. Rosenthal and R. K. Brown, Pulp & Paper Mag. Can., 51, No. 6, 99 (1950).
w A. Tribot and J. ChSdin, Mint, services chim. Slat (Paris), 32, 135 (1945).
DC. DERIVATIVES OF CELLULOSE 719
composition of the mixed acid. The extent of nitration is affected to a
lesser degree by the ratio of mixed acid to cellulose, the final nitrogen con-
tent being increased, within limits, by an increase in the ratio.30 The effect
of mixed acid composition on the extent of nitration was studied in great
detail by Lunge and coworkers.8l~88 In one of their first series of experi-
ments (Table 5) they attempted to prepare nitrocellulose with nitrogen
TABLE 6
Effect of Mixed Acid Composition on Attainment of Maximum Nitration of Cellulose
(Lunge and Cpworkers88)
Composition of mixed acid
Nitrogen in
nitrocellulose,
%
HtSOi, %
HNO., %
H,0, %
60.00
27.43
12.57
13.62
62.10
25.79
12.11
13.75
62.95
24.95
12.10
13.83
63.72
25.31
10.97
13.75
64.56
24.65
10.79
13.71
68.02
25.28
5.70
13.76
64.55
26.55
8.88
13.72
63.35
25.31
11.34
13.92
75.33
22.80
1.87
13.53
74.16
22.12
3.72
13.51
72.97
21.63
5.40
13.57
69.90
20.45
9.65
13.64
68.31
20.49
11.20
13.61
67.43
19.37
13.20
13.25
67.32
32.53
0.15
13.62
65.41
31.34
3.25
13.57
63.75
30.80
5.45
13.63
70.68
29.31
10.01
13.68
content as close as possible to the theoretical maximum of 14.14%. Nearly
all of the products they obtained fell in the range of 13.5 to 13.8% N even
though some of their acid mixtures contained as much as 12.5% water.
Maximum nitrogen content was obtained with HsSO^HNOs ratios be-
tween 0.25:1 and 3:1. With ratios above 8:1, the product always con-
tained some unnitrated cellulose and the fiber structure was attacked.
In other experiments (Table 6) the ratio of H^SOi to HNOs was kept nearly
» J. O. Small and C. A. Higgins, Chemical Age, 28, 211 (1920).
81 G. Lunge and E. Weintraub, Z. angew. Chern., 12, 441, 467 (1899).
" G. Lunge and J. Bebte, Z. angew. Chem., 14, 483, 507, 537, 561 (1901).
88 G. Lunge, /. Am. Chem. Soc.t 23, 527 (1901); this article is a summary of references
30 and 32 above.
720 CELLULOSE
constant at slightly less than one while the proportion of water was varied;
it was found that the nitrogen content of the nitrocellulose decreased regu-
larly with increase in the proportion of water in the mixed acid. Similar
results were obtained by Schiemann and Kiihne34 and Demougin.36
TABLE 6
Effect of Water Content of Mixed Acid on the Degree of Nitration of Cellulose
(Lunge and Coworkers83)
Solubility in
ether-alcohol,
Composition of mixed acid
Nitrogen in
nitrocellulose,
HiSOi, %
HNOs, %
H2O, %
45.31
49.07
5.62
13.65
42.61
46.01
11.38
13.21
41.03
44 . 45
14.52
12 76
40.66
43.85
15 49
12.58
40.14
43 25
16.61
12.31
39.45
42.73
17.82
12.05
38.95
42.15
18.90
11 59
38.43
41.31
20 26
10 93
' 37.20
40.30
22 50
9.76
36.72
39.78
23.50
9 31
35.87
38.83
25 30
8 40
34.41
37 17
28 42
6 50
1.50
5.40
22.00
60.00
99.14
99 84
100.02
99.82
74.22
1 15
0 61
1.73
Berl, Aridress, and Escales36 studied critically the literature on experi-
mental nitrations with different mixed acids and found good agreement
between the results of Lunge and coworkers,31""33 Boltenstern,87 and De-
mougin.85 Since the final nitrogen content is determined by the composi-
tion of the spent acid in contact with the nitrocellulose rather than by that
of the original dipping acid, Berl and coworkers calculated by the graphical
methods of Demougin the composition of the spent acids from the composi-
tion of the dipping acids in cases where only the latter values were given.
They then prepared a triangular coordinate diagram of the spent acid
compositions in which lines were drawn connecting compositions which
produced the same nitrogen content (see Fig. 6). Miles and Milbourn38
have drawn a somewhat similar diagram in which the compositions shown
are those of the dipping acids (see Fig. 7).
84 G. Schiemann and S. Kiihne, Cellulosechemie, 15, 78 (1934).
46 P. Demougin, Him. poudtes, 23, 262 (1928).
S6 E, Berl, K. R. Andress, and E. Escales, Beitrdge zur Kenntnis der Mischsaure, J. F.
Lehmann, Munich, 1937, p. 32.
87 W. von Boltenstern, Dissertation, Darmstadt, 1921, p. 86.
38 F. D. Miles and M. Milbourn, /. Phys. Chem., 34, 2598 (1930).
IX. DERIVATIVES OF CELLULOSE
721
Despite the wide variety of HNOs-^SO-r-HfoO compositions (shown in
Fig. 6) which can yield nitrocellulose of a given nitrogen content, the selec-
tion for industrial nitrations is fairly limited by considerations of cost and
because many of the compositions exert a marked swelling or dissolving
action on the fibers. In actual practice30 (see Table 7) the different mixed
acids used to produce the usual range of nitrocelluloses, with from about
30
fO
30
60
i/V\
/oo
so
70
so
30 ZO
/a
too
Fig. 6. Effect of spent acid composition on the nitrogen content of nitrocellulose (Berl,
Andress, and Escales36).
/O SO 3O 40 jro GO
/OO <9O SO 70 60
/OO
Fig. 7. Effect of dipping acid composition on the nitrogen content of nitrocellulose
(Miles and Milbourn38).
10 to 13.5% N, do not vary widely in composition from each other or from
the H2SC>4 : HNO3 ratio of 3 : 1 originally used by Schonbein a century ago.
Various investigators have recommended mixed acid compositions for
the preparation of specific types of nitrocellulose.4 Thus, Baker, Morgan,
and Quaid39 reported that nitrocellulose with 11.85-12.00% N and which
89 L. S. Baker, I. H. Morgan, and R. J. Quaid (to E. I. du Pont de Nemours & Co.),
U. S. Patent 2,105,627 (Jan. 18, 1938).
722 CELLULOSE
is completely soluble in methanol is obtained by nitrating chemical cotton
for 20 min. at 45°C. with 50 parts of mixed acid containing 27 to 32%
HNO3. With a mixed acid containing 30.7% HNO3, 51.7% H2SO4, and
17.6% H2O, they obtained nitrocellulose with 11.98% N.
TABLE 7
Composition of Commercial Nitrating Acids and the Nitrogen Content of Nitrocelluloses
Produced by Them (Small and Higgins80)
Nitrogen,
HNOs,
H,SO<,
HNOSOt,
H2O,
10.52
22.30
51.97
5.25
20.48
10.74
22.08
52.50
5.10
20.32
10.86
21.60
53.13
4.80
20.47
10.96
22.13
51.99
5.60
20.28
11.05
23.86
51.24
5.17
19.73
11.10
22.74
52.04
5.57
19.65
11.14
24.81
53.60
1.37
20.22
11.19
23.80
50.40
6.55
19.25
. 11.91
20.18
55.93
6.00
17.89
11.92
20.15
55.80
6.20
17.85
11.96
20.30
56.13
5.80
17.77
11.98
19.98
55.95
6.25
17.82
12.06
20.50
55.45
6.40
17.65
12.75
19.54
60.92
4.57
14.97
13.00
20.00
59.64
5.30
14.16
13.30
24.37
59.54
4.86
11.23
Sillick40 reported that nitrocellulose with improved adhesion is obtained
by nitrating oxidized cellulose containing 0.05 to 6% carboxyl groups in the
primary position of the anhydroglucose units. He nitrated the oxidized
cellulose with mixed acid containing 80 to 98% acid (40 to 60% HNO*
and 40 to 60% HzSOO and treated the product with a mild alkaline solution
(e.g., NaOH, calcium acetate, t>r calcium hydroxide) to impart insolubility
in acetone.
The nature of the cellulosic material being nitrated also plays a part in
determining the composition of mixed acid to be used. Thus, Schrimpff12
found that mixed acid must contain more nitric acid in order to nitrate
wood pulp to the same extent as cotton.
The rate of cellulose nitration increases with rise in temperature (see
Table 8), but the drawback to much higher nitration temperatures than are
normally employed (20 to 40°C.) is excessive degradation with resultant
40 W. E. Sillick (to Eastman Kodak Co.), U. S. Patent 2,544,902 (Mar. 13, 1951);
Chem. Abstract*, 45, 4927 (1951).
IX. DERIVATIVES OF CELLULOSE
TABLE 8
Effect of Temperature and Reaction Time on the Nitration of Cellulose
(Lunge and Coworkers18)
723
~»p.
Time,
hr.
Nitrogen,
%
Yield,
Yield
(calculated),
0
0.5
10.71
152.3
153
0
7
13.19
173.3
174
10
7
13.37
175.8
176
15
7
13.38
175.6
176
19
0.5
12.72
166.1
170
19
7
13.39
175.6
176
40
0.5
13.07
172.3
173
40
7
13.06
169.6
173
60
0.5
13.08
169.2
173
60
4.5
13.07
162.1
173
80
0.25
13.07
161.2
173
80
0.5
13.12
125.2
173
80
3
13.12
81.5
173
Notes: Nitration was effected with a 3 : 1 mixture of H2SO4 :
Yields are based on the weight of cellulose employed ; that is, the % yield is the weight
of nitrocellulose obtained from 100 parts of cellulose.
16
ROGEN CONTENT, %
t O» o> o ru *
7*
H
z
0 $ 10 15 20 Z5 30 35 40 45 50 55 6<
TIME IN MINUTES
Fig. & Rate of nitration of cellulose in a laboratory dipping pot with mixed acid
containing 21.0% HN08, 61.5% H»SO4, and 17.5% H2O (Spalding").
724 CELLULOSE
drop in viscosity and yield.80*88 The nitration rate decreases markedly
with increase in the HaSC^: HNOs ratio.83 The nitration of cellulose takes
place much more rapidly than the other cellulose reactions which are prac-
ticed industrially;34 for most commercial types, nitration is effected in
about 30 min. The extreme rapidity of cellulose nitration,41 particularly
in the first few minutes, is shown well in Figure 8.
(5) Nitration with Other Reagents
Considerable attention has been given to the use of nitration reagents
other than HNOa-HaSOr-H^O, but none of them have attained industrial
importance. The simplest of these reagents is aqueous nitric acid. • The
following data obtained by Vieille42 indicate the extent of nitration that can
be obtained with various concentrations of nitric acid :
%HN08 77.3 80.8 83.5 87.0 89.6 92.1 95.1
% N in nitrocellulose 6 85 8 . 07 8 . 78 10 . 33 1 1 . 53 12 . 23 12 . 68
Practically no nitration occurred with 75% HNOs. In the approximate
range of 80 to 85% HNOs, the fibers dissolved completely and the nitro-
cellulose could be precipitated by adding water; other concentrations of
HNOa swelled and gelatinized the fibers.
Bouchonnet and coworkers43 nitrated cellulose with pure HNO3 and re-
ported that the acid in the interior of the fibers is diluted by water formed
in the reaction, which, together with the heat liberated, causes gelatiniza-
tion or solution of some of the nitrocellulose. This partial or complete de-
struction of the fibers is a major drawback to the use of HNOs as the sole
reagent in the industrial nitration of cellulose. The addition of dehydrat-
ing salts to pure HNO3 prevents the hardening and contraction of fibers
which occur when HNOa is used alone. A maximum nitrogen content of
13.87% was obtained when 15. to 20% of NH4NO3 or 30% of KNO3 was
added to pure HNO3 as compared with 13.3% when HNO3 was used alone.
K2S04, KH2PO4, and NH4H2PO4 showed a similar action. The use of 97%
HNOs was less satisfactory than pure HN03.
Ch&lin and Tribot44 reported results similar to those obtained by
41 E. B. Spaulding, Hercules Powder Co., Parlin, N. J., private communication.
" P. Vieille, Mem. poudres, 2, 212 (1884-9).
41 A. Bouchonnet, F. Trombe, and G. Petitpas, Compt. rend., 197, 332 (1933); Mem.
poudres, 28, 277 (1938); Bull. soc. chim., [5], 4, 894 (1937); see also T. Urbanski and
W. Szypowski, Roczniki Ghent., 10, 387 (1939); Chem. Abstracts. 34, 4566 (1940).
44 J. Ch£din and A. Tribot, Mtm. services chim. Mat (Paris\ 34, 277 (1948).
IX. DERIVATIVES OF CELLULOSE 725
Bouchonnet and coworkers. In addition, Ch&iin and Tribot found that
nitrocellulose with 13.2% N, obtained in one nitration with 100% HNO»,
could be made amenable to further nitration with 100% HNOa by dissolv-
ing the gelatinized fibers in acetone and precipitating with water. They
obtained nitrocellulose with 13.7 to 13.8% N by repeated reprecipitation
and treatment with 100% HNO3.
Attempts have been made to simplify the nitration of cellulose and to
overcome the disadvantages in the use of nitric acid solutions by nitrating
cellulose with HNOg vapor.45 Rogovin and Tikhonov46 suspended cotton
over 96% HNO3 in a closed flask. After 4 hrs. at 20°C., the nitrogen con-
tent was only 4.96%, after 24 hrs. 11.96%, and after 96 hrs. 13.4%. Rais-
ing the temperature markedly accelerated the nitration. At 35°C.,
11.05% N was attained in 4 hrs., 12.88% N in 24 hrs., and 13.44% N in
48 hrs. The nitration of cellulose with HNO3 vapors at reduced pressures
was studied extensively by a group of French investigators.47""62 They
designed special equipment to prevent condensation of HNO? and water
vapor in the fibers.47'48 In one series of small-scale nitrations (4 g. of
cellulose) with 100% HNO3 at 84 mm. pressure, 11.12% N was attained
in 15 min. and 13.59% N in 75 min. In larger-scale nitrations61'52 (100 g.
cellulose), it was found that the mass of cellulose was not uniformly ni-
trated, the nitrogen content decreasing with increasing distance from the
point of entry of the HNO3 vapors. Fairly uniform products of high nitro-
gen content (about 13.7% N) were obtained on prolonged nitration. The
preparation of uniform nitrocellulose corresponding to industrial types
(10.7 to 12.3% N) was not demonstrated. In a study of the kinetics of the
nitration of ramie cellulose with HN*O3 vapors, Wilson and Miles58 showed
by means of fractionation and solubility tests that homogeneous products
were not obtained by this process.
Nitrogen oxides have been used experimentally to nitrate cellulose.
Products with 1 1.30 to 11.95% N were obtained on nitrating cellulose with
46 Deutsche Celluloidfabrik, German Patent 269,246 (Jan. 16, 1914); Chem. Abstracts,
8, 2060 (1914).
46 Z. A. Rogovin and K. Tikhonov, Cellulosechemie, 15, 102 (1934).
47 A. Bouchonnet and F. Trombe, Bull. soc. chirn., [5], 5, 715 (1938).
48 F. Trombe, Bull. soc. chirn., [5], 9, 526 (1942).
49 A. Bouchonnet, F. Trombe, and G. Petitpas, M6m. poudres, 28, 308 (1938).
60G. Champetier and M. Foe'x, Compt. rend., 211, 468 (1940); Bull. soc. chirn., [5,
8, 115 (1941); Bull. soc. chim., [5], 9, 711 (1942).
61 M. Foe'x, Bull. soc. chim., [5], 8, 381, 390 (1941).
62 F. Trombe, M. Foex, and G. Champetier, Ann. chim., [12], 4, 745 (1949).
" G. L. Wilson and F. D. Miles, Trans. Faraday Soc., 40, 150 (1944).
726 CELLULOSE
liquid NO8 at 10°C. for 48 hrs.64 Pinck85 found that nitrocellulose of any
desired nitrogen content could be prepared with a nitrating bath in which
HNOa had been replaced by ^(X An increase in the amount of N2O4
in the usual nitration mixture (HNOs-HjSO^HkO) decreases the yield58
and nitrogen content66'67 of the nitrocellulose. The degree of nitration
with ^OHHNOs mixtures reaches a maximum at 30% NjOi.68 Addition
of nitrogen oxides to nitric acid (in absence of sulfuric acid) causes a higher
degree of nitration (a maximum with N2O6) ; N2Oa and N^C^ cause greater
depolymerization of cellulose than N2O6. Addition of 5% ^Os to 95%
HNO3 yields nitrocellulose with 13.7% N compared to 13.0% N with 96%
HNOs alone. The more vigorous action of NjOs in comparison with P2Oa
can be explained by the more rapid diffusion into the interior of the fiber
and greater esterification in the interior by the HNOa resulting from the
combination of N2O5 with water.69 Dalmon and coworkers50 treated cellu-
lose with N2(>6 dissolved in carbon tetrachloride at 130°C. for 6 hrs. in the
dark and obtained a high yield of nitrocellulose with about 14% N. Simi-
lar results were obtained by Caesar.61 Treatment of dry cotton with pure
N^Ob in a current of dry air or oxygen yielded pure cellulose trinitrate
with 14. 1 2% N ,62 X-ray diagrams taken at frequent intervals during nitra-
tion with N2Os vapor showed that the structure progressed through all the
stages between cellulose and cellulose trinitrate.68
The replacement of sulfuric acid by phosphoric acid in nitrating acid
has interested many investigators because nitrocelluloses of high stability
are readily obtained. Krauz and Blechta64 nitrated cellulose for 1 hr. at
20°C. with a mixed acid containing 50% H8PO4, 48.2% HNO8, and 0.34%
water and obtained nitrocellulose with 12.93% N. When they increased
the water content of the mixed acid they observed a rapid drop in nitrogen
64 C. J. Staud and J. T. Fuess (to Eastman Kodak Co.), U. S. Patent 1,917,400 (July
11, 1933).
» L. A. Pinck, 2nd. Eng. Chejn., 22, 1241 (1930); U. S. Patent 1,784,945 (Dec. 16,
(1930).
" S. S. Mindlin and L. I. Kuz'mina, /. Applied Chem. ( U. S. S. R.), 8, 1415 (1935).
*7 L. Brissaud, Mem. poudres, 30, 217 (1948).
w P. P. Shorygin and E. V. Khait, /. Gen. Chem. ( U. S. S. R.), 7, 188 (1937).
M Z. A. Rogovin and K. Tikhonov, Cellulosechemie, 16, 11 (1935).
10 R. Dalmon, J. Ch6din, and L. Brissaud, Compt. rend., 201, 664 (1935).
61 G. V. Caesar (to Stein, Hall & Co., Inc.), U. S. Patent 2,400,287 (May 14. 1946);
Chem. Abstracts. 40, 4487 (1946); U. S. Patent 2,432,280 (Dec. 9, 1947).
w R. Dalmon, Compt. rend.t 201, 1123 (1935); see also T. Urbanski and Z. Janisze-
wski, Roczniki Chem.. 17, 349 (1937).
" M. Mathieu, Compt. rend.. 212, 80 (1941).
" C. K. Krauz and F. J. Blechta, Chem. News, 134, 1, 17 (1927).
IX. DERIVATIVES OF CELLULOSE 727
content and a sharp increase in the dissolving action of the nitrating acid on
the nitrocellulose. With essentially anhydrous HNO3-H8PO4 mixtures,
they found that the nitrogen content dropped from 12.93% N with a 1:1
ratio of HNO3 : H3PO4 to 1 1 .36% N with a 1 : 8 ratio of HNO3 : H3PO4.
Berl and Rueff68 nitrated dry cellulose for 4 hrs. at 19°C. with H3PO4-
HNO3 (1:1) containing a little P2Os and obtained a product with 14% N.
The nitrocelluloses they prepared with H3PO4~HNO3 mixtures (anhydrous
or containing some P2O6) had much higher viscosity and nitrogen content
than those obtained with HNO3~H2SO4 mixtures. Under certain conditions,
nitration with HNO3-H3PO4-P2O& was very rapid and a nitrocellulose with
13.5% N was obtained in 1.5 min.ft6 In experiments aimed at obtaining
maximum nitration, Lenze and Rubens67 obtained nitrocellulose with 14%
N by nitrating with HNO3~-P2O5 mixtures; the optimum proportion of
P2O& was 2.5%. Davidson68 carried out numerous nitrations with a mix-
ture of 48% HNO3, 50% H3PO4, and 2% P2O5 for 4 hrs. at 0°C. ; the nitro-
celluloses contained 13.7% N and about 0.3% P.
The ability of HNO3-H3PO4-Pi>O5 mixtures to yield nitrocelluloses with
little or no degradation68"70 has resulted in their use in scientific work for
the conversion of cellulose samples to a form which is soluble in organic
solvents and which can therefore be used in the determination of molecular
weights and molecular weight distribution.
Bouchonnet and coworkers71 investigated a wide range of nitrating acid
compositions based on phosphoric acid with the object of determining their
usefulness for industrial nitration. They concluded that H3PO4~HNO3
mixtures could be used for the preparation of nitrocelluloses72 with 11.7-
12.2% N and 12.9-13.4% N. Products with satisfactory stability were
obtained after only a boiling water treatment. A comprehensive study of
the same subject carried out at Poudrerie Nationale d'Angoul£me73 in
France led to a somewhat different conclusion. In their opinion, H3PO4-
66 E. Berl and G. Rueff, Ber.t 63B, 3212 (1930); Cellnlosechemie, 12, 53 (1931); see
also E. Berl, U. S. Patent 2,384.415 (Sept 4, 1945); Chem. Abstracts, 40, 206 H946).
86 E. Berl and G. Rueff, Cellulosechemic, 14, 109 (1933).
67 F. Lenze and E. Rubens, Z. ges. Schiess- u. Sprengsto/w., 26, 4 (1931) ; 27, 114, 154
(1932).
68 G. F. Davidson, /. Textile Inst., 29, T195 (1938).
89 E. Berl and G. Rueff, Cellulosechemie, 14, 115 (1933).
70 H. Staudinger and R. Mohr, Ber.t 70B, 2296 (1937).
71 A. Bouchonnet, F. Trombe, and G. Petitpas, Bull. soc. chim., [5], 4, 560 (1937).
72 The common French designations for these nitrocelluloses are: CP2 guncotton for
11.7-12,2% N content and CP, guncotton for 12.9-13.4% N content.
73 Poudrerie Nationale d'Angouleme, M6m. poudres, 28, 82 (1938X
728 CELLULOSE
HNO3 mixtures are not suitable for producing nitroeelluloses with about
12% N because the composition of the nitrating mixture must be kept
very close to 20% HNO3, 56% P2O5, and 24% H2O. Any slight deviations
from this composition yield mixtures which crystallize or which harden the
nitrocellulose fibers. Greater latitude is possible with the acid mixtures
producing nitroeelluloses with about 13% N. In the case of both 12 and
13% N nitrocellulose, slight deviations in the water content of the nitrating
acids produce significant variations in nitrogen content. This would
cause difficulty in controlling the degree of nitration in industrial nitration.
The best use for HaPO4-HNO;s mixtures seems to be in the production of
nitroeelluloses with high nitrogen content. Thus HaPCV-HNOa composi-
tions melting at'0°C. are available which yield nitrocellulose with 13.75%
N ; these compositions do not attack the fiber structure and should there-
fore permit the ready separation of the nitrocellulose from the spent acid.
Nitration with the conventional nitric-sulfuric mixed acids containing
organic diluents has also been investigated. A mixed acid emulsion con-
taining CCU or pentane and stabilized with a few drops of naphthalene-
sulfonic acid yielded products with about 11% N and fairly good solubility;
the acid mixture contained much less HNO3 and H2SO4 than usual.74
Brissaud75 reported that the CC^-mixed acid emulsion breaks on contact
with the fibers and that the mixed acid is absorbed by the fibers. The
mixed acid retention after centrifuging is the same as in normal nitration
and no economy in acid consumption is obtained. To avoid the gelation
of fibers which occurs on nitrating cellulose with concentrated HNOa
alone, Trogus76 used a mixture of HNOg and glacial acetic acid. To ob-
tain nitroeelluloses with over 12.5% N, it was advantageous to have present
a small amount of water-binding agent such as acetic anhydride or P^CV
Watanabe77 found that cellulose was nitrated rapidly by mixtures of CC14
and HNO»; with mixtures containing CCU, HNO;i, and acetic anhydride,
the maximum nitrogen content of over 13% was obtained only when the
HNO3 concentration was 45 to 50%. Darzens78 nitrated cellulose with a
homogeneous mixture of HNOs-acetic anhydride-chloroform and obtained
a product with 13.7% N. Further study of this process showed that nitra-
74 Z. A. Rogovin and P. Paradnya, Cellulosechemie, 15, 32 (1934).
76 L. Brissaud, M6m. poudrcs, 30, 205 (1948).
76 C. Trogus, Ber., 64B, 405 (1931); see also G. Petitpas, Mini, services chim. Mat
(Paris), 30, 248 (1943).
77 S. Watanabe, /. Soc. Chem. Ind.t Japan. 45, 829, 832 (1942); Chem. Abstracts, 43,.
1967 (1949).
78 G. Darzens, Mem. poudres, 25, 437 (1932-3).
IX. DERIVATIVES OF CELLULOSE 729
tion is effected very slowly and that the products are insoluble in ether-
alcohol.79
In experiments aimed at obtaining maximum nitration, nitrocelluloses
with about 14% N were obtained by nitration with mixtures containing
pure HNO3 and acetic, propionic, or butyric acid. Nitration with a mix-
ture containing 50% HNO3, 25%) acetic acid, and 25% acetic anhydride
for 5 hrs. at 15°C. yielded nitrocellulose with 14.08% N. By extraction
of this product with boiling ethyl alcohol, cellulose trinitrate (14.14% N)
with good stability was obtained.80 Chedin81 and Tribot*2 studied the ni-
tration of cellulose with a variety of HNO3~acetic anhydride mixtures and
with acetyl nitrate, which is a strong nitrating agent. They found that
dry cellulose fibers were nitrated to only a slight extent (about 1% N) by
acetyl nitrate and by HNO;^-acetic anhydride mixtures containing less
than 20% HNO:{ in which an appreciable amount of acetyl nitrate is formed.
A high degree of nitration (up to 13.7%) N) was obtained when the cellulose
fibers were first activated by steeping them successively in water, glacial
acetic acid, and acetic anhydride. Petitpas and coworkers83'84 nitrated
cellulose with mixtures containing HNO;>, and a diluent (ether, /?-dichloro-
ethyl ether, acetic acid, propionic acid, acetic anhydride, chloroform, or
phosphoric acid) with the object of determining the effect on nitration
of the basicity of the oxygen in the diluent. No nitration occurred with
ether as the diluent. With the other diluents, nitration was progres-
sively increased as the basicity of the oxygen in the diluent decreased.
It should be pointed out that the use of mixtures of HNO3 and acetic
anhydride can be very dangerous because of the formation of acetyl nitrate
which explodes at elevated temperatures.
The nitration of cellulose with HNO?-methyl nitrate mixtures under a
wide variety of conditions was studied by Rogovin and coworkers.85
Nitration for 2 hrs. at 35°C. with a 2:3 ratio of HNO3 (100%) to methyl
nitrate yielded nitrocellulose with 13.7 to 14.0% N; nitration with 95%
HNO3 under the same conditions resulted in 10.1% N. Most of the nitro-
79 L. Brissaud, Mem. poudres, 25, 440 (1932-3).
80 A. Bouchonnet, F. Trombe, and G. Petitpas, Compt. rend., 197, 63 (1933); Bull,
soc. chim., [51, 4, 1085 (1937); Mim. poudres, 28, 295 (1938).
81 J. Chedin, Kottoid-Z., 125, 65 (1952).
82 J. Ch6din and A. Tribot, Bull, assoi . tech. ind. papetiere, 5, 435 (1951); Mim.
services chim. etat (Paris), 36, 31. (1951).
83 J. Desmaroux, R. Dalmon, and G. Petitpas, Compt. rend., 212, 396 (1941).
84 G. Petitpas, Mim. services chim. etat (Paris), 30, 243 (1943).
8fi Z. A. Rogovin, K. Tikhonov, and A. Maslova, /. Applied Chem, (U. S. S. R.), 19,
659 (1946); Chem. Abstracts, 41, 6044 (1947).
730 CELLULOSE
celluloses with over 10% N dissolved completely in the nitrating mixture
and were precipitated by pouring into water. Replacement of up to 60%
of the methyl nitrate by methylene chloride or ethylene dichloride resulted
in nitrocelluloses with over 10% N and with improved solubility.
(4) Mechanism of Nitration
According to evidence presented by Berl and coworkers86 and other
investigators,87'88 cellulose nitration involves the usual principle of esterifi-
cation ; that is, when molecular quantities of an acid and an alcohol react,
an ester is formed and water is split out. Since the reverse reaction also
takes place at the same time, an equilibrium will be set up among acid,
alcohol, water, and ester. The reaction is forced to completion by dis-
placement of the equilibrium when an excess of one component (in this case
the HNOs) is used, and H2SO4 is added to combine with the water formed.
This equilibrium between the nitrocellulose and the nitrating medium has
been found to occur over a wide range of substitution, the nitrogen content
of the nitrocellulose being determined by the final composition of the nitrat-
ing bath rather than by the composition of the original mixed acid. The
evidence of the existence of this equilibrium includes the observation that
the equilibrium can also be approached from the other direction; that is,
long contact of nitrocellulose with relatively dilute mixed acid causes par-
tial denitration and production of a nitrocellulose with the same nitrogen
content as would be obtained by direct action of this mixed acid on cellulose.
Kagawa89 has claimed that the reaction temperature has no effect on the
nitrogen content attained at equilibrium.
Various investigators have attempted to determine the properties of
mixed acids which influence their nitrating capacity. Sapozhnikov90
determined the nitric acid vapor pressure of various mixed acids and the
nitrogen content of the nitrocelluloses produced by them. He observed a
close relation between these values and found that the highest nitrogen
content was attained with the mixed acid having the highest HNO* vapor
pressure. Berl and coworkers86 confirmed and extended the work of
86 E. Berl and R. Klaye, Z. ges. Schiess- u. Sprengstoffw., 2, 403 (1907); E. Berl and
E. Berkenfeld, Z. angew. Chem., 41, 130 (1928); E. Berl and O. Hefter, Cellulosechemie,
14, 65 (1933).
w P. Demougin and Bonnet, Mem. poudres, 24, 147 (1931).
M K. Fabel and H. Fritzsche, Nitrocellulose, 10, 3, 24 (1939).
M 1. Kagawa, /. Soc. Chem. Ind., Japan, 44, Suppl. binding, 130 (1941).
80 A. Sapozhnikov, Z. ges. Schiess- u. Sprengstoffw., 1, 453 (1906); 4, 441, 462 (1909);
see also C. Kullgren, Z. ges. Schiess- u. Sprengstoffw., 3, 146 (1908).
IX. DERIVATIVES OF CELLULOSE
731
P .ipozhnikov. They pointed out that, according to their triangular dia-
gram (Fig. 6) in which the points corresponding to HNO3 vapor-pressure
maxima are joined by the broken line, there is an increase in nitrogen con-
tent with decrease in water content on approaching the broken line and
that the maximum nitration is attained with a spent acid whose composi-
tion lies on this line.
HNO3
H2so4
H,O
Fig. 9. Composition of mixed acids and esterincation effectiveness
(Chedin916). Zone 1, nitrocellulose obtained approaches trinitrate; zone
2, no nitration; the cellulose undergoes degradation; zone 3, degree of
nitration at equilibrium varies inversely with degree of hydration of HNCV
Ch6din and co workers91"95 carried out an extensive investigation of the
composition of HNOa-H^SC^-H^O mixtures by means of Raman spectrog-
raphy. On the basis of their results, they divided the ternary diagram
of mixed acid compositions into three zones as shown in Figure 9. In
Zone 2, in which the mixed acids are made up with high concentrations of
H2SO4 and low concentrations of HNO3 and H2O, all of the HNO? is present
in the form of nitronium ions (NO 2). Mixed acids in this zone cause con-
siderable degradation of the cellulose and effect little or no nitration.
In Zone 1, the original mixed acids are richer in nitric acid; all the nitric
acid is present in the form of HNO3 and NO 2 ions with none being hydrated,
and the H2SO4 is more or less hydrated. All of the mixed acids in Zone 1
91 J. Ch6din, Mem. services chim. etat (Paris), (a) 31, 154 (1944); (b) 32, 108 (1945).
92 J. Chedin, S. Fen6ant, and G. Watelle, Mem. services chim. Mat (Paris), 33, 127
(?947); Compt. rend., 224, 1008 (1947).
93 J. Chedin and A. Tribot, Mem. services chim. etat (Paris), 33, 143 (1947).
94 J. Chddin, A. Tribot, and S. Fen6ant, Compt. rend., 226, 2068 (1948).
96 J. Chedin, Chimie & industrie, 61, 571 (1949); this paper reviews the work covered
in references 91 to 94.
732 CELLULOSE
yield nitrocelluloses with the sanje high nitrogen content (about 13.75%)
approaching that of cellulose trinitrate. The failure to attain the theo-
retical nitrogen content of 14.14% N (that is, the completely trisubstituted
product) with HNOg-H^SOr-HkO mixtures can be best explained by the
ability of H2SO4 to enter into the esterification reaction in competition with
HNOs and the resultant formation of sulfuric-nitric mixed esters.96 In
Zone 3, nitronium ions could not be detected spectrographically. Ch&Iin
and coworkers determined by spectrographic and calorimetric examination
of the mixed acid in Zone 3 that the water is partitioned between HNO3
and H2SO4. They also determined that the degree of nitration attained at
equilibrium by mixed acids in Zone 3 decreases with increasing degree of
hydration of HNO3. Mixed acids with different compositions but with the
same degree of hydration of HNO8 yield nitrocelluloses with equal nitrogen
contents.
The formation of intermediate compounds during the nitration of cellu-
lose has been studied by various investigators. Katz, Hess, and
Trogus97""99 found that the [CeHioCVHNOsJx addition compound reported
by Knecht100 has a constant composition (as indicated by the formula) and a
characteristic x-ray diagram; on washing with water, this addition com-
pound yields a regenerated cellulose with the x-ray diagram of mercerized
cellulose. Miles101 found that the x-ray diagram of fibers obtained by
denitrating nitrocelluloses of Classes I and II (so designated by Miles and
Craik102) shows them to consist almost entirely of mercerized cellulose.
It is therefore probable that some constituent of the nitrating mixture must
have penetrated the whole structure. The results of Katz and Hess97
suggest that it is the HN08 that must penetrate the micelle. Carri&re108
found a higher HNOa.'KkSC^ ratio in the water in which nitrocellulose is
drowned after nitration than in the wringer acid and suggested that a
"pernitrocellulose," having more than the theoretical nitrogen content for
cellulose trinitrate, is formed during nitration and decomposed on contact
with water. Demougin anfl Bonnet104 studied this phenomenon and ob-
tained results which indicate that the excess HNOs is selectively adsorbed
98 K. A. Hofmann, Sitzber. preuss. Akad. Wiss.t Physik.-math. Klasse, 1933, 800.
w J. R. Katz and K. Hess, Z. physik. Chem., 122, 126 (1926).
« C. Trogus, Cettulosechemie, 15, 104 (1934).
M C. Trogus and K. Hess, Z. Elektrochem., 40, 193 (1934).
wo E. Knecht, Ber., 37, 649 (1904).
»l F. D. MOes, Trans. Faraday Sue., 29, 110 (1933).
»« F. D. Miles and J. Craik, Nature, 123, 82 (1929); /. Phys. Chem., 34, 2607 (1930).
w« E. Carrtere, Bull. soc. ckim., [4], 39, 438 (1926).
"< P. Demougin and Bonnet, Mim. poudres, 24, 157 (1931).
IK. DERIVATIVES OF CELLULOSE 733
by the nitrocellulose, and is not present as a chemical compound. Berl
and Rueff105 treated nitrocellulose with pure HNOa or mixed acids and ob-
tained an adsorption product (14.8% N) similar to that of Knecht100;
water decomposed it to nitrocellulose and HNO«. X-ray diagrams, after
treatment of cellulose with dry HNOa, showed spots due to cellulose, nitro-
cellulose, and Knecht's compound.
Various investigators have sought to determine the effective nitrating
agent in nitrations with HNOr-Hj^Or-HjO mixtures. Farmer108 proposed
that esterification of cellulose with HNOa proceeds through its "pseudo"
form (C^N-OH). Other investigators107""110 have obtained evidence which
shows that the NO 2 ion is the effective nitrating agent in the nitration of
aromatic molecules with mixed acids. By analogy, it has been suggested
that in the nitration of starch there is an electrophilic attack on the hydroxyl
oxygen by the NO2 ion followed by elimination of a proton.111 A similar
hypothesis for the mechanism of nitration of cellulose is supported by the
observation that nitration of ordinary cellulose with O18-enriched mixed
acid yields nitrocellulose with nitrate groups in which only two of the three
nitrate oxygens are O18-enriched.112
Although various explanations have been offered of the role of sulfuric
acid in the nitration reaction, there is no reason to believe that it performs
any other useful function than that of lowering the activity of water. This
view is supported by the knowledge that cellulose can be nitrated equally
well by nitric acid containing other dehydrating agents such as nitric
anhydride, phosphoric anhydride, or acetic anhydride. In fact, cellulose
can be nitrated by nitric acid alone (as was shown above), and, if steps are
taken to remove the water formed in the reaction, as is the case when cellu-
lose is treated with HNOg vapors under reduced pressure, products with
13.6% N can readily be obtained.49
Two general concepts have been advanced to explain the manner in which
nitration proceeds in cellulose fibers.118 According to the first of these con-
cepts, the nitration reagent penetrates between the cellulose micelles and
nitration proceeds progressively from theoutside to the interior of the micelle
105 E. Berl and G. Rueff, Cellulosechemie, 14, 97 (1933).
™ R. C. Farmer, /. Soc. Cham. Ind. (London), SO, 75T (1931).
107 G. M. Bennett, J. C. D. Brand, and G. Williams, /. Chem. Soc., 1946, 869.
w F. W. Westheimer and M. S. Kharasch, /. Am. Chem. Soc.. 68, 1871 (1946).
109 G. Williams and A. M. Lowen, /. Chem. Soc., 1050, 3312.
110 A. M. Lowen, M. A. Murray, and G. Williams, /, Chem. Soc., 1950, 3318.
l» S. Isradashvili, Nature, 165, 686 (1950).
112 R. Klein and M. Mentser, /. Am. Chem. Soc., 73, 6888 (1951).
118 K. Fabel, Nitrocellulose, 11, 223 (1940); 12, 3 (1941).
734 CELLULOSE
(micellar heterogeneous reaction). The second concept, which has re-
ceived widest acceptance, suggests that the nitration reagent penetrates
uniformly into all parts of the cellulose fiber and all the molecules are
nitrated at approximately the same time (homogeneous or permutoid re-
action).
On the basis of x-ray examination of the structure of different specimens
of nitrocellulose, Herzog and von Ndray-Szab6,114 and von Susich116
claimed that all nitrocelluloses (below the trinitrate) are made up of mix-
tures of cellulose trinitrate and unchanged cellulose, and they suggested
that the nitration of cellulose is a heterogeneous reaction. Hess and
Trogus,99'116 who have been the chief supporters of this theory, expressed
the further belief that all esterification reactions of cellulose, including nitra-
tion, are micellar heterogeneous and are not of the homogeneous type.
Sakurada and co workers117'118 proposed the following equation to cover all
micellar heterogeneous reactions including cellulose nitration :
x = kzm (1)
in which x = amount of cellulose reacted, z = reaction time, and k and m =
constants (see Section A of this Chapter IX) . Chemical and x-ray analyses
confirm, in their opinion, the view that the kinetics of cellulose nitration
are determined by the velocity of diffusion rather than by the velocity of
nitration. Tomonari119 has claimed that the mechanism of the reaction is
dependent on the composition of the mixed acid. With mixed acid very
low in water content, the reaction is micellar heterogeneous because the
rate of nitration is greater than the rate of diffusion; with acid mixes of
high water content, the reaction approaches the permutoid type because
of the greater speed of diffusion. Ch&iin and Tribot120 concluded from a
study of nitration rates that the reaction is made up of a fast reaction with
readily available OH groups on the surface and a slow reaction with OH
groups in the interior of micelles.
The overwhelming mass of evidence on this subject favors the permutoid
114 R. O. Herzog and S. von Ndray-Szabo, Z. physik. Chem., 130, 616 (1927).
1W S. von Niray-Szab6 and G. von Susich, Z. physik. Chem., 134, 264 (1928).
"• K. Hess and C. Trogus, Z. physik. Chem., B15, 157 (1931).
117 1. Sakurada and M. Shojino, /. Soc. Chem. Ind., Japan, 35, Suppl. binding, 287
(1932); 36, Suppl. binding, 306 (1933).
118 T. Nakashima, H. Nakahaj*, and I. Sakurada, J. Soc. Chem. Ind., Japan, 39,
Suppl. binding, 51 (1936).
119 T. Tomonari, J. Soc. Chem. Ind., Japan, 37, Suppl. binding, 478, 654, 772, 775
(1934); 38, Suppl. binding, 58, 62 (1935).
w J. CWdin and A. Tribot, Mtm. services chim. iM (Paris), 31, 128 (1944).
DC. DERIVATIVES OF CELLULOSE 735
theory of cellulose nitration. This evidence includes the conclusion, based
on the work of Berl and coworkers86 and others,87'88 that an equilibrium ex-
ists between the nitrating medium and the nitrocellulose over a large range
of composition. Despite the fact that equilibrium may be reached slowly
when proceeding from a higher to a lower nitrogen content, the final
nitrogen content is independent of time and can be changed to a higher or
lower constant value by immersion in another mixed acid. The existence
of this equilibrium is in harmony with the permutoid theory of nitration
only.
Opposition to the micellar heterogeneous theory of cellulose nitration has
also resulted from the failure to find any appreciable amount of unchanged
cellulose in the examination of hundreds of nitrocellulose samples.121 In
other fractionation experiments with numerous nitrocellulose specimens,
none of the fractions varied more than 0.3% in nitrogen content from the
starting material; also no evidence was obtained of the presence of cellulose
trinitrate in any of the samples.101 Adam122 has shown that nitrocellulose
solutions spread to form monomolecular films of nitrocellulose on 2 N
NaOH solution. In discussing these results, Miles101 has pointed out that,
if nitrocelluloses are constituted as claimed by Hess and Trogus,118 either
the micelle must solvate or disperse as such, in which case it seems impos-
sible to account for the spreading to the thickness of one molecule; or
the cellulose trinitrate must separate from the cellulose core since there is no
reason for believing that the cellulose would disperse. If the latter case
were true, only the cellulose trinitrate would spread and the measured
film area for nitrocellulose of 10.4% N would be 35% less than was actually
found, an error which could hardly be undetected.
Miles and Craik101-102 examined by x-ray diffraction methods a series of
nitrocelluloses, prepared with various HNOa-H^SOi-H^O mixtures, and the
corresponding denitrated products. They divided the products into
three classes depending on the position on the diagram (Fig. 7) of the mixed
acid from which they were prepared: (I) N content less than 7.5%, fiber
structure unimpaired, insoluble in all organic solvents; (II) N content
7.5-10.5%, fiber structure more or less impaired, swell or dissolve in ace-
tone; (III) N content over 10.5%, fibrous, dissolve completely in acetone,
comprise all technical nitrocelluloses. Miles and Craik were unable to
confirm the conclusion of Herzog and coworkers that nitrocelluloses are to
be regarded as mixtures of cellulose trinitrate and unchanged cellulose.
121 H. Brunswig, Z. ges. Schiess- u. Sprengstoffw., 23, 337, 384 (1928).
122 N. K. Adam, Trans. Faraday Soc.t 29, 95 (1933).
736 CELLULOSE
To account for their results, Miles and Craik presented the following pic-
ture of cellulose nitration. In the first stage of nitration (Class I) only the
diffractions of almost completely mercerized cellulose were observed, and
they therefore suggested that the OH groups in the "inner surface" of the
fiber were nitrated while those in the interior of the micelles were not
nitrated; that is, the reaction is heterogeneous. With a more concen-
trated nitration mixture, a greater area of the surface was affected so that
in nitrations of Class II the number of unchanged cellulose chains was too
small for definite diffraction and the micellar arrangement was almost
entirely broken down. In Class III, the swelling and micellar disarrange-
ment decreased as the concentration of the acid increased. With increase
in concentration of the acid, the number of nitrate groups introduced in-
creased, but with only approximate regularity, so that definite crystalliza-
tion was difficult at first but became less so with increasing substitution,
until at about 12.8% N the cellulose trinitrate diagram appeared. The
essential feature of this picture is that in nitrations of Classes II and III
the nitrate groups are so distributed among the chain molecules that there
is little likelihood of any one molecule being completely nitrated to the tri-
nitrate before the other molecules; that is, the reaction is permutoid.
Miles101 also pointed out that Hess and Trogus116 based their belief that
cellulose nitration is a micellar heterogeneous reaction on the examination
of nitrocelluloses (7 to 13% N) prepared by the somewhat unusual pro-
cedure of nitrating ramie for various periods (1 to 480 min.) with a single
anhydrous nitrating mixture containing 29.7% HNO3 and 70.3% H2SO4.
It seems doubtful that the examination of these nitrocelluloses would give
information of much value about the nature of the reaction which takes
place in commercial nitration, because in the latter case the reaction is
carried out practically to completion and the nitrogen content is controlled
primarily by the mixed acid composition and not by the reaction time.
In an extension of the work of Demougin and Bonnet,104 Ch&lin and
Tribot91'98'96 investigated the composition of the spent nitrating acid re-
moved by vigorous centrifuging and that retained in the nitrated fibers.
They found that the retained acid is richer in HNOs. They concluded
from their extensive results that the crystalline portions of cellulose fibers
are more readily penetrated by the small planar HNO3 molecules than by
the relatively large H2SO4 molecules. At the end of nitration, the degree
of hydration of HNOs in the crystalline areas of the fibers is the same as
that in the noncrystalline areas. According to Ch&lin and Tribot, this
explains why the degree of nitration is uniform throughout the fibers at
equilibrium.
DC. DERIVATIVES OF CELLULOSE 737
Kagawa128 determined that the total heat liberated on nitration of
cellulose to 10.4% N is 77.5 calories per gram of cellulose. From this he
computed that the actual heat of nitration of cellulose is 2.11 kcal. per
mole of OH nitrated. In a similar investigation, Calvet124 obtained a value
for the heat of nitration of 2.05 kcal. per mole of OH nitrated. Jessup and
Prosen125 derived the following equations expressing heat of combustion
(Ajff°) and heat of formation (A£Z/) of nitrocellulose as functions of nitro-
gen content :
= 4176.70 + 14126/' cal./g. nitrocellulose
AflJ = - 5896.88 + 26178/' cal./g. nitrocellulose
/' = mass fraction of nitrogen in nitrocellulose (corrected for impurity in
the nitrocellulose) in the range 0.115 to 0.135.
(5) Stability and Stabilization
A point of major importance in the chemistry of nitrocellulose is the fact
that its usefulness in explosives or in ordinary articles of commerce depends
a great deal upon its stability, that is, upon its ability to resist decomposi-
tion after long periods of time under various conditions. Thus, in gun-
cotton, the decomposition of unstable nitrocellulose during storage can
accelerate to the point at which the nitrocellulose will explode. The use
of unstable nitrocellulose will result in gradual embrittlement and disinte-
gration of films; in coated cloth (artificial leather) the coatings will em-
brittle, then crack and peel off; in lacquers the instability will cause cor-
rosion of the metal on which the lacquer has been applied.80'126
In the first years of the development of the manufacture of nitrocellulose,
which were devoted almost exclusively to the production of guncotton, a
number of disastrous explosions occurred which were caused by the in-
stability of the product. The search for the causes of this instability re-
sulted in the discovery by Abel,127 nearly 100 years ago, that retention of
traces of the nitrating acid by the nitrocellulose can cause instability.
Wiggam128 showed that the stability of nitrocellulose is lowered to an ex-
123 1. Kagawa, J. Soc. Chem. Ind., Japan, 42, Suppl. binding, 236 (1939).
124 E. Calvet, Mem. services Mm. etat (Paris), 34, 179 (1948).
126 R. S. Jessup and E. J. Prosen, /. Research Nail Bur. Standards, 44, 387 (1950).
128 J. B. Wiesel, Paint, Oil Chem. Rev., 80, No. 10, 8 (1925).
127 F. Abel, Brit. Patent 1102 (1865); see also E. C. Worden, Technology of Cellulose
Esters, Vol. I, Part 3, p. 1604 et seq., E. C. Worden, Millburn, N. J., 1921.
128 D. R. Wiggam, /. Phys. Chem., 35, 536 (1931).
738 CELLULOSE
tent dependent upon the amount of acid remaining absorbed. The for-
mation of sulfuric acid esters of nitrocellulose during nitration was recog-
nized early in this century.129-130 Hake and Lewis131 were among the first
to suggest that the instability of nitrocellulose can also be caused by the
gradual splitting off of the unstable sulfate groups. The formation of
these sulfuric acid esters, and hence the stability, was found to vary with
the composition of the nitrating acid, mixture. The combined sulfate
content was found to increase with increase in (a) nitration time,130 (b)
H^SO^HNOs ratio in the mixed acid,130'132'133 and (c) nitrosylsulfuric acid
content of the mixed acid.66 An increase in the water content of the mixed
acid increased the stability of the nitrocellulose produced.66-134 It is inter-
esting to note that when the formation of sulfuric acid esters of cellulose is
avoided, as in nitration with HNO3 vapors, nitrocelluloses of high stability
are obtained by merely washing the products with cold water.49
Various methods have been developed to improve the stability of nitro-
cellulose. Abel undertook the study and manufacture of guncotton for
the British Government in 1863. Following his discovery of the effect of
retained acid on stability, he found that pulping of guncotton in a paper
beater considerably reduced the fiber length of the nitrated cotton and
thus permitted ready removal of the retained acid by washing.127 In the
Abel process, the washing was carried out with alkaline water in order to
bring the nitrocellulose to neutrality as soon as possible.
Reeves and Giddens136 found that treatment of incompletely purified
nitrocellulose with cold dilute aqueous ammonia caused a marked improve-
ment in stability tests without reducing the appreciable sulfate content of
the original nitrocellulose. They attributed the improvement in stability
to the neutralization of sulfuric acid trapped within the fibers. They
also observed that nitrocellulose which had been stabilized by long boiling
and beating treatments was not further improved in stability by their
ammonia treatment. There is reluctance in the nitrocellulose industry to
rely on ammonia neutralization of bound sulfate groups, because it is
129 C. F. Cross, E. J. Bevan, and R. L. Jenks, Ber.t 34, 2496 (1901).
180 C. N. Hake and M. Bell, /. Soc. Chem. Ind. (London), 28, 457 (1909).
131 C. N. Hake and R. J. Lewis, /. Soc. Chem. Ind. (London), 24, 374, 914 (1905).
182 A. Hervd, Le Moniteur Scientifique Quesneville, [5], 8, 193 (1918).
<„)» L. Brissaud, Mem. poudres, 26, 158 (1934-5).
184 T. Tomonari, /. Soc. Chem, Ind., Japan, 37, Suppl. binding, 511 (1934); 38,
Suppl. binding, 326 (1935).
1M R. E. Reeves and J. E. Giddens, Ind. Eng. Chem.t 39, 1303, 1306 (1947); (to the
U. S. Secretary of Agriculture), U. S. Patents 2,404,887 (July 30, 1946) and 2,471,583
(May 31, 1949); Chem. Abstracts, 43, 5950 (1949).
IX. DERIVATIVES OF CELLULOSE 739
feared that these groups will hydrolyze in the course of time, leading to
acidity and rapid decomposition.
According to Mohr,136 squeezing water- wet nitrocellulose in a hydraulic
press aids in the removal of retained acid and thereby improves the stability.
The foundation for an important part of the stabilization process was
laid in 1906 when Robertson187 discovered that the unstable sulfuric acid
esters of nitrocellulose are more readily decomposed by steeping the nitro-
cellulose in boiling water while it still retains a small amount of mixed acid,
than by Abel's treatment with boiling alkaline water. At equal concen-
trations the stabilizing action of the following acids in the wash water de-
creases in this order: HN03, HC1, and H2SO4.138 According to Milliken,139
the stabilization process can be shortened appreciably by heating the nitro-
cellulose in water at above 100°C. under pressure, for example, at 116°C.
under a hydrostatic pressure of 15 Ib./sq. in. If nitrocellulose which has
been stabilized by an acid boil is treated with water containing CaCOa (for
example, hard water), a further improvement in stability results.140""142
This behavior is explained by the fact that the residual combined H2SO4
in nitrocellulose exists largely in the form RcenO-SO2OH which is readily
hydrolyzed by water or acids but which becomes RceiiO • SO2OM in presence
of inorganic bases.143 This salt is stable to inorganic bases143 and even to
dilute acids.141 Kridel and Rogers144 stabilized nitrocelluloses containing
relatively large amounts of combined sulfate (0.3 to 3.0%) by steeping
them in an aqueous solution of an alkali metal hydroxide or alkali metal
salt of a weak acid, such as sodium acetate.
Kullgren145""147 carried out an extensive study of the combined sulfate
groups in nitrocellulose. He found that the cation exchange which occurs
when nitrocellulose is suspended in salt solutions is due to RcellO-SO2OH
186 R. Mohr, Makromol. Chem., 4, 55 (1949).
187 R. Robertson, /. Soc. Chem. Ind. (London), 25, 624 (1906).
188 P. Demougin and M. Landon, Mint, poudres, 27, 182 (1937).
18» M. G. Milliken (to Hercules Powder Co.), U. S. Patents 2,103,592 (Dec. 28, 1937)
and 2,103,593 (Dec. 28, 1937); Chem. Abstracts, 32, 1932 (1938).
140 J. Goujon, Mem. artillerie frangaise, 8, 837 (1929).
141 M. Landon, Mim. poudres , 27, 190 (1937).
142 A. Breguet and A. Caille, Bull. soc. chim.t [4], 35, 680 (1924).
148 J. F. Briggs, /. Soc. Chem. Ind. (London), 25, 626 (1906).
144 D. J. Kridel and W. J. Rogers (to Eastman Kodak Co.), U S. Patent 2,604,471
(July 22, 1952); Chem. Abstracts, 46, 9844 (1952). .
146 C. Kullgren, Svensk Kern. Tid., 53, 233 (1941); Chem. Abstracts, 36, 262 (1942).
148 C. Kullgren, Ing. Vetenskaps Akad., Handl., No. 165, 5 (1942)4, Chem. Abstracts.
41,5714(1947).
147 C. Kullgren, Svensk Kem. Tid., 56, 221 (1944); Chem. Abstracts, 40, 2630 (1946).
740 CELLULOSE
groups. Not all of the combined H2SO4 is present in this form. Some of it
is present as the neutral sulfate — RceiiO • SO2ORcen, which is decomposed
very slowly in boiling water and somewhat more readily if the boiling water
is acidified slightly. Nitrocellulose which has not been stabilized by boiling
evolves acid even after long-continued washing; this is due to the slow
decomposition of combined sulfate groups. Kullgren interpreted this
behavior to indicate that the hydrolysis of RceiiO-SC^OH groups in un-
stabilized nitrocellulose, which occurs when the latter is stored in water,
results in some of these groups being converted to RceiiOH and others to
RcellO • SC^ORceii. This postulated formation of neutral sulfate ester groups
in the presence of excess water is improbable. Regardless of the mechanism
involved, Kullgren's contention that the stabilization treatment should be
carried out as soon as the nitrating acids have been removed is probably
correct.
Improvement in the stability of nitrocellulose has been effected by ex-
traction with alcohols under various conditions,148"""158 including the com-
mercial dehydration of nitrocellulose with ethyl alcohol.80 Berl and Delpy154
removed material of low stability from nitrocellulose by alcohol extraction
and suggested that it was highly degraded nitrocellulose Kullgren146
found that H2SO4 is formed more rapidly on treatment of nitrocellulose
with methanol than with water. His results indicated that the methanol
treatment decomposes both RceUO-SO2OH and RcellO • SO2ORceU groups.
Ch&iin and Tribot155 dissolved in acetone a nitrocellulose which had been
thoroughly washed with cold water after nitration and then precipitated the
nitrocellulose by pouring the solution into water. This treatment re-
duced the combined H2SO4 content from 0.7% to 0.1%.
Parallel with the development of stabilization and purification treat-
ments, there has occurred the development of materials (stabilizers) to be
incorporated with nitrocellulose plastics, coatings, and propellant powders
in order to prolong their life -by absorbing the products of nitrocellulose
14» H. Muraour, Bull soc. chim., [4], 51, 1089 (1932).
149 F. Langenscheidt, Z. ges. Schiess- u. Sprengstoffw., 9, 54 (1914).
.15° G. Centola, Ann. chim. appplicata, 31, 539 (1941); Chem. Abstracts, 39, 1539 (1945).
W1 R. W. Scharf, U. S. Dept. of Commerce, Washington 25, D. C., OTS, PB Rept.
51091 (1942).
1M Sadayoshi Watanabe, /. Soc. Chem. Ind.9 Japan. 46, 505 (1943); Chem. Abstracts,
43, 1965 (1949).
»» L. Brissaud, M6m. poudres, 31, 81 (1949).
184 E. Berl and M. Delpy, Z. ges. Schiess- u. Sprengstoffw., 8, 129 (1913).
m J. Ch&iin and A. Tribot, Him. services chim. etat (Paris), 32, 157 (1945); Chem.
Abstracts. 42, 4746 (1948).
IX. DERIVATIVES OF CELLULOSE 741
decomposition.166 Ideally, the stabilizer should be nonvolatile, completely
compatible with the nitrocellulose, and chemically inert toward it. Basic
inorganic compounds, such as CaCOa, were among the first materials used
as stabilizers. It should be noted, however, that CaCO» will react with ni-
trocellulose under certain conditions and actually reduce its stability.167'168
Many organic compounds have been proposed and evaluated as stabili-
zers for nitrocellulose.169"170 Compounds most widely used for this pur-
pose are weak organic bases and include diphenylamine, syw-diethyldi-
phenylurea (N,N'-diethylcarbanilide; Centralite I), and 1,1-diphenylurea
(acardite). According to Gilbert,171 the color stability of nitrocellulose
lacquers and base solutions is improved by incorporating in the nitrocellu-
lose a small amount (0.01 to 0.20% calculated as phosphoric acid) of
phosphoric, citric, or tartaric acid.
When considering the possibility of adding a "stabilizer" to nitrocellulose,
it should always be remembered that no known substance can reduce the
inherent, very slow rate of spontaneous degradation. On the contrary,
many substances that have been proposed as stabilizers, such as the organic
nitrogen compounds mentioned above, urea, and calcium carbonate,167-168
158 A. Marshall, Explosives, P. Blakiston's Son & Co., New York, Vol. II, 1917, p. 640;
Vol. Ill, 1932, p. 210. A review.
167 A. Koehler and M. Marqueyrol, Him. poudres, 23, 11 (1928).
IM P. Demougin, Mem. poudres, 26, 119 (1934-5).
«9 M. Marqueyrol, Mem. poudres, 23, 158 (1928).
160 M. Tonegutti, Atti V congr. nazL chim. pura ed applicata, Rome, 1935, Part II, 899
(1936); Chem. Abstracts, 31, 8198 (1937).
«> H. Muraour, Bull. soc. chim., [5], 3, 2240 (1936).
182 R. Dalbert, Him. poudres, 27, 117 (1937); 28, 147 (1938).
168 T. Urbanski, B. Kwiatowski, and W. Miladowski, Z. ges. Schiess- u. Sprengstoffw.,
32, 1, 29, 57, 85 (1937).
164 T. Urbanski and W. Miladowski, Z. ges. Schiess- u. Sprengstoffw., 33, 247 (1938).
185 H. Ficheroulle, Mem. poudres, 31, 167 (1949).
iee V. R. Grassie, L. Mitchell, J. M. Pepper, and C. B. Purves, Can. J. Research, 28B,
468 (1950).
"7 L. M6dard, Mem. poudres, 32, 305 (1950).
"• W. A. Schroeder, B. Keilin, and R. M. Lemmon, Ind. Eng. Chem., 43, 939 (1951).
"* J. A. Wyler and R. N. Boyd (to Trojan Powder Co.), U. S. Patent 2,297,734 (Oct.
6. 1942); Chem. Abstracts, 37, 1622 (1943).
170 D. R. Swan and J. M. Calhoun (to Eastman Kodak Co.), U. S. Patents 2,311,098
(Feb. 16, 1943); Chem. Abstracts, 37, 4316 (1943); 2,378,594 (June 19, 1945); Chem.
Abstracts, 39, 4222 (1945); and 2,407,209 (Sept. 3, 1946); Chem. Abstracts, 40, 7625
(1946).
171 C. B. Gilbert (to Hercules Powder Co.), U. S. Patent 2,260,248 (Oct. 21, 1941);
Chem. Abstracts, 36, 903 (1942).
742 CELLULOSE
actually increase the rate of decomposition and are therefore deleterious
in most applications. For example, phosphoric acid is slightly harmful
(Chapter X-A, Fig. 16). However, phosphoric acid is much less harmful
than traces of base derived from hard water or ingredients of a lacquer
formulation. Therefore, it is sometimes advisable to use a small amount
of phosphoric acid.
One exception to the general rule that no "stabilizer" should be used is
furnished by smokeless powder. In this case, the material is handled in
massive quantities and may be stored for decades. There is little chance
for the escape of any nitric acid that may accumulate, and it is desirable
to add substances that will react with the nitric acid, and yet are not basic
enough to catalyze decomposition. The following discussion should be
considered as being applicable only to smokeless powder problems.
It is apparent from the results of many investigations on the mechanism
of decomposition of nitrocellulose that this decomposition takes place in
two stages. In the first stage, the reaction is relatively slow; in the second
stage, the initial decomposition products react with the nitrocellulose, and
the reaction becomes autocatalytic and therefore accelerates rapidly. If
the products of decomposition in the first stage are removed as rapidly as
formed (for example, by a stabilizer), the reactions of the second stage are
prevented, and the decomposition is held down to a relatively slow
rate.169-172
Since nitrocellulose decomposes rather slowly at room temperature, a
wide variety of stability tests have been devised in which the nitrocellulose
is heated to accelerate decomposition; the results thus give an advance
indication of the stability behavior of a particular lot of nitrocellulose on
storage. 173~175 Wide differences in the bases of these tests have led to
differences in the interpretation of stability behavior.
In the decomposition of nitrocellulose at room temperature, hydrolysis is
caused by traces of acid whioh have not been completely washed out or
which result from the decomposition of unstable ester groups. The hy-
drolysis results in scission of the cellulose chain and the splitting off of
nitrate groups. If the nitrogen oxide decomposition products are removed
172 G. de Bruin, £tudes sur la decomposition spontanie de la pouare sans fum6et //,
Communiqu6 de la Soc, Anon. Fabriques N6erlandaises D'Explosifs, No. 2, Amsterdam,
May, 1924.
178 A. Marshall, Explosives, P. Blakiston's Son & Co., Philadelphia, Vol. II, 1917, p.
644, and Vol. Ill, 1932, p. 213.
174 D. R. Wiggam and E. S. Goodyear, Ind. Eng. Chem., Anal. Ed., 4, 73 (1932).
175 M. Tonegutti and E. Brandimarte, Z. ges. Schiess- u. Sprengstoffw., 35, 52, 76,
100,124.145,169(1940).
IX. DERIVATIVES OF CELLULOSE 743
as rapidly as formed (for example, by a stabilizer) and thereby are prevented
from reacting with the nitrocellulose, the decomposition proceeds rather
slowly and not with the increasing velocity that otherwise results.176
At elevated temperatures, another reaction takes place in the decomposi-
tion of nitrocellulose; this is the internal combustion of some of the an-
hydroglucose groups with resultant scission of the cellulose chain. If the
nitrogen oxides are not removed, they form with the water present a solu-
tion of nitric and nitrous acids which comprises the medium for the follow-
ing secondary reactions: (a) oxidation of glucose units, (b) hydrolysis of
nitrate groups, and (c) hydrolytic scission of the cellulose chain.177'178
Moisture content plays an important part in the decomposition of
nitrocellulose by hydrolysis. According to Muraour,179 water exerts no
action at 15° to 20°C. on suitably prepared nitrocellulose even after 20 to 30
years. At 50° to 75°C. in a humid atmosphere, the deterioration of nitro-
cellulose is rapid; the initial hydrolysis of the ester causes rapid decomposi-
tion of nitrocellulose. The action of HNOa on nitrocellulose is strongly
dependent on its concentration. A decrease in the moisture content (but
not to complete dryness) of smokeless powder containing even a trace of
free acid may cause an increase in the rate of decomposition. A large
amount of water dilutes the HNOs and thereby retards or completely stops
the decomposition. A minimum concentration of HNOa is necessary to
produce hydrolysis.
(6) Viscosity
The degree of polymerization (D.P.) of nitrocellulose is important in all
of its uses. An increase in D.P. is usually favorable from the standpoint of
physical properties, such as strength and flexibility. However, this in-
crease in D.P. is also accompanied by an increase in the viscosity of solu-
tions. A high D.P. therefore makes it difficult to handle nitrocellulose
solutions, as, for example, the application of lacquers. Thus, in actual use,
it is necessary to make a compromise between favorable physical properties
and low viscosity in solution.
17«H. Muraour, Chimie & Industrie, 20, 610 (1928); Bull soc. chim., [4], 47, 1259
(1930).
177 J. Desmaroux, Compt. rend.. 194, 1649 (1932); 196, 1394 (1933).
178 J. Desmaroux, R. Vandoni, L. Brissaud, and T. Petitpas, M6m. poudres, 29, 134
(1939).
17« H. Muraour, Bull. soc. chim., [4], 49, 276 (1931); 51, 1094 (1932); see also G. de
Bruin and P. F. M. de Pauw, £tudes sur la decomposition spontanee de la poudre sans
fumte, IV, Communiqu6 de la Soc. Anon. Fabriques N6erlandaises D'Explosifs, No. 4f
Amsterdam, Oct., 1925.
744 CELLULOSE
The utility of viscosity measurements for the characterization of nitro-
cellulose was recognized long before there was wide acceptance of the fact
that nitrocellulose is a high polymer and that solution viscosity is connected
with D.P. Consequently, a number of empirical tests were devised to
characterize viscosity (see Chapter XII and also Appendix for a comparison
of various tests and D.P.), and attention was directed to methods for the
control of viscosity.
The viscosity of nitrocellulose may be influenced by a number of factors
including the viscosity of the original cellulose180*181 and the nitration con-
ditions.182 An increase in the H^SCVHNOs ratio or a reduction in the
ratio of mixed acid to cellulose (in the range from 80: 1 to 30: 1) will result
in nitrocellulose of lower viscosity.80 Viscosity will also be reduced by an
increase in nitration time,30'188 nitration temperature,80'184 or in the nitro-
sylsulfuric acid content of the mixed acid.30 The use of HNOs-HsPOr-
P2O6 mixtures for nitration yields nitrocelluloses of very high viscosity.
This procedure is believed to cause less decrease in molecular chain length
(and thus, D.P.) than any other nitration method68""70 and, as indicated
previously, is frequently used in scientific work in the determination of
D.P. and the distribution of molecular weights of the original cellulose
(see Chapter X-D).
For certain types of nitrocellulose the viscosity can be adequately regu-
lated by controlling the various factors mentioned above. On the other
hand, a considerable proportion of the nitrocellulose manufactured is of the
low-viscosity type designed to meet the need of the lacquer industry for
increasing the concentration of nitrocellulose in lacquers without increas-
ing their viscosity. The increase in nitrocellulose concentration permits
the deposition of a thicker lacquer film in a single application and reduces
the cost. Processes have been developed for reducing the viscosity of
cellulose for the production of low-viscosity nitrocelluloses.14 However,
most of the low-viscosity nitxocelluloses on the market are made by proc-
esses which include some special viscosity-reducing treatment after nitra-
tion. 18B~187 The nitrocellulose may be treated with hot dilute solutions of
180 F. Olsen, Ind. Eng. Chem., 21, 354 (1929).
181 H. Aaronson, U. S. Dept. of Commerce, Washington 25, D. C.f OTS, PB Kept.
53806 (1932).
w> A. Tribot and A. Marsaudon, M6m. services chim. itot (Paris). 32, 145 (1945).
*« K. Atsuki and M. Ishiwara, Proc. Imp. Acad. (Tokyo). 4, 386 (1928).
m K. Atsuki and M. Ishiwara, /. Soc. Chem. Ind., Japan, 31, Suppl. binding, 268
(1928).
1M R. Gabillion, Rev. &n. mat. plastiques, 7, 259, 323 (1931); Chem. Abstracts, 25, 5555
(1931).
DC. DERIVATIVES OF CELLULOSE
745
acids, bases,188'189 or oxidizing agents190'191 or it may be digested in water
under pressure.192'193 The rate of viscosity reduction on heating nitro-
cellulose in water at 132°C. under pressure192 is shown in Figure 10.
40
36
32
26
u24
ui
z 20
t 16
</)
O
*..
15 SEC. (12.17 %
SEC. (l2.l5V*N)
i/2 SEC. (12.05% N)
2 4 6 8 10
DIGESTION TIME AT I32°C. IN MRS.
Fig. 10. Rate of viscosity reduction of nitrocellulose on digestion
in water at 132 °C. (Milliken192). The viscosities were determined
in 12.2% solution by the Hercules method.5
186 C. Stark, Kunststoffe, 21, 151, 201 (1931); Chem. Abstracts, 25, 5760 (1931).
187 M. Pavlik, Congr. chim. ind., Compt. rend. 17 e Congr., Paris, Sept.-Oct., 1937, 1058;
Chem. Abstracts, 32, 7719 (1938).
188 S. B. Luce (to Monsanto Chemical Co.), U. S. Patent 2,467,324 (April 12, 1949);
Chem. Abstracts, 43, 4853 (1949).
189 N. Moreau and Y. Lacroix, Mem. poudres, 32, 443 (1950).
190 K. Thinius (to Deutsche Celluloid-Fabrik), U. S. Patent 2, 104,957 (Jan. 11, 1938);
Chem. Abstracts, 32, 2352 (1938).
191 J. R. Buckley (to Canadian Industries Ltd.), Canadian Patent 442,207 (June 17,
1947); Chem. Abstracts, 41, 6718 (1947).
"* M. G. Milliken, Ind. Eng. Chem.. 22, 326 (1930).
19« M. G. Milliken (to Hercules Powder Co.), U. S. Patents 1,818,733 (Aug. 11, 1931);
1,911.201 (May 30, 1933).
746 CELLULOSE
If the viscosity data in Figure 10 are converted to intrinsic fluidities
(by means of the graph given in the Appendix) and the latter are plotted
against time, a straight-line relation is found. This behavior is consistent
with the idea expressed in Section C of Chapter III that random scission of
the nitrocellulose chain molecules occurs in the viscosity reduction process,
Lawton and Nason194 found that exposure to ultraviolet light causes a
decrease in the intrinsic viscosity of nitrocellulose. The viscosity reduc-
tion is less in a nitrogen atmosphere than in air or oxygen.
(b) COMMERCIAL MANUFACTURE OF NITROCELLULOSE196'196
In the direct dipping process, which was used in the early days of nitro-
cellulose manufacture, the cotton was dipped in the mixed acid and, after
nitration was complete, the nitrated product was transferred by hand
to a centrifuge where the spent acid was removed. In a later development,
nitration was carried out directly in a centrifuge which was rotated slowly
to cause circulation of the acid; when nitration was complete, the centri-
fuge was rotated more rapidly and the spent acid was removed. The
nitrocellulose was then transferred with aluminum forks to washing tubs.197
The Thomson displacement process attained large-scale operation in
England, where it was developed for the manufacture of guncotton. In this
process, pans with perforated false bottoms are filled with mixed acid;
then cotton is added and covered with a perforated plate so that the cotton
is completely submerged. At the end of nitration, the spent acid is drained
off slowly and is displaced at the same time with water. The advantages
claimed for this process include: (a) low power consumption; (b) no
moving parts, hence low investment in equipment; (c) low maintenance
cost; and (d) practically complete recovery of all acid.197'198
The most widely used process is still the so-called mechanical dipper
process,199""208 which was developed many years ago by the du Pont Com-
pany. A: sketch of a nitration building and equipment is shown in Figure
11. The dippers consist of cylindrical tanks made of an acid-resistant steel
alloy and fitted with two stirrers. A pipe of large diameter, closed by a
valve, leads from the bottom of the dipper to a centrifuge on the floor below.
191 T. S. Lawton, Jr., and H. K. Nason, Ind. Eng. Chem.t 36, 1128 (1944).
195 G. Bonwitt, Das Celluloid und seine Ersatzstoffe, Union Deutsche Verlagsges., Ber-
lin, 1933.
198 K. Fabel, Nitrocellulose; Herstellung und Eigenschaften, Ferdinand Enke Verlag,
Stuttgart,, 1950, p. 91. See this reference, p. 74, for a review of patents on nitrocellulose
manufacture which supplements the list given by Faust (ref. 3).
w F. L. Nathan, J. Soc. Chem. Ind. (London), 28, 177 (1909).
"• Anon., Chem. & Met. Eng., 50, No. 10, 130 (1943).
IX. DERIVATIVES OF CELLULOSE
747
Air connection for
blowing out fume
outlets
Quick opening, /*ff=\F
jote valve ' \ T
Fume line y
WarmV
Dipping tanks - -^-
Air connection for
blowing chokts
cetMoft nitnrfe htackr. / *.,
. 11. Nitrating building and equipment for the mechanical dipper process of
nitration (du Pont199).
748 CELLULOSE
Four dippers and one centrifuge form a unit in this process. The cellulose
to be nitrated, in suitable physical form, is dried to a moisture content of
less than 1%. The drying operation is important because excessive mois-
ture in the cellulose results in : (a) lower degree of nitration than planned
due to dilution of the mixed acid, (b) poor solubility because of irregular
nitration, and (c) temperature rise on dipping which causes disintegration
of fibers and loss in yield.30'199 About 1600 Ib. of mixed acid of selected
composition is charged at a definite temperature into the dipper from a
measuring tank, and about 30 Ib. of dried cellulose is then added. The
latter is quickly wetted and submerged as a result of the mechanical agita-
tion. After nitration has continued about 30 min., the bottom valve is
opened and the mixture of spent acid and nitrocellulose is discharged
directly into the centrifuge, where most of the spent acid is removed in
about 5 min. In general, about 1 Ib. of acid is retained by each pound of
nitrocellulose in the centrifuge. The nitrocellulose is discharged through
an opening in the bottom of the centrifuge into a volume of water which
thoroughly drowns the nitrocellulose and is large enough to prevent any
appreciable rise in temperature. The nitrocellulose; is then floated through
a pipe (nitrocellulose header) to the purification area. The spent acid
leaving the centrifuge is accumulated and brought back to proper strength
for re-use in nitration by the addition of strong nitric and sulfuric acids.
This process yields a very uniform product because of the agitation during
nitration and requires a minimum amount of manual labor.
Purification and stabilization of the nitrocellulose is carried out in large
wooden tubs lined with chrome-steel sheet and fitted with agitators.
These tubs, which can hold as much as 1*2,000 Ib. of nitrocellulose, are also
equipped with a perforated false bottom under which steam lines are
uniformly distributed. After the nitrocellulose has been washed to a pre-
determined acidity, the tubs are filled with water which is heated to boiling
by admitting live steam. This stabilization treatment may consist of one
continuous boil or it may consist of a series of short boils with frequent
changes of water. Depending on the type of nitrocellulose being manu-
factured, boiling and agitation are continued for 10 to 60 hrs.204 For
199 J. R. du P<mt.-Chem. & Met. Eng., 26, 11 (1922).
200 A. B. Nixon, Hercules Mixer t 8, 55 (1926).
201 E. P. Partridge, Ind. Eng. Chern., 21, 1044 (1929).
202 L. Sheldon, U. S. Dept. of Commerce, Washington 25, D. C., OTS, PB Kept, 12662
(1945).
*°» E. F., Thoenges, Manufacture of Cellulose Nitrate, in P. H. Groggtns, editor, Unit
Processes in Organic Synthesis, 4th ed., McGraw-Hill, New York, 1952, pp. 648-650.
204 C. G. Dunkle, U. S. Dept. of Commerce, Washington 25 D. C., OTS, PB Kept.
3055(1941).
IX. DERIVATIVES OF CELLULOSE 749
some types, the nitrocellulose is also pulped in machines similar to paper
beaters in order to reduce the fibers to a fine state of division and to remove
occluded acid. The nitrocellulose is then treated with dilute Na2COs
solutions and finally washed to neutrality.
For certain industrial uses, such as the manufacture of high-grade plas-
tics, nitrocellulose is often bleached to remove traces of color. The nitro-
cellulose may be bleached with a solution containing 2 Ib. of KMn(>4 per
100 Ib. of nitrocellulose and made slightly acid with H2SO4. When reducr
tion of the permanganate is complete, the nitrocellulose is washed and the
brown color removed by a solution of oxalic acid or SOa. The nitrocellulose
may also be bleached by a 30- to 60-min. treatment with a 1% solution of
chlorine or bleaching powder. The nitrocellulose is then washed, treated
with an antichlor, such as Na2SOs, to remove all traces of chlorine, and
washed again.199
As was indicated previously, special treatment is required to produce the
very low viscosity types of nitrocellulose required for lacquers. In early
industrial practice, the usual procedure186 involved batch digestion during
which nitrocellulose suspended in water was heated under pressure in welded
steel, brick-lined autoclaves holding 4000 Ib. of nitrocellulose and 60,000 Ib.
of water per charge. With digesters of this size, difficulty was sometimes
-encountered in removing adequately the gases resulting from the slight
denitration of nitrocellulose, and explosions occurred^ These explosions
were probably due to the entrainment of gases in the fibrous mass, which
•caused dry spots in the batch. A marked advance in this phase of nitro-
cellulose manufacture was achieved by Milliken192"193 in the development
of a continuous digester which avoids the hazards encountered in batch
pressure digestion. In the continuous process, a suspension of purified
nitrocellulose in water is fed by centrifugal pumps into a coiled, acid-resist-
ing alloy tube 4 in. in diameter and 4000 ft. long (see Fig. 12), which is
heated by a steam jacket for a distance at the entrance, heat-insulated along
most of its length, and cooled by a water jacket for a distance near the exit
end. To prevent flashing of the water in the coil, it is subjected to a back
pressure by connecting the exit to a standpipe 200 ft. high. The rate of
flow of the mixture and its temperature are controlled so that any practical
degree of viscosity reduction can be uniformly maintained. The contin-
uous digester is so designed that the small amounts of gas resulting from
denitration move along the tube with the nitrocellulose from which they
were evolved so that they have no opportunity to form gas pockets. The
products of this treatment are very uniform as evidenced by their high
solution clarity ; their stability is also very good .
CELLULOSE
Practically all types of nitrocellulose are dehydrated before use. The
loosely held water can be removed by draining or ccntrifuging. The wet
nitrocellulose is then transferred to hydraulic presses where it is formed into
cylindrical blocks with a pressure of about 250 Ib./sq. in. Denatured ethyl
A
Fig. 12. Continuous digester for nitrocellulose viscosity reduction (Milliken193).
alcohol is then forced through the block in the proportion of about 2 Ib.
of alcohol per pound of nitrocellulose. The excess alcohol is removed by
applying a pressure of 3000 Ib./sq. in. to the block. The compressed block,
which now contains about 30% by weight of ethyl alcohol, is broken up and
packed in steel drums for shipment. Commercial nitrocellulose has also
been dehydrated with isopropyl alcohol or butyl alcohol for certain uses.
DC. DERIVATIVES OF CELLULOSE 751
According to Brissaud,206 the ease of dehydration is affected primarily
by solubility and swelling in the alcohol; these properties, in turn, are a
function of the nitrogen content and uniformity of nitration.
(c) REACTIONS OF NITROCELLULOSE
Although nitrocellulose is an ester, treatment of it with the usual saponi-
fication agents (aqueous alkalies) does not yield cellulose and the corres-
ponding alkali nitrate. Instead, marked decomposition takes place with
the formation of the alkali nitrite and decomposition products of cellulose.
The products reported by various investigators as having been formed by
the action of alkalies on nitrocellulose include inorganic nitrates and
nitrites j206'207 ammonia;207 cyanide;208 carbon dioxide;209 oxalic acid;206'210'211
malic, glycolic, trihydroxyglutaric, dihydroxybutyric, malonic, andtartronic
acids;210 sugars;207'212 modified celluloses207 and their nitrates;218 partially
denitrated nitrocdluloses;212'214'215 and a product which is believed to be
either hydroxypyruvic acid216 (CH2OH-CO-COOH) or 2-hydroxy-3-oxo-
propionic acid217 (CHO-CHOH-COOH). Quantitative measurements
have been made of the formation of nitrites218 and of the decomposition of
nitrocellulose219'220 as shown by alkali consumption. The most extensive
quantitative investigation of the alkaline decomposition of nitrocellulose
has been carried out by Kenyon and Gray,221 who found that relatively
small amounts of COa are produced and relatively large proportions of the
nitrate groups are reduced to nitrite. The production of CC>2 and reducing
206 L. Brissaud, Mem. poudres, 31, 145 (1949).
206 E. Hadow, /. Chem. Soc., 7, 201 (1855).
207 A. Bechamp, Compt. rend., 41, 817 (1855).
208 W. Will. Ber., 24, 400 (1891).
209 H. Vohl, Dinglers Polytech. /., 112, 236 (1849).
110 E. Berl and A. Fodor, Z. ges. Schiess- u. Sprengstoffw., 5, 296 (1910).
211 W. G. Mixter, Am. Chem. J., 13, 507 (1891).
212 A. B6champ, Bull. soc. chim.t [1], 3, 289 (1863).
218 E. Berl and A. Fodor, Z. ges. Schiess- u. Sprengstoffw., 5, 2tr* VA^IU;.
214 A. Bechamp, Ann. chim. phys.f [3], 46, 338 (1856).
21* J. M. Eder, Ber.. 13, 169 (1880).
218 J. H. Aberson, Z. physik. Chem.t 31, 17 (1899).
217 C. Neuberg and M. Silbermann, Z. physiol. Chem., 44, 134 (1905).
218 L. Vignon and I. Bay, Compt. rend.t 135, 507 (1902); T. Carlson, Arkiv. Kemi
Mineral. Geol., 3, Art. 8, 1-15; Ber.t 40, 4191 (1907).
219 O. Silberrad and R. C, Farmer, /. Chem. Soc.t 89, 1759 (1906).
220 C. Piest, Z. angew. Chem.t 23, 1009 (1910).
221 W. O. Kenyon and H. LeB. Gray, /. Am. Chem. Soc.t 58, 1422 (1936); this refer-
ence contains a good review of the literature on the alkaline decomposition of nitro-
cellulose.
752 CELLULOSE
substances appears to be related to time, concentration of alkali, ratio of
alkali to nitrocellulose, and temperature.
In attempting to explain the mechanism of alkaline decomposition of
nitric acid esters, Berthelot222 suggested that it may take place in the fol-
lowing way:
RCH2ONO2 + NaOH -* RCHO + NaNO2 + H2O (2)
Klason and Carlson223 elaborated on this hypothesis and suggested that the
primary products are a peroxide and a nitrite :
RCH2ONO2 + NaOH -* RCH2OOH + NaNO2 (3)
In the ordinary course of events, the peroxide will decompose into an alde-
hyde and water :
RCH2OOH -* RCHO + H2O (4)
but when a reducing agent is present, the peroxide is reduced to the alcohol.
The assumption that simultaneous oxidation and reduction take place in
the treatment of nitrocellulose (and other nitric acid esters, such as nitro-
glycerin) with alkali, makes it easy to account for many of the products
actually obtained.224 However, Farmer226 favors the view of Berl and
Delpy226 that the initial reaction is hydrolysis and suggests that nitric
acid and alcohol may be formed momentarily and that the alcohol is im-
mediately oxidized by the nitric acid. Lucas and Hammett227 studied the
formation of benzaldehyde and sodium nitrite from benzyl nitrate in the
presence of NaOH. They felt that evidence was insufficient to suggest a
mechanism for this oxidation -reduction reaction. (See Section G of this
Chapter IX.)
The ability to saponify nitrocellulose to cellulose with alkalies in the pres-
ence of reducing agents has been applied commercially in the manufacture
of sausage casings and rayon by the ' 'nitrocellulose" process. In this
process, the nitrocellulose fibers and casings are denitrated by treatment
with alkaline hydrosulfides. Reichel and Graver228 recommended carry-
ing out the denitration on swollen nitrocellulose with a solution containing
not over 5% of an alkali hydrosulfide at a temperature not exceeding 20°C,
222 M. Berthelot, Compt. rend., 131, 519 (1900).
228 P. Klason and T. Carlson, Ber.t 39, 2752 (1906).
224 T. M. Lowry, K. C. Browning, and J. W. Farmery, J. Chem. Soc., 117, 552 (1920)
226 R. C. Farmer, /. Chem. Soc., 117, 806 (1920).
226 E. Berl and M. Delpy, Ber.t 43, 1421 (1910).
227 G. R. Lucas and L. P. Hammett, /. Am. Chem. Soc., 64, 1928 (1942).
228 F. H. Reichel and A. E. Graver (to Sylvania Industrial Corp.), U. vS. Patent 2,289,-
520 (July 14, 1942); Chem. Abstracts, 37, 532 (1943).
IX. DERIVATIVES' OF CELLULOSE 753
Reichel and Cornwall229 proposed that denitration of sausage tubings and
filaments containing nitrocellulose should be carried out with an alkaline
solution of an alkali hydrosulfide maintained at a pH below 12 by addition
of a buffer consisting of the salt of a strong acid and a weak base.
Denitration of nitrocellulose also takes place on treatment with acids,
but the reaction is much slower than when alkalies are used.226 Acid de-
nitration has been exemplified in the treatment of nitrocellulose with mixed
acid containing more water than the acid used to produce this nitrocellulose;
in this case, the esterification equilibrium shifts in the direction of lower
nitrogen content. One practical aspect of this behavior is observed in the
denitration of nitrocellulose which occurs while wringing out the spent
acid. This denitration is caused by dilution of the spent acid with mois-
ture from humid air; the superficial denitration which occurs decreases the
solubility of the product.230'231 Desmaroux232 observed that the effective-
ness of denitration of nitrocellulose with dilute solutions of HNOs, HC1,
and H2SO4 decreases in the order named and concluded that the denitrating
action decreases with increase in size of the acid anion.
Thinius233 denitrated water-wet nitrocellulose suspended in an inert
organic liquid, such as carbon tetrachloride, by treatment with an inorganic
acid chloride, such as PC13, POC13, or SOC12. Using this method, he ob-
tained from nitrocellulose with 13.4% N a product with good solubility
and a 12.7% N content.
Treatment of nitrocellulose with lithium aluminum hydride causes com-
plete denitration which is accompanied by extensive depolymerization.284
Simultaneous denitration and acetylation result when nitrocellulose dis-
solved in acetic anhydride is treated with zinc dust and anhydrous hy-
drogen chloride.236 The resulting cellulose acetate has a much lower D.P.
than the original nitrocellulose.
Treatment of nitrocellulose (11.7% N) with a solution of sodium,
sodamide, or potassium amide in liquid ammonia yielded a product with
229 F. H. Reichel and R. T. K. Cornwell (to American Viscose Corp.), U. S. Patent
2,421,391 (June 3, 1947); Chem. Abstracts, 41, 5306 (1947).
280 1. M. Naiman, N. D. Troitzkaya, and Ya. G. Danyushevskii, Rev. g6n. mat. plas-
tiques, 11, 273, 303, 333 (1935).
281 K. Fabel, Nitrocellulose, 12, 143 (1941).
282 J. Desmaroux, Compt. rend., 206, 1483 (1938).
283 K. Thinius (to Deutsche Celluloid-Fabrik Akt.-Ges.), German Patent 723,628
(Aug. 7, 1942); Chem. Abstracts, 37, 5590 (1943).
284 L. M. Softer, E. W. Parrotta, and J. DiDomenico, /. Am. Chem. Soc., 74, 5301
(1952).
286 D. O. Hoffman, R. S. Bower, and M. L. Wolfrom, /. Am. Chem. Soc.t 69, 249 (1947).
754 CELLULOSE
8.1% N which was claimed by the authors to be a cellulose amine.286
When the treatment was carried out with sodium acetylide in liquid am-
monia, an aminocellulose acetylide was obtained.237 In the light of pres-
ent-day knowledge, it is likely that products such as the above were so
highly degraded that they could hardly be called cellulose compounds.
Segall and Purves288 investigated the reaction of nitrocellulose (13.9% N;
D.S. 2.92) with hydroxylamine, 0-methoxyamine, and their hydrochlorides
in dry pyridine. In all four cases, the nitrate substitution was reduced
to about 1.7. The degree of substitution of oxime groups in the product
was as follows: hydroxylamine — 0.08, 0-methoxyamine — 0.02, hydroxyl-
amine hydrochloride — 1, 0-methoxyamine hydrochloride — 1.
(d) CONSTITUTION OF NITROCELLULOSE
Since the constitution of cellulose and its derivatives is discussed in
Chapter III, reference will be made here only to some findings which deal
specifically with nitrocellulose.
The degree of polymerization (that is, the number of anhydroglucose
units in the molecule or D.P.), determined viscometrically, of a number of
nitrocelluloses was found by Staudinger and Sohn289 to be appreciably
higher than that of the original celluloses calculated from viscosity meas-
urements in cuprammonium hydroxide solution. On the other hand,
when the celluloses were dissolved in cuprammonium hydroxide solution,
recovered by precipitation, and then nitrated, the D.P. of the nitrocelluloses
was the same as that of the original celluloses. According to Staudinger
and Sohn, this behavior may be explained by the assumption that (a)
ester linkages are formed in the cellulose chain by oxidation (particularly
with acidic agents) which are stable in the nitrating acid but are readily
hydrolyzed by alkali, or (b) end groups of a broken chain molecule in the
cellulose fiber are condensed and linked together by the dehydrating action
of the nitrating acid.
Although there is no evidence showing the existence of the ester linkages
proposed by Staudinger and Sohn, there is evidence favoring the view that
the viscosity behavior discussed above is due to the presence of alkali-
sensitive groups in the cellulose rather than to a condensation reaction tak-
•* P. C. Scherer and J. M. Feild, Rayon Textile Monthly, 22, 607 (1941).
*w P. C. Scherer and J. M. Saul, Rayon Textile Monthly, 28, 474, 637 (1947).
»» G. H. Segall and C. B. Purves, Can. J. Chem., 30, 860 (1952).
»«H. Staudinger and A. W. Sohn, Naturwissenschaften, 27, 648 (1939); Ber.9 72B,
1709 (1939); J. prakt. Chem., 155, 177 (1940); Melliand Textilber., 21, 206 (1940);
Cellulosechemie, 18, 26 (1940).
IX. DERIVATIVES OF CELLULOSE 755
ing place during nitration. Thus, in an investigation of the oxidation of
cellulose, Davidson240 found that oxycellulose produced by neutral or acid
oxidation suffered a loss in nitrocellulose viscosity (that is, the viscosity of a
solution of nitrocellulose prepared from a given cellulose) on treatment
with alkali. He suggested that certain types of oxidation do not result in
direct scission of the cellulose chain molecule but produce alkali-sensitive
groupings in the molecule.
An attempt241 has been made to determine the distribution of nitrate
groups in partially nitrated cellulose based on the knowledge that sodium
iodide replaces primary nitrate groups with iodine in simple sugar nitrates.
The method was unsuitable for highly nitrated nitrocellulose (13.2% N)
because of oxidative side reactions. On treatment of other nitrocelluloses
(2.5-9.0% N) with sodium iodide, about one-half of the nitrate groups were
replaced by iodine, about one-third remained unchanged, and the rest
were removed. The results are taken to indicate that at least one-half
of the nitrate groups in the low-substitution nitrocelluloses are in the pri-
mary position.
Wide variations have been noted in the electrochemical activity of
membranes prepared from nitrocellulose solutions (collodion) . The electro-
chemical activity is believed to be due to the presence in the nitrocellulose
molecules of carboxyl groups which are formed as the result of oxidation.242
Champetier and Cl&nent243 measured the infrared absorption spectrum
of nitrocellulose with the object of determining the structure of the ester
group. Their results led them to propose that the nitrate group in nitro-
cellulose has the following structure :
2. Cellulose Sulfate
The ability of concentrated sulfuric acid (preferably about 72% strength)
to dissolve cellulose has been known for over a century. If the cellulose is
precipitated immediately after dissolving, it contains little or no combined
sulfate. Esterification will take place if the cellulose is left in solution for a
240 G. F. Davidson, J. Textile /«$*., 29, T195 (1938); 31, T81 (1940).
141 G. E. Murray and C. B. Purves, J. Am. Chem. Soc.t 62, 3194 (1940).
"*K. Sollner, I. Abrams, and C. W. Carr, /. Gen. PhysioL, 24, 467 (1941); 25, 7
(1941).
248 G Champetier and P. L. Ctement, Compt. rend., 224, 199 (1946).
756 CELLULOSE
time. Reference has already been made to the formation of mixed sulf uric-
nitric acid esters of cellulose during nitration with H^SC^-HNOs mixtures.
A great many methods for preparing cellulose sulf ate and its salts have
been reported, as will be evident from the discussion which follows. How-
ever, the processes244'245 which have most closely approached commercial
success are based on the use of a mixture of sulfuric acid and isopropyl al-
cohol as the sulfating agent. In the process described by Frank244 fibrous
water-soluble cellulose sulfate is prepared by treating cellulose with sul-
furic acid and a suitable amount of an aliphatic alcohol. Malm and
Crane245 also developed a fibrous process for making cellulose sulfate in
which they treated cellulose with sulfuric acid in presence of an aliphatic
alcohol, ammonium sulfate, and an inert diluent such as kerosene, ben-
zene, or toluene.
A high-quality sodium cellulose sulfate, with a degree of substitution
(D.S.) of about 1, was introduced commercially in 1950246 but was sub-
sequently withdrawn from the market. This product is soluble in hot or
cold water and yields clear, highly viscous, neutral aqueous solutions at
relatively low concentrations. Sodium cellulose sulfate has been recom-
mended for use as a thickening agent in water and emulsion paints, food
products, and textile printing pastes; as a sizing and finishing agent for
textile fibers; and as a greaseproof and oilproof coating and a sizing agent
for paper.
Braconnot247 reported in 1819 that linen dissolves in cold concentrated
sulfuric acid and that on dilution with water a clear solution is obtained
which contains, in addition to H2SO4, another acid which he called "acide
v6g6to sulphurique." The barium salt of this acid, which remained in
solution, could be precipitated by addition of alcohol. Analysis of the salt
by de Carolles248 showed that the amount of Ba combined with the cellulose
increased with time of contact between the concentrated H2SO4 and the
cellulose; the product obtained after the solution was allowed to stand for
24 hrs. before dilution was reported to have a composition corresponding
to C4H8O4(SO3)2BaO-2H2O. A barium salt249 with the composition
244 G. Frank, U. S. Patent 2,559,914 (July 10, 1951); Chem. Abstracts, 45, 8770 (1951).
245 C. J. Malm and C. L. Crane (to Eastman Kodak Co.), U. S. Patent 2,539,451
(Jan. 30, 1951); Chem. Abstracts, 45, 4453 (1951).
246 Tennessee Eastman Corp., Sodium Cellulose Sulfate, Tennessee Eastman Corp.,
Kingsport, Tenn., 1950, 4 pp.
247 H. Braconnot, Ann. chim. phys., [2], 12, 185 (1819).
248 B. de Carolles, Ann., 52, 412 (1844).
249 H. Fehling, Ann., 53, 135 (1845).
IX. DERIVATIVES OF CELLULOSE 757
C9oHi8oO9oBaO(SO3)2 and a lime salt260 corresponding to CasHkg
have also been reported.
Honig and Schubert251 dissolved cellulose in concentrated H2SO4 at
room temperature and prepared Cu, Pb, and Ba salts of the cellulose sulfate.
They confirmed de Carolles* observation that the ratio of Ba to H2SO4 in the
salts is always 1 :2. Stern282 neutralized a sulfuric acid solution of cellulose
with baryta and decanted the clear liquid after BaSC>4 had settled out.
He concentrated the clear liquid on a water bath and then under vacuum ;
the solution remained neutral. The Ba salt was precipitated by adding
alcohol, and after purification and drying, was obtained in 48% yield as a
white hygroscopic powder with a composition corresponding to CeHgOs-
(SO^Ba. Kagawa263 reported that a better product (sodium cellulose
sulfate) is obtained by neutralizing the sulfuric acid solution of cellulose
with Na2COs and removing the sodium sulfate by dialysis. Champetier
and Bonnet254 found that aqueous solutions containing up to 54% H^SCX
yield H^SC^ addition compounds with cellulose which are decomposed on
washing with water. Higher concentrations of H2SO4 yield cellulose
sulfate; with 64% H2SO4 a cellulose sulfate corresponding to CeHgCVSOsH
is obtained.
Traube,255 Blaser, and Grunert256 found that well-dried cellulose combines
with SO3 vapors without charring to form cellulose trisulf ate ; the formation
of lower sulfates could not be detected. When sufficient SO3 had been
added for complete esterification, the product was dissolved in water and
treated with a slight excess of PbCO?. After removal of the Pb with H2S,
the product was neutralized with KOH. On cooling the solution, separa-
tion of " potassium cellulose trisulf ate A" occurred; this product, which
was obtained in 65% yield had a composition corresponding to CeHrCV
(SO4K)3. Concentration of the mother liquor yielded more material of the
same composition but with different specific rotation.
Gebauer-Fiilnegg,267 Stevens, and Dingier258 esterified cellulose with a
260 R. Marchand, J. prakt. Chem., 35, 199 (1845).
261 M. Honig and S. Schubert, Monatsh., 6, 708 (1885).
262 A. L. Stern, /. Chem. Soc., 67, 74 (1895).
*" 1. Kagawa, /. Soc. Textile and Cellulose Ind., Japan, 1, 681 (1945); through Che,.
Abstracts, 44, 7633 (1950).
264 G. Champetier and J. Bonnet, Bull. soc. chim., [5], 10, 585 (1943).
866 W. Traube, French Patent 657,204 (Jan. 11, 1929).
386 W. Traube, B. Blaser, and C. Grunert, Ber., 61B, 754 (1928).
887 E. Gebauer-Fiilnegg (to E. I. du Pont de Nemours & Co.), U. S. Patent 1,734,291
(Nov. 5, 1929); Chem. Abstracts, 24, 498 (1930).
268 E. Gebauer-Fiilnegg, W. H. Stevens, and O. Dingier, Ber.t 61B, 2000 (1928).
758 CELLULOSE
mixture of chlorosulfonic acid (C1SO8H) and pyridine at 100°C. until
complete solution occurred. On addition of alcohol, the pyridine salt of
cellulose sulfate was precipitated; with alcohol containing NaOH or NaCl,
the sodium salt was obtained. Addition of Ba salts to solutions of the
other salts yielded a Ba salt of the cellulose sulfate which was insoluble in
water and very resistant to acid hydrolysis. Analysis showed that the
products prepared by this process were almost completely esterified;
that is, the D.S. was nearly 3.0. In order to determine the extent of deg-
radation that had occurred in the esterification, they investigated methods
for removing the sulfate without excessive decomposition. This was
finally accomplished by treating the cellulose sulfate with methanol con-
taining 0.5% HC1 for 6 hrs. under reflux. The regenerated, sulfur-free
cellulose showed a specific rotation in cuprammonium hydroxide solution
similar to that of native cellulose. It could also be converted to trimethyl
cellulose and cellulose triacetate.269 However, no data indicative of molec-
ular weight were presented to prove that the product had not been strongly
degraded.
Traube and coworkers260 studied the ClSO3H-pyridine esterification
process and found that the product ordinarily obtained had a D.S. of about
2.6; prolonged esterification yielded a product with D.S. 2.9. A compara-
tive study of the SOa process showed that it is advantageous to use a carbon
disulfide solution of SOa in place of gaseous SO3 and to have an excess of
cellulose present. The reaction mixture is worked up by pouring it into an
excess of KOH solution, filtering off the unesterified cellulose, and recover-
ing the potassium cellulose trisulfate as before. The viscosity of a 1%
aqueous solution of sodium cellulose sulfate (D.S. 2.6) prepared by the
CISOsH-pyridine method was found to be much higher than that of a
similar solution of potassium cellulose trisulfate prepared by the SOa
process. This was taken to indicate that appreciable depolymerization
had taken place in the latter case. Hagedorn and Guehring281 have
claimed the preparation of cellulose sulfate by reacting alkali cellulose with
an oxygen-containing chloride of sulfuric acid (for example, SO2Cl2) in the
presence of benzene. Haskins262 prepared a cellulose ester containing sub-
stantially only sulfate groups by treating cellulose with a mixture of sulfuric
acid, pyridine, and an organic acid anhydride in which there is not a molar
2B» E. Gebauer-Fiilnegg and O. Dingier, /. Am. Chem. Soc., 52, 2849 (1930).
tec \y. Traube, B. Blaser, and E. Lindemann, Ber.t 65B, 603 (1932).
M1 M. Hagedorn and E. Guehring (to I. G. Farbenindustrie Akt.-Ges.), U. S. Patent
1,848,524 (Mar. 8, 1932).
MI J. F. Haskins (to Du Pont Rayon Co.), U. S. Patent 1,866,632 (July 12, 1932).
IX. DERIVATIVES OF CELLULOSE 759
excess of anhydride over sulfuric acid in the form of its pyridine salt.
Schulze268 prepared cellulose sulfate by treatment of cellulose with SOa in
the presence of a tertiary amine such as pyridine. Rigby esterified cellu-
lose with (a) pyrosulfuric acid (1128207) or its equivalent in the presence of
a tertiary amine like pyridine,264 and (b) a mixture of a substance contain-
ing the radical — O — S(>2 (for example, chlorosulfonic acid), a tertiary
amine, and an inert diluent (for example, toluene).266 Rubenstein266
prepared cellulose sulfate insoluble in water but soluble in cold dilute NaOH
solution by treating cellulose with up to 0.5 mole of a sulfating agent in
presence of NaOH or a tertiary cyclic amine.
Crane267 obtained stable, water-soluble cellulose acetate sulfates, in the
form of salts of sodium or other metals, by treating cellulose with relatively
large quantities of H2SO4 in presence of acetic acid and acetic anhydride
and neutralizing the mixed esters with a metal salt. One of the products
contained 20.7% sulfur and 1.8% acetyl. Crane268 also prepared cellulose
acetate sulfates with high sulfur content by treating cellulose with acetic
acid, acetic anhydride, and sulfuric acid in the presence of a substantial
amount of a bisulf ate which prevents the product from dissolving in the
reaction mixture. Araki269 prepared a mixed cellulose ester (for example,
cellulose acetate sulfate) and treated it with alkali to remove selectively the
organic acid groups and leave a pure cellulose sulfate.
Thomas270 obtained cellulose sulfate with D.S. less than 0.5 by heating
cellulose impregnated with an aqueous solution of sulfamic acid (HaN-SCV
OH) and urea. The ammonium cellulose sulfate was insoluble in water
but soluble in 10% NaOH solution at -10°C.
868 F. Schulze (to E. I. du Pont de Nemours & Co.), U. S. Patent 2,016,299 (Oct. 8,
1935).
864 G. W. Rigby (to E. I. du Pont de Nemours & Co.), U. S. Patent 2,025,073 (Dec.
24, 1935).
m G. W. Rigby (to E. I. du Pont de Nemours & Co.), U. S. Patent 2,033,787 (Mar.
10, 1936).
266 L. Rubenstein (to Imperial Chemical Industries Ltd.), U. S. Patent 2,042,484
(June 2, 1936); Chem. Abstracts, 30, 5037 (1936).
887 C. L. Crane (to Eastman Kodak Co.), U. S. Patent 2,582,009 (Jan. 8, 1952);
Chem. Abstracts, 46, 3275 (1952).
8MC. L. Crane (to Eastman Kodak Co.), U. S. Patent 2,622,079 (Dec. 16, 1952);
Chem. Abstracts, 47, 3565 (1953).
*** T. Araki (to Tokyo Industrial Research Inst.), Japanese Patent 176,243 (May 18,
1948); Chem. Abstracts. 45, 5406 (1951).
«TO jt c. Thomas (to E. I. du Pont de Nemours & Co.), U. S. Patent 2,511,229 (June
13, 1950); Chem. Abstracts. 44, 8657 (1950).
760 CELLULOSE
Jullander271 prepared cellulose sulfate by treating cellulose with
(in admixture with a lower alcohol) in the presence of an immiscible diluent
(toluene or ethylene dichloride) and a small amount of a saturated or un-
saturated fatty acid as an emulsifying agent.
A water-soluble ammonium cellulose sulfate containing not less than 25%
combined H2SO4 was obtained by Ward and Tallis272 by heating cellulose
impregnated with an aqueous solution containing H2SC>4, cyanamide, and
ammonium sulfate or an alkali metal sulfate. The product from this proc-
ess could be made insoluble in water and solutions of alkali by impregnat-
ing it with an aqueous alcoholic solution of the salt of a polyvalent metal
(CaCk, AlCla, ZnCl2) and drying the impregnated material.278
Caille274 treated strips of cotton cloth with a mixture of equal parts of
H2SO4 and acetic acid at 45°C. for 30 min. The product, which retained
its original fibrous structure, could be washed neutral with alcohol or cal-
careous water; washing with ordinary water caused hydrolysis. Heat
tests showed that the stability of the product increases with increasing
neutralization of the combined acid groups by the alkaline constituents of
the wash water. The cotton cloth had an affinity for basic dyes. Numer-
ous patents have been granted covering superficial treatments of cellulose
with compounds of sulfur to change its dyeing properties and its resist-
ance to water.275
3. Cellulose Phosphate
Champetier276 has reported the formation of an addition compound hav-
ing the composition SCeHioCVHsPCX on treatment of cellulose with aqueous
phosphoric acid solutions. The addition compound was destroyed by
washing with water; the fibrous cellulose was recovered apparently un-
changed. Cellulose phosphate (containing about 16% PC^) has been pre-
pared by treating cellulose with an oxygen-containing chloride of phos-
271 1. Jullander (to Mo och Domsjo Aktiebolag), Swedish Patent 137,018 (Aug. 26,
1952); Chem. Abstracts, 47, 4607 (1963).
272 F. Ward and E. E. Tallis (to Courtaulds Ltd.), U. S. Patent 2,603,551 (July 15.
1952); Chem. Abstracts, 46, 8371 (1952).
178 F. Ward (to Courtaulds Ltd. and F. Ward), Brit. Patent 670,346 (Apr. 16, 1952);
Chem. Abstracts, 46, 8371 (1952).
174 A. Caille, Chimie & Industrie, 15, 189 (1926).
276 O. Faust, Cettuloseverbindungen und besonders wichtige Verwendungsgebiete, J
Springer, Berlin, 1935, p. 518.
276 G. Champetier, Compt. rend., 196, 930 (1933); Ann. chim., [10], 20, 5 (1933).
IX. DERIVATIVES OF CELLULOSE 761
phoric acid in the presence of benzene.261 Tanner277 prepared a product
containing about 17% phosphorus (cellulose triphosphate would have
23% P) by treating cellulose with a mixture containing concentrated sul-
furic and phosphoric acids and a small amount of a weak acid catalyst
(for example, glacial acetic, boric, or formic acid). Cellulose phosphate
may be prepared by treating cellulose with moderately concentrated phos-
phoric acid and POCU with or without the addition of a diluent.278 In a
modification of this process, cellulose is treated with concentrated phos-
phoric acid and a small amount of another inorganic acid (for example,
HjjSCU) ; the cellulose ester is precipitated by addition of water to the re-
action mixture.279
Malm and Fordyce280 prepared an insoluble phosphorus compound by
treating cellulose with an acid chloride of a phosphoric acid ester in which
one chlorine atom is attached to the phosphorus atom of each molecule, for
example, dicresyl chlorophosphate. Cellulose phosphodiamide was ob-
tained by treating cellulose with phosphoryl chloride and then treating
the product with ammonia.281 Daul and Reid282 prepared the pentaeryth-
ritol phosphoric acid ester of cellulose by heating cellulose impregnated
with pentaerythritol phosphoric acid. By similar treatment with poly-
vinyl phosphoric acid, they obtained the corresponding cellulose ester.283
Much of the work on the preparation of cellulose phosphates has been
done with the object of developing a treatment for flameproofing textiles
made from cellulose fibers. In a comprehensive study of this subject,
Coppick and Hall284 reported that treatment of cellulose at 150° to 200°C.
with phosphoric acid and urea yielded a cellulose phosphate without caus-
277 W. L. Tanner (to National Chemical and Mfg. Co.), U. S. Patent 1,896,725 (Feb.
7, 1933).
278 1. G. Farbenindustrie Akt.-Ges", German Patent 547,812 (Mar. 29, 1932).
279 1. G. Farbenindustrie Akt.-Ges., German Patent 556,590 (Aug. 17, 1932); Chem.
Abstracts ,27,413(1933).
280 C. J. Malm and C. R. Fordyce (to Eastman Kodak Co.), U. S. Patent 2,008,986
(July 23, 1935); Chem. Abstracts, 29, 6055 (1935).
281 C. A. Thomas and G. Kosolapoff (to Monsanto Chemical Co.), U. S. Patent 2,401,-
440 (June 4, 1946); Chem. Abstracts, 40, 5250 (1946).
282 G. C. Daul and J. D. Reid (to the U. S. Secretary of Agriculture), U. S. Patent
2,592,544 (April 15, 1952); Chem. Abstracts, 46, 7768 (1952).
288 G. C. Daul and J. D. Reid (to the U. S. Secretary of Agriculture), U. S. Patent
2,610,953 (Sept. 16, 1952); Chem. Abstracts, 47, 1389 (1953).
284 S. Coppick and W. P. Hall, in R. W. Little, editor, Flameproofing Textile Fabrics,
Reinhold, New York, 1947, p. 179; see also A. C. Nuessle, /. Soc. Dyers Colourists, 64,
342 (1948).
CELLULOSE
ing excessive damage to the fibers. Reid and Mazzeno285 obtained prod-
ucts containing phosphorus and chlorine by treating cotton cloth with
POCU and pyridine. Patents have also been issued covering the modi-
fication of the dyeing and other properties of cellulose by superficial treat-
ment with a variety of phosphorus compounds.276
4. Cellulose Esters of Other Inorganic Acids
Treatment of cotton with thionyl chloride (SOCk) in presence of pyri-
dine resulted in the introduction of one atom of chlorine for each CeHioOs
group; the reaction product was dark in color and almost a powder.2*6
An attempt has been made to esterify cellulose by dissolving it in fluoro-
sulfonic acid (FSOsH), but results on identification of the cellulose ester
formed are inconclusive.287 Sulfur was introduced into cellulose by treat-
ing hydrocellulose with sulfur monochloride in hydrochloric acid.288 It
will be noted that in none of the above cases was a high polymer derivative
obtained with certainty.
*M J. D. Reid and L. W. Mazzeno, Jr., Ind. Eng. Chem., 41, 2828 (1949); see also J. D.
Reid, L. W. Mazzeno, Jr., and E. M. Buras, Jr., Ind. Eng. Chem., 41, 2831 (1949).
286 P. Carr6 and P. Mauclere, Compt. rend., 192, 1567 (1931).
"7 C. H. Mdllering, J. prakt. Chem., 134, 209 (1932).
*** R. Sthamer, German Patent 137,206 (Dec. 2, 1902)
C. ORGANIC ESTERS
CARL J. MALM AND GORDON D. HIATT
Cellulose as a polyhydroxy alcohol offers the possibility of ester forma-
tion with any organic acid. The complexity of the cellulose molecule,
however, makes it difficult to obtain a high degree of esterification with
many acids, and also limits the variety of reaction methods which can be
successfully applied. Esters of a high degree of esterification have, there-
fore, been confined largely to the normal fatty acid series and to the simpler
aromatic acids.
Derivatives of cellulose are usually prepared with the objective of ob-
taining physical properties not possessed by cellulose itself. Treatments
which result in only slightly modified cellulose, yielding no soluble deriva-
tives or no greatly changed physical properties, are therefore of compara-
tively minor interest. It is also important that the derivative be pro-
duced under conditions which do not excessively degrade the cellulose.
Methods for preparation of cellulose esters may in general be classified
as esterification by reaction of (a) acids, (b) anhydrides, and (c) acid
chlorides.
Appreciable esterification by direct treatment of cellulose with an or-
ganic acid is possible only in the case of the formate.1 Room temperature
treatment of cellulose with formic acid results in introduction of a small
amount of formyl; by addition of a catalyst this may be increased to a
sufficiently high degree of esterification to produce soluble products.
At room temperature other organic acids react only to a very slight de-
gree with cellulose. Clarke and Malm2 have shown that acetic, propionic,
and butyric acids at their boiling points produce minor degrees of esteri-
fication, the exact extent depending upon the reactivity of the cellulose.
Treatment with higher acids results in only low degrees of reaction. The
use of catalysts to promote reaction usually results in severe degradation
and has not been a useful esterification procedure.
Acids can be made to react with cellulose by using substituted anhydrides,
1 1. Sakurada, /. Soc. Chem. Ind., Japan, 35, B123 (1932).
2 H. T. Clarke and C. J. Malm, /. Am. Chem. Soc., 51, 274 (1929).
763
764 CELLULOSE
which in themselves do not esterify cellulose, as impelling agents to bring
about reaction. This process has been found to have wide applicability
and may be used as a general method of esterification, particularly with
high molecular weight acids. As impelling agents, anhydrides of halogen-
substituted3 ~B or alkoxy-substituted6 acetic acid are effective, mono-
chloroacetic anhydride being the most useful material for general applica-
tion. A reaction mixture of 1 part of cellulose with 5 parts of chloroacetic
anhydride, 0.01 part of magnesium perchlorate as catalyst, and slightly
more than the calculated amount of the acid to be esterified may be reacted
at 60-70°C., resulting in a solution of the cellulose ester within a few hours.
The impelling agent in this process brings about conversion of the esteri-
fication acid to its anhydride. The reaction is thus an esterification by
anhydride, but it is only necessary to use a slight excess of the calculated
amount of acyl to react with the cellulose, whereas usual anhydride re-
actions require at least double that amount. High molecular weight acids
are poor solvents for their reaction product, are difficult to wash out, and
are, therefore, undesirable components of a reaction mixture.
Esterification by acid anhydrides in the presence of mineral acid catalysts
is the most widely known process of esterification because of its commercial
use for cellulose acetate manufacture. The same general process may be
made to operate with propionic and butyric anhydrides, but for aromatic
anhydrides or aliphatic anhydrides of higher molecular weight the reactiv-
ity is not sufficient to produce high degrees of esterification. Soluble
cellulose derivatives containing free hydroxyl groups, such as acetone-
soluble cellulose acetate, may, however, be treated with higher anhydrides
in the presence of acid catalysts to esterify the free hydroxyl groups and
yield mixed esters.
Ketene has been suggested as an acetylation reagent.7'8 It does not react
directly with cellulose but through formation of acetic anhydride. Con-
3H. T. Clarke and C. J. Malm (to Eastman Kodak Co.), U. S. Patent 1,880,808
(Oct. 4, 1932); Chem. Abstracts, 27, 600 (1933); Brit. Patent 313,408 (Aug. 27, 1929);
Chem. Abstracts, 24, 1217 (1930).
4 A. Gundlach and T. Becker (to I.G. Farbenindustrie Akt.-Ges.), German Patent
516,250 (Jan. 20, 1931); Chem. Abstracts, 25, 1993 (1931).
fi M. Stacey, E. J. Bourne, J. C. Tatlow, and J. M. Tedder, Nature, 164, 705 (1949).
«H. T. Clarke and C. J. Malm (to Eastman Kodak Co.), U. S. Patent 1,987,053
(Jan. 8, 1935); Chem. Abstracts, 29, 1634 (1935).
7 D. A. Nightingale (to Ketoid Co.), U. S. Patent 1,604,471 (Oct. 12, 1926); Chem.
Abstracts, 21, 176 (1927).
8 G. D. Graves (to E. I. du Pont de Nemours & Co.), U. S. Patent, 1,990,483 (Feb. 12.
1935); Chem. Abstracts, 29, 1983 (1935),
IX. DERIVATIVES OP CELLULOSE 765
siderable color accompanies its use. It is more desirable to prepare and
purify the anhydride as a separate operation.
Tertiary organic bases, such as pyridine, may be used as reaction media
for organic anhydrides, but even at reflux temperatures the activity is not
sufficient to produce complete esterification, except for special cases, such
as the action of acetic anhydride on exceptionally reactive forms of cellulose.
This type of reaction is, therefore, generally useful only for partial esteri-
fication. Use of pyridine hydrochloride as a reaction medium for anhy-
drides provides a much stronger reaction condition.9 The greater activity
may be explained by formation of acid chlorides as the effective esterification
agents.
Acid chlorides of nearly all types of organic acids may be depended upon
to react with cellulose in the presence of pyridine as a reaction medium.
The reaction is carried out at elevated temperatures, usually at 100°C.
In certain cases, such as with acetyl chloride and toluenesulfonyl chloride,
secondary reactions result in the introduction of appreciable quantities of
combined nitrogen and chlorine into the cellulose derivative, making
special precautions necessary.
Aromatic and high molecular weight aliphatic acid chlorides have been
reacted with alkali cellulose to produce partial esterification. Under these
conditions it is necessary to employ excessive amounts of reagent, and the
process is, in general, inferior to the pyridine reaction.
Various types of cellulose have been employed as starting materials for
cellulose ester preparation, the most commonly used being bleached and
purified cotton linters and purified wood cellulose. As a by-product of the
cotton industry, linters have been available at reasonable prices and of
quality satisfactory for chemical use. The viscosity of solutions of native
cotton linters is higher than necessary for esterification and may be ad-
justed during purification to a suitable value, which may vary according
to the esterification process to be used. Wood cellulose for esterification
processes must be properly treated to yield a quality approaching that of
cotton linters, at the same time preserving uniform chemical reactivity
and sufficient viscosity so that esterification conditions will not reduce the
viscosity to too low a figure.
Analytical characteristics typical of cotton linters and wood pulp suitable
for acetylation are as follows :
f J. F. Haskins and F. Schulze (to Du Pont Rayon Co.), U. S. Patent 1,967,405
(July 24, 1934); Ckem. Abstracts, 28, 5982 (1934).
766 CELLULOSE
Cotton linters Wood pulp
Alpha-cellulose 99% 96%
Solubility in 7. 14% NaOH 1.8% 6%
Pentosans 0.8% 1.5%
Ash 0.08% 0.08%
Viscosity in 2.5% cupram- 5000 cps. 2000 cps.
monium hydroxide
For preparation of special cellulose esters and, to a considerable extent
in experimental investigations, chemically modified celluloses have been
used as starting materials because of their increased chemical activity as
compared with native celluloses.
Cellulose regenerated from cuprammonium or viscose solutions provides
a useful source of raw material lower in viscosity than the native celluloses,
and useful for mild reaction conditions. Another form of cellulose very
useful for experimental study is that regenerated from cellulose acetate by
suspension of the material in aqueous ammonia (15%) for two or three
days at room temperature to bring about complete deacetylation. This
cellulose has the advantage that it has the chain length of a typical cellulose
ester, and can be esterified under mild experimental conditions to produce
soluble derivatives.
1. Aliphatic Fatty Acid Esters
(a) CELLULOSE FORMATE
Cellulose formates have been of little interest, either for theoretical study
or for commercial use, because of their instability. They are hydrolyzed
readily by hot water, and they show a strong tendency toward develop-
ment of acidity at room temperature in the presence of moisture. Further-
more, the fonnyl groups are removed completely by dry heating at ele-
vated temperatures, and cellulose is regenerated on such a treatment.10
Complete esterification with formic acid is very difficult to attain, al-
though it is claimed that, with zinc chloride as a catalyst, the triformate
can be made.11 The usual reaction 'products contain 20-30% formyl,
10 G. Tocco, Giorn. chim. ind. applicata, 13, 325, 414 (1931).
11 1. G. Farbenindustrie Akt.-Ges., German Patent 636,307 (Dec. 18. 1936); Chem.
Abstracts, 31, 868 (1937).
IX. DERIVATIVES OP CELLULOSE 76
corresponding to approximately a diformate.12""14 To obtain these prod
ucts, the use of swelling agents, such as ZnClz or CaBr2, and of catalysts
such as HC1 gas, HaSCX and P2O6, are recommended; also metal perchlo
rates, especially magnesium perchlorate, are said to have catalytic activity.1
The temperature used varies with the catalysts employed, but to preven
degradation of the cellulose, Elod13 and Jurling16 recommended tempera
tures of 5°C. and below. The physical properties of films and filament
made from these formates have been investigated by Ueda.17 Mixed ace
tate formates18 have been made by reaction of cellulose in the presence o
sulfuric acid with the mixed anhydride of acetic and formic acids.
The formation of cellulose formate containing low percentages of formy
groups is said to be useful as a means of activating cellulose for furthe
esterification with other acids.19 These formates contain 6-10% formy
and are easily made by soaking the cellulose at room temperature in formi
acid of at least 85% strength. Formic acid has been recommended in pre
treatment of cellulose prior to acetylation.20
(b) CELLULOSE ACETATE
Cellulose acetate is the most widely known organic acid ester of cellulose
It is made at the present time in large commercial quantities for a variety
12 Y. Ueda, S. Nakamura, and T. Simada, Cellulose Ind. (Tokyo), 15, 212 (1939)
references to earlier papers are included.
18 E. E16d, U. S. Patent 1,900,599 (Mar. 7, 1933); Chem. Abstracts, 27, 3073 (1933)
Brit. Patent 275,641 (Aug. 9, 1926); Chem. Abstracts, 22, 2273 (1928); French Paten
638,431 (Aug. 29, 1927); Chem. Abstracts, 23, 278 (1929); German Patent 528,14
(June 25, 1931); Chem. Abstracts, 25, 4706 (1931).
14 W. Konig (to R. Koepp & Co. Chemische Fabrik Akt.-Ges.), German Patent 657,
874 (Mar. 16, 1938); Chem. Abstracts, 32, 6057 (1938).
» A. Dobry, French Patent 780,775 (May 3, 1935); Chem. Abstracts, 29, 6056 (1935)
18 J- G. Jurling, U. S. Patent 1,656,119 (Jan. 10, 1928); Chem. Abstracts, 22, 1041
(1928).
17 Y. Ueda and T. Simada, Cellulose Ind. (Tokyo), 15, 426 (1939); references t<
earlier papers are included.
M G. W. Seymour and B. B. White (to Celanese Corp. of America), U. S. Paten
2,363,091 (Nov. 21, 1944); Chem. Abstracts, 39, 3158 (1945).
19 P. A. Chevalet, Brit. Patent 264,181 (Jan. 11, 1926); H. Dreyfus, French Patem
642,329 (Sept. 16, 1927); Chem. Abstracts, 23, 1504 (1929); H. Dreyfus and C. I. Haney
Brit. Patent 288,657 (Apr. 10, 1928); Chem. Abstracts, 23, 703 (1929); I. G. Farben
industrle Akt.-Ges., Brit. Patent 305,601 (Feb. 5, 1929); Chem. Abstracts, 23, 4811
(1929); R. O. Herzog and G. Frank, French Patent 700,165 (Aug. 7, 1930); Chem
Abstracts, 25, 3481 (1931).
20 C. I. Haney and M. E. Martin (to Celanese Corp. of America), U. S. Patent 2,391,
569 (Dec. 25, 1945); Chem. Abstracts, 40, 2983 (1946).
res
CELLULOSE
>f uses including the manufacture of cellulose acetate yarn, photographic
ilms, transparent and pigmented sheeting, and plastic compositions such
LS those used for compression, extrusion, and injection molding, and, to a
nore limited extent, surface coatings. Annual production of cellulose
icetate in the United States has increased rapidly since 1920, as shown in
he curve of Figure 13, plotted from estimated data. The product most
ised is acetone-soluble acetate made by partial hydrolysis of the triester, as
550
500
g 45°
^ 400
V> 350
I 3
E 850-
c
o 20°
o
-g '50
e
Q- 100
50
Price
0)
Q.
CO
1.50.2
-8
100 c
CD
.50 .o
1920 24 28 32 36 40 44 48 52
Fig. 13. Production and price of cellulose acetate in the United
States since 1920.
rst described by Miles in 1905. 21 It is common practice today to refer to
he substantially fully esterified ester as the triacetate or primary acetate
nd to the hydrolyzed product as acetone-soluble acetate or secondary
cetate.
The most outstanding technical developments within recent decades
ave been in the direction of greatly reduced cost of manufacture (Fig. 13),
f improved control of processing conditions resulting in better clarity and
niformity of product, and of development of mild conditions of acetylation
o prevent excessive degradation of the cellulose, retaining the high molec-
lar structure of the cellulose molecule which contributes to the strength
nd durability of the product.
« G. Miles, French Patent 358,079 (Dec. 5, 1905).
IX. DERIVATIVES OP CELLULOSE 761
A continual trend toward milder esterification conditions has been notice
able throughout the history of cellulose acetate. The first reported acet
ylation of cellulose was accomplished by Schiitzenberger in 186522 by the
action of acetic anhydride in a sealed tube at 180°C. The amorphous
powder which resulted was soluble in alcohol. Franchimont in 18792J
used various catalysts to improve the esterification reaction and obtained
two types of products distinguished by their solubility in hot or cold alcohol
Later investigations, including those of Cross and Bevan and Miles during
the period of 1900-1905, still employed comparatively high acetylatioc
temperatures, although under somewhat more mild conditions than those
of Franchimont. The products obtained under these conditions were in-
soluble in alcohol but soluble in chloroform. Miles distinguished between
fully esterified acetate and his new partially hydrolyzed product as, re
spectively, chloroform-soluble and acetone-soluble derivatives. Later,
lower temperatures of esterification were employed, such as those described
by Dreyfus,24 in which the initial esterification mixture is cooled to bekrw
0°C. and during reaction is allowed to increase in temperature to a maxi-
mum of 20~30°C., giving a cellulose triacetate insoluble either in chloro-
form or chloroform-alcohol mixtures. Tetrachloroethane and mixtures oJ
methylene chloride with alcohol are among the few good solvents for this
product. Most commercial processes today use esterification. temperatures
not exceeding 50-60°C.
Numerous catalysts have been recommended for cellulose acetate manu-
facture, but none has offered outstanding advantages over sulfuric acid
This acid was employed by Franchimont23 in his early work on cellulose
acetylation and has since become the most standard agent for commercial
use. A great many patents have been issued describing its use in acetyla-
tion procedures. The quantity of catalyst used varies widely, depending
upon the temperature of esterification and reactivity of the cellulose em-
ployed. Ost26 compared sulfuric acid and zinc chloride as acetylation cata-
lysts. Quantities of 50-100% of the weight of the cellulose and compara-
tively high reaction temperatures were required with zinc chloride, whereas
sulfuric acid gave good results under mild temperature conditions with
5-10% of the weight of the cellulose. Patents covering the use of these
catalysts for cellulose acetate manufacture agree in general with these data,
28 P. Schutzenberger, Compt. rend., 61, 485 (1865).
88 A. Franchimont, Compt. rend., 89, 711 (1879).
" H. Dreyfus, U. S. Patent 1,280,975 (Oct. 8, 1918).
» H. Ost, Z. angew. Chem., 32, 66 (1919).
770 CELLULOSE
Perchloric acid is an exceptionally strong esterification catalyst and has
been patented for use in acetylation of cellulose.26 Commercial use of this
catalyst, however, has thus far been restricted to the manufacture of
fibrous cellulose triacetate. The acid has a purely catalytic action, and
does not combine with the cellulose. The fully esterified cellulose acetate
is, therefore, of good stability. In solution processes there are no indica-
tions of perchloric acid replacing sulfuric acid in commercial use. It is
highly corrosive to the metal equipment now used for the preparation of
cellulose esters.
Methanesulfonic acid in quantities of at least 50% of the weight of the
cellulose27 is reported to be an effective catalyst for esterification with an-
hydrides of acids as high as eight carbon atoms. Aromatic sulfonic acids
have also been used as esterification catalysts28 but are not as effective as
sulfuric acid.
The effective acidity of a catalyst is the predominant factor in its activ-
ity. Since the acetylation reaction is carried out in an anhydrous medium,
evaluation of catalysts must be considered from the point of view of non-
aqueous systems. The work of Br0nsted29 has shown that ionization must
be regarded as related to the solvent system employed. Conant and
Hall80 found that in acetic acid solution a number of amides and other ma-
terials, which in aqueous media do not form salts with acids, exhibit salt-
forming behavior; these present a system of acid-base equilibria which can
be titrated. Measurements of acidities of several mineral acids in glacial
acetic acid agree remarkably with their behavior as acetylation catalysts.
Sulfuric and perchloric acids in acetic acid solution were termed super-
acid solutions because of their exceptional strength as compared with other
acids. Hydrochloric, phosphoric, and nitric acids in glacial acetic are
comparatively weak, as are the organic sulfonic acids.
Sulfuric acid undergoes chemical changes in the presence of acetic an-
hydride. Franchimont81 reported that sulfoacetic acid is formed by action
of acetic anhydride and assumed that acetylsulfuric acid is an intermediate
*• C. J. Malm (to Eastman Kodak Co.), U. S. Patent 1,645,915 (Oct. 18, 1927);
Chem. Abstracts, 22, 164 (1928); D. Krtiger and F. Hohn, German Patent 519,877
(Nov. 5, 1931); Chem. Abstracts, 25, 3481 (1931).
87 Soci6t6 des usines chimiques Rh6ne-Poulenc, French Patent 705,546 (June 9, 1931);
Chem. Abstracts, 25, 5287 (1931).
» H. S. Mork, A. D. Little, and W. H. Walker, U. S. Patent 709,922 (Sept. 30, 1902);
French Patent 324,862 (Dec. 27, 1902).
*• J. N. Brjztasted, Ber.t 61B, 2049 (1928).
» J. B. Conant and N. F. Hall, /. Am. Chem. Soc., 49, 3047 (1927).
11 A. Franchimont, Compt. rend., 92, 1054 (1881).
IX. DERIVATIVES OF CELLULOSE
771
product. Stillich82 showed that acetylsulfuric acid is readily formed at low
temperatures; at elevated temperatures, in the presence of an excess of
acetic anhydride, quantitative conversion to sulfoacetic acid takes place.
Van Peski88 studied the reactions of acetic and sulfuric acids in detail and
prepared acetylsulfuric acid.
ABODE
F 6
2.5
o
r
o
ui
(A
«
tr
£
ACETYLATION SCHEDULE
A MIXER CHARGED WITH LINTERS AND AcOH
B MINOR PORTION OF CATALYST ADDED
C BEGAN COOLING TO 65* F;
THEORET.CAL MAXIMUM 2.27% ° Ac20 ADDED. CONTINUED COOLING TO 60* *
" • E MAJOR PORTION OF CATALYST ADDED
F-G WATER ADDED DURING ONE HOUR
to
05
02
NO WATER ADDED
TIME IN HOURS
AFTER ADDITION OF WATER- TIME IN DAYS
123 5
HOURS FROM BEGINNING
OF MIXER CYCLE
12345
DAYS HYDROLYSIS AT 100* F.
Fig. 14. Combined sulfur during preparation of cellulose acetate (Malm, Tanghe, and
Laird").
Under normal acetylation conditions there is practically no conversion
of acetylsulfuric acid to sulfoacetic acid. In the first part of the acetylation,
when the sulfuric acid and acetic anhydride concentrations are high enough
for this reaction, the temperature is low. Also, in a short time the sulfuric
acid combines quantitatively with the cellulose, as has been shown by Malm,
Tanghe, and Laird.*4 During the acetylation reaction it is replaced by
acetic acid through transesterification and reacts with other available
hydroxyl groups in the cellulose. After all the hydroxyl groups have been
M O. Stillich, Ber., 38, 1241 (1905).
" A. J. Van Peski, Rec. trav. chirn., 40, 103 (1921).
" C. J. Malm, L. J. Tanghe, and B. C. Laird, Ind. Eng. Chem., 38, 77 (1946).
772
CELLULOSE
esterified, the transesterification continues and the amount of combined
sulfuric acid decreases. This is shown in Figure 14.
The sulfation reaction is preferential to the primary OH groups.84*
At the end of the esterification most of the bound sulfuric acid is combined
with primary OH groups. This is shown84a (Fig. 15) by extending an esteri-
fication and comparing the combined sulfur of samples with the primary
and secondary OH content (determined by tritylation and carbanilation)
of samples from which the combined sulfuric acid had been removed by
selective hydrolysis.
0 5 10 15 20 25
Hours, after completion of esterification
Fig. 15. Preferential combination of sulfuric acid
with primary hydroxyl groups in cellulose acetate.
Combined sulfuric acid results in very poor stability of cellulose acetate
and must be reduced to an insignificant quantity in the final product.
In the manufacture of cellulose acetate the acetylation is interrupted when
only a small amount of the combined sulfuric acid has been replaced by
acetyl through transesterification. This is done by adding water in the
form of aqueous acetic acid of 50-75% strength. Most of the combined
sulfuric acid can be removed at this point by adding the water at a high
enough temperature and at a slow enough rate. This is shown in Figures
16andl7.84
If the cellulose acetate is to be used as the substantially fully esterified
product, this addition of water is very critical and determines the stability
of the ultimate product. If, however, the acetate is to be hydrolyzed to an
8te C. J. Malm and L. J. Tanghe, paper presented before the XHIth International
Congress of Pure and Applied Chemistry, Stockholm, Sweden, August, 1953.
IX. DERIVATIVES OP CELLULOSE
773^
acetone-soluble product, the combined sulfuric acid is also removed during
the hydrolysis, and the amount combined in the final product is influenced
by the sulfate-ion concentration during the hydrolysis. To keep this con-
centration low when large quantities of sulfuric acid have been used as
catalyst during the acetylation, metal ions such as magnesium can be added
015
UJ
3
D
o
o
£T
D
O
UJ
z
CD
2
O
o
0,10
005
002
001
ADDITION AT END OF
ESTERIFICATION
PARTS %AcOH
3 67
3 67
3 67
3 67
4 50
TIME
FOR
ADDITION
2 HOURS
I HOUR
30 MINUTES
10 MINUTES
10 MINUTES
5 10
20
40 60 6O
HOURS HYDROLYSIS AT IOO* F.
IOO
120
Fig. 16. Sulfur content during hydrolysis of cellulose acetate after addition of water at
different rates (Malm, Tanghe, and Laird84).
with the water. These form insoluble sulfates with part of the sulfuric
acid, thereby removing some of the sulfate ions from the solution.35 If the
combined sulfuric acid content still is too high for satisfactory stability at
the time of precipitation, several methods can be used for its removal.
Boiling water has often been recommended as an effective stabilization
treatment. Modifications designed to give improved results include use
of superheated steam36 instead of boiling water, or treatment in an auto-
» C. L. Fletcher (to Eastman Kodak Co.), U. S. Patent 2,259,462 (Oct. 21, 1941);
Chem. Abstracts, 36, 655 (1942); H. Dreyfus, Brit. Patent 566,863 (Jan. 17, 1945);
Chem. Abstracts. 41, 1435 (1947).
w C. Dreyfus and H. Martin (to Celanese Corp. of America), U. S. Patent 2,071,333
(Feb. 23, 1937); Chem. Abstracts, 31, 2816 (1937).
774
CELLULOSE
clave to permit use of higher temperatures.87 Boiling water acidified with
either organic or mineral acids has also been recommended as a stabilization
treatment.88 In all of these treatments, the essential requirement is re-
moval of sulfate without appreciable hydrolysis of acetyl groups. The
sulf ate groups are quite readily removed by boiling water if the combined
D
UJ
u
o
o ,
UJ
CO
DC
D
U.
0 10
005
m
o
u
0.02
0.01
A WATER ADDED AT 75' 82* F
B WATER ADDED AT 97-106° F
C WATER ADDED AT IIO-I24*F
5 10
20
40 60 60
HOURS HYDROLYSIS AT IOO* F.
too
120
Fig. 17. Sulfur content during hydrolysis of cellulose acetate after addition of water at
different temperatures (Malm, Tanghe, and Laird84).
sulfuric acid is in the form of an acid rather than a salt. Should the un-
stable product be given an opportunity to react with sodium or potassium
salts before stabilization, the action of boiling water is comparatively in-
effective, since alkali metal salts of the cellulose sulfate are very resistant
to hydrolysis. If left- in the final product the sulfuric acid, even if partially
neutralized with alkali metals, leads to instability toward heat.
Salts of alkaline earth or heavier metals may be used to neutralize the
acidity of the combined sulfate and give stable products. Magnesium
17 A. Aktien and H. Schulz (to Wolff and Co.), German Patents 440,844 (Feb. 12,
1927); 511,020 (July 10, 1930); Chem. Abstracts, 25, 1379 (1931).
*I. G. Farbenindustrie Akt.-Ges., Brit. Patent 299,326 (Mar. 24, 1930); Chem.
Abstracts, 23, 3343 (1929).
IX. DERIVATIVES OF CELLULOSE 775
salts have been recommended, as have those of calcium, barium, and alu-
minum. These may be used in just sufficient quantity to neutralize the
sulfate, or may be added in excess and allowed to remain on the product.
Such treatments result in increased ash content of the product and in some
cases affect the clarity of its solution. They also have a distinct effect on
the viscosity of the acetate in nonpolar solvents.89
(I) Solution Process
Except for a small amount of cellulose triacetate which is made by the
fibrous process, all cellulose acetate is manufactured by the solution proc-
ess. Sulfuric acid is the most commonly used catalyst and the discussion
below will be confined to the use of this catalyst.
Acetylation of cellulose is a heterogenous reaction, the cellulose being
suspended in the reaction medium and passing gradually into solution after
esterification has taken place. The course of the reaction has been found by
Sakurada40 to be largely controlled by rates of diffusion of reactant and cata-
lyst into the cellulose fiber. Hess and Trogus41 have observed changes in
x-ray diagrams from that of cellulose to that of the reaction product.
After the initial surface esterification, the reaction depends upon diffusion
of reactant and catalyst into the inner part of the fiber. The physical
condition of the cellulosic material is thus of primary importance to the
quality of the product. Only under conditions of uniform reaction can
products free from insoluble particles, unesterified fiber, and hazy appear-
ance in solution be obtained. Ideal esterification conditions are obtained
if the cellulose is not dried but the water is removed with acetic acid.
This might, however, not be economically practical. If the cellulose is
dried prior to acetylation, the drying should be carried out to avoid high
temperatures and local over-drying. Removal of the last few per cent of
moisture increases greatly the resistance of the fibers to acetylation.
(a) Pretreatment. The cellulose is usually given an activation treat-
ment before the acetylation. Of the many modifications which have been
recommended, the most common is treatment with acetic acid either alone
or in the presence of part or all of the esterification catalyst. In an early
patent, Wohl42 described steeping cellulose in acetic acid and then pressing
* C. J. Malm, L. J. Tanghe, and G. D. Smith, Ind. Eng. Chem., 42, 730 (1950).
<° I. Sakurada, /. Soc. Chem. Ind.t Japan, 35, Suppl. binding 3, 283 (1932).
41 K. Hess and C. Trogus, Z. physik. Chcm., B15, 157 (1931).
<*A. Wohl, Brit. Patent 20,527 (July 17, 1913); French Patent 448,072 (Jan. 22,
1913).
776 CELLULOSE
out the excess liquid before acetylation. Hot acetic acid43 and its vapors44
have also been patented as pretreatment agents. Use of catalysts in the
pretreatment step is included in many process modifications. Aqueous
acetic acid is more effective than glacial in producing a reactive cellulose,48
but, inasmuch as any water remaining from pretreatment must later be
destroyed by additional anhydride, glacial acid is usually used. The
quantity of pretreatment acid may vary widely, according to the conven-
ience of the process employed.
Although the temperature of pretreatment may range from 20°C. to the
boiling point of acetic acid (118°C.) or above, it is usually maintained be-
low 50°C. to avoid losses through evaporation. In the absence of catalyst,
this treatment causes only slight reduction of viscosity of the cellulose. In
the presence of catalyst, the cellulose viscosity drops rapidly. Cellulose
employed for acetylation has usually many times the viscosity it is to pos-
sess after acetylation, and at some step in the process the viscosity will be
greatly reduced. Werner46 has studied the reduction of viscosity during
esterification. This action is brought about by the catalyst; it will occur
mainly in the pretreatment step if catalyst is present, and will also take
place simultaneously with acetylation.
The amount of activation which is necessary depends upon the previous
history of the cellulose and the amount of catalyst used in the acetylation.
Increases in temperature and time during the activation with acetic acid
increase the efficiency of the activation. Decrease in ratio of acetic acid
to cellulose will also increase the efficiency. This is due to the greater effect
of the moisture in the cellulose on the concentration of the acetic acid.
(b) Acetylation. High reactivity of cellulose toward acetylation is ob-
tained when there is rapid and uniform sorption of sulfuric acid. The
amount of sorbed catalyst is not as important as the uniformity of its dis-
tribution in the cellulose.46* The acetylation takes place much slower than
the sulfation. Here again it is found that the primary hydroxyl groups
react more readily than the secondary.461* The acetylation of cellulose
should be allowed to proceed sufficiently slowly so that the reaction tem-
48 H. T. Clarke and C. J. Malm (to Eastman Kodak Co.), U. S. Patents 1,668,944-5
(May 9, 1928); Chem. Abstracts, 22, 2272 (1928).
44 H. Dreyfus, Brit. Patent 263,939 (Jan. 6, 1927); Chem. Abstracts, 22, 164 (1928).
46 H. Dreyfus, Brit. Patent 343,986 (Feb. 16, 1931); Chem. Abstracts, 26, 303 (1932).
49 K. Werner, Cellulosechemie, 12, 320 (1931).
«* C. J. Malm, K. T. Barkey, D. C. May, and E. B. Lefferts, Ind. Eng. Chem.t 44,
2904 (1952).
«*> C. J. Malm, L. J. Tanghe, B. C. Laird, and G. D. Smith, /. Am. Chem. Soc.t 75, 80
(1953).
IX. DERIVATIVES OP CELLULOSE 777
perature can be controlled, and, since the reaction is accompanied by evolu-
tion of a considerable quantity of heat, the esterifying anhydride must be
added at a comparatively low temperature. This will vary in different
processes, the degree of cooling required being in general dependent upon
the amount of catalyst used and the desired viscosity of the final product.
An amount of acetic anhydride, somewhat in excess of that necessary to
esterify the cellulose and to take care of any moisture present at the end of
the pretreatment, is added after the cooling operation has brought the tem-
perature to the proper point, and, if the catalyst has not been previously
added to aid in pretreatment, it must also be introduced at this time.
During the early stages of acetylation, the reaction should be well con-
trolled by external cooling to prevent too rapid a temperature rise. As the
reaction proceeds, it may be allowed to increase gradually in temperature
to a maximum, which should be well controlled to insure proper viscosity
of product. Proper correlation of the initial speed of reaction, maximum
temperature, and total time of esterification is important in production
control and must be maintained in balance.
Complete solution of the cellulose indicates the end of the esterification
reaction, and the temperature is then held constant until the solution vis-
cosity, which drops slowly under these conditions, is found by control test
to be at the proper point. Care must be taken not to allow the reaction to
continue for too long a time after acetylation is completed if small catalyst
concentrations are employed, since gelling will take place.
This gelling is caused by the insolubility in acetic acid or acetic acid-
acetic anhydride of a triacetate of high molecular weight cellulose. It can
be avoided by uniformly breaking down the cellulose before the acetylation
Dr by using large enough quantities of sulfuric acid as catalyst. A fully
*sterified cellulose acetate sulfate of high combined sulfuric acid content
is soluble in the acetylation mixture. However, as the combined sulfuric
icid is being replaced by acetyl through transesterification at the end of the
ssterification, the solubility decreases and ultimately gelling occurs.
(c) Hydrolysis. The acetylation reaction is stopped by the addition of
sufficient water to destroy any acetic anhydride remaining in the reaction
mixture and to bring the water content of the solution usually to 5-30%.
The water is best added in the form of aqueous acetic acid to prevent pre-
cipitation of the cellulose acetate. During the addition of this solution,
there is considerable removal of combined sulfate. This is especially effec-
:ive if the temperature is above 40°C. and the addition is slow. (See Figs.
16 and 17.) There is a critical water concentration at which the combined
778
CELLULOSE
sulfate is removed most efficiently.460 The exact optimum concentration
depends upon the amount of free sulfuric acid present and changes as the
sulfuric acid is split off.8**
Further quantities of sulfuric acid may be added with the aqueous acetic
solution, if this is desirable to speed up the hydrolysis step which is to fol-
low. Neutralizing agents may be added if large amounts of sulfuric acid
were employed in the esterification.
30
20
10
.01 .02 .03 04 05
% Sulfur in product
06
Fig. 18. Relation between soluble sulfate and combined
sulfur in cellulose acetate.
The amount of combined sulfur in the product depends upon the concen-
tration of soluble sulfate in the hydrolysis mixture34* (Fig. 18) and the de-
gree of hydrolysis (Fig. 16). The combined sulfur content passes through
a minimum during the hydrolysis and increases then slowly due to the in-
crease in the concentration of cellulose hydroxyl groups in the hydrolysis
mixture.84*
The reaction solution, as obtained after addition of the aqueous acetic
acid, undergoes an acid hydrolysis to remove some of the acetyl groups.
En order that this be a uniform reaction, it is necessary to maintain uniform
temperature until the desired degree of hydrolysis has been obtained. This
temperature may be chosen to suit the equipment and convenience of the
process, the rate of reaction for a given concentration of catalyst being
"aster, the higher the temperature is.
The higher the percentage of water in the hydrolysis solution, the less will
460 T. Araki, Textile Research /., 20, 631 (1952).
IX, DERIVATIVES OF CELLULOSE
779
be the degradation at a given temperature. The susceptibility of the ace-
tate to degradation increases with decrease in acetyl content; therefore if a
rapid hydrolysis with large quantities of sulfuric acid or at high tempera-
tures is desired, water should be added during the process.47 Hydrolysis
under ordinary conditions may be carried to as low as §0% acetyl. At
about this point the usual reaction solutions containing approximately 10%
5 40
•
t»
•»
° 30
o
">.
S
•I 20
>.
w
* 10
u 10 20 30
% H20 in Hydrolysis Both
Fig. 19. Effect of water content on per cent primary hydroxyl
(Malm, Tanghe, and Laird40).
water become poor solvents for the acetate, and gelatinous precipitation
begins to take place. By addition of larger quantities of water to reduce
the acetic acid strength to 60 or 70%, the hydrolysis may be continued to
acetyl values below 20%, at which point the product becomes completely
water-soluble (see Fig. 25) and must be isolated by precipitation into or-
ganic solvents such as acetone or alcohol.48 Continued hydrolysis below
about 13% acetyl results in regeneration of essentially deacetylated cellu-
lose, which is insoluble in water and organic solvents.
The percentage of water in the hydrolysis bath influences the ratio of
primary to secondary hydroxyl groups in the acetate.49 (See Fig. 19.)
47 C. J. Malm and C. L. Fletcher (to Eastman Kodak Co.), U. S. Patent 2,013,830
(Sept. 10, 1935); Chem. Abstracts. 29, 7074 (1935).
41 C. R. Fordyce (to Eastman Kodak Co.), U. S. Patent 2,129,052 (Sept. 6, 1938);
Chem. Abstracts, 32, 8777 (1938).
*• C. J. Malm, L. J. Tanghe, and B. C. Laird, /. Am. Chem. Soc.9 72, 2674 (1950).
780 CELLULOSE
This is due to partial preferential re-esterification of primary hydroxyl
groups during the hydrolysis. An increase in the acetic acid concentration
promotes the re-esterification and results in a product of lower primary
hydroxyl content. 49a-40b
During the course of the hydrolysis, test samples may be taken, by which
the time of hydrolysis to give any desired acetyl value may be determined.
If an ester of high purity and clarity is desired, the solution can be filtered
before precipitation. This can be done by neutralizing the sulfuric acid
with magnesium acetate under conditions which form the right-size mag-
nesium sulf ate crystals and using these as a filter aid.50
(d) Purification. At the time when control tests have indicated that the
proper degree of hydrolysis has been reached, the cellulose acetate is pre-
cipitated from solution.
The most important objects in this operation are to produce a precipi-
tated material which is readily penetrated by water so that it may be easily
washed to remove all uncombined acids, and to obtain, for recovery,
acetic acid of as high concentration as possible. A stream of viscous
solution poured into water will usually result in formation of "skins" of
precipitated material containing entrapped solution into which it is difficult
for water to penetrate. For this reason, if solutions are precipitated
directly into water, they must first be diluted to a low viscosity. A more
practical procedure for commercial use is to precipitate the solution into
dilute acetic acid, keeping the concentration of the precipitation acid within
a range of 25 to 35% acetic acid by simultaneous addition of reaction solu-
tion and dilute acid.61
If the ester is desired in a powder form, the precipitation can be carried
out by diluting the solution with an acetic acid-water mixture and finally
adding an acid-water mixture of sufficient water concentration to cause
precipitation. The amount of dilution, strength of acid, and temperature
** C. J. Malm, L. J. Tanghe, B.-C. Laird, and G. D. Smith, /. Am. Chem. Soc., 74,
4105 (1952).
«b L. A. Hiller, Jr., / Polymer Sci., 10, 385 (1953).
60 H. G. Reed and J. K. Beasley (to Eastman Kodak Co.), U. S. Patent 2,494,143
(Jan. 10, 1950); Chem. Abstracts, 44, 2806 (1950); M. E. Martin and T. M. Andrews
(to Celanese Corp. of America), U. S. Patent 2,522,580 (Sept. 19, 1950); Chem. Abstracts,
45, 1343 (1951).
81 H. T. Clarke (to Eastman Kodak Co.), U. S. Patent 1,823,348 (Sept. 15, 1931);
Chem. Abstracts, 26, 304 (1932); French Patent 693,133 (Nov. 17, 1930); Chem. Ab-
stracts, 25, 1671 (1931); Brit. Patent 342,596 (Feb. 5, 1931); Chem. Abstracts, 25, 4401
(1931).
IX. DERIVATIVES OF CELLULOSE 781
of precipitation have to be varied, depending upon the viscosity of the ester
and its acetyl content, to obtain optimum results.62
The process of washing must be given careful consideration in commer-
cial operation because it is necessary to recover the acetic acid. The cost
of recovery increases rapidly with increased dilution. The original pre-
cipitation liquor is first drained off, and means are then employed to remove
as much as possible of the acid held by the cellulose acetate with as little
further dilution of the acid as possible. After this has been accomplished,
the product is thoroughly washed with water to remove the last traces
of acetic acid and catalyst.
The quality of water employed for washing is important to the quality
of the product. Good clarity and freedom from color can be obtained
only by use of clear water free from iron and other color-causing substances.
The final operation in cellulose acetate manufacture is drying. The last
wash water is removed by a centrifuge, press, or similar equipment, and the
product is dried.
The recovery of used acetic acid, mentioned above as an essential step
in cellulose acetate manufacture from an economic point of view, is rendered
difficult by the fact that vapor pressure relationships prevent easy separa-
tion of water and acid by fractional distillation. Early recovery methods
involved evaporation to obtain sodium acetate. Present processes, how-
ever, use extraction or azeotropic distillation.
The following procedure84 will serve as an example of the solution proc-
ess:
One part of cellulose, moisture content about 5%, is added to 2.4 parts of acetic acid
in a Werner and Pfleiderer type stainless steel mixer and the mixer is run for one hour at
37.8°C. (100°F.). Four parts of acetic acid and 0.88% H2SO4, based on the weight of
the cellulose, are added, and the mixing is continued at the same temperature for 45
minutes before cooling to 18.3 °C. (65°F.). Next, 2.7 parts of 98% acetic anhydride are
added, and the cooling is continued to 15.6°C. (60°F.). Then 6.12% H2SO4, based on
the weight of the cellulose, diluted with an equal weight of acetic acid is added. The
temperature is permitted to rise gradually to 32-35°C. (90-95°F.). during an interval of
1.5 to 2 hrs. At this stage the reaction mixture is very viscous and free from fibers. A
mixture of one part of water and two parts of acetic acid is then added during an interval
of 1 hr. The reaction of the excess anhydride with the water gives a sharp temperature
rise of about 5°C. (10°F.). After the solution has been thoroughly mixed and the
temperature adjusted to 37.8°C. (100°F.), the solution is transferred to a hydrolysis
vessel and held at 37.8°C. (100°F.) until an ester of the desired acetyl content is ob-
tained.
" C. J. Malm and C. L. Crane (to Eastman Kodak Co.), U. S. Patent 2,469,396
(May 10, 1949); Chem. Abstracts, 43, 5592 (1949).
782 CELLULOSE
(2) Modified Solution Processes
Numerous variations of procedure for the manufacture of cellulose ace-
tate have appeared in the patent literature. A great many of these are
concerned with details of processes essentially as described above. A few
involve major changes and are worthy of note.
Zinc chloride as catalyst has been employed on a production scale. Its
advantages lie in better solubility of the triacetate in an acetic acid-zinc
chloride mixture, which prevents gelling, and in good stability of the re-
sulting product. The presence of zinc chloride in the precipitation liquor,
however, complicates the recovery of acetic acid.
Several solvents other than acetic acid have been recommended for
acetylation media, but of these only the chlorinated hydrocarbons88 and
sulfur dioxide are of commercial interest. Of the chlorinated solvents,
methylene chloride is being used on production scale. Its excellent solvent
power for cellulose triacetate makes it possible to obtain products of very
high viscosity with low catalyst concentration, which in an acetic acid
reaction solution would result in gelling. The small amount of catalyst is
advantageous in giving only minor quantities of combined sulfate in the
product. Methylene chloride, because of its low boiling point, can be used
to control the acetylation reaction temperature by evaporation.54 This
makes it possible to use much larger batch sizes than are possible when
the reaction temperature is controlled by cooling of the mixer jacket only.
The following procedure56 will serve as an example of the process which
uses methylene chloride as a solvent :
The acetylation is carried out in a horizontal cylindrical vessel provided with stirring
blades on a horizontal shaft. This acetylator is fully charged with cellulose; then the
reaction mixture, previously cooled to 15-20°C., is added in two or more portions. The
reaction mixture consists of 1 part H^SO^ 400 parts of methylene chloride, and 300
parts of acetic anhydride (95% or over) per 100 parts of cellulose. The temperature is
not allowed to exceed 50 °C. Under these conditions esterification to triacetate is com-
pleted in a reaction time of 5 to 6 hrs.
" R. Hofmann (to Hercules Powder Co.), U. S. Patent 2,126,190 (Aug. 9, 1938),
Chem. Abstracts, 32, 7723 (1938); Brit. Patent 337,366 (Oct. 27, 1930); Chem. Abstracts;
25, 2288 (1931); French Patent 697,156 (Jan. 13, 1931); Chem. Abstracts, 25, 2848
(1931); German Patent 526,479 (June 6, 1931); Chem. Abstracts, 25, 4401 (1931); H.
LeB. Gray (to Eastman Kodak Co.), U. S. Patent 1,823,359 (Sept. 15, 1931); Chem.
Abstracts, 26, 304 (1932).
M L. E. Clement (to Eastman Kodak Co.), U. S. Patent 2,104,023 (Jan. 4, 1938);
Chem. Abstracts, 32, 1931 (1938).
M Office of the Publication Board, PB Report 377, Continuous and Staple Fiber
Plants of Germany, 1945.
DC. DERIVATIVES OF CELLULOSE 783
The contents of the acetylator are then discharged into the hydrolyzing and precipi-
tating vessel. There is then added 50 parts of water and 6.5 parts of sulfuric acid per 100
parts of the original cellulose and the ester is hydrolyzed to the desired extent at 55° to
60*C. Hydrolysis is stopped by addition of 20 parts of a 30% solution of sodium acetate,
whereupon the methylene chloride is removed by distillation and recovered. The
temperature is ultimately raised to 80 °C. in order to maintain the residue, a 30% solution
of cellulose acetate in acetic acid, in a fluid state. At this temperature the acetate is
precipitated by adding 800 parts of an acetic acid solution of 5 to 12% concentration,
the exact concentration depending upon the type of acetate made. The acetate is then
washed and dried in the customary way.
Sulfur dioxide is a good solvent for cellulose acetate and has been em-
ployed as a reaction solvent in a process which has been tried on a commer-
cial scale.66 Into a closed reaction vessel containing the cellulose, a solution
of acetic anhydride, liquid sulfur dioxide, and sulfuric acid catalyst is intro-
duced under suitable pressure. The reaction temperature is conveniently
controlled by pressure variation to permit necessary evaporation of sulfur
dioxide, which is returned to the reaction chamber by reflux condensation.
Excessive heat of reaction is readily absorbed by this type of "internal re-
frigeration." Hydrolysis to acetone solubility47 may be carried out in the
customary way by adding enough water to destroy excess acetic anhydride
and to supply 3-10% water to the reaction mixture, and, if desired, by
supplying additional sulfuric acid to increase the speed of hydrolysis.
(3) Fibrous Acetylation Processes
If a sufficient quantity of a liquid component which is not a solvent for
cellulose acetate is included in an acetylation bath, the reaction product
may be prevented from dissolving, so that a fibrous esterification results.
Such reactions proceed in much the same manner as in the solution methods.
Similar temperatures and catalyst concentrations are used, but usually
larger proportions of liquid to cellulose are necessary because of the bulk
of the fibrous material as compared with equal amounts in viscous solution.
In an early process of this type, Lederer58 employed carbon tetrachloride
as the inert component. Aromatic hydrocarbons are the liquids most
commonly recommended for use, although other inert diluents such as
ethers and aliphatic hydrocarbons, which do not react with acetic anhydride
" L. M. Burghart (to U. S. Industrial Alcohol Co.), U. S. Patent 1,816,664 (July 28,
1931); Chem. Abstracts, 25, 5557 (1931).
" L. M. Burghart (to U. S. Industrial Alcohol Co.), U. S. Patent 1,822,563 (Sept. 8,
1931); Chem. Abstracts, 25, 5990 (1931).
« L. Lederer, U. S. Patent 999,236 (Aug. 1, 1911); Chem. Abstracts, 5,3156(1911);
Brit. Patent 3103 (Feb. 7, 1907); Chem. Abstracts, 1, 2649 (1907).
784 CELLULOSE
or interfere with the action of the catalyst, may be employed. The value
of a nonsolvent depends largely upon its ability to maintain the cellulose
in a swollen and easily penetrable condition throughout the esterification,
while at the same time preventing dissolution. Thus, when aliphatic hy-
drocarbons are used, an auxiliary solvent, such as methylene chloride,
ethylene chloride, or sulfur dioxide, improves the speed of the reaction and
the uniformity of the resulting product.
The fibrous process is of interest only if a fully esterified cellulose acetate
is desired. Due to the difficulties of removing combined sulf uric acid from
a fully esterified product, perchloric acid is the preferred catalyst in this
process.
No method of hydrolysis in suspension has yet proved sufficiently suc-
cessful to achieve commercial use. Fibrous products which are somewhat
less than fully esterified may be prepared by processes of Sindl and Frank59
in which large quantities of sulfuric acid catalyst are employed under condi-
tions which encourage high cellulose sulfate formation. The product is
then treated with esters, such as ethyl acetate, at elevated temperatures to
remove the sulfate, leaving unesterified hydroxyl groups.
The following esterification procedure60 serves as an example of the fibrous
acetylation process:
The esterification is carried out in a perforated stainless steel drum which rotates inside
a stainless steel container. Into this drum is fed 313 kg. (dry weight) of cellulose condi-
tioned to a moisture content not greater than 7%. Then 15 times this weight of acetic
acid containing 14% acetic anhydride is added. The internal drum is meanwhile rotat-
ing at 6 r.p.m. and continues to do so for 2 hrs. The excess liquid is then removed by
rotating the drum at 370 r.p.m.; 750 kg. of the pretreatment liquors are left on the
cellulose.
The acetylation liquor consists of 65% acetic acid, 18.5% benzene, and 16.5% acetic
anhydride.
The catalyst is 70% perchloric acid, and between 1 and 2% (based upon the weight of
the cellulose) is used depending upon the quality of the cellulose. Into the drum
rotating at 6 r.p.m. is run 3200 liters of the acetylation liquor previously cooled to — 18°
to — 25 °C. The temperature is allowed to rise slowly and reaches 27-29 °C. in about 1-
1.5 hrs. The conditions are adjusted so that the acetylation is complete in about 6 hrs.
For low viscosity the maximum temperature is 35° C.; for medium viscosity, 30 °C.;
and for high viscosity, 26°C. Throughout the whole acetylation the acetylating liquors
are kept circulating.
As soon as a sample just dissolves in a mixture of methylene chloride-methanol (9 : 1 by
volume), 600 liters of benzene containing 3% acetic acid is run in. This addition is made
« (X Sindl and G. Frank, U. S. Patent 2,134,332 (Jan. 10, 1939).
M British Intelligence Objectives Subcommittee, London, B.I.O.S. Final Report 1859,
Item 21, Manufacture of Cellulose Triacetate, Yarn and Films (Feb. and March, 1948).
IX. DERIVATIVES OF CELLULOSE 785
about 1 hr. after the commencement of the acetylation and prevents the acetylated
product from swelling and partially dissolving during the later stages.
After the reaction is completed, another 1200 liters of benzene is added and the
catalyst is "killed" by the addition of 3 kg. of potassium carbonate in 30 kg. of acetic
acid. The circulation is continued about 0.5 hr. until the specific gravity of the liquid
becomes constant, indicating that the potassium carbonate is thoroughly distributed
throughout the batch. The excess liquid is then removed by rotating the drum at 370
r.p.m.
The batch is washed with benzene, and the benzene is removed with steam distillation.
The product is then washed with water and dried.
The fibrous process has been used for the esterification of wood and cellu-
lose to study the effects of partial substitution. Bletzinger61 found that
rag stock acetylated in the range of 9-25% acetyl with acetic anhydride
and pyridine gave a poor paper stock due to the increased water resistance
of the fibers. Acetyl values below 6% separate the cellulose molecules and
allow greater ease of hydration. Similar work has been done by Stamm62
who worked with thin layers of wood. Treatment with acetic anhydride-
pyridine vapors gave products of about 21% acetyl content having good
strength with reduced shrinkage and resistance to decay. Cotton fibers
which have been surface acetylated to the extent of about one acetyl
group per glucose unit with acetic anhydride and perchloric acid catalyst
have outstanding properties.68
The strength is about the same as before acetylation, the sensitivity to
moisture is less, the heat stability is greatly improved, and there is almost
complete protection against attack by microorganisms as long as the acetyl
groups are not removed by hydrolysis. X-ray examination shows that the
crystalline portions of the fibers are unchanged cellulose.64
(c) CELLULOSE PROPIONATE
Cellulose propionate may be prepared by esterification with propionic
anhydride in the presence of acid catalysts66 under conditions similar to
those used for cellulose acetate manufacture. The anhydride is somewhat
less reactive than acetic anhydride, thus requiring special consideration for
the conditions of pretreatment and catalyst concentration.
81 J. C. Bletzinger, Ind. Eng. Chem.t 35, 474 (1943).
88 H. Tarkow, A. J. Stamm, and E. C. O. Erickson, U. S. Forest Products Laboratory,
Report 1593 (1946).
61 C. F. Goldthwait, E. M. Buras, and A. S. Cooper, TeacKle Research J., 21, 831
(1951).
84 F. Happey, /. Soc. Dyers Colourists, 66, 14 (1950).
88 C. Dreyfus and G. Schneider (to Celanese Corp. of America), U. S. Patent 1,824,877
(Sept. 29, 1931); Chem. Abstracts, 26, 305 (1932).
786 CELLULOSE
Mild hydrolysis of cellulose tripropionate results in a product soluble in
benzene and in butyl acetate. Further hydrolysis of cellulose propionate
is reported by Fothergill66 to give products of particular interest for plastics
or film use. The hydrolysis is carried out to give a product of about 47%
prppionyl content which is insoluble in benzene but soluble in butyl acetate
and acetone. Still further hydrolysis results in products insoluble in both
benzene and butyl acetate but soluble in acetone, methyl Cellosolve, and
dioxane. A particularly good solvent combination for making films and
plastics from this cellulose propionate is claimed to be a mixture of 62 parts
of acetone, 21 parts of butyl acetate or other nonsolvent component to act
as a residual swelling agent, and 17 parts of ethyl alcohol.
Cellulose tripropionate is considerably softer than either the triacetate
or its hydrolysis product. Cellulose propionate has been manufactured
in limited quantities for use in plastics.
(d) CELLULOSE BUTYRATE
Cellulose butyrate, similarly to the propionate, may be prepared by
esterification with the anhydride and a catalyst, such as sulfuric acid,
provided the reaction conditions are adjusted to permit an efficient pre-
treatment and well-controlled esterification. Esselen and Mork67 in 1922
recommended the use of small quantities of water with butyric acid as a
pretreatment step.
Hydrolysis of cellulose tributyrate has been described by Gault and
Angla68 who carried out the reaction in butyric acid of 76-78% strength at a
temperature of 45-50°C. Increased hydrolysis results in changes of solu-
bility, the least hydrolyzed materials being soluble in benzene but insolu-
ble in methyl alcohol. Further reaction yields products soluble in both
benzene and methyl alcohol and, finally, products soluble in ethyl alcohol
but insoluble in benzene.
Herzog and Frank69 have^ described a process for the preparation of
hydrolyzed cellulose butyrates involving preliminary treatment of cellu-
lose with 87% formic acid at 20°C. This product, which contains a small
w R. E. Fothergill (to E. I. du Pont de Nemours & Co.), U. S. Patent 2,076,556
(Apr. 13, 1937); Chem. Abstracts, 31, 4119 (1937); Brit. Patent 449,182 (June 22, 1936).
w G. J. Esselen and H. S. Mork, U. S. Patent 1,425,580 (Aug. 15, 1922); Chem.
Abstracts, 16, 3393 (1922).
* HE. Gault and B. Angla (to Soci6t6 de$ usines chimiqucs Rh6ne-Poulenc), U. S.
Patent 1,912,189 (May 30, 1933); Chem. Abstracts, 27, 4077 (1933); French Patent
685.637 (July 12, 1930); Chem. Abstracts, 24, 6013 (1930).
WK. O. Herzog and G. Frank, French Patent 700,165 (Aug. 7, 1930); Chem. Ab-
stracts, 25, 3481 (1931); G. Frank and H. Goto, Cettubsechemie, 12, 68 (1931).
DC. DERIVATIVES OF CELLULOSE 787
amount of combined formic acid, is then washed and dried and thereaftei
esterified with a butyric anhydride — butyric acid mixture and zinc chloride
as a catalyst. The reaction product is isolated and then hydrolyzed in
95% butyric acid at 90° C. for 32 hrs. It is soluble in acetone and in mix-
tures of benzene and alcohol.
Cellulose tributyrate melts at a lower temperature than either the ace-
tate or propionate and is considerably softer than those esters. The prod-
uct has not been made in commercial quantities up to the present time.
(e) MIXED ESTERS OF LOWER ALIPHATIC ACIDS
Cellulose mixed esters containing, in addition to acetyl, either propionyl
or butyryl groups offer the opportunity of obtaining products with certain
physical properties improved over those of cellulose acetate, at the same
time being free from the disadvantages of softness, low strength, and the
difficulty of manufacture of cellulose propionate and butyrate.
These mixed esters are commercially manufactured by including the acyl
components in the esterification bath in the form of acids or anhydrides,
Clarke and Malm70 have pointed out that acyl groups from propionic or
butyric acid may be introduced into a cellulose ester without employing
the anhydride if these acids are present in an esterification mixture in which
some other active anhydride is present. Propionic acid may in this way be
incorporated with acetic anhydride or a mixture of acetic anhydride and
acetic acid to produce a uniform product of good quality containing both
acyl groups. It is also possible to esterify with propionic or butyric an-
hydride with acetic acid present in the reaction mixture to produce a mixed
ester. The ratio of acetic to propionic or butyric acid obtained in the cellu-
lose derivative will be proportional to the concentration of the corre-
sponding acyl groups in the esterification mixture, making possible accurate
control in manufacture (Fig. 20).
It has been found that the relationship shown in Figure 20 exists between
the per cent higher acyl of total acyl in the esterification mixture and the
per cent higher acyl in the resulting fully esterified cellulose mixed ester
This relationship varies slightly with extreme variations in ratios of esteri-
fication mixture to cellulose. For economic reasons it is, of course, bettei
to use as much acetic anhydride and as little higher anhydride as possible
however, for the activation of the cellulose prior to esterification, as mud:
acetic and as little higher acid as possible should be employed to obtaii
optimum results. The proper proportions of these chemicals must tx
TO H. T. Clarke and C. J. Malm (to Eastman Kodak Cp.), U. S. Patent 2,048,68!
(July 28, 1936).
788
CELLULOSE
selected in every case and depend upon the composition and viscosity of the
ester to be prepared and the reactivity of the cellulose used.
In the manufacture of cellulose mixed esters, sulfuric acid is the most
practical catalyst; however, with increase in the higher acyl content of the
esterification mixture, the catalytic efficiency of sulfuric acid decreases
because it is less readily sorbed by the cellulose from the esterification mix-
ture.71 Figure 21 shows the difference in sulfuric acid sorption when cotton
linters are kept in a mixture of acetic acid containing 5% acetic anhydride
and 0.5% sulfuric acid and in a mixture of butyric acid containing 5%
acetic anhydride and 0.5% sulfuric acid. The treatment was carried out
60-
Sf 40
c £ 30
o c
!t!
.?! 10
10 20 30 40 50 60 70 80 90 100
Percent propionyl or butyryl of totol ocyl in
esterification bath
Fig. 20. Relationship between composition of esterification bath
and per cent propionyl or butyryl introduced into product.
at 25°C., with 20 parts of liquid for 1 part of cellulose.
Butyric and propionic anhydrides react less readily with cellulose than
acetic anhydride when sulfuric acid is used as catalyst. The lower sul-
furic acid sorption from the higher acids under anhydrous condition and
the lower reactivity of the higher anhydrides result in a slower reaction
and cause more breakdown-of the cellulose before it is protected by acyl
groups.
To overcome these handicaps it has been found advisable to decrease the
liquid-to-cellulose ratio in the esterification mixture, thereby increasing
the anhydride and catalyst concentrations but still keeping their ratio to
the cellulose the same. This speeds up the reaction. The cellulose is
esterified and protected by acyl groups before it becomes too much de-
graded, resulting in products of higher viscosity and improved physical
properties.72 Pretreatment is an important step in the preparation of
cellulose mixed esters of satisfactory quality and should be so adjusted as to
71 C. J. Malm (to Eastman Kodak Co.), U. S. Patent 2,097,954 (Nov. 2, 1937);
Chem. Abstracts, 32, 363 (1938); U. S. Patent 2,173,964 (Sept. 26, 1939).
» L, W. Blanchard, Jr. (to Eastman Kodak Co.), U. S. Patent 2,208,669 (July 23,
1940); Chem. Abstracts, 35, 316 (1941).
IX. DERIVATIVES OF CELLULOSE
789
insure uniform esterification and freedom from haze and undissolved fiber
in the reaction product.
A procedure for preparation of cellulose acetate propionates of high vis-
cosity has been described by Malm.73 In a process for manufacturing
mixed esters containing moderate amounts of butyryl, Billing and Tinsley74
have recommended pretreatment with acetic acid followed by esterifica-
tion with butyric acid and acetic anhydride together with a suitable catalyst.
CHsCOOH+5% (CHaOOfeO
C3H7 COOH + 5 % (CHsCO)eO
j I
8
10
234567
Mou^s
Fig. 21. Sorption of sulfuric acid by cotton linters in
the presence of acetic and butyric acids.
Mixed esters containing very high proportions of propionyl or butyryl
groups require special procedures for their manufacture. Gardner76 has
described a method calling for from 5-20% water, based upon the weight of
the cellulose, in pretreatment mixtures in which acids of three to four
carbon atoms predominate. The presence of this amount of moisture is
very effective in bringing about a uniform esterification reaction.
The hydrolysis of fully esterified mixed esters containing at least 15%
propionyl or butyryl groups has been reported by Malm and Fletcher76 to
78 C. J. Malm (to Eastman Kodak Co.), U. S. Patent 2,026,986 (Jan. 7, 1936); Chem.
Abstracts, 30, 1559 (1936).
7< W. M. Billing and J. S. Tinsley (to Hercules Powder Co.), U. S. Patent 1,973,693
(Sept. 18, 1934); Chem. Abstracts, 28, 7013 (1934).
» H. S. Gardner, Jr. (to Eastman Kodak Co.), U. S. Patent 2,113,301 (Apr. 5, 1938);
Chem. Abstracts, 32, 4335 (1938).
7fl C. J. Malm and C. L. Fletcher (to Eastman Kodak Co.), U. S. Patent 2,026,583
(Jan. 7, 1936); Chem. Abstracts, 30, 1230 (1936).
790 CELLULOSE
bring about changes in solubility and physical properties which make the
products more useful than the unhydrolyzed esters.
(f) HIGHER ALIPHATIC ACID ESTERS
Anhydrides of organic acids containing more than four carbon atoms can-
not readily be made to esterify cellulose by use of acid catalysts, and these
esters therefore require other methods of preparation. Grim and Wittka77
treated cellulose with lauryl and stearyl chlorides in pyridine at "the tem-
perature of a water bath," and obtained partially esterified products which
were not changed in appearance from the original cellulose. Gault and
Ehrmann78 described the preparation of several cellulose higher esters using
benzene as a diluent for the pyridine and acid chloride reaction. Mono-,
di-, and tri-esters of lauric, palmitic, and stearic acids were made. Prod-
ucts from unmodified cellulose were found to be insoluble; modified cellu-
lose, such as that regenerated from viscose or cuprammonium solution,
gave soluble esters. The use of benzene or toluene as a diluent for the
pyridine-acid chloride reaction mixture offers a considerable improvement
in operation, the hydrocarbon acting at elevated temperatures as a solvent
for the intermediate product between the acid chloride and the tertiary
base. A still better diluent for this purpose is chlorobenzene, which has
been used by Hagedorn.79
Kita and others80 have shown that alkali cellulose reacts with higher
acid chlorides to give degrees of substitution varying with alkali concen-
tration. Sakurada and Nakashima,81 by repeated treatment with alkali
and stearyl chloride, obtained a degree of substitution of 2.1.
Clarke and Malm8 used chloroacetic anhydride as an impelling agent to
bring about the esterification of cellulose with higher acids. A series of
esters from the acetate through the stearate was prepared and the proper-
ties studied.
Malm and coworkers82 used the acid chloride-pyridine method to pre-
77 A. Grim and F. Wittka, Z. angew. Chem., 34, 645 (1921).
78 H. Gault and P. Ehrmann, Compt. rend., 177, 124 (1923); Chimie & Industrie,
Special No. 574 (May, 1924); Bull. soc. chim.t [4], 39, 873 (1926).
79 M. Hagedorn and O. Reichert (to I. G. Farbenindustrie Akt.-Ges.), U. S. Patent
1,832,381 (Nov. 17, 1931); Chem. Abstracts, 26, 1438 (1932); M. Hagedorn and G.
Hingst (to I. G. Farbenindustrie Akt.-Ges.), U. S. Patent 1,940,589 (Dec. 19, 1933);
Chem. Abstracts, 28, 1532 (1934).
80 G. Kita, I. Sakurada, and T. Nakashima, Cellulose Ind. (Tokyo), 2, 30 (1926);
J. Leibowitz, Cettulosechemie, 9, 125 (1928).
81 1. Sakurada and T. Nakashima, Sci. Papers Inst. Phys. Chem. Research (Tokyo), 6,
197 (1927).
M C. J. Malm, J. W. Mench, D. L. Kendall, and G. D. Hiatt, Ind. Eng. Chem., 43, 684
(1951).
IX. DERIVATIVES OF CELLULOSE
791
SB*
Be
5 Si.
<|i
.3 .So
000
CO Ci
CO i-< f *H O O O O O
J5 2 2 fe 8 § 8
O O
1C 00 O CO
»O CO C^l
(M
ioooooooo
00 Oi
^ <N I 00 CO
c c
I co co co co co c^ co co co co co
A A A A A
§ S 2 S S3 §3 So S So S
<M rH rH
c
r,' 0)
c
8
S
I "8
S'"
TO O
s
I*
Si
1!
792
CELLULOSE
pare a series of fully esterified esters, essentially free from contaminating
groups and with each member of the series having as nearly as possible
the same degree of polymerization. A cellulose regenerated from a com-
mercial cellulose acetate served as the starting material. Reaction con-
ditions were studied to find the best ratio of acid chloride to pyridine and
300
280
260
240
220
o
V 200
*o
Q.
£ '80
S 160
140
120
100
80
0123456789 10 12 14
Number of carbon atoms in esterifying acid
16
Fig. 22. Melting points of triesters (Malm, Mench, Kendall, and Hiatt82).
the preferred time, temperature, and reaction diluent. Esters were put
through the reaction a second time and checked by analysis for complete
esterification. Portions of the finished esters were de-esterified and the
recovered celluloses found to have similar intrinsic viscosities in cupram-
monium solution. A summary of the properties of these esters is given in
Table 9.
Melting points were found to fall off sharply, then rise toward the end of
the series, as shown in Figure 22.
Water tolerance values, a photometric1 measure of the amount of water
required to produce a given level of haze in a dilute solution of the ester in
IX. DERIVATIVES OF CELLULOSE
793
0123456789 10 12 14
Number of carbon otoms in esterifying acid
16
Fig. 23. Per cent moisture sorption of triesters at
25°C. (Malm, Mench, Kendall, and Hiatt82). See text p. 796.
0123456789 10 12 14
Number of carbon atoms in esterifying acid
Fig. 24. Tensile strengths of triesters (Malm, Mench, Kendall, and Hiatt11),
See text p. 796,
794 CELLULOSE
TABLE 10
Solubilities of the Cellulose Triesters of the w-Fatty Acids
(Malm, Mench, Kendall, and Hiatt")
Solvent-solid ratio 9:1 by weight. See text p. 796.
Celluloee Trtetter
24
SOLVENT
C Atoms
in Acid
Alcohols
Methanol-
Ethanoa—
1 -Propano
n-Butanol.
F-Butaaol.
I-Ethylhexanol
Diecetone Alcohol.
Tetrahydrofurfuryl Alcohol -
Ether-Alcohole
9 t -Methoxyethanoi _
10 t-Ethoxyethanol —
1 1 9 -Butoxyethanol —
12 Ethyl Carbitol
13 Butyl Carbitol
Ethers and Acetals
14 Ethyl Eth«r.
15 Uopropyl Eth«r_
16 Diothyl C«llo»olv..
17
18 T«tr»hydrofur»n-.
1° EthyUn* Formal .
K«to
2! KUthyl Ethyl K«ton«
22 Methyl Uobutyl K«ton«.
23 Cy el ohexmnone __
•ophoron*
Eaters
25 Methyl Formate
26 Methyl Acetate
27 Ethyl Acetate ___
28 Uopropyl Aeetate__
29 n-Butyl Acetate.
30 n-Amyl Acetete__
31 F-Methoxyethyl Acetate.
32 »-Cthoxyethyl Acetate
33 Ethyl Laetate.
34 t-Hydroxyethyl Acetate.
Haloteoated Compoundt
hlortde.
35 Methylene Chi
36 Ethylene Chloride —
37 Propylene Chloride.
38 Chloroform,
39 Carbon Tetrachlo:
40 TrichloroethyUne
41 ••TetrachloroetheJie.
• a-A»yi Chloride—
45 Ethylen«ChlorohydHn_
46 |,K-OUWorPethxl Btbf r.
10
16
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
13
34
3$
36
37
38
39
40
41
42
43
44
45
46
IX. DERIVATIVES OF CELLULOSE
TABLE 10 (Continued)
795
Cellulose Trieste r
«
«
£
Bttty rate
•I
1
*3
^
Caproate
«
"x
CapryUte
|
«
2
9
5
I
I
C Atoms
in Acid
10
12
14
16
Nitrogen Compound*
40 B-^^^M.
Acids
Hydrocarbon*
Mixtures
M A *..**«.. W.t.r (4.1)
68 • ' (1:4)
71 * (1:4)
76 " (1:4)
79 ' {»<•)
^SS/A
• i A^M* A*M.W»*.»(A.I)
87 - " (1:4)
•8 |. ) -Dichloroethyl Ether :Meth*nol — 1
(4:1)
•
m
•
•
•
U
14
16
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
Grain-free solution et 25°C. Q
Grainy or swollen at 25°C.; train -fret at 100 C. or at B.P. of solvent.
Insoluble at 25°C. and at 10CTC. or at B.P. of solvent.
796 CELLULOSE
acetone, became extremely low for the high esters. Likewise, the moisture
sorption values at different relative humidities fell rapidly and were well
below 1% for most of the esters (see Fig. 23, p. 793).
Tensile strength measurements on films cast from chloroform solutions
fell regularly and then levelled off roughly between Cj and Ci«, as shown in
Figure 24 (p. 793).
Solubilities were determined in 88 different solvents and mixtures shown
in Table 10. The most widely soluble esters were the valerate and capro-
ate. The chlorinated solvents methylene chloride, chloroform, and s-
tetrachloroethane were the only compounds tested which were solvents
throughout the entire series.
Bartell and Ray*2a»82b measured the contact angles of water and various
organic solvents on several simple and mixed cellulose esters. From these
measurements an order of hydrophobicity was established for esters with
various substituent groups.
Cellulose mixed higher esters, particularly those containing a substantial
amount of acetyl and a higher molecular weight acid, have very interesting
properties. These mixed esters may be readily prepared, for example, from
acetone-soluble cellulose acetate by esterification of the free hydroxyl
groups with the desired high molecular weight acid. This may be accom-
plished either by heating a pyridine solution of the cellulose acetate with a
large excess of the higher acid chloride88 or by the use of chloroacetic an-
hydride and the higher acid. The resulting products show a wide range of
solubility in ketones, esters, and mixtures of toluene with alcohol. They
are highly compatible with a large variety of resins, fats, and oils. The
products are highly resistant to sorption of moisture.
Cellulose esters of naphthenic acids have been prepared by Kita, Mazume,
Sakrada, and Nakashima84 by a variety of reaction processes. Anhydrides
gave only low degrees of esterification; acid chlorides yielded products
between the di- and tri-esters. Complete esterification was not obtained.
Patents have described the action of naphthenic acid chlorides upon cellu-
lose in the presence of picoline and chlorobenzene at a temperature of
135°C.86 Mixed esters have also been prepared by treatment of alkali
cellulose88 with mixtures of aliphatic and naphthenic acid chlorides or by
"* F. E. Bartell and B. R. Ray, /. Am. Chem. Soc., 74, 778 (1952).
Mb B. R. Ray and F. E. Bartell, /. Phys. Chem., 57, 49 (1953).
M H. Gault and P. Ehrmann, Caoutchouc 6f gutta-percha, 24, 13748, 13824 (1927).
M G. Kita, T. M&zume, J. Sakrada, and T. Nakashima, Kunststoffe, 16, 167 (1926).
M I. G. Farbcnindustrie Akt.-Ges., Brit. Patent 305,947 (June 11, 1930).
M M. Hagedorn (to I. G. Farbenindustrie Akt.-Ges.), U. S. Patent 1,994,608 (Mar. 19,
1035); Chem. Abstracts, 29, 3157 (1935).
IX. DERIVATIVES OP CELLULOSE 797
esterification of cellulose with these chlorides in the presence of a tertiary
base.87
(g) USES OF ALIPHATIC FATTY ACID ESTERS
Since the physical properties of cellulose esters determine their end uses,
an understanding of property trends is desirable. The commercially
available cellulose esters include cellulose acetates of several acetyl levels
and a few acetate butyrate esters covering a wide composition range.
Propionate and acetate propionate esters will not be considered, since their
properties lie between those of the acetates and the acetate butyrates, and
accordingly can be duplicated by proper choice from these groups.
An attempt has been made in Figure 25 to summarize the property
trends in the acetates and acetate butyrates between the tri- and di-ester
composition lines. Commercial acetates are available in the 38-44%
(a-6) acetyl range. The small letters c, d, e, and / locate commercially
available mixed esters containing acetyl and with butyryl levels at 16, 27,
38, and 48%, respectively. Tensile strength and stiffness are highest with
the simple acetates; flexibility improves with introduction of the higher
acyl group. Moisture sorption is lowest with high degrees of butyryl
substitution and with fully esterified products, the uptake increasing with
relative humidity and the hydroxyl content of the ester, as shown in Figure
26aand26b.88
The simple acetates are higher melting than the mixed esters. With any
given triester the melting point drops, then rises again with continued
hydrolysis (see Fig. 27).
Hydrolysis of the simple and mixed triesters improves solubility in polar
solvents or mixtures containing polar solvents and reduces solubility in,
and tolerance for, nonpolar solvents. The behavior of polar solvents is
shown by j8-methoxyethyl alcohol and by a 1:1 mixture of toluene :meth-
anol, while nonpolar solvents are represented by propylene chloride and
methyl isobutyl ketone. Increase in the butyryl content of an ester im-
proves solubility (Figs. 28a, 28b, 28c, and 28d).
The same considerations affecting solvent solubility hold true for plasti-
cizer solubility but to a lesser degree. This is shown in Figure 29a, 29b,
and 29c by improved solubility in ethyl phthalate and tripropionin as the
ester is hydrolyzed. Solubilities in octyl phthalate show the influence of
high butyryl content.
87 M. Hagedorn and P. M6Uer (to I. G. Farbenindustrie Akt.-Ocs.), U. S. Patent
1,975,897 (Oct. 9, 1934); Chem. Abstracts, 28, 7628 (1934).
88 C. J. Malm, C. R. Fordyce, and H. A. Tanner, Ind. Eng. Chem,, 34, 430 (1942).
798
Fig. 25. Effects of composition on physical properties. Apices: A — acetyl; B «
butyryl; C «• cellulose. 1, increased tensile strength, stiffness; 2, decreased moisture
sorption; 3, increased melting point; 4, increased plasticizer compatibility; 5, increased
solubilities in polar solvents; 6, increased solubilities in nonpolar solvents; 7, increased
flexibility; 8, decreased density.
Fig. 26a. Per cent sorption of moisture by cellulose esters of acetic acid at 25% relative
humidity (Malm, Fordyce, and Tanner18).
DC. DERIVATIVES OF CELLULOSE
C
799
Fig. 26b. Per cent sorption of moisture by cellulose esters of acetic acid at 90% relative
humidity (Malm, Fordyce, and Tanner88).
Acetic -Butyric Esters
Fig. 27. Relation of composition of cellulose esters of acetic and butyric acids to melting
point in °C. (Malm, Fordyce, and Tanner81).
800
/3-Methoxyethyl Alcohol
Fig. 28a. Relation of solubility in 0-methoxyethyl alcohol to degree of hydrolysis
(Malm, Fordyce, and Tanner88).
Toluene: Methanol (|:|)
Fig. 28b. Relation of solubility in toluene : methanol (1:1) to degree of hydrolysis
(Malm, Fordyce, and Tanner88).
Cellulose ester density decreases as the amount of butyryl is increased in
the mixed esters as shown in Figure 30.
The foregoing general property trends help to explain the uses and limita-
tions in applications of the cellulose esters.
Cellulose triacetate, the first of the acetates made, has only recently
become of commercial importance. Its limited solvent solubility has
IX. DERIVATIVES OF CELLULOSE
C
801
Propylene Chloride
Fig. 28c. Relation of solubility in propylene chloride to degree of hydrolysis (Malm,
Fordyce, and Tanner88).
Fig. 28d.
50
Methyl Isobutyl Ketone
Relation of solubility in methyl isobutyl ketone to degree of hydrolysis
(Malm, Fordyce, and Tanner88).
restricted its use; most emphasis has been placed on the hydrolyzed ace-
tates in the acetone-soluble range. Because it has a higher melting point
and better moisture resistance than hydrolyzed acetates, "triacetate" has
found a place in the photographic industry in replacing cellulose nitrate as
a Cin6 film base. (The "triacetate" in use is 1-2% lower in acetyl content
than the 44.8% acetyl of the triacetate but is still considerably removed
802
Ethyl Phtholote
Fig. 29a. Relation of cellulose ester composition to solubility in ethyl phthalate at 25 c
and 180 °C. (Malm. Fordyce, and Tanner18).
Tripropiorvin
fig. 29b. Relation of cellulose ester composition to solubility in tripropionin at 25°
and 180 °C. (Malm, Fordyce, and Tanner18)
IX. DERIVATIVES OP CELLULOSE
C
803
Octyl Phtholote
Fig. 29c. Relation of cellulose ester composition to solubility in octyl phthalate at 25°
and 180 °C. (Malm, Fordyce, and Tanner88).
Acetic- Butyric Esters
Fig. 30. Relation of cellulose ester composition and density (Malm. Fordyce, and
Tanner8*).
804 CELLULOSE
from the acetone-soluble range.) Sheeting compositions with higher
amounts of plasticizer are being used for drawing and forming of plastic
shapes. The triacetate has also been recommended as an electrical
insulating material89 because of its resistance to extended heating. Methyl-
ene chloride is still the commonest coating solvent, which means that sol-
vent recovery and good ventilation are necessary economic and safety
requirements.
Hydrolysis of cellulose triacetate produces esters which are acetone-
soluble in the 36-42% acetyl range, have lower melting points, and have
greater susceptibility to moisture and to textile dyes. Acetone has been
the standard solvent for most uses, but several other readily available
compounds, such as methyl acetate, ethyl lactate, diacetone alcohol,
ethylene glycol monomethyl ether, and mixtures of ethylene dichloride with
alcohols, are also used, as well as appreciable quantities of diluents, such
as ethyl acetate, ethyl methyl ketone, and methyl, ethyl, and butyl alcohols.
Cellulose acetate yarn, manufactured by dry spinning from acetone solu-
tion, has been largely responsible for the great growth of acetate manu-
facture. Concentrated acetone solutions are forced through spinnerets to
form filaments which are cured by circulated air, then gathered and passed
over guide rolls to apply controlled twisting and stretching. The earliest
descriptions of dry spinning of cellulose acetate were given in patents of
Bayer and Co.,90 although before that time wet-spinning processes, in-
volving spinning of acetic acid or chloroform solutions into precipitating
liquids, were described by Wagner91 and by Little, Walker, and Mork.91
The first large-scale manufacture of cellulose acetate rayon in the United
States was in 1926. Rapid expansion continued from that time, resulting
in an increase of from 2,620,000 pounds of acetate rayon manufactured in
1926 to nearly 10,000,000 pounds manufactured in 1930. The 1950 ace-
tate yarn production was 443,000,000 pounds (326,000,000 pounds con-
tinuous filament and 117,000,000 pounds staple fiber).
New products have appearecl on the market which are made by the in-
corporation of the coloring agent in the spinning solution. This coloring
method allows a wider choice of light-stable dyes and yields products which
are colored completely through the fiber.
Photographic film base made from acetone-soluble cellulose acetate,
M W. Schrfider, Kunststoffe, 32, 82 (1942); Bull Inst. Paper Chem., 13, 185 (1942-3).
10 Bayer and Co., Brit. Patent 28,733 (Nov. 2, 1905); French Patent 350,422 (Jan. 4,
1906).
"A. Wagner, German Patents 137,255 (June 5, 1901); 152,452 (May 4, 1901);
A. D. Little, W. H. Walker, and H. S. Mork, U. S. Patents 712,000 (Oct. 28, 1902);
792,149 (June 13, 1905).
IX. DERIVATIVES OF CELLULOSE 805
which replaced the hazardous cellulose nitrate for x-ray, portrait, and ama-
teur Cin£, is now being replaced by the triacetate or by mixed esters.
Acetone-soluble acetate is also used for the production of sheeting of high
clarity and uniformity. This product is manufactured by flowing care-
fully filtered viscous solutions of the acetate in acetone onto a moving pol-
ished surface, and, after evaporation of solvent, continuously removing the
film. Thicknesses down to 0.0002-0.0005 inch may be made. Films of
greater thickness are made by solvent-laminating thinner gages, or by
extrusion of pastes or plasticized pellets. Minimum amounts of solvent
are used in paste extrusion of sheeting; the solvent must then be well
removed to secure a dimensionally stable product. High-temperature
extrusion is now widely used where the plasticizer serves as the solvent at
temperatures of 250-450°F. Sheets, tubes, rods, and articles of various
cross sections are made by this process. Plasticizers, such as methyl,
ethyl, or methoxyethyl phthalates, which are solvents for the acetate at
high temperatures, must be used. Sheet stock below 0.001 inch may be
used as condenser dielectrics, thicker material (0.001-0.003 inch) for deco-
rative wrapping and protective lamination, and still heavier gages (0.005-
0.020 inch) for fabricating and hot-drawing of containers. Gages up to
0.125 inch are prepared by extrusion or skiving and may be used in forming
processes, such as vacuum drawing.
Cellulose acetate compositions are commonly used in the thermoplastic
molding field. Powdered ester is intimately mixed with plasticizer and
pigment, and this mixture is fluxed on hot rolls. The cooled composition
is granulated to furnish molding pellets which are generally handled by
injection or extrusion molding. Since the acetates are high melting, fairly
large amounts of plasticizer are required to produce compositions of reason-
able flows. The higher acetyl esters of the acetone-soluble range give
compositions of increased heat and moisture resistance. Molding of the
triacetate has not been reduced to commercial practice.
Only minor quantities of cellulose acetate have been used in lacquers,
mainly because of the limited solubility of the cellulose ester in suitable
solvents and its incompatibility with most resins in quantities necessary
to produce good adhesion and surface hardness.
Except in the textile field, mixed esters such as the acetate butyrates are
more generally useful than the simple acetates. Against this fact must
be balanced their higher cost. For photographic films, advantages are to be
gained by the use of mixed esters containing sufficient quantities of higher
acids to prevent undesirable curl with changes in humidity, at the same time
keeping the higher acyl content within the range having satisfactory
806
CELLULOSE
rigidity and tensile strength. The wider solubility of the mixed esters al-
lows a balancing of solvents and nonsolvents in the coating composition so
that film of excellent physical properties can be coated at high speeds.92
For manufacture of plastics, behavior toward plasticizers is a most im-
portant quality. Cellulose acetate requires comparatively high concen-
trations of active plasticizers for injection-molding operation, and those
480
440
400
360
320
"I280
<S 240
i
>»
| 200
1 ,60
120
80
40
• AB-272-3
• AB-36I-I
D AB-500-1
10 20 30 40 50 60 70
% Alcohol in solvent mixture
80 90
Fig. 31. Viscosity versus solvent composition, toluene-ethyl alcohol mixtures,
10% solids at 25 °C. (Malm and Smith97). The first two numbers in the three-
digit number following AB (acetate butyrate) refer to per cent butyryl. See text
p. 808.
plasticizers which produce satisfactory flow characteristics are not retained
as well as desired.91 Here the mixed esters, particularly cellulose acetate
butyrates containing a substantial amount of butyryl, offer distinct im-
provements; they possess greater compatibilities with plasticizers which
show better retention characteristics, and they exhibit suitable flow with very
M C. R. Fordyce and W. F. Hunter, Jr. (to Eastman Kodak Co.), U. S. Patents
2,319,051, 2,310,052, and 2,319,055 (May 11, 1943); Chem. Abstracts, 37, 6129 (1943).
w C. R. Fordyce and L. W. A. Meyer, Ind. Eng. Chem., 32, 1053 (1940).
DC. DERIVATIVES OF CELLULOSE
807
small plasticizer concentrations. A cellulose acetate molding composition
containing 49 parts of diethyl phthalate may thus be duplicated in flow
characteristics by a cellulose acetate butyrate of 13% acetyl and 37%
butyryl content with 15 parts of diamyl phthalate, the latter composition
being much more permanent and dependable.
Compositions of high-butyryl cellulose acetate butyrate have been han-
dled as molten compositions in coating paper and cloth.94 This application
2,000,000
1,000,000
800,000
600,000
400,000
300,000
200,000
UJ
5 100,000
80,000
60,000
40,000
30,000
20,000
OL
p
z
UJ
o
I
O
o
10,000
8,000
6,000
4,000
3,000
2,000
1,000
30
50
70 90 110
TEMPERATURE -°F
130 150
Fig. 32. Viscosity versus temperature for various lacquers (Malm and Smith98).
1, typical gel lacquer prepared from 38% butyryl ester; 2, nongelling lacquer prepared
from 38% butyryl ester; 3, typical nitrocellulose solution in active solvents. See text
p. 809.
requires esters of a high degree of heat stability compounded with modify-
ing agents which are solvents at high temperature but which yield non-
tacky, compatible surfaces on cooling. Somewhat similar molten com-
positions have been used to give metal parts a melt dip coating for protec-
" C. J. Malm, M. Salo, and H. F. Vivian, Ind. Eng. Chem., 39, 168 (1947).
808
CELLULOSE
tion against corrosion and abrasion.95 Malm, Kaul, and Hiatt96 studied
the melt- viscosity curves of high-butyryl ester compositions and prepared
quick-setting formulations which were used in melt-casting plastic articles
without the use of pressure or volatile solvent.
The wider choice of volatile solvents for lacquers and film-forming solu-
tions offered by mixed esters is, for many uses, an important factor.
Esters, higher ketones, and mixtures of toluene with alcohols are among
the common solvents used.
TABLE 11
Trends in Compatibility of Synthetic Resins with Cellulose Acetate Butyrate
(Malm and Smith97)
AB « acetate butyrate. The first two numbers in the three-digit numbers following AB
refer to per cent butyryl.
Type of Resin
AB-161-2
AB-272-3
AB-381-1
AB-500-1
Alkyds
Phenolics
Ureas
Maleics
Polyesters]
Aryl sulfonamide-formaldehydel
condensates j
Modified hydrocarbons j
Chlorinated biphenyls
Acrylates
Rosin derivatives
Polyvinyl acetates
Poly vinyl chlorides 1
Polyvinyl chloride-acetates }
Polyvinyl acetals J
Unmodified hydrocarbons 1
Melamines \
Furfurals J
Alkyd resin plasticizers
Compatibility depends on modifiers used in resins.
Many resins of these types designed for lacquer use
are compatible, particularly with type AB-381-1
ester.
-Compatibility very good-
Good »
Fair Good
Fair »
Fair Very good
< Very good
Very good Fair
« Generally good —
Good Poor
Generally not compatible •
Generally not compatible
Fair
Generally good
As the butyryl content is increased, the solubility in inexpensive lacquer
solvents increases as does the tolerance for common diluents.97 Figure 31
shows the viscosity changes as ethyl alcohol is added to three acetate
butyrates in toluene. The higher butyryl esters give low-viscosity solu-
tions which tolerate considerable alcohol without appreciable viscosity
« C. J. Malm, H. B. Nelson, and G. D. Hiatt, Ind. Eng. Chem., 41, 1065 (1949).
« C. J. Malm, O. W. Kaul, and G. D. Hiatt, Ind. Eng. Chem., 43, 1094 (1951).
w C. J. Malm and H. L. Smith, Jr., Ind. Eng. Chem., 41, 2325 (1949).
IX. DERIVATIVES OF CELLULOSE 809
increase. An acetate butyrate ester in the viscosijy range of half-second
nitrate has recently been put on the market for coating formulations.
The 38% butyryl ester has been worked into compositions which gel as
the temperature is lowered ("gel lacquers' ')> which allow the application
of heavy plastic coatings by a single dip into a warm lacquer.98 Figure 32
shows the change in viscosity with temperature for (1) a gel-inducing sol-
vent mixture, (2) a nongelling composition, and (3) a standard nitro-
cellulose formulation.
Reinhart and Kline" have found that a small amount of hydrolysis is
desirable in cellulose mixed esters used for aircraft fabrics because of the
more desirable effect of solvent mixtures on the physical properties of the
products. Completely esterified esters became brittle upon exposure, while
those with 0.2 to 0.4 free hydroxyl group for each glucose unit withstood
exposure tests exceptionally well. Higher free hydroxyl group contents
gave greater tautness fluctuation, causing coated fabrics to become slack in
the rain.
Compatibility with resins is related to the amount of combined butyryl
of the cellulose ester, the degree of hydrolysis, and the polarity and solu-
bility of the resin class. Trends in compatibility are shown in Table II07
where the higher butyryl products show better properties except with the
somewhat polar poly vinyl acetate.
Cellulose esters of higher aliphatic, substituted aliphatic, and aromatic
acids, and mixed cellulose esters of these acids with lower aliphatic acids
have received considerable attention, especially in patent literature.
Products with very interesting properties have been described, but for
economic reasons they have found no broad industrial application.
2. Other Aliphatic Esters
(a) UNSATURATED ESTERS
Preparation of cellulose esters of unsaturated aliphatic acids requires in
general the same processes used for the corresponding saturated acids.
Certain exceptions exist, however, in cases in which the unsaturated link-
age interferes with normal behavior of the esterifying acid. Acrylic and
methacrylic acids are difficult to employ as esterifying agents because of the
ease with which these materials polymerize. Maxwell100 has reported the
98 C. J. Malm and H. L. Smith, Jr., Ind. Eng. Chem.t 38, 937 (1946).
99 F. Reinhart and G. M. Kline, Ind. Eng. Chem., 32, 185 (1940).
100 R. W. Maxwell (to E. I. du Pont de Nemours & Co.), U. S. Patent 2,175,357 (Oct.
10, 1939); Chem. Abstracts, 34, 886 (1940).
810 CELLULOSE
preparation of mixed esters by reaction of methacrylic anhydride with
cellulose acetate in the presence of sodium acetate at least equivalent in
quantity to the methacrylic acid formed by the reaction.
Cellulose acetate methacrylate esters have been described100* in a prepa-
ration involving the treatment of an activated cellulose with methacrylic
anhydride and sulfuric acid followed by an acetylation step. These mixed
esters gel in solution when heated with benzoyl peroxide and form infusible
copolymers when heated with acrylic monomers.
Crotonic acid tends to polymerize less readily and so can be used in a
normal fashion to esterify cellulose. The triester may be prepared by treat-
ing purified cotton linters with acetic acid at room temperature, removing
the activating acid with methylene chloride, and esterifying with crotonic
anhydride in the presence of a sulfuric acid catalyst. The ester recovered
by precipitation and washing in water is somewhat unstable to aging.
Products held at room temperature for three months show loss in solubility.
The addition of hydroquinone to the last wash prior to drying retards
polymerization or cross-linking.
Mixed acetate crotonates are readily prepared either indirectly by treat-
ing a hydrolyzed cellulose acetate with crotonic anhydride and a catalyst
or directly by treating cellulose with a mixed acid-anhydride bath. Cro-
tonic acid is less reactive than butyric acid and in competition with acetic
acid combines in smaller proportions than does butyric acid. Mixed
acid-anhydride baths which in the acetate butyrate series would give
25, 38, and 50% butyryl contents give, respectively, 15-18, 29-35, and
42-47% crotonyl contents. Esters in the range of 40-50% crotonyl are
quite resistant to acid hydrolysis. Acetate crotonates with 15-30% cro-
tonyl are fairly readily hydrolyzed, but the change is largely at the ex-
pense of acetyl groups. In a typical case, as the weight per cent of acetyl
changed from 29% to 19% (2 to 1.2 groups per glucose unit), the crotonyl
changed from 18% to 21%rcorresponding to 0.8 group per glucose unit in
both cases.
The solubilities of the simple and mixed crotonate esters are similar to
those of the corresponding butyrate esters. The compatibility with plasti-
cizers is comparable, although the crotonate esters are less soluble in plasti-
cizers than similar butyrate esters. Acetone solutions show little change in
viscosity on standing even when benzoyl peroxide has been added. Films
cast from these solutions are flexible but become brittle as the ester con-
"* A. A. Berlin and T. A. Makarova, Zhur. Obshchei Khim. (J. Gen. Chem.), 21, 1267
(1951).
IX. DERIVATIVES OP CELLULOSE 811
verts to an insoluble form. Heat or ultraviolet light101 increases the con-
version to the insoluble, cross-linked state. Esters high in crotonyl be-
come insoluble after 24 hrs. at 100°C. Small quantities of benzoyl per-
oxide speed up the conversion. As the crotonyl content in mixed esters is
lowered insolubilization by heat or light becomes slower.
The vulcanizing of mixed esters containing methacrylyl and crotonyl
groups has been described102; temperatures were in the range of 135-
185°C., and sulfur and organic rubber accelerators were used. Textile
fibers spun from these compositions and heated showed an improvement
over acetate rayon in elastic recovery and resistance to dry-cleaning sol-
vents.
Cellulose sorbates and acetate sorbates are prepared by the impeller
method. Mixed esters containing combined sorboyl may also be made by
including sorbic acid in ordinary esterification reactions. These dried esters
lose solubility rapidly but are easily handled after treatment with hydro-
quinone before drying.
Cellulose undecylenate has been prepared by Gault and Urban108 by
treating hydrocellulose with the acid chloride and pyridine in the presence
of toluene at 110-120°C. The impeller method (use of chloroacetic
anhydride)104 has also been used to prepare the simple ester; mixed acetate
undecylenate esters may also be made from hydrolyzed cellulose acetate
by this method. The double bonds of these esters are extremely reactive
so that gelling of the reaction mixture is not unusual.
Higher unsaturated acids, such as oleic acid or mixed acids from vege-
table oil saponification, may be used in the impeller or acid chloride re-
action. Oleic acid esters are readily prepared, especially acetate oleate
esters from acetone-soluble cellulose acetate as starting material. Al-
though the finished esters are but slowly affected by heat or sunlight, the
reaction mixture itself is subject to gelling. To secure esters with hydrol-
carbon solubility, stearic acid may be used as an ingredient in the unsatu-
rated acid mixture to reduce the gelling tendency during ester preparation.
A mixture of stearic acid and 20-50% of linseed oil acids reacted with a
hydrolyzed cellulose acetate yields an ester showing excellent solvent
101 C. J. Malm and C. R. Fordyce (to Eastman Kodak Co.), U. S. Patent 1,973,493
(Sept. 11, 1934); Chem. Abstracts. 28, 7013 (1934).
MJ M. L. Ernsberger and A. S. Gregory (to E. I. du Pont de Nemours & Co.), U. S.
Patent 2,396,165 (Mar. 5, 1946); Chem. Abstracts, 40, 2984 (1946).
10» H. Gault and M. Urban, Compt. rend., 179, 333 (1924).
M« H. T. Clarke and C. J. Malm (to Eastman Kodak Co.), U. S. Patent 1,690,620
(Nov. 6, 1928); Chem. Abstracts, 23, 512 (1929); U. S. Patent 1,739,210 (Dec. 10, 1929);
Chem. Abstracts, 24, 962 (1930).
812 CELLULOSE
solubility and high sensitivity to heat or ultraviolet light. The use of
carbon dioxide or nitrogen over the esterification reaction is an added pre-
caution in preventing gelling.106
Certain addition reactions are possible with the double bonds in these
unsaturated cellulose esters. Chlorine and bromine will add to the cro-
tonates dissolved in chloroform. About one-half of the theoretical amount
adds in reactions in which excess of halogen is used.
Sulfur-containing derivatives have been made by heating a suspension
of cellulose crotonate and sodium bisulfite in water or aqueous dioxane.
The addition of sufficient groups leads to water-soluble derivatives.106
(b) HALOGEN-SUBSTITUTED ESTERS
The simple halogen-substituted cellulose esters are prepared with some
difficulty. Cellulose treated with chloroacetic acid and anhydride in the
presence of a zinc chloride catalyst slowly dissolves to give an almost
completely substituted ester which is, however, degraded and of high color.
Barnett,107 using chloroacetyl chloride and a tertiary base, recovered simi-
larly degraded derivatives. The di- and tri-chloroacetic acids are also re-
luctant to esterify cellulose. This characteristic behavior leads to the use
of a-halogenated acid anhydrides as impelling agents in bringing about
esterification of other carboxylic acids. Halogenated acids of more than
five carbon atoms are more reactive with cellulose than the lower mem-
bers, giving esters in the normal manner. Thus, a-halogen stearic acids,
when reacted in the presence of chloroacetic anhydride,108 readily esterify
cellulose or its derivatives containing free hydroxyl groups.
The introduction of chlorine into cellulose acetate is much easier. Phos-
phorus pentachloride acts by substituting chlorine for some of the hydro-
gens of the acetyl groups.109 Chloroacetyl may be added to acetone-
soluble cellulose acetate by (a) direct reaction with chloroacetic anhydride
and catalyst with or without an added solvent, (b) heating with chloro-
acetic acid, or (c) hydrolyzing in the presence of chloroacetic acid. The
first reaction leads to products of minimum degradation and a high degree
106 C. J. Malm and G. D. Hiatt (to Eastman Kodak Co.), U. S. Patent 2,241,226
(May 6, 1941); Chem. Abstracts, 35, 6315 (1941).
** H. Dreyfus (to Celanese Corp. of America), U. S. Patent 2,321,069 (June 8, 1943);
Chem. Abstracts, 37, 6893 (1943).
«» W. L. Barnett, /. Soc. Chem. Ind., 40, 253T (1921).
«» H. T. Clarke and C. J. Malm (to Eastman Kodak Co.), U. S. Patent 1,698,049 (Jan.
8, 1929); Chem. Abstracts, 23, 1267 (1929).
*•!. G. Farbenindustrie Akt.-Ges., Brit. Patent 306,132 (Feb. 17, 1928); Chem.
Abstracts, 23, 5040 (1929).
IX. DERIVATIVES OF CELLULOSE 813
of substitution; the second is a degrading reaction and not efficient in intro-
ducing chloroacetyl; the third gives products low in both acetyl and
chloroacetyl. Almost any desired cellulose acetate chloroacetate may be
prepared by proper choice of the acetate starting material.110 The re-
sulting products are stable to heating at 180°C. for several hours. Their
solubilities and compatibilities are comparable with those of acetate
butyrates of similar composition. Since the chloroacetyl group is more
resistant to hydrolysis than is the acetyl group, hydrolyzing the mixed ester
in acid solution yields products relatively higher in chloroacetyl. A ma-
terial analyzing 1.7 acetyl and 1.3 chloroacetyl groups per glucose unit was
hydrolyzed to a product containing 0.4 acetyl and 0.7 chloroacetyl.110
The acetate chloroacetates react readily in suspension or in solution with
amines, such as pyridine, to form water-soluble quaternary salts. The
reaction requires anhydrous conditions, and with large excesses of pyridine
is essentially quantitative. This reaction on derivatives with a low amount
of chloroacetyl has been claimed to improve acid-dyeing properties.111
Izard and Morgan112 described the preparation of an addition salt by
treating acetate chloroacetates with sodium thiosulfate. Mild oxidation
led to splitting out of sodium hydrogen sulfate with the formation of a
disulfide linkage between two carbon atoms originally having the chlorine
atom. Thiourea gave a similar addition reaction. Attempts to convert
this intermediate to the disulfide failed because the acyl groups hydrolyzed
under the required alkaline conditions.
Other halogen-containing esters may be prepared by adding chlorine or
bromine to unsaturated esters dissolved in chloroform or tetrachloro-
ethane.118 When halogenated, cellulose crotonates, oleates, or esters con-
taining unsaturated vegetable oil acids become more widely soluble and
show an increase in moisture resistance.
(c) HYDROXY, KETO, AND AMINO ESTERS
The common low molecular weight carboxylic acids with a second func-
tional group such as hydroxy, keto, or amino have not been successfully
used to produce high degrees of esterification of cellulose. Hydroxy acids
cannot be employed in ordinary esterification procedures without at the
"o C. J. Malm, J. W. Mench, R. F. Williams, Jr., and G. D. Hiatt, Ind. Eng. Chern., 42,
1547 (1950).
111 H. C. Olpin, S. A. Gibson, and J. E. Jones (to Celanese Corp. of America), U. S.
Patent 2,348,305 (May 9, 1944); Chcm. Abstracts, 39, 617 (1945).
»« E. F. Izard and P. W. Morgan, Ind. Eng. Chem.t 41, 617 (1949).
"»H. T. Clarke and C. J. Malm (to Eastman Kodak Co!), U. S. Patent 1.687,060
(Oct. 9f 1928); Chem. Abstracts, 22, 4816 (1928).
814 CELLULOSE
same time esterifying the hydroxyl group. Lactones, such as /3-propio-
lactone, are reluctant to add to cellulose to form either an ether or an ester
linkage. In general, a- and £-keto acids are unstable in the presence of
anhydrides or catalysts used for esterification. Caldwell114 described the
preparation of cellulose acetate acetoacetates by treating a partially hy-
drolyzed acetate with diketene in the presence of pyridine. Higher keto
acids, such as levulinic, may be used with the usual reaction methods em-
ployed for acids of comparable molecular weight. Although cellulose esters
containing amino groups have been prepared indirectly by replacement of
substituted halogen or tosyl (£-toluenesulfonyl), only Gardner116 described
a direct reaction. Hydrolyzed cellulose acetate is treated with N-acet-
ylated amino acids in the presence of chloroacetic anhydride to yield
derivatives whose nitrogen content makes them more susceptible to dyeing.
Hydrolyzed cellulose esters containing minor quantities of combined
hydroxy or keto acyl groups may be prepared by employing these acids as
reaction media for hydrolysis of cellulose acetate.116 This hydrolysis fol-
lows the behavior of equilibrium reactions, permitting a certain amount of
interchange of acyl groups between the reaction solvent and the cellulose
esters. The result is that some of the acyl groups from the hydrolysis sol-
vent combine with cellulose, while a continual removal of acetyl groups
from the cellulose ester takes place. The reaction may be carried out in the
presence of a catalyst at a moderate temperature or with no added catalytic
agent at higher temperatures. Cellulose acetate lactates, tartrates, py-
ruvates, and citrates may be made in this way. Several of the products
are soluble in water. By the time much of the interchange acid has added,
the total amount of combined acyl is quite low.
(d) ALKOXYACYL ESTERS
Acid chlorides of alkoxy fatty acids may be made to react with cellulose
in the presence of pyridine to give soluble products, although there is usu-
ally difficulty in obtaining complete esterification. Cellulose derivatives
with free hydroxyl groups may be quite readily esterified by reaction of
either the acid chlorides or anhydrides of these acids in pyridine.117
»« J. R. Caldwell (to Eastman Kodak Co.), U. S. Patent 2,521,897 (Sept. 12, 1950);
Chem. Abstracts, 45, 860 (1951).
u* T. S. Gardner (to Eastman Kodak Co.), U. S. Patent 2,461,152 (Feb. 8, 1949);
Chem. Abstracts, 43, 3616 (1949).
IM C. J. Staud and C. S. Webber (to Eastman Kodak Co.), U. S. Patent 1,900,871
(Mar. 7, 1933); Chem. Abstracts, 27, 3073 (1933).
UT C. J. Malm and J. D. Coleman (to Eastman Kodak Co.), U. S. Patent 2,028,792
(Jan. 28, 1936); Chem. Abstracts. 30, 1998 (1936).
DC. DERIVATIVES OF CELLULOSE 815
In the presence of acid catalysts, anhydrides of alkoxyacetic acid are
similar to halogen-substituted acetic acids in that they do not react to
give cellulose derivatives. For this reason they may be employed similarly
to chloroacetic anhydride as impelling agents to form esters of other organic
acids.6 As with hydroxyacetyl groups, small amounts of alkoxyacyl groups
can be introduced by hydrolyzing a cellulose ester with an alkoxy fatty
acid as solvent.117*
3* Miscellaneous Esters
(a) AROMATIC ACID ESTERS
There has been no commercial interest in cellulose esters of aromatic
acids, investigation of their properties having shown no distinct advantages
over the more readily prepared aliphatic esters. Cross and Bevan,118 in
the course of their researches on cellulose, prepared esters of benzoic acid
by reaction of the acid chloride on alkali cellulose. Incomplete esterifica-
tion was obtained, lower concentrations of reagents giving a fibrous mono-
benzoate whereas under more severe reaction conditions a solution of the
dibenzoate resulted.
Cellulose tribenzoate was prepared by Wohl,119 who used an excess of
benzoyl chloride in pyridine at 110-130°C. with nitrobenzene as a diluent.
Ost and Klein120 investigated both caustic alkali and pyridine as reaction
media for benzoylation. A pyridine reaction mixture was recommended as
more satisfactory, giving products with 73% combined benzoic acid.
Atsuki and Shimoyama121 prepared cellulose dibenzoate from both
ordinary chemical cellulose and regenerated cellulose. After treatment
with 35% alkali, the cellulose was aged 24 hrs. at ordinary temperature,
then reacted with a benzene solution of benzoyl chloride at 50-60°C. for
117a Editors1 Note: The most noteworthy example of the mixed ester-ether category
of cellulose derivatives is acetylated hydroxyethyl cellulose (I. G. Farbenindustrie Akt.-
Ges., U. S. Patents 1,876,920 (Sept. 13, 1932) and 1,994,038 (Mar. 12, 1935); Carbide
and Carbon Chemicals Corp., U. S. Patents 2,327,397 (Aug. 24, 1943) and 2,330,263
(Sept. 28, 1943)). At a hydroxyethoxyl substitution level of greater than 0.3, the tri-
acetate derivatives have acetone solubility and exhibit lower water sensitivity than that
of conventional acetone-soluble types of cellulose acetate. In general, hydroxyethyl
cellulose is more reactive than cellulose itself in esterification reactions.
118 C. F. Cross and E. J. Bevan, Researches on Cellulose 1895-1900, Longmans, Green,
London, p. 34.
"• A. Wohl, Z. angew. Chem., 25, 285 (1903).
l» H. Ost and F. Klein, Z. angew. Chem., 26, 437 (1913).
181 K. Atsuki and K. Shimoyama, Cellulose Ind. (Tokyo), 2, 336 (1926); Kunstseide,
10, 250 (1928).
816 CELLULOSE
1-2 hrs. The product from regenerated cellulose was completely soluble
in chloroform and acetone; that from normal cellulose gave poor solutions.
Benzoic acid in the presence of chloroacetic anhydride reacts to give
cellulose tribenzoate. A recommended procedure is the reaction of 3
parts of cellulose, 15 parts of benzoic acid, 20 parts of chloroacetic anhy-
dride, and 0.05 part of magnesium perchlorate for 8 hrs. at 60-70°C. The
same reaction conditions may be employed for various substituted acids,
including chloro-, nitro-, and methoxy-benzoic acids. The nitro- and halo-
gen-substituted acids react with greater difficulty than benzoic; the 0-
methoxybenzoic derivative of cellulose is readily formed.122 Cellulose
cinnamate has been prepared both by reaction of the acid chloride in the
presence of pyridine128 and by esterification with the acid in the presence of
chloroacetic anhydride.8 Mixed esters have been prepared by the action
of acetylvanillic acid and chloroacetic anhydride on a partially hydrolyzed
cellulose acetate.128* Acetylvanillyl was introduced to the extent of 0.1
group per glucose unit.
Phenylacetyl chloride, when reacted upon cellulose in a mixture of pyri-
dine and chlorobenzene at 80-120°C., yields an ester of 77% combined
phenylacetic acid. The ester melts at 140°C.86
(b) DIBASIC ACID ESTERS
Cellulose may be esterified by one or both carboxyl groups of dibasic
acids. Products of the latter type, such as are obtained by treatment of a
pyridine solution of hydrolyzed cellulose acetate with a dibasic acid chlo-
ride or by using a mineral acid catalyst with a dibasic acid anhydride,1235 are
insoluble (cross-linked) and the products are of comparatively little com-
mercial interest. If only one carboxyl group is combined with cellulose,
the other may be converted to a salt, esterified by some other organic
radical, or left in the free acid form. Products of all these types have been
made.
Frank and Caro124 prepared cellulose oxalic acid esters by reaction of
acid chlorides of half -esters of oxalic acid with cellulose in the presence of
l" H. T. Clarke and C. J. Malm (to Eastman Kodak Co.), U. S. Patent 1,704,283
(Mar. 5, 1929); Chem. Abstracts, 23, 2033 (1929); Brit. Patent 313,408 (Aug. 27, 1929);
Chem. Abstracts, 24, 1217 (1930); Societe Kodak-Path^, French Patent 653,742 (Dec. 31,
1929); Chem. Abstracts, 23, 3807 (1929).
1M G. Frank and H. Mendrzyk, Ber., 63B, 875 (1930).
lf* B. B. White and E. Barabash (to Celanese Corp. of America), U. & Patent
2,581,565 (Jan. 8, 1952).
li* R. Rigamonti and V. Riccio, Ann. Mm.. 42, 283 (1952).
** G. Frank and W. Caro, Ber., 63B, 1532 (1930).
IX. DERIVATIVES OP CELLULOSE 817
pyridine, with nitrobenzene as a diluent. The lower alkyl cellulose oxa-
lates were found to be widely soluble in organic solvents. The cetyl ester
was of more limited solubility and the menthyl ester was only partly solu-
ble in all solvents.
Similar products may be prepared by reaction of cellulose with the acid
form of the half -esters with chloroacetic anhydride as an impelling agent.126
Cellulose alkyl succinates and phthalates have melting points below 200°C.
and are soluble in a wide variety of solvents126 (Table 12).
Mixed esters prepared by esterification of the free hydroxyl groups of hy-
drolyzed cellulose acetate with dibasic half-esters also show wide solu-
bilities (Table 13) and exhibit better film-forming properties than the simple
esters.
Cellulose half -esters of dibasic acids with one carboxyl group in the acid
form are best prepared by treating cellulose or cellulose acetates with a
dibasic acid anhydride and a tertiary organic base. The reaction may be
carried out on cellulose or partially hydrolyzed cellulose acetate.127 Re-
generated cellulose presoaked in water and then dewatered with pyridine
is a satisfactory starting material for the simple ester. Acetate phthalates
can be prepared by treating partially hydrolyzed cellulose acetates in a
solvent such as acetone or dioxane with one and one-half times the theo-
retical quantity of phthalic anhydride and about two times the theoretical
quantity of pyridine. These quantities are sufficient to add phthalyl to
60-70% of the available hydroxyl groups. Greater excesses of reagent
will increase the amounts of phthalyl somewhat, but preparation of fully
substituted derivatives is difficult even on repeated phthalation. The
phthalate ester is recovered by diluting the viscous reaction solution and
pouring it with stirring into acidified water which decomposes the pyridine
salt.
Products of somewhat lower viscosity and combined dicarboxylic acid
content may be made by heating cellulose acetate with a dicarboxylic acid
anhydride in an inert solvent.128 Melt reactions are also possible in which
125 R. L. Stinchfield (to Eastman Kodak Co.), U. S. Patent 1,704,306 (Mar. 5, 1929);
Chem. Abstracts, 23, 2033 (1929).
188 C. J. Malm and C. R. Fordyce, Ind. Eng. Chem.. 32, 405 (1940).
127 C. J. Malm and C. R. Fordyce (to Eastman Kodak Co.), U. S. Patent 2,024,238
(Dec. 17, 1935); Chem. Abstracts. 30, 1230 (1936); F. Schulze (to E. I. du Pont de
Nemours & Co.), U. S. Patent 2,069,974 (Feb. 9, 1937); Chem. Abstracts. 31, 2430
(1937); C. J. Malm and C. E. Waring (to Eastman Kodak Co.), U. S. Patents 2,093,462
(Sept. 21, 1937) and 2,093,464 (Sept. 21, 1937); Chem. Abstracts. 31, 8194 (1937).
"»L. B. Genung (to Eastman Kodak Co.), U. S. Patent 2,126,460 (Aug. 9, 1938);
Chem. Abstracts. 32, 7723 (1938); L. W. Blanchard, Jr., and C. L. Crane (to Eastman
Kodak Co.), U. S. Patent 2,183,982 (Dec. 19, 1939); Chem. Abstracts. 34. 2602 (1940).
818
CELLULOSE
j
++++++++
1
+ + ++ +
-2
5
5
Si
i
jT
"^
. a
it
58<N
M M +
r
2&
a
?
1 1 ++ 1 1 ++
^
1
fr
u
t
*s
b
v "•
0
S,
•d
ctf
0 0
IJ2
1 I + + +
fe
T3
V
1
Benzen<
i +++ 1 +++
1
V
s
g
B
i!
«-*
£
§
M M +
§
a o
8
V HI
tf)
a
g
4^
2^
fl
STJ
g
d
1
.S
|'§
++++++++
S
0
.2
1
M M +
d
o
O
S3
SIS
1
*?
••<
6
|
«
*>>
«**
Solubi
Butyl
acetate
1 1 ++ 1 +++
S
Q
i
Bthylene
chloride
+ + + + +
1
jd
bo
.S
Toluene
S
1
„
d
I
V
•8
sS
+
bfi
If
1 1 + + +
aJ
i
1
V
y 8
*S
0
W
«
d
"3
« w
V
+-»
w
*R
I
13
, , , , ,
^
1
!g
++++++++
o
s
S|
4->
43
OH
fi
sl
$
+3j4
M
"R— w
l>
1
S
w
1
S*l
+ + + + +
?
11
o
1
§
4}
t
*
0
0
+ + + + H T + +
§
0
0
1
*
3
'1
+ + + + +
3
P
8
S
c5
§*
0
o
O
*o
P-
£
r
SSo8S?2Ng
•g
%
§ K N : 8
|jjj
iH
a
* CO
4J
1
s'
Or)
&
S«j
&3
«fl
C-
J3
u
TABL1
er of cellulose acetate wit
|a|s|j||
wSf3OQ,l!i*9^
g M a *•* 43 CX »5
•RS^S^a- ft
Ullllll
TABLE 13. Pro
of cellulose acetate wit!
onomethyl succinate
onobenzyl succinate
onomethyl phthalate
onoethyl phthalate
onobutyl phthalate
TABLE 14. P
^10
S°
5*
58^
a
§d d d d d d d
o o o o o o o
sssssggs
1
M
sssss
£s
I + + +
I I + +
111 +
1 + + +
I I I I I
MM
+ M 1
^ JH 00
IX. DERIVATIVES OF CELLULOSE 819
the hydrolyzed cellulose acetate is heated directly with an excess of the
anhydride. Two parts of maleic anhydride melted with one part of an
acetate of 32% acetyl content gives an acetate maleate containing 20%
maleyl.129
The simple esters such as cellulose succinate or phthalate are readily
soluble in organic solvents containing 5-20% of water and in various aque-
ous bases. Mixed esters prepared from cellulose acetate become increas-
ingly soluble in organic solvents and in aqueous bases as the combined
dibasic acid content is increased (Table 14). A product with about 1.4
acetyl, 0.8 phthalyl, and 0.8 hydroxyl degrees of substitution per glucose
unit is soluble in a wide range of organic solvents and forms viscous aque-
ous solutions when dissolved by the gradual addition of the calculated
quantity of sodium hydroxide or bicarbonate.130 These solutions are most
stable at a slightly alkaline pH. The viscosity may be increased by the
careful addition of ions such as aluminum, zinc, magnesium, or calcium, or
the ester may be completely precipitated by the addition of copper or mer-
cury salt. The sodium salt may be precipitated by pouring the aqueous
solution into excess acetone.
The acid esters are useful for application of water-insoluble surface coat-
ings or sizings which are later to be removed by treatment with dilute
aqueous alkali. The same property is used in the application of cellulose
acetate phthalate as an antihalation backing in photography where the
colored backing layer is removed as a water-soluble salt after treatment in
the alkaline developer solution.131 These esters are also being used in
enteric coatings for medical purposes, 182-182a since they are resistant to
conditions encountered in the stomach but are soluble in the more alkaline
medium of the intestinal tract.
Cellulose acetates containing small amounts of phthalyl or adipyl and
treated with difunctional amines or with glycols are described as having
better textile properties than the parent acetate.188 Similar products
129 G. D. Hiatt and J. Emerson (to Eastman Kodak Co.), U. S. Patent 2,352,261
(June 27, 1944); Chem. Abstracts, 38, 5671 (1944).
180 C. J. Malm and C. R. Fordyce (to Eastman Kodak Co.), U. S. Patent 2,082,804
(June 8, 1937); Chem. Abstracts, 31, 5578 (1937).
111 C. J. Staud (to Eastman Kodak Co.), U. S. Patent 1,954,337 (Apr. 10, 1934);
Chem. Abstracts, 28, 3676 (1934).
182 G. D. Hiatt (to Eastman Kodak Co.), U. S. Patent 2,196,768 (Apr. 9, 1940);
Chem. Abstracts, 34, 5602 (1940).
iaf» C. J. Malm, J. Emerson, and G. D. Hiatt, /. Am. Pharm. Assoc., Sci. Ed., 40, 520
(1951).
i" H. Dreyfus (to Celanese Corp. of America), U. S. Patent 2,302,752 (Nov. 24, 1942);
Chem. Abstracts, 37, 2575 (1943).
820 CELLULOSE
containing small amounts of combined succinic, adipic, or sebacic acid
introduced during the acetylation process are reported as giving improved
elasticity.184
(c) SULFONIC ACID ESTERS
Cellulose may be esterified by acid chlorides of organic sulfonic acids
either in the presence of alkali or a tertiary organic base. Under acid
conditions of reaction, the sulfonic acids do not esterify cellulose; they are
on the other hand, suitable catalysts for certain esterification procedures.
Regenerated cellulose has been treated with a large excess of pyridine
and benzenesulfonyl chloride to yield a soluble product.1344 After 22 hrs.
of reaction, the recovered ester contained two benzenesulfonyl groups per
glucose unit. Only a small amount of combined chlorine was found.
Most published work on these derivatives has been concerned with esters
of ^-toluenesulfonic acid. Sakurada and Nakashima186 studied the re-
action of alkali cellulose with ^-toluenesulfonyl chloride, obtaining best
results at temperatures under 30°C. The degree of esterification for a
single treatment corresponded to somewhat less than a monoester, al-
though by repeated esterification products between mono- and di-esters
were obtained.
Hess and Ljubitsch,186 using pyridine as a reaction medium, obtained a
product containing 12% chlorine and 1% nitrogen from reaction of cellulose
and £-toluenesulfonyl chloride at 70°C. By use of cuprammonium rayon
as a starting material and reaction at 15-20°C., an ester containing two
tosyl (/>-toluenesulfonyl) groups for each glucose unit and only 0.2%
chlorine and 0.7% nitrogen was obtained. Hayes and Lin187 heated
a cellulose acetate £-toluenesulfonate in pyridine, 3-picoline, and isoquino-
line and obtained about a 70% conversion of tosyl to the quaternary salt
indicating that about 70% of the tosyl was combined with primary groups.
Bernoulli and Stauffer138~attempted to avoid the side reactions by using
the anhydride of £-toluenesulfonic acid in pyridine under a variety of
conditions but obtained only a minor degree of esterification.
184 G. A. Richter, Jr. (to American Viscose Corp.), U. S. Patent 2,534,371 (Dec. 19,
1950); Chem. Abstracts, 45, 2206 (1951).
"<• I. V. Nemilova, Zhur. Priklad. Khim., 25, 1107 (1952).
m I. Sakurada and T. Nakashima, Sri. Papers Inst. Phys. Chem. Research (Tokyo), 6,
214 (1927).
1W K. Hess and N. Ljubitsch, Ann., 507, 62 (1933).
l« F. N. Hayes and C. H. Lin, /. Am. Chem. Soc., 71, 3843 (1949).
l* A. L. Bernoulli and H. S. Stauffer, Helv. Chim. Acta, 23, 627 (1940).
IX. DERIVATIVES OF CELLULOSE 821
Rigby189 has described a modified method for reaction of £-toluenesul-
fonyl chloride on alkali cellulose, employing first a temperature not above
20°C., then an elevated temperature up to 120°C. Esterification corre-
sponding to 1.5 to 1.7 tosyl groups for each glucose unit was obtained.
Large quantities of tosyl chloride and alkali were necessary.
Malm and Nadeau140 prepared mixed esters by reaction of cellulose de-
rivatives containing free hydroxyl groups with aromatic sulfonyl chlorides
in the presence of pyridine. The products have better solubilities in or-
ganic solvents than the parent derivatives, and are highly resistant to
moisture.
In recent years the £-toluenesulfonyl (tosyl) group has been used as a
tool in studying the configuration of cellulose derivatives. Cramer and
Purves141 have shown that the tosyl group reacts most readily with the
primary hydroxyl group of the cellulose, esterification beyond that point
being at a considerably slower rate. The primary and secondary hydroxyl
groups are distinguished by the relatively selective reaction of sodium
iodide in acetone solution replacing the tosyl groups on primary hydroxyls
by iodine. By preparation of tosyl derivatives of acetone-soluble cellulose
acetate and treatment with sodium iodide, it was found that of the free
hydroxyl groups in the cellulose acetate at least one-third were primary.
Mahoney and Purves,142 working with commercial ethyl celluloses, com-
bined data from lead tetraacetate and periodate oxidation with tosylation
followed by iodination values, and concluded that the first-order rate con-
stants for tosylation of the unsubstituted hydroxyls were in a ratio of 2.3
for the second, 0.07 for the third, and 15 for the sixth position. In similar
work on hydrolyzed cellulose acetate, Purves and Gardner148 concluded that
the available hydroxyls reacted in ratios of 2.16, 0.106, and 23.4 for the
hydroxyls in the second, third, and sixth positions, respectively.
Malm, Tanghe, and Laird144 studied the tosylation of hydrolyzed cellu-
lose acetates. Extended reaction times gave increased amounts of tosyl
but at a slower rate. Examination of a reaction curve allowed the choice
of a point approximating the reaction of the primary hydroxyl. Hydro-
1M G. W. Rigby (to E. I. du Pont de Nemours & Co.), U. S. Patent 2,123,806 (July 12,
1938); Chem. Abstracts, 32, 7263 (1938).
140 C. J. Malm and G. F. Nadeau (to Eastman Kodak Co.), U. S. Patent 2,036,423
(Apr. 7, 1936); Chem. Abstracts, 30, 3647 (1936).
141 F. B. Cramer and C. B. Purves, /. Am. Chem. Soc., 61, 3458 (1939).
148 J. F. Mahoney and C. B. Purves, /. Am. Chem. Soc., 64, 9 (1942).
148 C, B. Purves and T. S, Gardner, /. Am. Chem. Soc., 64, 1639 (1942).
"4 C. J. Malm, L. J. Tanghe, and B. C. Laird, /. Am. Chem. Soc., 70, 2740 (1948).
822 CELLULOSE
lyzed cellulose acetates showed a considerable variation in ratio of primary
to secondary hydroxyl depending on the method of preparation.
Heuser and coworkers145 reported on the tosylation of cellulose regener-
ated from the xanthate and concluded from iodination results that the rate
constant for substitution on the primary hydroxyl was 5.8 times that for
the secondary position.
Modification of the dyeing properties of cotton through introduction
of tosyl groups followed by treatment with ammonia or amines was sug-
gested by Karrer and Wehrli148 since the tosyl radical is replaced by amino
groups.
Hess and Ljubitsch186 by treatment of ditosyl (di-£-toluenesulfonyl)
cellulose with ammonia or ethylamine were not able to introduce over 2%
nitrogen. Treatment of cellulose sulfonic esters with aliphatic primary
or secondary amines to produce derivatives soluble in dilute acid has been
patented by Haskins.147
The reaction solution with an amine as the solvent was allowed to stand
at 20° to 65°C. for several days, yielding products which contained from
0.5 to 1 amino group for each glucose unit. The products obtained were
insoluble in water but soluble in dilute acids which form amine salts.
These reactions are similar to those with pyridine described above.187
Heating of an acetate £-toluenesulfonate, analyzing 0.196 primary and
0.054 secondary tosyls per glucose unit, in methanol with sodium methylate
essentially removed the acetyl and ^-toluenesulfonyl groups and produced
0.183 anhydro units per glucose unit in the cellulose. Reacetylation of the
product supported this structure.148
Wolfrom, Sowden, and Metcalf 149 have prepared cellulose esters of meth-
anesulfonic acid. Mercerized cotton or cellulose regenerated from the
acetate treated in pyridine with methanesulfonyl (mesyl) chloride yielded
products between the mono- and di-ester. They did not dissolve in the
reaction mixture during preparation. Cellulose acetate of low acetyl
content (1.7 acetyl groups) was reacted to introduce one mesyl group.
This product, when treated with sodium iodide, gave a lower degree of
iodine replacement than did £-toluenesulfonyl derivatives, contrary to
that which might be expected from behavior of hexoses. The mesyl
145 E. Heuser, M. Heath, and M. H. Shockley, /. Am. Chem. Soc.t 72, 670 (1950).
148 P. Karrer and W. Wehrli, Z. angew. Chem.t 39, 1509 (1926).
147 J. F. Haskins (to E. I. du Pont de Nemours & Co.), U. S. Patent 2,136,299 (Nov. 8,
1938); Chem. Abstracts. 33, 1495 (1939).
148 C. B. Purves and T. S. Gardner, J. Am. Chem. Soc.t 65, 444 (1943).
149 M. L. Wolfrom, J. C. Sowden, and E. A. Metcalf, /. Am. Chem. Soc.t 63, 1688
(1941).
IX. DERIVATIVES OF CELLULOSE 823
derivatives were also comparatively inactive toward replacement by
ammonia.
Mench150 treated hydrolyzed cellulose acetates with N-alkyl and aryl
sulfamyl chlorides and obtained products analyzing from 0.1 to 0.8 sul-
famyl groups per glucose unit depending on the acetyl content of the start-
ing acetate. In certain ranges of substitution water-soluble products
were obtained.
(d) CARBAMIC ACID ESTERS
Carbamates of cellulose may be prepared by reaction of organic iso-
cyanates upon dried cellulosic material in the presence of a tertiary base
such as pyridine. Completely anhydrous reaction conditions are neces-
sary; other arise the water, in a rather violent reaction, will consume two
molecular equivalents of isocyanate to form substituted ureas which are
generally quite difficult to remove from the substituted cellulose.
The reaction of cellulose and partially hydrolyzed cellulose acetates
with alkyl and aryl isocyanates has been studied by Hearon, Hiatt, and
Fordyce.161 The reaction of carbamic acid itself gave products of low
nitrogen content. Methyl and ethyl isocyanates gave partial reaction with
hydrolyzed cellulose acetates but failed to give fully esterified products in
reactions at 50° and 100°C. Aromatic isocyanates reacted readily to give
fully esterified products and, with the use of less than the theoretical
quantities of phenyl isocyanate, reacted quantitatively. The reaction of
cellulose with excess isocyanate was slower but gave, after 48 hrs. at 100°C.,
soluble products which were relatively undegraded and proved on analysis
to be the tricarbamates.
A second paper162 reported that hydrolysis experiments were carried out
to remove acetyl groups from cellulose acetate carbamates to produce
simple cellulose carbamates with some hydroxyl groups. This was possible
because of the much greater alkali stability of the carbamate group. Series
of cellulose carbanilates and a-naphthyl carbamates were prepared and
their solubilities described.
Schneebeli152* studied the reaction of phenyl isocyanate with cellulose in
pyridine as a function of time and temperature. X-ray studies showed
that reaction did not occur within the crystalline portion until one-quarter
"« J. W. Mench (to Eastman Kodak Co.), U. S. Patent 2,518,706 (Aug. 15, 1950);
Chem. Abstracts, 45, 1768 (1951).
U1 W. M. Hearon, G. D. Hiatt, and C. R. Fordyce, /. Am. Chem. Soc.t 65, 829 (1943).
1M W. M. Hearon, G. D. Hiatt, and C. R. Fordyce, J. Am. Chem. Soc.. 65, 833 (1943).
™* P. Schneebeli, Compt. rend., 234, 738 (1952).
824 CELLULOSE
to one-third of the hydroxyl groups had reacted. Dyer and McConnick158
added o- and £-chlorophenyl isocyanates to methyl cellulose and got com-
plete substitution in 6 hrs. at 65 °C. They reported only a small amount
of degradation.
The reaction of hydrolyzed cellulose acetate with different aryl iso-
cyanates was examined by Hearon and Lobsitz.154 Rated in order of in-
creasing reactivity, these materials were 0-tolyl, £-tolyl, phenyl, a-naphthyl,
0-chlorophenyl, and £-bromophenyl isocyanate.
Some attempts have been made to improve cellulose acetate products
by treatment with diisocyanates. Hydrolyzed acetates treated with
hexamethylene diisocyanate are claimed to show an improvement for
textile uses because of their decreased organic solvent solubility and in-
creased ironing temperature.156
High molecular weight isocyanates such as those from hydrogenated rosin
have been added to cellulose derivatives to effect an improvement in solu-
bility, compatibility, and resistance to hydrolysis.166
Breslow167 has produced acid-soluble derivatives by treating ethyl cellu-
lose, cellulose acetate, and hydroxyethyl celluloses with isocyanates or
azides of tertiary bases. Nicotinyl azide reacted under nitrogen in the
presence of pyridine to give products soluble in dilute acetic acid. Simi-
lar solubility resulted from treating hydroxyethyl cellulose with £-(di-
methylamino)phenyl isocyanate.
i" E. Dyer and K. L. McConnick, /. Am. Chem. Soc., 68, 986 (1946).
154 W. M. Hearon and J. L. Lobsitz, /. Am. Chem. Soc., 70, 296 (1948).
165 D. D. Coffman and J. S. Reese (to E. I. du Pont de Nemours & Co.), Brit. Patent
548,807 (Oct. 26, 1942); Chem. Abstracts, 38, 485 (1944).
168 D. S. Breslow (to Hercules Powder Co.), U. S. Patent 2,492,928 (Dec. 27, 1949);
Chem. Abstracts, 44, 2238 (1950).
167 D. S. Breslow, /. Am. Chem. Soc., 72, 4244 (1950).
D. ALKALI AND OTHER METAL DERIVATIVES
W. D. NICOLL, N. L. Cox, AND R. F. CONAWAY
The treatment of cellulose fibers with alkalies is one of the oldest and
most important commercial reactions of cellulose. Although this broad
field has been the subject of investigation for more than 100 years, the
effects of alkali on cellulose fibers are not yet fully understood. There are
two important objectives in treating cellulose with alkalies, namely, to
modify the physical and chemical properties of natural fibers such as in
mercerization, and to obtain intermediates such as alkali cellulose for con-
version to cellulose derivatives.
The simplest classification of the alkali metal--cellulose complexes is the
following; alkali cellulose, the true metal alcoholates of cellulose, cupram-
monium- and cupriethylenediamine-cellulose complexes, and cellulose-
organic base complexes. The organic nitrogen bases by definition are out
of place in this chapter but are included since their action on cellulose is
similar to that of the inorganic bases.
1. Alkali Cellulose
Strong alkali solutions acting on cellulose at low or room temperatures
produce alkali cellulose. This treatment of cellulose has been known by the
textile industry for a very long time as "mercerization,"1 so called after
John Mercer who discovered the process in 1844.2
The technical use of alkali cellulose dates from the discovery of Lowe8
in 1889 that tension on cellulosic fibers during treatment with strong
alkali increased luster, and from the even more important discovery of
Cross, Bevan, and Beadle4 in 1892 that treatment of alkali cellulose with
carbon disulfide produced cellulose xanthate which could be dissolved in
water or dilute alkali to form an orange-yellow viscous solution which they
termed " viscose. " As a result of this discovery, the preparation of alkali
* D. A. Clibbens, /. Textile Inst., 14, T217 (1923).
* E. A. Parnell, Life and Labors of John Mercer, Longmans, Green & Co., London,
1886.
» H. A. Lowe, Brit. Patents 20,314 (1889) ; 4462 (Dec. 22, 1890).
« C. F. Cross, E. J. Bevan, and C. Beadle, /. Chem. Soc.t 63, 837 (1893).
826 CELLULOSE
cellulose (or "soda cellulose1' as it is sometimes named) became the first
step in the manufacturing operations of the important group of viscose
industries from which modern civilization derives such varied products as
rayon, cellophane, cellulose sponges, cellulose sausage casings, and certain
types of bottle closures.
In addition to its use in the preparation of viscose, alkali cellulose is an
important intermediate product in the preparation of certain cellulose
ethers, such as the ethyl, methyl, and benzyl ethers, which find application
in the preparation of films and many types of plastic products. 4 Merceriza-
tion is also used as a finishing treatment for textiles, and in the preparation
of certain types of highly absorptive papers and paperboard products now
employed in the manufacture of such materials as shoe lasts and blotting
papers. This discussion on the nature of alkali cellulose and its applica-
tions in industry will be presented mainly from the standpoints of the vis-
cose process and the mercerization of textile fibers.
(a) STRUCTURE OF ALKALI CELLULOSE
Mercer assumed that the formation of alkali cellulose is a result of chem-
ical combination between the cellulose and sodium hydroxide, and that the
resulting compound is susceptible to decomposition under the action of
water. Karrer5 in support of this view cited analogous cases of simple
and polyhydric alcohols and pointed out that many of these alcohols form
compounds of a characteristically complex nature with alkali. Other
workers, however, have held that the changes involved when cellulose is
treated with alkali are purely physical and similar to those produced in the
swelling of many other colloidal substances. Emphasis in recent years has
been given to the methods of x-ray analysis as a means of determining the
changes occurring when alkali cellulose is formed, and this method of in-
vestigation appears to give definite support to the view that cellulose may
form not only one, but several, compounds with strong alkalies. The par-
ticular compound formed depends on the type of cellulose, the concentra-
tion of alkali in solution, and the temperature. But regardless of the type
of compound formed, alkali celluloses decompose readily on washing with
water, and the caustic can be easily and completely removed.
Assuming a chemical compound as the more probable form for alkali
cellulose, two views are possible. Either the cellulose, like the simple
alcohols and certain other polyhydric compounds, forms a true alcoholate
with a metal ion (M) substituting for a hydrogen ion, thus :
IWOH) + MOH > R^n(OM) + H2O
• P. Karrer, CeUulosechemie, 2, 125 (1921).
IX. DERIVATIVES OF CELLULOSE 827
or the alkali is present as an addition complex such as is believed to form
between certain sugars and the alkaline earth hydroxides. In the latter
case, alkali cellulose would have the composition represented by Rceu(OH)r •
(MOH)y. Although the literature indicates differences in opinion on this
point, the majority of investigators in this field seem to agree that, at least
in the presence of water, the addition complex is to be preferred over the
alcoholate type of structure. It is interesting to note that using one of the
latest methods for studying this question, Makolkin6 was able to follow
the isotope exchange of oxygen 18 between water and alkali cellulose and
decided that mercerization proceeds according to the equation:
Roeii(OH) + NaOH > RMn(OH).NaOH
Thus, alkali cellulose is a product of the addition of alkali to cellulose in the
form of a weak electrolyte and not in the form of an alcoholate.
Another controversial point concerns the question of whether more than
one addition complex may exist, with many investigators agreeing on
(CeHioOsVNaOH as correctly representing the composition of the alkali
cellulose commonly formed in alkali solutions of 12 to 18% concentration.
A second compound has been given the formula CeHioCVNaOH and
variously reported as existing in solutions containing 20 to 40% NaOH.
X-ray evidence strongly supports the existence of a new crystalline struc-
ture in caustic soda solutions of about 21% concentration, but the results
from other methods used in examining these compounds have not been
consistent.
In this connection it should be kept in mind that different celluloses con-
tain different ratios of crystalline to noncrystalline material. This affects
the number of cellulose chains available for reaction, and, therefore, in-
fluences not only x-ray results but also those obtained by other methods.
Important reasons for the divergent views which have been expressed in
regard to the structure of alkali cellulose include differences in the types of
celluloses investigated, the colloidal nature of alkali cellulose, and the
widely different concepts held in regard to the basic nature of the swelling
process. Following the discovery of the preferential absorption of alkali
by cellulose in contact with solutions containing more than 12% of NaOH,
early investigators concentrated attention on this fact and tended to dis-
regard swelling phenomena and the role played by water in the formation
of alkali cellulose. More recently, it was recognized that considerable
quantities of water accompany the alkali into the solid cellulose phase,
• I. A. Makolkin, /. Gen. Chem. ( U. S. S. R.), 12, 365 (1943); Ckem. Abstracts, 37,
3418(1943).
828 CELLULOSE
and emphasis shifted to the ratio of water to the total amount of alkali
present, as well as to the degree of swelling and accompanying increase in
internal surface. For example, studies by Saito7 have helped to clarify
complex relationships existing between the total amounts of alkali and water
absorbed and how these relationships are influenced by such variables as
the type of cellulose, the kind and concentration of alkali, the temperature,
and even the addition of an inert substance such as sodium chloride. This
and similar work will be discussed in more detail later. In general, support
is given to the view that physical factors probably have a greater influence
on the properties of alkali cellulose than does any stoichiometric relationship
which may exist between cellulose and combined alkali. However, be-
cause of the immense amount of work done to acquire exact knowledge of
the quantitative relationship in the system cellulose-sodium hydroxide-
water, a brief review of some of the methods which have been used to study
this problem seems justified.
(b) METHODS FOR DETERMINING ALKALI ABSORBED
(1) The Direct Method
The first investigator to attempt to determine the amount of caustic
soda taken up by cellulose was Gladstone,8 who in 1852 treated cotton with
various concentrations of sodium hydroxide and then washed the samples
with cold absolute alcohol or hot alcohol of about 92.5% concentration until
constant weight was obtained. Similar experiments were made with
potassium hydroxide, and Gladstone concluded that each of these alkalies
forms definite compounds with the cellulose. The method of washing
with alcohol has since been reinvestigated by several other experimenters
who have used either the same or slightly improved modifications of Glad-
stone's original method. Rassow and Schwarze,9 for example, found that
results obtained by washing with alcohol depend to a great extent on the
alkali content of the solution, the water content of the alcohol, and the
selection of the proper indicator to determine the completion of the wash-
ing operation. The conclusion was reached that a compound containing
12.3 g. of NaOH per 100 g. of cellulose and corresponding to the formula
is formed in alkali solutions containing from 18 to 40%
7 G. Saito, /. Sac. Chem. Ind., Japan. 43, B126, B133, B160, B194 (1940); 44, B89
(1941); Cellulosechemie, 18, 106(1940).
8 J. H. Gladstone, /. Chem. Soc., 5, 17 (1852).
9 B. Rassow and K. Schwarze, Papier-Fabr., 28, Tech.-wiss. Tl., 693, 746 (1930).
DC. DERIVATIVES OF CELLULOSE 829
NaOH. Still more recently Rogovin10 substituted butyl or propyl alcohol
for ethyl alcohol because these solvents reduce hydrolysis of "soda cellu-
lose." He concluded that the amount of alkali combined is greater than
that indicated above.
Bancroft and Calkin11 have criticized Gladstone's method, stating that it
is not proof of compound formation since the use of alcohol only introduces
another variable, and the results merely indicate how much sodium hy-
droxide is retained by the sample under the conditions of washing used.
If this conclusion is correct, it is surprising that the results obtained by
Gladstone and certain other workers agree so well with those from at least
two other methods, namely, the change-in-titer and change-in-x-ray-
diffraction-pattern, both of which have been interpreted to indicate the
formation of a cellulose-alkali compound.
(2) The Indirect or Change-in-Titer Method
The most widely studied method of determining the composition of
alkali cellulose is that introduced by Vieweg12 in 1907. It consists in deter-
mining by titration the change in alkali concentration when a known weight
of cellulose is placed in either a known volume or known weight of alkali
solution. The method requires careful standardization of the experimental
conditions in order to obtain reproducible results. Heuser and Nietham-
mer18 recommended a ratio of at least 10: 1 for alkali solution to cellulose,
and Rassow and Schwarze9 showed that absolutely dry cellulose should
not be used in measuring absorption because the process will then be non-
uniform.
Probably the most persistent criticism which has been made of the in-
direct method is that made first by Leighton,14 namely, that no attention is
given to the very considerable swelling and hydration which occur when
cellulose is treated with strong solutions of alkali. This criticism, however,
has generally been made by those interested in the total quantity of alkali
absorbed by the cellulose and is not believed justified when only the pref-
erentially absorbed alkali is to be determined. This point is further dis-
cussed below in connection with the work of Sharkov and Saito.
10 Z. A. Rogovin, Trudy Konferentsti Vysokomolekulyar. Soedineniyam, Akad. Nauk
S. S. S. R., Otdel. Khim. Nauk i Otdel. Fiz.-Mat. Nauk. 1, 33 (1943) (Pub. 1945) ; through
Chem. Abstracts, 40, 457 (1946).
" W. D. Bancroft and J. B. Calkin, Textile Research, 4, 119, 159 (1934); /. Phys.
Chem., 39, 1 (1935).
» W. Vieweg, Ber.t 40, 3876 (1907) ; 41, 3269 (1908) ; 578, 1917 (1924).
18 E. Heuser and W. Niethammer, Cellulosechemie, 6, 13 (1925).
14 A. Leighton, /. Phys. Chem., 20, 32, 188 (1916).
830
CELLULOSE
The results obtained by use of the Vieweg indirect method may most
readily be discussed by referring to the curve formed by plotting the ap-
parent change in alkali absorption against the caustic soda content of the
treating solution. This curve, generally known as the Vieweg curve,16 has
the following characteristics (Fig. 33) : It shows a gradual increase in alkali
absorption with increased concentration of caustic soda from 0 to 13-16%
NaOH ; at this point there occurs a definite break in the curve followed by a
region of constant absorption up to a concentration of approximately 21%
NaOH; above this the curve again rises gradually but flattens off at a new
300
o
O 200
d
6
8
CD
Q 100
0 10 20 30 40 50
CONCENTRATION OF EQUILIBRIUM SOLUTION, % NaOH
Fig. 33. Alkali preferentially absorbed by cellulose as determined by various investi-
gators using the change-in-titer method (d'Ans and Jager15). Curves: la, Vieweg I;
Ib, Vieweg II; 2, Liepatoff; 3, Rassow and Wadewitz; 4, Dehnert and Konig; 5,
Heuser, Niethammer, and Bartunek; 6, Karrer and Nishida.
level in the range of 35-40% NaOH. Richardson and Maass16 made a
study of the absorption of alkali from concentrated solutions, and, on the
basis of their work and that of Rumbold17 and Vieweg, constructed a curve
which shows the absorption of alkali over the whole range of 0 to 50%
concentration (Fig. 34) .
The amount of alkali absorbed at the first break in the Vieweg curve has
been determined by different investigators and may be said to average about
13 g. of NaOH per 100 g. of cellulose. The absorption at the second break
is approximately 22.5 g. of NaOH per 100 g. of cellulose. Vieweg, and
many other investigators who have used the method more recently, con-
« J. d'Ans and A. Jfiger, Cellulosechemie, 6, 137 (1925).
M R. Richardson and O. Maass, /. Pkys. Chern., 36, 3064 (1932).
" J. S. Rumbold, /. Am. Chem. Soc., 52, 1013 (1930).
IX. DERIVATIVES OF CELLULOSE
831
eluded that the first break in the curve denoted the formation of a definite
compound, (CeHioQsVNaOH, which is stable throughout the range of
alkali concentration represented by the first flat portion of the curve.
Evidence for the formation of the second compound, CeHioCVNaOH, in
concentrations of caustic soda above 35% is not so clear, and several investi-
gators following Vieweg were unable to obtain any absorption values suffici-
ently high to substantiate the existence of this compound.
The theory of compound formation and the use of the Vieweg method for
determining alkali absorbed by cellulose has been vigorously attacked from
a theoretical standpoint on the grounds that the curve does not conform to
phase-rule requirements. Ltighton14 was the first to point this out, and
it has since been discussed by Bancroft tod Calkin.11 According to these
400
300
200
100
1
10 20 30 40
CONCENTRATION OF EQUILIBRIUM
SOLUTION, % NaOH
50
Fig. 34. Composite 'curve for alkali preferentially absorbed by cotton at 20°C. (Richard-
son and Maass18). Curves: 1, Rumbold; 2, Vieweg; 3, Richardson and Maass.
investigators, the alkali absorption curve should show a distinct vertical
portion rather than a horizontal portion, as evidence for compound for-
mation. As ordinarily determined, the curve does not show this, but rather
shows a gradual rise to the inflection point. An important factor not recog-
nized by the critics of the Vieweg curve is that the system under investiga-
tion is one containing a high-molecular-weight material having character-
istically colloidal properties. In such a system swelling is not uniform,
and one should expect that as swelling progresses there would be a gradual
increase in the amount of reactive surface available for compound forma-
tion and that this would preclude the possibility of obtaining a sharp verti-
cal rise in the absorption curve. These ideas are in accord with a view
832 CELLULOSE
recently expressed by G. and T. Petitpas.18 Even if all of the cellulose
chains are equally available, the extent of reaction will be a continuous
function of the alkali concentration, and the alkali absorption curve will
have the same shape as the titration curve of a polybasic acid if more than
one type of addition complex of NaOH with cellulose is found. Calkin,19
independently, applied the x-ray method to the system and found that the
change in x-ray diffraction diagram from native cellulose to that charac-
teristic of alkali cellulose takes place only gradually, starting in a caustic
soda solution of about 13% concentration and becoming complete at about
16% NaOH. This range, it will be recognized, coincides closely with that
immediately preceding the break in the Vieweg curve. It seems likely,
therefore, that the change in x-ray diffraction pattern and the gradual
rise in the absorption curve for alkali may be due to one and the same
phenomenon.
(3) The Blotting or Centrifuging Method
Another method which has been used for investigating the cellulose-
caustic soda-water system is to blot or to centrifuge off the excess alkali
solution and to determine the total alkali and water taken up by the cellu-
lose. A general criticism of these methods is that it is difficult to deter-
mine accurately the point at which excess alkali solution clinging to the sur-
faces of the fibers is removed. It is also difficult to avoid loss of water
vapor which conceivably would affect the final results. Both the blotting
and centrifuging methods fail to take into account water absorbed by the
cellulose itself as differentiated from water accompanying the alkali as
solvent. Nevertheless, useful information regarding both the absorption
and swelling processes has been gained from this sort of work.
Beadle and Stevens,20 among the first to use the blotting method,
studied particularly the effects of temperature and the concentration of
the caustic soda solution. These experimenters used regenerated cellulose
in the form of cuprammonium rayon, and their results (Fig. 35) show
that, for temperatures between 5° and 40°C., points of maximum hydra-
tion and maximum alkali absorption occur, and that these maxima are
greater at lower temperatures. For solutions of 3 to 9% alkali content,
the effects of temperature on hydration and on caustic soda absorbed
bear a close resemblance to each other (the shapes of the curves are rela-
» G. and T. Petitpas, Compt. rend., 226, 672 (1948).
" J. B. Calkin, /. Phys. Chem., 40, 27 (1936).
» C. Beadle and H. P. Stevens, Orig. Com. 8th Intern. Congr. Appl. Chem., 13, 25
(1912).
IX. DERIVATIVES OF CELLULOSE
833
tively much alike), indicating that under these conditions the solutions in
which the cellulose was immersed were absorbed without change in com-
position. For the more concentrated solutions (10-25% NaOH) the curves
2800
2400
§2000
8 1600
9
o 1200
*tf 80°
400
0
280
0240
UJ
I 200
g!60
o!20
z
40
0
II
40 °C
0 4 8 12 16 20 24 28
% NaOH IN EQUILIBRIUM SOLUTION
1 5 9 13 17 21 25
% NaOH IN EQUILIBRIUM SOLUTION
2800
2400
2000
1600
1200
800
400
0
III
r
4%
280
240 Q
UJ
200 I
160 |
120 §
80
40
0
1900
1700
2 1500
§ 1300
O1100
cu
x 900
^ 700
500
300
100
0 10 20 30 40 5 15 25 35 45
TEMPERATURE OF SOLUTION.°C.
210
190
170
150
130
110
90
70
50
30
10
10 20 30 40 5 15 25 35 45
TEMPERATURE OF SOLUTION.°C.
Fig. 35. Effects of temperature and concentration on total alkali and water absorbed
by regenerated cellulose (Beadle and Stevens20).
are not similar, and this has been taken to mean that preferential absorp-
tion or chemical reaction occurs.
Neale21 also employed the blotting method with small pieces of regener-
ated cellulose sheeting. His results (recalculated and graphically pre-
" S. M. Neale, /. Textile Inst., 20, T373 (1929); 21, T225 (1930); 22, T320, T349
834
CELLULOSE
sented by Bancroft and Calkin) give the total quantities of sodium hydrox-
ide and water absorbed from solutions of varying concentrations. The
maximum water absorption (600%) occurs at about 12% alkali concentra-
tion (Fig. 36). On close study it is found that, if the amounts of alkali
preferentially absorbed as indicated by the Vieweg change-in-titer curve
(curve 2) are deducted from the total absorption (curve 1), the remaining
quantities of alkali and water are present in the cellulose in almost exactly
lite same proportions as in the mother liquors. The differences in total
absorption may be assumed, therefore, to represent mainly differences in
the degree of swelling.
0 2 46 8 10 12 14
MOLALITY OF ALKALI IN FINAL SOLUTION
Fig. 36. NaOH and H2O absorbed by cellophane at 25°C. (Neale's
data recalculated by Bancroft and Calkin11). Curves: 1, total alkali
absorbed; 2, preferentially absorbed alkali; 3, water absorbed.
Leighton14 tried to determine the alkali absorbed by cellulose by using a
centrifuge to remove excess alkali. The values he obtained produced a
smooth absorption curve, and he concluded that no compound formation
takes place. Similar conclusions were reached by Coward and Spencer,22
and Bancroft and Calkin11 who also used the centrifuge method. The last
workers used an indirect method for determining the point at which all
surface liquid was lost. Samples of the alkali cellulose were analyzed after
various centrifuging intervals. The time required to remove surface liquid
was assumed to be that at which no further change took place in the ratio
of caustic to water leaving the cellulose.
Champetier28'28* has used a somewhat different method in which the
alkali celluloses, instead of the centrifuged liquors, were analyzed after
« H. F. Coward and L. Spencer, /, Textile Inst.9 14, T28, T32 (1923).
*' G. Champetier, Ann. chim., 20, 5 (1933).
*** G. Champetier and K. Ashar, Makromol. Chem., 6, 85 (1951).
IX. DERIVATIVES OP CELLULOSE
835
progressive centrifuging or pressing. The quantity of preferentially ab-
sorbed alkali (Fig. 37) was obtained by plotting the ratios of alkali-to-
cellulose against water-to-cellulose and extrapolating to the zero axis for
water. Champetier concluded that cellulose may form four stoichiometric
compounds with either sodium or potassium hydroxides. No evidence of
compound formation with lithium hydroxide was found. Probably the
most serious objection that can be made to Champetier's method is that the
extrapolations involved are large and the likelihood of error fairly great.
0 10 20 30
MOLES H20 PER BASE MOLE OF CELLULOSE
Fig. 37. Changes in the composition of the residues obtained by centrifuging to
different degrees the alkali celluloses produced in NaOH solutions of various concentra-
tions (Champetier23). The numbers on the curves represent the grams of alkali per
liter.
(4) The Neutral Salt Method
Still another method used to study the formation of alkali cellulose has
been to add a neutral salt such as sodium chloride to the alkaline treating
836
CELLULOSE
solution. This method, originally used by Schwarzkopf,24 is based on the
assumption that if no salt is absorbed by the cellulose it should be possible
to calculate the amount of water taken up by analyzing the mother liquor
for chlorine before and after coming in contact with cellulose. Shar-
kov,26 who also used this method, investigated its validity and concluded
that it produced reasonable results provided the concentration of salt was
low (1% or under). Bancroft and Calkin,11 on the other hand, rejected
the method as unsound, but their conclusion was based on the use of fairly
concentrated salt solution.
Sharkov25 and Saito7 used the neutral salt method in conjunction with
the Vieweg change-in-titer method to investigate the action of not only
1.5
IV
ijj tjj
Q. CO
Q -J
Ul 3
OQ -j
CC -1
~ 1 1 1
1.0
a
»o
'0.5
15
DQ U.
< O
O UJ
*°
UJ uj
0%
SCQ
0 2 4 6 8 10
MOLALITY OF NaOH IN EQUILIBRIUM SOLUTION
Fig. 38. Minimum true absorption of NaOH and H^O by different cellulose fibers at
20 °C. (Saito7). Minimum NaOH absorption; minimum H2O absorption.
sodium hydroxide but lithium and potassium hydroxides as well. The
data obtained do not agree absolutely but the general conclusions reached
are similar.
Both Sharkov and Saito studied more than one cellulose and more than
one alkali. Figure 38 shows data by Saito indicating that with sodium
hydroxide there is a sharp increase in the amount of alkali absorbed by
cellulose in the range of 2 to 4 N NaOH. Within this range, the amounts
of sodium hydroxide absorbed vary considerably with the type of cellulose.
Regenerated cellulose begins to absorb alkali strongly from solutions as
weak as 2 J\T, similar absorption with ramie begins at about 3 N9 and that
for cotton at about 4 N. Between 4 and 6 N NaOH the absorption of
« O. Schwarzkopf, Z. Ekktrochem., 38, 363 (1932).
*» V. I. Sharkov, Iskussfrcnnoe Volokno, 2, 2 (1931).
DC. DERIVATIVES OF CELLULOSE 837
alkali increases only very slightly, but above 6 TV" the increase occurs again.
Saito showed that, for the celluloses investigated, the absorption taking
place in 6 N alkali is independent of the type of cellulose used and is ap-
proximately 1 mole of alkali per mole of cellulose. In the case of water
absorption, the amount for each cellulose passes through a maximum, but
the absorption of water by regenerated cellulose at the maximum is more
than twice that reached by cotton.
On comparing the various methods for determining the absorption of
alkali and water by cellulose it is interesting to note that they do not all
give the same total (alkali + water) absorption. The neutral salt method,
for example, produces only about half the absorption shown by either the
blotting or centrifuge methods. This difference may be explained if cellu-
lose, in the form of films or fibers, can be assumed to have a porous structure
capable of holding relatively large amounts of mother liquor which by the
blotting method would be included along with any liquor more intimately
associated with the cellulose itself. Presumably the same would not be
true for water absorption determined by the neutral salt method, and one
should therefore expect to obtain lower values by this method.
(c) SWELLING OF ALKALI CELLULOSE
The physical swelling phenomena which all celluloses undergo when
treated with a solution of an alkali hydroxide have as much theoretical
interest and probably are more important practically than the purely
chemical aspects of the changes involved in the formation of alkali cellulose.
The two phenomena, chemical and physical, appear to be so completely
interdependent that one cannot properly be considered without the other.
Mercer, in his study of the changes produced in cotton on treating with
alkalies, observed generally the effects which swelling had on hygroscopicity,
tensile strength, and dye affinity of the washed yarn. Crum,26 in 1863, was
the first to attempt a quantitative examination of the swollen condition
of these fibers by measuring the cross section and determining the longitu-
dinal shrinkage of cotton fibers. Crum also introduced the use of the micro-
scope as a tool for studying swelling and described in detail how the fibers
lose their flat, ribbon-like form and assume a nearly cylindrical appearance.
On the basis of modern concepts concerning the fine structure of cellulose,
changes in length and diameter of the fiber depend on the fact that the net-
work of cellulose chains and crystallites is most highly oriented in the direc-
tion of the fiber axis. The alkali solution, entering the network, pushes the
units apart and loosens interconnecting secondary bonds, thus allowing the
* W. Crum, J. Chem. Soc., 16, 404 (1863).
838 CELLULOSE
structure to assume a less oriented condition. However, the rupture of
the secondary bonds is probably incomplete, so that the structure is still
held together at certain points, and the shape of the fiber must change
anisotropically, that is, more in one direction than in the other.
Swelling differences observed in caustic solutions of varying concentra-
tions can perhaps best be explained on the basis that penetration of alkali
can be either inter- or intra-crystalline as first pointed out by Katz.27
Since x-ray evidence indicates no change in spacing between individual
cellulose chains when cellulose is brought into contact with solutions of
caustic soda below 12% concentration, it may be assumed that in these
solutions swelling takes place almost entirely between the crystallites. In
solutions above 12% caustic soda concentration, the swelling process
changes to the intracrystalline type, since alkali is preferentially absorbed
or combined with the cellulose and a definite change occurs in the crystal
lattice as determined by x-rays.
In addition to the important factors of solution concentration and tem-
perature, already mentioned as influencing the degree of swelling and alkali
absorption, the following variables may also be considered when cfellulose
is treated with a solution of an alkali: (1) the nature of the cellulose, (2)
the time of contact, (3) hysteresis and equilibrium, (4) changes in solvent,
and (5) presence of salts.
Purified cotton or ramie cellulose is less readily swollen by alkalies than
is either mercerized or regenerated cellulose. Wood pulps usually occupy
an intermediate position. The differences in ability to swell appear to
be associated with the degree of orientation and crystallinity of the cellu-
lose chain molecules and micellar aggregates composing the various cellu-
lose structures. In addition, one must also consider the previous history
of the sample, such as treatments with hydrolyzing or swelling agents. To
illustrate, Figure 38 shows typical data covering the swelling effects pro-
duced by different concentrations of sodium hydroxide on ramie, cotton,
and regenerated cellulose. It will be seen that the regenerated cellulose
both reaches a maximum in swelling at a lower concentration of alkali and
swells much more highly than either of the native celluloses. The con-
clusion may be drawn that the ease and degree of swelling increase with
diminishing structural organization of the cellulose.
Heuser and Niethammer" have reported that both hydrocellulose and
oxycellulose show an increase in degree of swelling and alkali absorption
over the original cellulose. Neale21 also studied the effects of alkali on
» J. R. Katz, Physik. Z., 25, 321 (1924).
DC. DERIVATIVES OF CELLULOSE 839
these degraded forms of cellulose but showed that, if these modified forms
of cellulose are treated with hot dilute alkali, they lose their increased sensi-
tivity toward swelling and alkali absorption and again behave as normal
cellulose. The latter result is due, presumably, to removal of degraded
celluloses present in the hydrocellulose or oxycellulose.
The time of exposure to alkali is not a critical factor in the swelling of
cellulose provided the cellulose is in the form of loose fibers or thin sheets.
Under most conditions, mercerization and swelling take place almost in-
stantaneously, but this is not necessarily true in the commercial viscose
process where relatively thick sheets of pulp are generally used. Under
the latter conditions, air may be entrapped between the fibers in the sheets,
and it is important that it be expelled in order to obtain rapid and uniform
swelling. In some cases, particularly in the mercerization of unsecured
cotton fibers or highly purified wood pulps, the use of certain surface-
active agents has been proposed to promote better wetting of the cellulose.
Various mixtures of phenolic and hydroxy aliphatic compounds, several of
which are described in the patent literature, are useful for this purpose.
(1) Neale' s Theory of Swelling
Among the theories which have been advanced to explain swelling, one
of the most interesting is that suggested by Katz28 and Pauli and Valko,29
and worked out by Neale.21 According to this theory, the action of alkali
on cellulose can best be represented as an osmotic effect. This is based On
the assumption that cellulose may be regarded as a weak monobasic add
which forms a sodium salt approximately according to the laws of mass
action. The reasons why a more exact adherence to the mass action laws
is not obtained are three. First, the reacting system is not homogeneous
because of the variation in hydroxyl availability through the cellulose fiber.
Second, the cellulose molecule, being a high polymer, is polybasic, and it
contains several different types of acid structures. Third, the alkalies
used are in extremely high concentration, and the activities of the ions are
no longer even roughly proportional to their concentrations. To help
explain this theory Neale gives the following diagram:
H Na+ H Na+ Na+
A A- A A- A-
in which A is the negatively charged unit or radical corresponding to each
acid hydrogen atom. In dilute sodium hydroxide solution certain of these
18 J. R. Katz, Micellartheorie und Quellung der Cellulose, in K. Hess, Die Chemie der
Zellulose, Akadem. Verlagsgesettschaft, Leipzig, 1928.
" W. Pauli and E. Valko, Elektrochemie der Kolloide, J. Springer, Vienna, 1929.
840 CELLULOSE
hydrogens are replaced by sodium. The cations may be imagined as con-
stantly moving within the network from point to point, replacing and being
replaced by others, so that tinder given conditions a state of dynamic
equilibrium is reached between sodium ion, hydrogen ion, or undissociated
cellulose hydroxyl groups. It is believed that a similar type of equilibrium
could exist between sodium hydroxide and a sodium hydroxide-cellulose
complex. Excess alkali may be thought of as diffusing into the cellulose in
an amount determined by Donnan's thermodynamic equation for mem-
brane equilibrium. The resulting unequal distribution of ions causes os-
mosis, imbibition of water, and swelling of the cellulose until the osmotic
pressure is balanced by forces arising from the cohesion of the gel. On
-i50
4 8 12 16 20
MOLALITY OF NaOH SOLUTION
24
Fig. 39. Osmotic pressure and water absorption of cellulose film (cellophane)
from NaOH solutions of different concentrations (NealeM).
washing with water, the cellulose sodium salt is hydrolyzed, the osmotic
pressure falls, and the cellulose is recovered unchanged chemically but dis-
torted physically by an amount depending upon the maximum osmotic pres-
sure reached. The value 1.84 X 10~14 was determined as the ionization
constant of the regenerated cellulose, and Neale calculated a theoretical
osmotic pressure curve which was found to agree fairly well with the ex-
perimentally determined water absorption curve (Fig. 39). Neale, on the
basis of this .theory, was able to account for the greater swelling of cellulose
in alkali solutions of intermediate concentration and also to explain why
swelling is greater and the optimum concentration producing it is lower for
lower temperatures.
IX. DERIVATIVES OF CELLULOSE
841
If cellulose is treated with a concentrated solution of alkali and then
immersed in water or dilute alkali, it undergoes a transient swelling which
is much greater than that produced by direct treatment with dilute alkali.
Neale explained this phenomenon on the basis that a sudden rise in os-
motic pressure causes imbibition of water and rapid swelling of the solid
phase.
(2) Other Theories of Swelling
Besides the theory of Neale, several other mechanisms have been sug-
gested to explain the swelling of cellulose in alkali, but in general these have
not been accepted by all workers in the field.
100
90
80
70
3 60
§50
Q
30
20
10
2345
NORMALITY OF NaOH
Fig. 40. Solubility of cellophane in NaOH solutions
at various temperatures (Davidson80).
One of these theories states that the swelling is a result of osmotic pres-
sure, but that the pressure is due to dissolved cellulose molecules in the
fiber. Davidson30'81 demonstrated that solubility, like swelling, passes
through a maximum and that both of these maxima occur at approxi-
mately the same alkali concentration. In Figure 40 the solubility of cello-
phane at various temperatures is shown as a function of sodium hydroxide
concentration. The solubility is expressed as the ratio of dissolved to
» G. F. Davidson, /. Textile Inst., 27, T112 (1936).
" G. F. Davidson, 7. Textile Inst., 25, T174 (1934).
842 CELLULOSE
undissolved material in a large excess of aqueous alkali. The solubility
in sodium hydroxide is much greater than in potassium hydroxide and this
corresponds also to the behavior observed with the swelling.
Another explanation is that the swelling is caused by hydration; that is,
by the cohesive forces between the alkali cellulose compound and the water
molecules. With increasing concentration of alkali hydroxide, there is an
increasing competition between the free Na+ and OH~ ions and the alkali
cellulose for the water molecules; consequently, the hydration of the alkali
cellulose compound diminishes. It is interesting to note in this connection
that, according to Hess and his collaborators,82 the crystallized alkali cellu-
lose contains water of hydration. At the lower alkali concentrations, the
stable alkali cellulose I contains 3 moles of water per CeHioOg compared
with 1 mole of water per CeHioOe in the alkali cellulose II, which is stable
only at higher alkali concentrations. However, the swelling, especially
that of the regenerated cellulose, is too extensive to be attributed solely
to the amount of water of crystallization.
A third explanation is that the swelling is caused by the electrostatic
repulsion between the cellulose particles which, on account of the ioniza-
tion of the alkali cellulose, possess an excess of negative charges. At
higher alkali concentrations the free ions exert a screening effect on the
electrostatic repulsion. This is, of course, associated with the reduction
of osmotic effect considered in the discussion of the theory of Neale.
(3) Heat of Swelling
When cellulose is treated with an alkali solution of sufficient strength to
produce swelling, heat is developed. Several investigators38*84 have at-
tempted to measure the heat liberated with the object of relating^the he£t of
reaction with £he degree of Swelling and compound formation. For caustic
solutions up to 10 M, thejresults have been in fairly good agreement for
cotton cellulose and indicate that the heat of mercerization increases with
the strength of the solution, but is not proportional to it. Barratt and
Lewis86 as well as Neale21 obtained a definite deflection between 3 and 4 M
concentration indicating a sharp rise in the heat produced (Fig. 41). It
will be recognized that this range of alkali concentration coincides perfectly
with that shown to produce mercerization and formation of an alkali cellu-
" H. Sobue, H. Kiessig, and K. Hess, Z. physik. Chem., B43, 309 (1939).
91 1. Okamura, Naturwissenschaften, 21, 393 (1933).
* J. L. Morrison, W. B. Campbell, and O. Maass, Can. J. Research, B18, 168 (1940).
15 T. Barratt and J. W. Lewis, /. Textile Inst., 13, T113 (1922).
IX. DERIVATIVES OF CELLULOSE
843
lose compound/ It is also in this range that the x-ray structure has been
observed to change markedly. Barratt and Lewis85 further found that a less
marked increase in heat evolved was obtained above 10 M concentration of
alkali, but Neale21 did not confirm this and has stated that the heat evolved
continues to rise rapidly for solutions containing up to 20 M concentration
of sodium hydroxide. Neale also found that the curve representing heqt
evolved by regenerated cellulose on swelling was different from that found
Ixl
0 4 8 12 16 20
MOLALITY OF NaOH SOLUTION
Fig. 41. Heat evolved on treating various celluloses with alkali
(Neale21).
Viscose rayon.
O Cotton, soda-boiled.
A Cotton, soda-boiled and mercerized.
for cotton. With the regenerated cellulose, the initial steep rise in the
curve was obtained in solutions of less than 2 M concentration as compared
with 3 to 4 M for cotton. This was explained by the assumption of an
"accessibility" factor which was considered to be greater for the relatively
unorganized structure of regenerated cellulose than for the highly organized
structure in cotton.
Neale's21 conclusions have been criticized by Bancroft11 on the basis that
"heat of reaction" may just as well be called "heat of absorption," since the
844 CELLULOSE
thermal data obtained could be said to fit one as well as the other. It
seems possible that these differences in points of view might be recon-
ciled if it were assumed that, at the higher concentrations of alkali, the
heat evolved can be due to both absorption and chemical combination
while at lower concentrations, where no changes in x-ray diagram can be
detected, the heat evolved is due to absorption only. As is the case else-
where in this book, it appears proper to re-emphasize the fact that the
magnitude of the forces binding sodium hydroxide to cellulose hydroxyls
does not vary along the chain, but that reaction with the disorganized
portions occurs at a lower concentration due to their greater availability.
This portion of the reaction, of course, does not affect the x-ray diagram.
(4) Effects of Added Solvent on Swelling
The addition of alcohol, alone or in combination with water, has been
used by several investigators as a means of determining the nature of alkali
cellulose. Both Vieweg12 and d'Ans and Jager16 found that for a given
concentration of alkali the apparent amount of sodium hydroxide taken up
by cellulose as determined by the change-in-titer method increases with
increase in the amount of alcohol added. Vieweg concluded that with
alcohol present no chemical combination occurred between celluloses and
caustic soda. d'Ans and Jager (Fig. 42), however, obtained flat portions
in the curves for 10% and 20% alcohol and showed that the absorption
curves change position in relation to the alkali concenlration and that the
amounts of alkali absorbed at the inflection points in these curves are
greater with increasing alcohol concentration. Bancroft and Calkin,11
also using the change-in-titer method, found considerably more alkali
absorbed from 95% alcohol than from an aqueous solution of the same alkali
concentration. Champetier85* has reported that the rate of fixation of
alkali is markedly faster in a water-ethyl alcohol mixture than in either
water or ethyl alcohol alone. Also, Legrand and Grund85b found that the
concentration of alcohol influences profoundly the relative proportions of
alkali celluloses I and II which exist after equilibrium is reached. Beadle
and Stevens,20 using the blotting method, noted, on the other hand, that
both the total sodium hydroxide and the water absorbed decreased on addi-
tion of absolute alcohol.
It is believed that a comprehensive explanation for these data might be
found if it is remembered that cellulose takes up water readily, but absolute
** G. Champetter, C. Legrand, and G. Gombaud, Compt. rend., 233, 1602 (1951).
* C. Legrand and A. Grund, /. Polymer Sci., 9, 527 (1952).
IX. DERIVATIVES OP CELLULOSE
845
alcohol slightly, if at all. Also, caustic soda is much less soluble in alcohol
than in water, and, in the case of a caustic soda-alcohol-water system,
the effective concentration of alkali in the water would be far greater than
if no alcohol were present. In such a system the apparent increase in the
alkali taken up by the cellulose would be greater by the change-in-titer
method of measurement, but, because of a decrease in swelling, the over-all
absorption of alkali as judged by the blotting method might actually be
LU
I
UJ
o
d
a:
LJ
a.
o
UJ
oo
CC
o
CO
CD
I
O
<T3
O
300
200
100
0 10 20 30
CONCENTRATION OF EQUILIBRIUM
SOLUTION, VOLUME % NaOH
Fig. 42. NaOH preferentially absorbed by cellulose in the presence
of alcohol (d'Ans and Jager15). Curves: 1, 0% alcohol, 23°C.; 2,
10% alcohol, 23°C.; 3, 20% alcohol, 23°C.; 4, 30% alcohol, 23°C.;
5, 40% alcohol, 23°C.; 6, 20% alcohol, 2°C.
less. Also, if the alkali cellulose were formed first and afterward treated
with alcohol, the water taken out of the cellulose by the alcohol might be
expected to carry out appreciable amounts of caustic soda, though any
alkali preferentially absorbed should not be affected to the same extent.
(5) Effects of Salts on Swelling
Neutraf salts, particularly sodium chloride, have been added to the cellu-
lose-caustic soda-water system with the object of influencing swelling
and the amount of alkali absorbed. The reported results again are not in
846
CELLULOSE
complete agreement, but most investigators have found that for^ concen-
trations from 5 to 20%, sodium chloride decreases swelling but increases the
amount of alkali absorbed by the cellulose. Schwarzkopf,24' d'Ans and
Jager,18 and others have claimed that sodium chloride is itself not absorbed
by cellulose in the presence of alkali; and Schwarzkopf, as we have seen,
used the presence of sodium chloride as a method for determining the
amount of water absorbed. Saito7 studied the effects produced on the
amount of alkali absorbed by increasing the concentration of sodium
chloride and found that, while the apparent absorption (change-in-titer
0.20
CD
CC
3
!0.15
0.10
0.60
2
£
0.50
50 100
NaCI CONTENT (G. PER LITER)
0.40
Fig. 43. Effects of NaCI on the total and preferential absorption of
NaOH and total absorption of H2O from NaOH solution (140 g. per liter) at
20° C. (Saito7). Curves: 1, minimum true NaOH absorption; 2, prefer-
ential NaOH absorption; 3, total absorption of water.
method) increases gradually, the total absorption of alkali and water
actually decreases slightly (Fig. 43). It is believed that these findings may
be explained on the same general basis as has been used to explain the effects
of alcohol on alkali cellulose.
Joyner*6 studied the effects of potassium salts and concluded that po-
tassium chloride has a more powerful effect than sodium chloride in in-
fluencing alkali absorption. Jimbo and coworkers87 tested the effect of
sodium chloride in the preparation of alkali cellulose for the viscose process.
They concluded that small quantities, up to 2%, may exert a slight bene-
* R. A. Joyner, /. Ckem. Soc., 121, 2395 (1922).
91 S. Jimbo, T. Takazawa, and K. Tanaka, J. Soc. Chem. Ind., Japan, 37, B395 (1934).
IX. DERIVATIVES OF CELLULOSE
847
ficial influence on the properties of the regenerated yarns, but that higher
concentrations of salt cause difficulty in filtration, probably due to a reduc-
tion in the degree of swelling achieved during steeping.
(d) METAL HYDROXIDES OTHER THAN SODIUM HYDROXIDE
Although most of the work covering the action of alkalies on cellulose
has been concerned with sodium hydroxide, some attention has also been
given to the hydroxides of the other alkali metals.
10 20 30 40 50
CONCENTRATION OF ALKALI, VOLUME %
60
Fig. 44. Comparison of (A) the alkali absorbed by and (B) the swelling produced
in cellulose on treatment with solutions of various alkali metal hydroxides (Heuser
and Bartunek38).
Curve
Alkali
Concentration of
maximum swelling
CsOH
40%
RbOH
38%
KOH
32%
NaOH
18%
LiOH
9.5%
Pseudostoichiometric
ratio
(C«HioOB),-CsOH
(C6HioO6)2-KOH
(C«HioO6)rNaOH
1
2
3
4
5
Heuser and Bartunek,38 using the hydroxides of lithium, sodium, po-
tassium, rubidium, and cesium, found that each produces an absorption
curve by the change-in-titer method which contains a break interpreted
as indicating the formation of a definite compound with the cellulose (Fig.
44A). In the case of lithium and potassium hydroxides, the compounds
38 E. Heuser and R. Bartunek, Cellulosechemie, 6, 19 (1925).
848 CELLULOSE
are strictly analogous to that formed with sodium hydroxide and may
therefore be represented by the general formula (CeHioOsVMOH where
M is the alkali metal. With rubidium and cesium hydroxides, the indi-
cated compounds contain 3 moles of cellulose for each mole of hydroxide and
would therefore be represented by (C6HioO5)3-MOH. The concentrations
of alkali in solution required for the formation of these compounds increase
in the same order as the atomic weights of the alkali metals, which is: Li
< Na < K < Rb < Cs. It is also to be noted that for the first three
hydroxides the molar concentration at which the break in the absorption
curve occurs is almost identical, although the per cent concentration
increases.
Fig. 45 Photograph showing the "dumbbell-like"
swelling of microscopic sections of cotton fibers in mercer-
izing solutions (Willows and Alexander39).
The degree of swelling, determined microscopically by observations on
the cross section of the treated fiber, has also been studied for the various
metal hydroxides. However, the examination of transverse sections can
lead to errors on account of the so-called "dumbbell-like" swelling, which
was first observed by Willows and Alexander39 (Fig. 45) . This occurs with
cotton at the mercerizing concentration due to the extrusion of the cellulose
from the ends of the section during the swelling process. More reliable
results can be obtained by measuring the width of the whole fiber.
Heuser and Bartunek (Fig. 44B) found that, for each hydroxide, swelling
of scoured cotton passes through a maximum at a concentration of alkali
which corresponds closely with that producing the break in the change-in-
titer curve. The degrees of swelling at the maxima for the various alkalies
39 R. S. Willows and A. C. Alexander, /. Textile Inst., 13, T237 (1922).
DC. DERIVATIVES OF CELLULOSE 849
are in the same relative order as the degrees of hydration of the re-
spective metal ions. The ions of lowest atomic volume are associated
with the greatest number of water molecules, and in solutions of the various
hydroxides in which the MOH:H2O ratios are the same as those corre-
sponding to the fully hydrated ions, maximum swelling occurs. Both
hydration of these hydroxides and swelling of alkali cellulose increase with
decrease in temperature.
Saito7 studied the absorption of alkali and water by cellulose from solu-
tions of lithium, sodium, and potassium hydroxides by using the alkali-
neutral salt method. The data obtained agree at least qualitatively with
those found by the change-in-titer method and indicate breaks in the
curves for preferentially absorbed alkali at about the same molar concen-
trations. Swelling at the maxima was found to decrease with increase in
the molecular weight of the alkali. Saito7 also studied effects obtained
through the use of mixtures of these alkalies, taken two at a time, the re-
sults being of interest chiefly because the values found did not always lie
intermediate between the values produced by either alkali alone. Mix-
tures of lithium and potassium hydroxides, for example, produce swelling
values which pass through a maximum at a ratio of 1:1. Mixtures of
lithium and sodium hydroxides, on the other hand, produce values showing
a minimum at approximately a 1 : 1 ratio. The only mixtures behaving
normally are those containing sodium and potassium hydroxides, for which
all swelling values are intermediate between those characteristic of each
alkali alone.
(e) EFFECTS PRODUCED BY DILUTION OR CONCENTRATION OF ALKALI
SOLUTIONS
The curve for alkali absorption determined by the change-in-titer
method is not retraced exactly when a strong solution of alkali in contact
with cellulose is diluted. In the case of cotton cellulose, for example, Ban-
croft and Calkin11 found that dilution of a 5 M alkali solution gave values
for alkali absorption which were considerably greater than those obtained
either by direct treatment with solutions of equivalent concentrations or
by progressively increasing the concentration of alkali in solution (Fig. 46).
It has been similarly shown by Saito7 that, with both ramie and cotton
cellulose, progressively decreasing the concentration of alkali in solution
produces a shift in the swelling maximum in the direction of lower alkali
concentration (Fig. 47). The extent of the shift appears to be dependent
on the initial concentration of the alkali used and, within limits, increases
with this initial concentration. Evidence points, therefore! to the con-
850
CELLULOSE
23456
MOLALITY OF NaOH SOLUTION
8
Fig. 46. Preferential absorption of alkali obtained on dilution and concentration
of solutions of NaOH (Bancroft and Calkin").
Curves
Concentration
1, O
2, V
3, •
Dilution
4,X
5.+
6.Q
Vieweg (no points)
100 200 300 400
CONCENTRATION OF NaOH
SOLUTION (G. PER LITER)
500
Fig. 47. Comparison of effects on the swelling of ramie cellulose obtained
by dilution and concentration of solutions of NaOH (Saito7).
DC DERIVATIVES OP CELLULOSE
851
elusion that alkali absorption and swelling are both incompletely reversible
phenomena, although the alkali can be entirely removed if the cellulose is
washed thoroughly with water.
Saito7 also obtained data for changes in swelling when the concentra-
tion of alkali is increased stepwise. Under such conditions fibers remain
greatly swollen at the higher concentrations, although normally they are
not so highly swollen if the cellulose is introduced directly. This investi-
gator also studied the effects of varying the temperature of the alkali
er
ui
10
U-S
£o5
<* P
o
UJ
0.12
0.10
UJ
0.08 g
CD
or
0.06
0.040
0,02
14 20
TEMPERATURE, °C.
31
Fig. 48. Effects of temperature on swelling and alkali absorption of ramie
cellulose in NaOH solutions (130 g. per liter) (Saito7).
O — Decreasing temperature.
• — Increasing temperature.
X — Direct introduction of cellulose.
solution over the range 14-3 1°C. before and after adding cellulose. Swell-
ing and alkali absorption data were obtained for conditions in which the
cellulose was added directly to the solution at different temperatures, and
for conditions in which the cellulose was put in contact with the alkali
solution and the temperature subsequently either increased or decreased.
From Figure 48 it is apparent that, if the cellulose is initially introduced
into a cold solution of alkali, the amount of alkali absorbed is not changed
852 CELLULOSE
appreciably by increasing the temperature, although this is not true if the
cellulose is introduced at a higher temperature and the solution allowed to
cool. To some degree the same relationships hold for swelling. These
observations, in general, may be considered to confirm the nonreversible
character of the swelling process.
(f) EXAMINATION OF ALKALI CELLULOSE BY X-RAYS
The application of x-rays to the study of alkali cellulose has been of tre-
mendous value in helping to clarify questions regarding the nature of the
material itself and the changes which take place when cellulose is treated
with alkali solutions of different concentrations. Evidence supplied by
Hess and Trogus,40 von Susich and Wolff,41 and others indicates that
certain definite changes take place in the crystalline structure of cellulose
when it is treated with sodium hydroxide solutions. The principal rela-
tionships found by these workers may be summarized by the following
scheme:
Hydrate cellulose
HiO
12.5% NaOH 21% NaOH
Native cellulose > Alkali cellulose I < Alkali cellulose II
HjO
+H2O
Alkali cellulose III
The changes in crystal structure, while strongly suggesting the formation
of definite compounds, do not occur sharply but show a gradual transition
over a range of several per cent alkali. This, as already noted, agrees with
observations made on the alkali absorbed by cellulose from caustic soda
solutions of different concentrations. Alkali cellulose I as defined by x-ray
studies may therefore be identified with the ordinary alkali cellulose ob-
tained in 16-18% alkali, and alkali cellulose II is probably identical with
the product of mercerization as normally carried out in the stronger alkali
solutions used in the treatment of textiles. Alkali cellulose III is a third
form obtained by dehydration of alkali cellulose I.
« K. Hess and C. Trogus, Z. physik. Chem., B4, 321 (1929) ; Bll, 381 (1930).
41 G. von Susich and W. W. Wolff, Z. physik, Ckem., B8, 221 (1930).
DC. DERIVATIVES OF CELLULOSE 853
More recently Neumann42 found that if alkali cellulose I is made from a
hydrated cellulose and the concentration of alkali in solution is gradually
decreased, there is formed below about 6% alkali concentration a fourth
modification in crystal structure, which he termed alkali cellulose IV.
Following this, Schramek and Gorg48 found that, if alkali cellulose I de-
rived from hydrate cellulose is treated with caustic soda of 25-28% con-
centration, it may be converted temporarily into still another crystal form
which was termed alkali cellulose V. The latter is not stable, however,
and on standing gradually reverts to alkali cellulose II. If the alkali
concentration in contact with alkali cellulose II is then reduced, further
changes occur as indicated by a gradual transition through alkali cellulose I
and, below 6% alkali, to alkali cellulose IV. The relationships between
alkali celluloses II, III, and V have been studied thermodynamically by
K. Lauer,48a who determined the heats evolved on treating cotton and
mercerized cotton with alkali solutions in the range 18.6-40.2% NaOH.
The data obtained were explained on the basis of varying proportions of
crystalline, semicrystalline and amorphous celluloses present. Sobue44
has shown that at low temperatures where swelling is very high, two fur-
ther compounds of cellulose and sodium hydroxide exist. These have been
designated as alkali cellulose VI and alkali cellulose Q. Sobue has sum-
marized the interrelationships of most of the known alkali cellulose com-
pounds as revealed by x-ray methods (Fig. 49) .
The importance of swelling in the formation of the various crystalline
modifications of alkali cellulose detectable by x-rays is shown by the fact
that if hydrate cellulose is used, the conversion to alkali cellulose I takes
place at a lower concentration of alkali than is the case with native cellu-
lose. Moreover, alkali cellulose I, made from hydrate cellulose, does not
convert to alkali cellulose II in 20-25% NaOH but only in solutions more
concentrated than 28% NaOH. Observed differences in the alkali con-
centration ranges which give rise to a given lattice structure under con-
ditions of increasing caustic concentration, as opposed to decreasing caustic
concentration, are also most readily explained on the basis of swelling.
Further evidence of the importance of swelling on change in crystal
structure is the fact that alkali cellulose I is not obtained in water-free
solutions of sodium hydroxide-methanol which are known to be poor
swelling agents for cellulose. Tension, which is also known to affect swell-
41 H. Neumann, Dissertation, Dresden, 1933.
« W. Schramek and H. Gdrg, KollM-Bcihcfle, 42, 302 (1935).
*» K. Lauer, Makromol. Chcm,, 7, 5 (1951).
« H. Sobue, /. Soc. Chem. Ind., Japan, 43, B24 (1940).
854
CELLULOSE
ing, has been found to affect changes in the x-ray diffraction diagram. On
the basis of this relationship, Hess and Trogus40 were able to show that ramie
under tension could be treated with 30-35% caustic soda and subsequently
washed without any evidence of a hydrate being formed. It may be
inferred, therefore, that swelling is absolutely necessary in order for the
alkali to penetrate into the cellulose lattice. Also, the degree of pre-
swelling definitely helps to determine the amount of alkali absorbed and
the crystal structure formed when cellulose is treated in an alkali solution
of a given concentration.
-20
0 10 20 30
CONCENTRATION OF EQUILIBRIUM
SOLUTION, WEIGHT % NaOH
Fig. 49. Changes in the x-ray diffraction diagrams of alkali cellulose due to tempera-
ture and concentration of alkali in solution (Sobue44).
® — Alkali cellulose III H — Native cellulose
O --Alkali cellulose I X —Alkali cellulose II
• —Alkali cellulose Q V— Alkali cellulose V.
The most recent analysis of the physical changes which occur in cellulose
as a result of mercerization is contained in three papers by R&nby.45
RSnby studied the mercerization of cotton linters versus wood pulp fibers
using water sorption, x-rays, and electron diffraction methods to establish
the different alkali concentration ranges required to produce marked
changes in water sorption and fine structure. He found that increased
water sorption precedes changes detectable by x-rays or electron diffraction
and that changes in wood pulps normally occur at concentrations of alkali
approximately 2% lower than those producing similar changes in cotton
cellulose. For native cellulose water sorption is probably a surface phe-
nomenon since water does not enter the lattice, but with mercerized cellu-
« B. G. R&nby, Acta Chem. Scand., 6, 101 (1952).
DC. DERIVATIVES OF CELLULOSE 855
lose, water is absorbed both intermicellarly and intramicellarly. The con-
clusion is drawn that the alkali concentration necessary for transition is
related to the water sorption of the original native cellulose fibers as both
processes involve reaction with hydroxyl groups. Differences observed be-
tween celluloses are thus in line with the hydroxyl accessibility in the native
state. R&nby also found that the micelle strings in wood cellulose are 10-
20 A. thinner than those of cotton cellulose, a factor which further increases
the active surface in favor of wood cellulose.
R&nby also obtained data on the adiabatic mercerization of cellulose and
concluded that the transformation of wet native cellulose to wet mercerized
cellulose is an exothermic and natural process at room temperature and
atmospheric pressure. Wet native cellulose is considered the unstable
state under these conditions assuming that the cellulose fibers contain both
crystalline and noncrystalline substances. The relative stability of native
versus mercerized cellulose is known to be affected by temperature. Al-
though cellulose II can readily be formed from cellulose I at ordinary tem-
peratures, the work of Kubo46 and Hermans47 has shown that the reverse
process can only occur at high temperatures. From these considerations,
R&nby has concluded that cellulose crystals are built up in vivo as the chains
and fibers grow, leading to lattice cellulose I, rather than synthesized as
individual chains which later aggregate and crystallize in vitro in the pres-
ence of water. He has further suggested that both chains and crystals form
under specific enzyme action.
(g) CHEMICAL REACTIONS OF ALKALI CELLULOSE
Although the consensus of opinion favors an addition compound as the
most likely structure for alkali cellulose, its properties, from a purely
chemical standpoint, are consistent with those of a metal alcoholqte derived
from a polymeric carbohydrate containing reactive hydroxyl groups.
Alkali cellulose reacts with halogen compounds to form ethers :
Rceii(OH)-NaOH + RC1 > Rceii(OR) + NaCl + H2O
Alkali cellulose also reacts with acid anhydrides and acid chlorides to form
esters :
Rceii(OH)-NaOH + ROC1 > Roeii(OOR) + NaCl + H2O
Undoubtedly, the most important ester prepared from alkali cellulose is
sodium cellulose xanthate obtained through reaction with carbon disulfide:
Rc.n(OH)-NaOH + C& > Rc.u(OCSS-Na) + H2O
<• T. Kubo, Kolloid-Z., 88, 62 (1939) ; 93, 338 (1940); 96, 41 (1941).
47 P. H. Hermans, Physics and Chemistry of Cellulose Fibres, Elsevier, New York,
1949, pp. 155-156.
856 CELLULOSE
Each of these reactions is important in itself, both chemically and from the
commercial point of view, but they will not be discussed in detail here as
they are described thoroughly in Sections E, C, and F, respectively, of this
Chapter IX.
(1) Aging of Alkali Cellulose
The reaction of alkali cellulose with oxygen or oxygen-producing com-
pounds is of a somewhat different character from its other reactions in that
the most important effect produced is a degradation of the cellulose brought
about by combination with oxygen and subsequent splitting of the chain
molecules. Cross, Bevan, and Beadle,4 the discoverers of the viscose proc-
ess, recognized that if pressed alkali cellulose is allowed to stand for any
appreciable length of time, changes take place resulting in a decrease in the
solution viscosity of the cellulose. This process, known as "aging," has been
the subject of much study from both the practical and theoretical points of
view.
Although certain early workers ascribed the effects accompanying aging
solely to the action of the alkali, most of the evidence has indicated that
oxygen of the air in the presence of the alkali is an essential condition of the
aging process. As early as 1906, Margosches,48 and later Ost49 (1911),
recognized oxidation as having an important part in the aging process with-
out, however, investigating in detail the mechanism involved.
In addition to the decrease in viscosity, other major changes which cellu-
lose undergoes during aging are an increase in the alkali solubility, a slight
increase in reducing power, and the formation of carboxylic acid groups.
The extent to which these changes occur depends primarily upon the oxygen
concentration, the temperature, and the duration of the aging, but is also
affected by the origin and previous history of the cellulose, the concentra-
tion of the steeping solution, the press ratio (weight of alkali cellulose to
cellulose), and the presence or absence of certain oxidation catalysts or
inhibitors.
In the past, some investigators have minimized the influence of oxygen
and assumed that aging is mainly a continuation of physical changes
initiated by swelling of the cellulose in a strong solution of caustic soda.
It has been postulated that swelling results in a breakdown of secondary
valence forces holding the micelles or cellulose chains together. However,
48 B. M. Margosches, Die Viskose, 2d ed., Verlag der Z. gts. Textil-Ind. (Klepzigs),
Leipzig, 1906.
49 H. Ost, F. Westhoff. and L. Gessner, Ann., 382, 340 (1911).
DC. DERIVATIVES OF CELLULOSE 857
it has been shown50 that the changes undergone during aging are very dif-
ferent in character from those occurring during the swelling of merceriza-
tion. For example, during mercerization the cellulose swells and, because
of the increase in internal surface thus created, becomes more reactive
toward all types of chemical and physical attack, but undergoes almost no
change in degree of polymerization. During aging, on the other hand,
chemical activity, as measured by rate of hydrolysis or xanthate formation
does not increase, although the solution viscosity is greatly affected.
New light on the mechanism of the aging process was obtained when
Waentig61 showed that no aging takes place if air is excluded by covering
the cellulose with liquid caustic soda. After this observation, various
studies were made with the object of determining more exactly the rate of
oxygen absorption and the role played by oxygen in the aging process.
The first investigators to measure carefully the amount of oxygen absorbed
were Weltzien and zum Tobel,52 who found that alkali cellulose prepared
from an 18% solution of caustic soda reached a maximum absorption of
140 cc. of oxygen per g. of cellulose after exposure for 360 hrs. at 60°C.
These investigators also found that the absorption of oxygen passes through
a maximum by steeping in solutions of about 25% caustic soda. The con-
ditions used in these experiments were, of course, extreme but indicate
clearly that cellulose can be oxidized to a considerable extent.
Among the products of oxidation found by Weltzien and zum Tobel were
several acidic compounds including considerable amounts of carbon dioxide.
These products neutralized much of the alkali originally present and caused
the oxygen consumption to slow down toward the end of the experiment.
This indicated that the concentration of alkali present is also of importance
in the aging process. The same investigators studied the aging of alkali
cellulose in an atmosphere of nitrogen or hydrogen, or covered with an
organic liquid which prevented contact with the atmosphere. Under these
conditions, oxidation was greatly retarded but aging was not eliminated
completely. Banderet and R&nby5* have used this type of experiment as
a means of studying cellulose structure. They observed a small but definite
decrease in the molecular weight of cellulose treated with alkali in the ab-
sence of oxygen and concluded that the glucose residues of the cellulosic
80 O. Faust and P. Karrer, Hclv. Chim. A eta, 12, 414 (1929) ; S. M. Lipatov and N. A.
Krotova, Iskusstvennoe Volokno, 7, 1112 (1930); S. M. Lipatov and E. Ya. Vinetskaya,
Iskusstvennoe Volokno, 5, 2; 6, 2 (1931) ; Z. A. Rogovin and M. Schlachover, Iskusstoen-
noe Volokno, $, 3 (1933).
il P. Waentig, Kvlloti-Z., 41, 152 (1927).
M W. Weltzien and G. zum Tobel, Ber.t 60B, 2024 (1927).
" A. Banderet and B. Rdnby, Hclv. Chim. Acta, 30, 1190 (1947).
858 CELLULOSE
chain are joined principally by /3-glucosidic bonds, but that cellulose must
also contain another type of bond, possibly ester groups which are readily
broken by alkali alone.
Staudinger and Jurisch64 showed that in a high vacuum practically no
aging of alkali cellulose occurred even after standing for 8 days at 20°C.
arid that aging was very slight after a similar treatment at 100°C. Stau-
dinger followed the changes occurring during aging by measuring the vis-
cosity of cellulose solutions in low-concentration cuprammonium hydroxide
solvent. The conclusion was drawn that the changes in alkali cellulose
produced by aging are due entirely to degradation of the cellulose molecule
by atmospheric oxygen. It Was also shown that the oxygen consumed
during normal aging increases with the amount present, but aging will pro-
ceed with very small concentrations. For a given concentration, the
amount of oxygen consumed is linear with time and is quite small for ordi-
nary alkali cellulose. The curve for aging time versus viscosity is a hyper-
bola, and the logarithm of the viscosity forms a straight line when plotted
against the logarithm of the time.65
The mechanism of the degradation of alkali cellulose by oxygen is so
complex that early explanations were necessarily incomplete or erroneous.
Now that the course of so many oxidations has been clarified in terms of a
free-radical mechanism, it appears logical to apply the same concepts to
alkali cellulose. A very promising start in this direction has been made by
Entwistle, Cole, and Wooding.56 They have found that the initial course
of the reaction is governed by the nature of the easily oxidizable end-groups
and impurities that may be present. After the initial period, all types of
cellulose tend toward the same rate of degradation. Both the initial and
the later stages of the oxidation can be influenced by the usual oxidation
catalysts and inhibitors. Of course, the strongly alkaline nature of the
medium has to be kept in mind when parallels with other oxidations are
drawn. (The kinetics of the aging of alkali cellulose are discussed in more
detail in Chapter III-C-1.)
(a) Practical Applications of the Aging Reaction. From the technical
standpoint, it is more important to appreciate the influences of certain
chemical and physical factors affecting the aging process than to under-
stand fully the chemical mechanism of the process. Aging is principally of
interest as a method of controlling solution viscosity, and changes in vis-
14 H. Staudinger and I. Jurisch, Zellstoffu. Papier, 18, 690 (1938).
» O. Eisenhut, /. prakt. Chem.t 157, 338 (1941).
* D. Entwistle, E. H. Cole, and N. S. Wooding, Textile Research /., 19, 527, 609
(1949).
IX. DERIVATIVES OF CELLULOSE
859
cosity are the criteria for estimating the degree of aging produced. The
degree of aging may be determined conveniently by a cuprammonium or
cupriethylenediamine viscosity measurement of the regenerated celluloses
or by xanthation of the aged alkali cellulose and measurement of the vis-
cosity of the resulting viscose. The effect of time on aging and viscosity
has already been mentioned and may conveniently be followed by means of
a log-log curve. Temperature, as might be expected, has a marked effect
on aging, and the rate of change in viscosity at 65°C., for example, is ap-
proximately twenty times that at 25°C. (Fig. 50).
240 r
"0 10 20 30 40 50 60 70
Fig. 50. Effect of temperature on rate of aging of alkali
cellulose (courtesy of Conaway).
Ordinate: Cuprammonium viscosity in centipoises.
Abscissa: Aging time in hours.
The type of cellulose affects the rate of aging. It appears probable that
the major reason is that differences in swelling characteristics of the dif-
ferent types of cellulose affect the amount of surface available for reaction
with oxygen. The presence in cellulose of certain metallic impurities is
also known to be important because they can act as catalysts for the oxida-
tion reaction. In any case, it is usually necessary in the viscose process to
determine the rate of aging for each new type of cellulose used. Recently
Mitchell57 has shown by use of an accelerated aging test (2 hrs. at 50°C. in
w R. L. Mitchell, Ind. Eng. Chem., 43, 1786 (1951).
860
CELLULOSE
oxygen) that differences in aging requirement of different pulps may be ex-
pressed in terms of two factors, the initial degree of polymerization and the
degradation rate. Figure 51 illustrates this in the form of straight-line
relationships on a semilogarithmic plot.
500
IOOO
NITRATE OP OF CELLULOSE
1500
Fig. 51. Relationship of degree of polymerization (D.P.) to aging requirement at
various degradation increase (D.I.) levels by accelerated test (Mitchell87). Degradation
increase occurs when the alkali cellulose prepared by steeping in 18% NaOH is aged 2
hrs. at 50°C. in oxygen.
DC. DERIVATIVES OF CELLULOSE
861
The concentration of alkali in the steeping liquor also affects aging, which,
as has been shown, increases with increase in the concentration of caustic
soda up to about 25% NaOH. However, under commercial conditions in
the viscose process, this variable is not generally encountered, since the
solution used almost always contains between 17 and 18% NaOH. As the
press ratio is decreased, the time required to reach a given viscosity in-
50
48
60
80
90
70,
72 96
AGING TIME IN HOURS, 21°C.
100 110
Fig. 52. Effect of iron in steeping caustic on rate of aging of alkali cellulose
(Hooker, Ritter, and MacLaren*8).
Curve 1234 56
Parts Fe per million parts NaOH 128 64 32 16 9.6 4
creases. It is not known whether this effect is due to the decrease in the
total alkali contained in the pressed alkali cellulose, or to the fact that the
alkali cellulose is generally shredded less efficiently at low press ratios.
Before this discussion of aging is concluded, mention should be made of
» A. H. Hooker, B. H. Ritter, and S. F. N. MacLaren (to Hooker Electrochemical
Co.), U. & Patent 2,079,120 (May 4, 1937); Chem. Abstracts. 31, 4495 (1937).
862 CELLULOSE
the effects produced by certain metal ions which may be present, either in
the pulp or in the steeping caustic, and which exert a catalytic effect on the
oxidation process. Iron has long been recognized as a factor affecting the
rate of aging, and Hooker and his coworkers68 give data on the quantitative
relation between iron content in the steeping caustic and the rate of vis-
cosity reduction during aging (Fig. 52). The effect of manganese is almost
ten times that of iron,69 and nickel has also been shown to catalyze aging;
copper inhibits the reaction slightly. Cowling60 found that manganese
apparently increases the efficiency of the oxidation reaction and produces a
greater loss in viscosity for a given amount of oxygen than can be obtained
without manganese present. Other substances known to accelerate the
rate of aging are sodium sulfide and certain materials of organic origin such
as polyhydric alcohols. Phenolic-type compounds, on the other hand, act
as antioxidants and are known to retard the aging process. Finally, vari-
ous active oxidizing agents such as peroxides and hypochlorites may be
added to the steeping caustic to accelerate the aging process.
(2) Industrial Preparation of Alkali Cellulose
Production of alkali cellulose in the viscose rayon industry is normally
accomplished by steeping the cellulose in the form of sheets set on edge in
a large horizontal tank equipped with a powerful hydraulic ram movable
lengthwise through the tank. Both the steeping and pressing out of the
excess caustic soda solution are accomplished in the one piece of equipment.
The time of steeping normally is from 30 min. to 1 hr., after which the ex-
cess liquor is drained off, the ram is started, and the sheets of alkali cellu-
lose are pressed to the desired degree, usually to about three times the origi-
nal weight of the cellulose.
After the alkali cellulose is formed, the next step in the process is to shred
or disintegrate the pressed sheets so as to produce a material having the
large surface essential for uniform reaction with carbon disulfide. Shredding
is normally done in large disintegrators of the so-called Weraer-Pfleiderer
type equipped with heavy rotating arms operating against a serrated saddle.
The shredders are jacketed for circulation of hot or cold water in order to
control temperature, because aging is already under way at this stage as a
result of removal of the excess caustic liquor by pressing. After the shred-
ding operation, the aging process proper is completed by storing the alkali
cellulose crumbs in metal or fiber containers under conditions of constant
» F. E. Bartefl and H. Cowling, Ind. Eng. Chem., 34, 607 (1942).
« H. Cowling, Dissertation, Michigan, 1939.
DC. DERIVATIVES OF CELLULOSE 863
temperature. For a fuller discussion of this process, see Section 4F of
Chapter IX.
Except for a general increase in the size of the equipment units iised in
the foregoing operations, few fundamental changes have been introduced
in the methods of making alkali cellulose since the early days of the indus-
try. Proposals have been made61 to modify the process to make it conform
more nearly to modern continuous methods of manufacture, but progress
in this direction is slow. One such method62 starts with the pulp in the
form of large rolls in place of sheets, and the continuous web of cellulose
is led through a long steeping tank at the end of which the excess liquor is
squeezed out between closely set rolls having a uniform ckarance. The
pressed alkali cellulose is then immediately passed in festoons through a
large enclosed chamber where the aging is accomplished quickly at a some-
what higher temperature than that used in the standard process. In the
final step of the process, the web of alkali cellulose is broken up by putting
it through a shredding machine consisting essentially of a barbed roller
which tears the pulp apart uniformly. An alternative procedure is to
shred the alkali cellulose before aging and to accomplish aging either by
passing the shredded material in the form of a loose blanket through an
oven, or by passing it slowly through a cylinder of large diameter which
rotates on a fixed axis somewhat after the manner of a cement kiln. Alkali
cellulose can also be produced by a slurry steeping process with either
water-wet pulp taken directly from the pulp mill or dried sheets which are
beaten up in a tank with a mercerizing solution of caustic soda. This
method offers the possible advantage that it insures a uniform reaction
between the cellulose and caustic soda. Removal of the excess alkali may
be accomplished conveniently by centrifuging, by passing the material
through spaced rollers, or by formation of a web such as can be done on a
paper machine. The pressed alkali cellulose may then be disintegrated by
any of the methods discussed above.
2. Mercerization of Cellulosic Textiles68
Mercerization, the term applied to the treatment of cotton yarns and
fabrics with concentrated aqueous alkali, is carried out by the textile indus-
try to effect desirable improvements in luster, dye affinity, and strength.
81 A. Faberj, Seta Artificial, 4, 3 (1932) ; O. Faust, Z. Ver. deut. Ing., 80, 981 (1936).
w F. Steimmig, German Patent 604,015 (Oct. 17, 1934) ; Chem. Abstracts, 29, 926
(1935).
88 This section is in part a revision of pertinent material in that on "Mercerization"
prepared by E. I. Valko for the first edition.
864 CELLULOSE
In addition, secondary changes in ''handle/1 uniformity, elasticity, and
chemic|J reactivity of material may often be important. Although several
of thSJl properties may be of interest to the mercerizer, it is not possible to
obtain tht maximum degree of improvement in all properties simultane-
ously. The greatest luster is obtained by mercerizing with cotton under
maximum tension, but this condition reduces the improvement which can
be expected in dye affinity and elasticity of the material.64
Mercerization continues to be an important commercial process, even
after many decades of operation. During this period several hundred in-
vestigations have been carried out to elucidate the mechanism of the
mercerization process. This work has enriched the knowledge of fiber struc-
ture but has not greatly influenced industrial practice. The literature on
mercerization has been reviewed by Clibbens,1 Valko,66 and Marsh,66 and
a bibliography has been compiled by Edelstein and Cady.67
(a) EFFECT OF MERCERIZATION ON COTTON FIBERS
Studies of mercerization processes in connection with (a) the mechanism
of the chemical interaction of cellulose and alkalies, and (b) changes of the
fiber during the treatment with caustic and during subsequent washing, are
discussed earlier in this section and in Chapter IV-B. Other aspects of the
mercerization process and permanent changes in the properties of the fiber
produced thereby will be given here.
(1) Dimensional Changes of Fibers in Alkali
When cotton fibers are treated with alkali, lateral swelling and longitu-
dinal shrinkage occur. Both of these effects pass through a maximum when
plotted against alkali concentration, and, if sodium hydroxide is the alkali,
the maxima are at approximately the same concentration. For other
alkalies, no general correlation between diameter and length changes of the
fiber has been established!68 The concentration of sodium hydroxide neces-
sary to obtain a given degree of swelling depends on the external surface of
the fiber. Calvert69 has shown that "scoured" cotton (pretreated with hot
dilute alkali) swells more than raw cotton whereas "rubbed" fibers having
" S. M. Edelstein, Am. Dyestuff Reptr., 25, 724 (1936).
M E. L Valko, KMoidchemische Grundlagen der Textilveredlung, J. Springer, Berlin,
1937.
* J. T. Marsh, Mercerizing, Chapman & Hall, London, 1941.
« S. M. Edelstein and W. H. Cady, Am. Dyestuff Reptr., 26, 447 (1937).
« G. E. Collins, /. Textile /»$*., 16, T123 (1925) .
« M. A. Calvert, /. Textile Inst., 21, T293 (1930).
IX. DERIVATIVES OF CELLULOSE
15 20
NaOH,
Fig. 53. Increase of the width of cotton fibers in NaOH
solutions (Calvert69). Lower curve, raw hairs; middle curve,
scoured hairs; upper curve, rubbed hairs.
13.3
22.3
26.8
17.8
NaOH, %
Fig. 54. Longitudinal shrinkage of scoured cotton fibers when treated
under various tensions with NaOH solutions (Calvert69).
the cuticle (exterior surface) partially removed by an abrasive, swell even
more (Fig. 53). Corresponding effects of alkali concentration on length
were also obtained by these modifications of the fiber.69 Usually, however,
866 CELLULOSE
in tfie Commercial process cellulose yarns and fabrics are treated under
tenxbn so that longitudinal shrinkage of the fiber is greatly restricted (Fig.
54).
(2) Molecular Changes and Crystal Structure
X-ray examination of native and mercerized cotton fibers indicates that
rnercerization changes the form of the crystallites (discussed in detail under
"Examination of Alkali Cellulose by X-Rays" above and in Chapter IV)
and increases the amount of amorphous cellulose at the expense of the
crystalline material. Hermans and Weidinger70 have developed a method
for the quantitative evaluation of the crystalline fraction in cellulose fibers
from intensity measurements on x-ray diffraction patterns. They have
shown that mercerized cellulose has considerably less crystalline material
than native cellulose. The effect of mercerization on degree of crystallinity
can also be demonstrated by chemical degradative methods.71 The degree
of orientation characteristic of the crystallites in native cellulose can be
either retained by application of tension as is usually done in commercial
practice or increased by stretching the fibers. Without tension, orientation
of the cellulose crystallites decreases.72
It is thought that most of the technically important changes in fiber
properties occurring during mercerization do not depend on basic differ-
ences in the lattice structure between native and mercerized cellulose.
Rather, they can be traced to the changes in the amount of crystalline ma-
terial and to changes in the orientation of the crystallites.
Degradation of the cellulose molecule occurs in strong alkali under cer-
tain conditions. It has been demonstrated,73 however, that fiber properties
dependent on the state of polymerization remain essentially unchanged by
mercerization. Staudinger74 found that the degree of polymerization does
change from 3000 to 2000 during mercerization, but this amount is in-
appreciable when considered in connection with mechanical properties of
fibers.
(3) Changes in Tenacity and Extensibility
Native cellulose fibers such as cotton, flax, and ramie have high break
strength and low extensibility, properties which are associated with high
70 P. H. Hermans and A. Weidinger, /. Polymer Sci.t 4, 135 (1949).
71 R. F. Nickerson, Ind. En°. Chem.t 34, 85, 1149, 1480 (1942); 39, 1507 (1947);
G. Goldfinger, H. Mark, and S. Siggia, Ind. Eng. Chem., 35, 1083 (1943); M. L. Nelson
and C. M. Conrad, Textile Research /., 18, 140 (1948).
72 G. L. Clark, Ind, En*. Chem., 22, 474 (1930).
78 B. P. Ridge, H. L. Parsons, and M. Corner, /. Textile Inst., 22, T117 (1931).
74 H. Staudinger and A. W. Sohn, /. prakt. Chem., 155, 177 (1940).
IX. DERIVATIVES OF CELLULOSE 867
orientation and high crystallinity. Mercerization can cause large changes
in these properties. Mercerization under tension causes an increase in
tensile strength of cellulose yarns and fabrics as Langer75 found when he
mercerized scoured cotton yarn and increased its break strength 34.8%.
Midgley™ obtained data which suggest that the increase in strength of
mercerized yarn is due to strengthening of individual cotton fibers and to
increased cohesion between fibers. Also Edelstein64 showed that variation
of tension above that required to prevent shrinkage caused no further sig-
nificant increase in break strength.
The quantitative data on the effect of mercerization without tension on
break strength are quite contradictory even in the case of cotton fibers
which does not involve the complicating effects of yarn and fabric construc-
tion. For example, Barratt77 found that mercerization of scoured Egyptian
cotton fiber without tension lowered breaking load from 7.2 g. to 6.7 g.
Greenwood78 found no conclusive evidence that fibers from mercerized
yarns lost strength. On the other hand, Clegg79 found increases in breaking
strength of 11.8% to 49.2%.
The effect of tension during mercerization on the extensibility of cellulose
yarns and fibers is quite marked. Barratt77 showed that extensibility of
cotton fibers mercerized without tension increased from 7.4% to 12.2%.
In the case of cotton yarn Edelstein64 found that increasing tension from the
minimum value preventing shrinkage to the maximum tension which could
be applied without causing breakage reduced extensibility from 5.2% to
3.5%.
(4) Change in Luster
The commercial mercerizing process usually has the improvement of
luster as its primary objective. The luster of cotton fabrics has been shown
to be closely connected with the shape of the fiber cross section, amount of
twist, alignment of the fibers in the yarn, fabric construction, and, above all,
tension of the material in mercerizing solution. In caustic solution native
cellulose fibers (cotton) swell immediately and their cross sections become
first elliptical and then almost circular. Simultaneously, they untwist,
and convolutions are eliminated. The cellulose wall swells inwardly, and
the lumen almost disappears. When the caustic is washed out the cross sec-
» K. Langer, MeiUand Textilber., 15, 165 (1934).
» E. Midgley, Textile World. 87, 1382 (1937).
77 T. Barratt, /. Textile Inst., 13, T21 (1922).
78 R. S. Greenwood, /. Textile Inst.t 10, T274 (1919).
79 G. G. Clegg, J. Textik Inst.. 15, T6 (1924).
868 CELLULOSE
tion shrinks but retains a circular form and small lumen. These stages of
swelling are shown in Figure 55.80 Adderley81 showed that dimensional
changes of this nature have an important influence on luster. He demon-
strated that the luster of cotton increases as the cross section of the fibers
approaches a circular form in the case of both native and mercerized cellu-
lose. *
No improvement in luster is obtained when mercerization is carried out
without tension despite the fact that the resulting fibers are smoother than
native cellulose and the convolutions have disappeared. The development
of luster under tension is apparently due to further constriction of the
a
O o
c d
Fig. 55. Changes of the form of the cross section of cotton fibers dur-
ing mercerizing (Calvert and Summers80). Key: (a) fully collapsed
hair; (b) same, swollen in 18% NaOH solution; (c) same, washed with
water; (d) same, washed and dried.
lumen, formation of a more perfectly cylindrical cross section, and elimina-
tion of surface imperfections from the fibers. It is possible that removal of
voids plays a part in this process. Voids (a few tenths of a micron in
diameter) are thought to cause dullness in viscose rayon yarn, and, if pres-
ent in or near the primary wall of cotton fibers, they would tend to collapse
during swelling under tension. This would produce a more lustrous effect.
In order to explain the changes in the fiber dimensions described above,
it seems necessary to assume that (a) the fiber skin exerts a mechanical
restriction on the extent and direction of swelling in alkali, and (b) the ar-
rangement of the cellulose chains, which was responsible for the collapse
» M. A. Calvert and F. Summers, /. textile Inst.. 16, T233 (1025).
w A. Adderley, /. Textile Inst.. 15, T195 (1924).
DC. DERIVATIVES OF CELLULOSE
of the native hair, is broken up. The mechanism by which the applied
tension during swelling prevents the folding of the fiber skin during sub-
sequent drying operations is not clear.
For materials to give the best luster after mercerization under tension,
it is essential that the yarn be properly constructed. * Corser and Turner88
showed that low twist appears to be essential for maximum luster in a yarn.
It is thought that double yarns, which are more lustrous than single yarns,
are superior in this respect because it is possible to prepare them with lower
twist and greater paralldization of the cotton fibers to the axis of the yarn.
(5) Changes in Absorptivity and Reactivity
Since the mercerization process decreases the amount of crystalline ma-
terial and increases the separation of the cellulose chains, the resulting fibers
have greater absorptive capacity and are more reactive. Moisture regain,
imbibition of water, absorption of metal hydroxides or dyes from dilute solu-
tions, as well as rates of attack by hydrolyzing and oxidizing agents, are
greatly increased. These effects depend on mercerization conditions such
as concentration of sodium hydroxide, tension, and temperature; conse-
quently many tests for determining the degree of mercerization are based
on absorptivity and reactivity.
Although it was long recognized that mercerized cellulose has a higher
equilibrium moisture content than native fibers, Urquhart and Williams88
for the first time made exact measurements under strictly comparable con-
ditions with native cotton (kier-boiled) and the same mercerized without
tension. They found the ratio of water contents of mercerized and native
cottons to be about 1.5 which was nearly constant for all relative humidities.
Figure 56 shows how this ratio depends on the tension during the merceriza-
tion and on the concentration of the sodium hydroxide solution.
Vieweg84 observed that sodium hydroxide is absorbed from dilute alkali
solution more strongly by mercerized cotton than by native. Neale85
measured the ratio of sodium ions absorbed from 0.5% NaOH by native
and mercerized celluloses in order to study the effects of specific merceriza-
tion variables. Edelstein88 and Neale85 also used barium hydroxide for
measuring absorption ratios. The copper number of the fiber after treat-
w EL K. Corser and A. J. Turner, /. Textile Inst., 14, T332 (1923).
"A. R. Urquhart and A. M. Williams, /. Textile Inst.. 16, T165 (1925); 18, T66
(1927).
" W. Vieweg, Ber., 40, 3876 (1907).
* S. M. Neale, /. Textile Inst., 22, T320, T349 (1931).
» S. M. Edelstein, Am. Dyestuff Reptr., 25, 1186 (1936).
870
O.S
CELLULOSE
a
r*~'
^ "^
S^
6
«£$SB*d
.
ff
j
-^^>l
•^s«
10 20 30 0 10
NaOH, %
20
30
40
Fig. 56. Ratio of the amount of water absorbed by cotton which was treated with
NaOH solutions of various concentrations, (a) without tension and (b) under tension,
to the amount of water absorbed by untreated cotton (Urquhart and Williams88).
Adsorption desorption
ment with (a) 5% H2SO4,87 or (b) alkali hypobromite solution88 is greatly
increased by the mercerization process and has been used to characterize
reactivity of fibers.
The affinity of cotton for dyestuffs is increased by mercerization.
Knecht89 found that mercerized cotton absorbed almost twice as much dye
from a given dye bath as native cotton. He also investigated the effect of
tension during mercerization and drying of the fiber before dyeing and
showed that these procedures reduce the effects of mercerization on the
affinity for dyestuffs. Boulton and Morton90 found that the rate of ab-
sorption of dyes is greatly increased by mercerization. Htibner91 stated
that the same shade of color is obtained if equal amounts of mercerized and
native cotton are treated with the same weight of dye. Many workers in
the field, however, feel that mercerized cotton requires less dye than raw
cotton to produce a given shade. No quantitative tests have been reported
since Hiibner, and it must be assumed the apparent economy in dyestuffs
is associated with the ability of mercerized cellulose to exhaust the dye bath
more than native cotton. Also the increased luster would help to produce
a higher brilliance of color.
(b) OPTIMUM CONDITIONS FOR MERCERIZATION
A summary of the variables which are of interest from the commercial
point of view and which affect the ratio of native to mercerized cellulose
in the finished product includes the previous treatment given the fibers,
time and temperature of treatment, concentration &f caustic, and tension
87 C. G. Schwalbe, 2. angew. Chem., 22, 197 (1909).
88 C. Biitwell, D. A. Clibbens, A. Geake, and B. P. Ridge, J. Textile Inst.t 21, T85,
(1930).
89 E. Knecht, /. Soc. Dyers Coburists, 24, 68, 107 (1908),
90 J. Boulton and T. H. Morton, /. Soc. Dyers Colourists, 56, 145 (1940).
91 J. Hfibner, /. Sac. Chem. Ind. (London), 27, 105 (1908),
IX. DERIVATIVES OF CELLULOSE 871
on the fibers while they are in the swollen stage. Under ordinary circum-
stances with concentrations of 20-25% NaOH there is no advantage in the
use of temperatures below 25°C.M'92 The fibers are generally not in con-
tact with the mercerizing solution more than one minute, and in order to
get complete penetration of the fabrics or yarns in such a short time, scoured
cotton is usually employed. If raw cotton is used, a wetting agent must
be added to the bath. The tension applied to the material in the caustic
solution is selected on the basis of the property (luster, dye affinity, strength,
or extensibility) particularly desired in the finished product.
3. Metal Alcoholates of Cellulose
(a) PREPARATION AND PROPERTIES
The discovery in 1931 by Scherer and Hussey93 of a method for preparing
the trisodium alcoholate of cellulose by treating cotton linters with metallic
sodium in liquid ammonia created considerable interest and speculation in
cellulose chemistry, since it opened a new frontier in the investigation of the
metal derivatives of cellulose. Kraus and White94 had shown previously
that when simple monohydric alcohols are treated with metallic sodium in
liquid ammonia, sodium alcoholates are formed with the liberation of an
atom of hydrogen for each atom of sodium consumed, according to the
following reaction :
ROH + Na > RONa + l/2 H2
The formation of the alcoholates of the simple alcohols involves reaction in
a homogeneous system, whereas the comparable reaction of fibrous cellulose
materials with sodium in liquid ammonia involves a heterogeneous system.
A quantitative study of the formation of the sodium alcoholate of cellulose
by Scherer and his students showed that a mole of dry cellulose reacts with
a maximum of three atoms of sodium, although any degr.ee of substitution
less than three can be obtained with the liberation of an equivalent amount
of hydrogen. Schorigin and Makarowa-Zemljanskaja95 have substantiated
Scherer's results. Schmid and his co workers,96*97 as well as Muskat,98 pre-
pared the metal alcoholates of simple carbohydrates such as glucose, inulin,
M A. R. Urquhart, /. Textile Inst., 18, T55 (1927).
M P. C. Scherer, Jr., and R. E. Hussey, /. Am. Chem. Soc.t 53, 2344 (1931).
94 C. A. Kraus and G. F. White, /. Am. Chem. Sac., 45, 768 (1923).
94 P. Schorigin and N. N. Makarowa-Zemljanskaja, Ber.t 69B, 1713 (1936).
98 L. Schmid and B. Becker, Ber., 58B, 1966 (1925).
97 L. Schmid, A. Waschkaw, and E. Ludwig, Monatsh.. 49, 107 (1928).
98 1. E. Muskat, /. Am. Ckem. Sac.t 56, 693, 2449 (1934).
872 CELLULOSE
and water-soluble starches. According to these investigators, potassium
and lithium form alcoholates with carbohydrates in the same manner as
sodium, but attempts to obtain an alcoholate with calcium were not suc-
cessful.
The trisodium alcoholate of cellulose is prepared by dissolving 3 moles of
sodium in an excess of liquid ammonia at temperatures of —33° to — 50°C.
and adding 1 mole of carefully dried cellulose to the blue solution. The
reaction is allowed to proceed in the absence of moisture from the air until
the characteristic blue color of the solution disappears. In the absence of
catalysts, this reaction requires several hours in which the rate of hydrogen
evolution diminishes slowly as the reaction proceeds. The reaction rate
varies with different types of cellulose and is probably determined by the
accessibility of the hydroxyl groups in the amorphous and crystalline por-
tions of the cellulose structure rather than by a specificity in the reactivity
of the three hydroxyl groups in the cellulose structure. Scherer and Gotsch"
have shown that the reaction time can be decreased from several hours to
a few minutes by the addition of 1% of a sodium halide to the liquid am-
monia system. The alcoholates of cellulose can probably be prepared from
sodamide in liquid ammonia, since Miller and Siehrs100 were successful in
forming the potassium alcoholates of the simple carbohydrates by this
method. Since sodamide is soluble in liquid ammonia to the extent of only
0.607 g. per 100 cc. at — 34°C., this procedure for preparing the sodium
alcoholates would probably be extremely slow.
Although attempts to prepare sodium and potassium alcoholates of cellu-
lose by an alcohol interchange with the simple alcoholates such as sodium
methylate have been unsuccessful, Harris and Purves101 have shown that
the thallium alcoholate of cellulose can be prepared by treating dry cellu-
lose with thallous ethylate in a solution of diethyl ether or benzene. It is
not possible to prepare the trithallium alcoholate of cellulose by this
method, as reaction does not proceed to completion. The large thallium
atoms cannot enter the crystalline portion of the cellulose structure so that
only the surface hydroxyl groups are available for reaction. This steric
factor enabled these workers to develop a new and ingenious method for
determining the area of the internal surface or the number of available
surface hydroxyl groups present in a cellulose structure by treating the thal-
lium alcoholate of cellulose with methyl iodide and determining the extent
of methylation. The internal surfaces of carefully purified ramie and of the
M P. C. Scherer and L. P. Gotsch, Bull. Virginia Pulytech. Inst.f 32, No. 11 (1939).
*• C. O. Miller and A. E. Siehrs, Proc. Soc. Exptl. Biol. Med.. 29, 635 (1931).
m C. A. Harris and C. B. Purves, Paper Trade /., 110, 29 (Feb. 8,, 1940).
DC. DERIVATIVES OF CELLULOSE 878
corresponding mercerized fibers as determined by this method are 3,7 X
104 and 2.9 X 108 cm.2 per g., respectively.
The trisodium alcoholates of chemical celluloses, such as cotton linters
and wood pulp, resemble the original fibrous cellulose in physical form.
These metal alcoholates are, however, unstable to moisture, oxygen, and
light. When the trisodium alcoholate is exposed to atmospheric conditions,
it discolors rapidly with the evolution of heat accompanied by a profound
degradation of the cellulose as determined by cuprammonium viscosity
measurements. Because of the high reactivity of these metal derivatives
of cellulose, it is necessary to carry out reactions with these materials under
carefully controlled conditions.
(b) REACTIONS OF METAL ALCOHOLATES OF CELLULOSE
In view of the chemical reactivity of the alcoholates of the simple alco-
hols in condensation reactions, it might be expected that the trisodium
alcoholate of cellulose would be an ideal intermediate for the preparation
of trisubstituted derivatives of cellulose, but this, in general, has not proved
true. Miller and Siehrs102 prepared the organic esters of cellulose by treat-
ing trisodium cellulosate with acid chlorides or anhydrides in the presence
of hydrocarbon diluents. Likewise, Peterson and Barry108 prepared various
cellulose ethers by treating the metal alcoholate of cellulose with alkyl
halides in the presence of hydrocarbon diluents.
Schorigin .and Makarowa-Zemljanskaja96 concluded from a study of the
xanthation of the sodium alcoholate of cellulose that alkali cellulose has the
structure of an alcoholate. On the other hand, Scherer and Gotsch"
showed that the reaction of carbon disulfide with the trisodium alcoholate
of cellulose is catalyzed by the presence of water and sodium nitrate. These
workers report the preparation of the trisodium xanthate of cellulose by
adding 2.65% of water based on the carbon disulfide to the reaction mixture
which, on the basis of their reported experiments, was sufficient water to
form an 11% alkali solution with the sodium present. Since the presence of
limited amounts of water seems to increase the reactivity of the metal
alcoholates of cellulose and since these compounds are known to be hydro-
lyzed readily by water, the conclusion is reached that, if alkali cellulose
reacts as an alcoholate, the excess molecular sodium hydroxide present is
extremely desirable and beneficial in the reaction system.
101 C. O. Miller and A. E. Siehrs (to North American Rayon Corp.)» U. S. Patent
2,181,906 (Dec. 5, 1939) ; Chem. Abstracts, 34, 2172 (1940).
wi p. c. Peterson and A. J. Barry (to the Dow Chemical Co.), U. S. Patent 2,157,083
(May 2, 1939) ; Chem. Abstracts, 33, 6595 (1939).
874 CELLULOSE
4. Cuprammonium-Cellulose Complexes
Schweizer104 observed in 1857 that ammoniacal solutions of copper hy-
droxide dissolve certain plant and animal fibers such as cotton, linen, and
silk. This important discovery is the basis of the cuprammonium or
Bemberg rayon industry as well as of the viscosity determination employed
in characterizing the celluloses used in the chemical and textile industries.
Although many solvents for cellulose have been investigated by various
workers since Schweizer's discovery, cuprammonium and cupriethylene-
diamine solutions are still the best solvents for all types of cellulose.
(a) PREPARATION AND PROPERTIES OF CUPRAMMONIUM SOLUTIONS
The compound formed when copper hydroxide is dissolved in aqueous
ammonia solutions has the formula Cu(NH3)4(OH)2, and the blue ammoni-
acal solutions of this compound are known as cuprammonium hydroxide
solutions or simply as cuprammonium solutions (often abbreviated cu-
pram). These solutions are usually prepared either by dissolving freshly
precipitated cupric hydroxide in aqueous ammonia or by bubbling air over
small pieces of metallic copper covered with aqueous ammonia solution.
The cuprammonium solution employed by the British investigators106
contains approximately 15 g. of copper and 240 g. of ammonia per liter,
whereas the solution recommended by the American Chemical Society106
contains 31 g. of copper and 165 g. of ammonia per liter (see Chapter XII).
Many investigators add to this solution 1 to 10 g. of sucrose per liter as a
stabilizing agent. In a comprehensive study, Browning, Sell, and Abel106a
found that the copper concentration must be above 25 g./liter if all types of
cellulose are to be dissolved. The ammonia concentration may vary from
125 to 250 g./liter. The viscosity of the solutions varies (in a complicated
manner) with solvent composition.
Cuprammonium solution is a strong reducing agent and, in addition, is
decomposed by light and by standing at room temperature for a period of
several days. Due to this instability, it is necessary to use relatively fresh
solutions as well as extreme precautions to protect the cellulose solutions
From oxygen and as much as possible from light. The stability of cupram-
™ E. Schweizer, /. prakt. Chern., 72, 109 (1857).
J<* Shirley Institute Test Leaflet, No. Chem. 7, 1st ed., Aug., 1948; Shirley Inst. Mem.t
IS, 25 (1936) ; /. Textile Inst.. 27, T285 (1936).
108 E. K. Carver and Committee, Ind. Eng. Chem.t Anal Erf., 1, 49 (1929).
** B. L. Browning, L. O. Sell, and W. Abel, Tappi, 37, 273 (1954).
IX. DERIVATIVES OP CELLULOSE
871
monium solutions is increased by dissolved cellulose. When it is necessary
to store cuprammonium for some time, it should be placed in a dark con-
tainer, blanketed with nitrogen or other inert gas, and kept at a tempera-
ture of 0-5°C. Under these conditions the solution is relatively stable foi
several weeks.
(b) COMPOSITION OF CUPRAMMONIUM-CELLULOSE COMPLEXES
Traube107 was one of the first investigators to study the composition oi
cuprammonium--cellulose complexes. He concluded that a compound was
formed in which part of the copper is bound in the cation and part in the
anion. His formula can be represented as follows:
Cu(NH3)4
— O— C*H702
OH
According to this formula the ratio of copper to cellulose in the complex is
1 : 1 although the copper is present in two different forms. Since Traube's
formula conforms neither to modern knowledge of copper coordination
complexes nor to the more precise information available on the copper-
ethylenediamine system, it is undoubtedly incorrect.
Hess and Messmer108 also concluded from optical rotation studies on
cuprammonium-cellulose solutions that the copper is combined chemically
with the cellulose but believed the compound could best be represented as
a salt with the structure (CeH7O6Cu)2-(Cu(NH3)4). Bauer109 analyzed
Hess's data from a different viewpoint and obtained typical absorption
curves. He pointed out that cuprammonium solutions are colloidal and
concluded that the solution of cellulose in this solvent is due only to a
physical peptizing action.
Neale110 advanced the theory that the cuprammonium-cellulose com-
plexes are colloidal electrolytes similar to soap, in which the strong cupram-
w W, Traube, Ber., 54B, 3220 (1921); 55B, 1899 (1922); 56B, 268 (1923).
108 K. Hess and E. Messmer, Kottoid-Z.t 36, 260 ( 1925) .
W9 E. Bauer, Kottoid-Z.t 36, 257 (1925).
"o S. M. Neale, /. Textile Inst., 16, T363 (1925).
876 CELLULOSE
monium base forms with cellulose, which is a weak acid, a soluble basic salt
of which the cation is crystalloidal and the anion is colloidal. Neale repre-
sented the reaction involved in the solution of cellulose in cupranunonium
as follows :
(CtHioOOn + n Cu(NH,)4(OH), >
n [Cu(NH,)4]++ + [C«H.Or]» + n [OR]' + n H2O
This formula differs from Traube's in that all of the copper is bound in the
cation. The ratio of copper to cellulose, however, is 1:1, and Neale con-
firmed this ratio by analyzing the precipitated and purified cuprammonium-
cellulose complex. Lieser111 has proposed still another formula which can
be considered as a modification of Traube's and Neale's proposals in which
1.5 moles of copper are combined with each cellulose or anhydroglucose
unit.
One of the most interesting investigations in this field was carried out by
Jolley112 in which cuprammonium solutions containing insufficient copper
to effect complete solution of the cellulose fibers were studied. Under these
conditions it was possible to determine the concentration of copper at
equilibrium in both the solid and dissolved phases. With this technique it
was possible to show that the equilibrium concentration of copper in the
cellulose solution is appreciably lower than in the original solvent due to
preferential absorption of the copper by the undissolved cellulose. Jolley
also showed that the addition of small amounts of sucrose to cuprammonium
solutions as a stabilizing agent decreases the solvent action of the solutions.
In view of the fact that the highest concentration of copper employed in
Jolley's tests was only 5 g. per liter and the dissolved phase contained 0.75
mole of copper per mole of cellulose, it is certainly reasonable that in con-
ventional solutions which contain from 15 to 31 g. of copper per liter, the
ratio of copper to cellulose in the complex would increase to unity as pro-
posed by Traube and other investigators. Stamm118 and Kraemer,114 in
determining the molecular weight of cellulose by the ultracentrif uge method,
assumed that in the cuprammonium-cellulose complex, 1 mole of copper is
combined with each mole of cellulose.
The difficulties in obtaining analytical information on the cuprammo-
nium-cellulose complex are due to the high vapor pressure of ammonia and
the great excess of ammonia that is present. According to the Traube
"l T. Lieser, Papier-Fabr.t 36, Tech.-wiss. TL, 272 (1938).
111 L. J. Jolley, /. Textile Inst., 30, T4 (1939).
"» A. J. Stamm, /. Am. Chem. Soc., 52, 3047, 3062 (1930).
114 E. O. Kraemer and W. D. Lansing, J. Phys. Chem., 39, 153 (1935) ; E. O. Kraemer,
Ind. Eng. Chem., 30, 1200 (1938).
DC. DERIVATIVES OF CELLULOSE 877
formula, ammonia should be liberated from the copper-ammonium hy-
droxide compound by reaction with cellulose. Attempts by various
workers such as Berl and Innes118 to determine if the vapor pressure of the
ammonia increases when cellulose is dissolved in cuprammonium solution
have not been entirely Successful. One method of eliminating this objection
is to replace the ammonia with a higher boiling amine such as ethylene-
diamine. In fact a considerable portion of the structural work on cupram-
monium complexes has been done with the copper hydroxide-ethylene-
diamine complex, which has the formula C^NHgC^CHaNEkMOH^ and
is usually represented as Cu(en)2(OH)2. A similar complex has been re-
ported by Jayme11Ba for cobalt and ethylenediamine. This complex was
also found to be a solvent for cellulose.
Jolley,118 in attempting to determine the structure of cuprammonium
complexes, also worked with the analogous diamine compounds. The con-
ditions employed were essentially the same as for cuprammonium in which
the concentration of copper was too low to effect complete solution of the
cellulose, so that the concentration of copper and diamine could be deter-
mined in both the solid and solution phases. Jolley confirmed the observa-
tion that the diamines are different from ammonia in that the addition of
excess ammonia to cuprammonium solutions increases the solvent action
of the solution, whereas the addition of excess diamine decreases the solvent
action. In copper hydroxide-ethylenediamine solutions in which the con-
centration of diamine was varied from 6 to 210 g. per liter, the ratio of
ethylenediamine to copper was not 2 : 1 but varied between 1 : 9 and 1.78 : 1.
Jolley concluded, as had Hoffmann,117 that both the monodiamine complex
[Cu(en)(OH)2] and the di-diamine complex [Cu(en)2(OH)2] were formed
in the copper hydroxide-diamine system. Recently, the existence of both
of these forms has been confirmed by Jonassen and Dexter118 through spec-
trographic analysis.
Levy and Muffat119 have pointed out that differences in cuprammonium
and cupriethylenediamine solutions of cellulose may be more apparent than
real. If the ratio of diamine to copper is high (approaching that in cupram-
monium solvent) precipitation does not occur on further addition of di-
amine. Dilution with water causes precipitation of cellulose just as in the
l» E. Berl and A. G. Innes, Z. angew. Chem., 23, 987 (1910).
i«* G. Jayme, Das Papier, 5, 244 (1951).
i" L. J. Jolley, /. Textile Inst., 30, T22 (1939).
»' H. Hoffmann and N. Bruch, Cettulosechemie, 14, 50 (1933).
iw H. B. Jonassen and T. H. Dexter, /. Am. Chem. Soc., 71, 1553 (1949).
»• R. M. Levy and P. Muffat, Paper Trade /., 118, 32 (Feb. 3f 1944),
878 CELLULOSE
case of cuprammonium solvent. Browning and eoworkersI06a have studied
the variation of solvent power and solution viscosity with solvent com-
position. Jolley106 showed that, in solutions of partially dissolved cotton,
the ratio of copper to diamine in the undissolved fibers was unity. As a
result of a careful study of this system, Jolley proposed the following mecha-
nism for the solution of cellulose in copper hydroxide diamine solutions:
/ ^Cu(en)++
Cellulose + Cu(en)2(OH)2 , Rceii— OH/* + 2[OH]~ + en
\OH
In this formulation the ratio of copper to diamine in the cellulose complex
is also unity.
The cuprammonium-cellulose system is complicated still further from
the viewpoint of establishing a reaction mechanism by the addition of
inorganic bases such as sodium, potassium, or lithium hydroxides to the
solution. It was shown by the work of Traube,107 Hess,108 and Trogus and
Sakurada120 that the addition of small amounts (1 to 3%) of alkali hydrox-
ides to cellulose increases its solubility in cuprammonium solutions. It is
not known exactly how the sodium ions enter the cuprammonium-cellulose
complex, but the most plausible mechanism is that they replace a portion
of the [Cu(NH3)4]++ ions in the cation part of the molecule. Additional
evidence for this view was obtained by Jolley,118 who showed that cellulose
reacts with copper hydroxide in the presence of sodium, lithium, and potas-
sium hydroxides to form a cellulose-copper complex but that the complex is
not soluble in aqueous solutions of these bases. Yet as pointed out by Ark-
hipov and Kharitonova,121 if the cellulose-copper complex is placed in
concentrated ammonia, solution takes place immediately.
However, the following tentative hypothesis may serve to rationalize
the above observations. First, it is apparent that cupric ion in an alkaline
medium forms a very sta&le complex with cellulose. This complex is in-
soluble and is therefore probably cross-linked, with the copper coordinated
with four hyd^oxyls in all. These hydroxyls may be assumed to be in posi-
tions 2 and 3 of two adjoining chains. As ammonia is added, it will tend
to break the cross-links by forming coordination complexes similar to
Jolley's picture of cupriethylenediamine-cellulose, or a similar zwitterion-
like structure formed by the ionization of the hydrogens from the two hy-
» C. Trogus and I. Sakurada, Ber.t 63B, 2174 (1930).
l" M. I. Arkhipov and V. P. Kharitonova, /. Applied Chem, (U. S. S. R.). 22, 1030
(1949) ; through Chem. Abstracts, 44, 2233 (1950),
IX. DR&lVATlVBS Off CfcLLtTLOStt 879
droxyls. A stronger complexing agent, such as ethyfenediamine, can break
the remaining bonds if present in excess, and thereby reduce the solubility
of cellulose. Since ammonia is a weak complexing agent, it may be under-
stood why the solubility in cuprammonium improves as the ammonia con-
centration is increased to the limit imposed by its solubility. In addition,
it must be remembered that ammonia itself is a relatively good solvent for
salt-like bodies such as the postulated complex, whereas ethylenediamine
in excess will tend to precipitate the cellulose in the same manner as alcohol.
The reaction of cuprammonium solution with cellulose fibers is certainly
heterogeneous in that the surface molecules of the structure are first
attacked. The solution of the surface molecules then exposes new surfaces
for further reaction. Although cuprammonium solution is unstable and is
relatively difficult to prepare, it is still one of the most useful tools for char-
acterizing cellulose.
5. Cellulose-Organic Base Complexes
The reaction of cellulose with organic bases has been primarily of aca-
demic interest. However, some of the tetraalkylammonium bases have
become commercially available. Dehnert and Konig122 showed that cer-
tain organic bases such as tetraalkylammonium bases, R4NOH, and guani-
dinium hydroxide, [(NflyaCJOH, are strong swelling agents for cellulose.
A number of other organic bases have been investigated. Shutt128 has
shown that the trialkylsulfonium hydroxides dissolve cellulose. Hess and
Trogus124 investigated the action of certain other organic bases, such as
hydrazine, ethylenediamine, and tetraethylenediamine, on cellulose and
concluded from x-ray studies that these bases form definite compounds with
cellulose.
Lieser and Leckzyck126 concluded from a study of the action of tetra-
alkylammonium bases on celluloses that those bases having a molecular
weight of approximately 150 or greater are solvents for cellulose, whereas
bases of lower molecular weight are only swelling agents. They showed
also that each base has a specific concentration at which it is most effective
as a solvent for cellulose, and that this concentration decreases as the
molecular weight of the base increases. Brownsett and Clibbens126 con-
188 F. Dehnert and W. Kdnig, Cellutosechcmie, 5, 107 (1924) ; 6, 1 (1925).
128 R. S. Shutt (to Battelle Memorial Inst.), U. S. Patent 2,371,359 (Mar. 13, 1945);
Chem. Abstracts, 39, 3667 (1945).
184 K. Hess and C. Trogus, Z. physik. Chern., B14, 387 (1931).
186 T. Lieser and E. Leckzyck, Ann., 522, 56 (1936).
1M T. Brownsett and D. A. Clibbens, /. Textile Inst.. 32, T32, T57 (1941).
CELLULOSE
100
10 20 30 40
CUPRAMMONIUM FLUIDITY
50
Fig. 57. Relation between fluidity of modified cottons and their fractional solubility
in different bases at their optimum concentrations (Brownsett and Clibbens186). Curves :
1, dibenzyldimethylammonium hydroxide at 20 °C.; 2, benzyltrimethylammonium
hydroxide at 20°C.; 3, tetramethylammonium hydroxide at 15°C.; 4, NaOH at 15°C.;
5, LiOH at 15°C.; 6, KOH at 15°C.; 7, NaOH at -5°C.
12345
NORMALITY OF BASE
Fig. 58. Relative viscosities of aqueous solutions of different bases at 20 °C.
(Brownsett and Clibbens126). Curves: 1, dibenzyldimethylammonium hy-
droxide (samples A and B); 2, benzyltrhnethylammonium hydroxide; 3, tetra-
methylammonium hydroxide; 4, NaOH; 5, LiOH; 6, KOH.
IX. DERIVATIVES OF CELLULOSE 881
firmed the work of Lieser and Leckzyck and showed that the maximum
solubility of cellulose in aqueous solutions of tetramethyl-, benzyltri-
methyl-, and dibenzyldimethyl-ammonium hydroxides occurs at 2.5 N,
2.15 N, and 1.9 N9 respectively. These workers also pointed out that there
is a relationship between solvent activity and the relative viscosity of the
organic bases. Thus, dibenzyldimethylammonium hydroxide, which has
the greatest solvent action on cellulose, also possesses the highest relative
viscosity (Figs. 57 and 58). Bases of high molecular weight, such as di-
benzyldimethylammonium hydroxide, approach cuprammonium hydrox-
ide in solvent action, since they are capable of dissolving relatively unde-
graded cotton linters and wood pulp fibers. Lovell127 developed a procedure
for using these solutions to determine the molecular weight of cellulose but
found that it was more time consuming than the cuprammonium method.
The trialkylsulfonium hydroxides described by Shutt123 as solvents for cel-
lulose must contain two hydrocarbon radicals of at least two carbon atoms
each. The solutions have been adapted to extrusion into coagulating baths
to produce shaped articles of regenerated cellulose. As in the case of the
quaternary ammonium bases, the utilization of the sulfonium bases has
been limited because of their high cost in comparison with inorganic bases
such as sodium hydroxide.
The reactions of cellulose in aqueous solutions of the organic bases are, in
general, similar to reactions in solutions of inorganic bases with the excep-
tion that in the organic bases the reactions can be carried out in homogene-
ous media. Lieser and Leckzyck125 were able to prepare the trixanthate of
cellulose by treating cellulose with carbon disulfide in a 3.7 M solution of
tetraethylammonium hydroxide. Likewise, Powers and Bock128 were able
to prepare the substituted acetals and ethers of cellulose by treating cellu-
lose dissolved in tetraalkylammonium bases with or-chloro ethers and alkyl
chlorides, respectively. One possible advantage of the organic over the
inorganic bases as reaction media for cellulose is that more uniform par-
tially substituted derivatives may be obtained, since the reactions in organic
bases can be carried out in a homogeneous system. Mahoney and Purves129
have confirmed this possibility by showing that a partially ethylated cellu-
lose prepared by a homogeneous reaction in an organic base is ethylated
more uniformly than are similarly etherified celluloses prepared in the regu-
lar heterogeneous reaction from alkali cellulose.
1J7 E. L. Lovell, Ind. Eng. Chern., Anal. Ed., 16, 683 (1944).
1M L. H. Bock and A. L. Houk (to Rohm & Haas), U. S. Patent 2,083,554 (June 15,
1937); Chem. Abstracts, 31, 5577 (1937); D. H. Powers, L. H. Bock, and A. L. Houk
(to Rohm & Haas), U. S. Patent 2,087,549 (July 20, 1937); Chem. Abstracts, 31, 6461
(1937).
1W J. F. Mahoney and C. B. Purves, /. Am. Chem. Soc., 64, 9 (1942).
E. ETHERS1
A. B. SAVAGE, A. E. YOUNG, AND A. T. MAASBERG
Etherification of the hydroxyls of cellulose yields technically useful prod-
ucts. Ethers of cellulose are organosoluble and thermoplastic, water-solu-
ble, or aqueous alkali-soluble depending upon the kind and the degree of
structural change effected.
This Section E contains, first, a review of the principles of cellulose etheri-
fication and, second, a discussion of cellulose ether manufacture in the
United States. The manufacture of ethyl cellulose, methyl cellulose, car-
boxymethyl cellulose, hydroxyalkyl celluloses, benzyl cellulose, and cyano-
ethyl cellulose is discussed.
The active hydroxyls of cellulose are etherified by organic halides, alkene
oxides, or olefins activated by polar substituent groups, in the presence of
alkali; commonly the cellulose is treated with sodium hydroxide and the
product, alkali cellulose, reacts with the etherifying reagent. Important
cellulose ethers and the reagents from which they are made are :
Ether Example Class reagent
Alkyl Ethyl cellulose Alkyl halides or sulfates
Carboxyalkyl Carboxymethyl cellulose w-Halocarboxylic acids or salts
Hydroxyalkyl Hydroxyethyl cellulose Alkene oxides or halohydrins
Aralkyl Benzyl cellulose Aralkyl halides
/3-Substituted alkyl Cyanoethyl cellulose Olefins activated by polar sub-
stituent groups
1. History
Suida2 in 1905 proposed the etherification of cellulose. He let dimethyl
sulfate react with an alkali-swollen cellulose to make inactive the surface
hydroxyls of the cellulose, but he did not obtain the cellulose ether as an
; ] 4 This section is a revision of that prepared by Shailer L. Bass, A. J. Barry, and A. E.
Young for the first edition of this book.
1 W. Suida, Monatsh., 26, 413 (1905).
882
DC. DERIVATIVES OF CELLULOSE 883
entity. Lilienfeld3 in 1912 and later4 outlined the field of cellulose ether
technology. He described organosoluble ethers, ethers soluble in cold but
not in hot water, mixed ethers, and ethers of variant degrees of substitution
(D.S.). Leuchs6 and Dreyfus6 in 1912 also claimed organosoluble ethers.
Worden7 described the resultant patent race.
Denham and Woodhouse8'9 prepared methyl cellulose as an aid to estab-
lishing the structure of cellulose (see Chapter III-B). Haworth10 prepared
methyl cellulose by the simultaneous hydrolysis and methylation of cellu-
lose acetate. Jansen11 described carboxymethyl cellulose in 1921; it was
manufactured12 in Germany prior to 1924. Hydroxyethyl cellulose was
mentioned by Hubert 13 in 1 920. Benzyl cellulose was described by Gomberg
and Buchler14; it is manufactured in Europe, but not in the United States.
2. Chemistry of the Etherification Reaction
The etherification of cellulose usually consists of the preparation of alkali
cellulose by the interaction of cellulose with a base and a solvating agent
and the reaction of the alkali cellulose with the etherifying reagent. Proc-
esses in which the alkali is not consumed are : methylation of the hydroxyls
of cellulose by diazomethane,16 addition of alkene oxides, and addition of
olefins activated by polar substituent groups, such as nitrile, carboxyl, or
sulfonyl, to the hydroxyls of cellulose.
*L. Lilienfeld, Brit. Patent 12,854 (Sept. 1, 1913); U. S. Patent 1,188,376 (June
20, 1916); Chem. Abstracts, 10, 2145 (1916).
4L. Lilienfeld, U. S. Patent 1,683,831 (Sept. 11, 1928); Chem. Abstracts, 22, 4246
(1928).
8 O. Leuchs (to F. Bayer and Co.), German Patent 322,586 (July 1, 1920); J. K.
Chowdhury, Biochem. Z., 148, 76 (1924).
6H. Dreyfus, French Patent 462,274 (Jan. 23, 1914); Chem. Abstracts, 8, 3859
(1914).
7 E. C. Worden, Technology of Cellulose Ethers, Vols. I and III, Worden Laboratory
and Library, Millburn, N. J.f 1933.
8 W. S. Denham and H. Woodhouse, 7. Chem. Soc., 103, 1735 (1913).
9 W. S. Denham and H. Woodhouse, /. Chem. Soc.t 105, 2357 (1914).
10 W. N. Haworth, E. L. Hirst, and H. A. Thomas, /. Chem. Soc.t 1931, 821.
11 E. Jansen (to Deutsche Celluloid- Fabrik, Eilenburg), German Patent 332,203
(Jan. 22, 1921) ; J. K. Chowdhury, Biochem. Z.t 148, 76 (1924).
12 J. K. Chowdhury, Biochem. Z., 148, 76 (1924).
18 E. Hubert (to F. Bayer and Co.), German Patent 363,192 (Nov. 4, 1922); /. Soc.
Chem. Ind. (London), 42, 348 (1921); E. Hubert and O. Leuchs (to F. Bayer and Co.),
German Patent 368,413 (Feb. 5, 1923); Chem. Zentr.t 1923, 755.
14 M. Gomberg and C. C. Buchler, /. Am. Chem. Soc.t 42, 2060 (1920); 43, 1904
(1921).
15 R. E Reeves and H. J. Thompson, Contrib. Boyce Thompson Inst., 11, 55 (1939).
884 CELLULOSE
(a) ETHERIFYING REAGENTS
The reagent used to prepare the cellulose ether influences the efficiency
of the reaction. Efficiency of etherification is defined as the percentage of
the reagent reacted that becomes substituent upon the cellulose. The
remaining reagent consumed is converted to by-products: alcohols, gly-
cols, ethers, or salts.
In ordinary single-phase chemical reactions, the order of reactivity of
the alkyl halides decreases from iodide to bromide to chloride.
The reaction with alkali cellulose is multiphase. In the early stages the
solid alkali cellulose is surrounded by a solvate that is usually aqueous so-
dium hydroxide; the reagent must diffuse through this solvate to the reac-
tion centers. Methyl iodide diffuses slowly, yet it is very reactive; side
reactions are more rapid than is the reaction with the cellulose. Methyl
iodide is less efficient than either methyl bromide or methyl chloride.
Methyl bromide is slightly less efficient than methyl chloride; the small
advantage in reactivity of the bromides is offset by thrtr greater cost.
Methyl chloride is used to manufacture methyl cellulose; ethyl chloride
is used to manufacture ethyl cellulose. Alkyl sulfates are not used to manu-
facture these ethers in the United States. Heuser16 has discussed the labora-
tory use of methyl sulfate.
The primary alkyl halides of three or more carbon atoms do not diffuse
readily to the reaction zone even at high temperature (140° to 150°C.) ;
instead, excessive by-product formation and cellulose degradation occur.
If, however, the cellulose is first ethylated or methylated, it then reacts
readily with the higher alkyl halides to form mixed ethers.17""19 Thus, ease
of etherification depends directly upon the rate of diffusion of the reagent
to the reaction centers,20-21 and this in turn depends upon the size of the
reagent molecule. Larger molecules can enter if small molecules are used
as opening wedges.
Branched-chain alkyl- halides (such as isopropyl chloride) react only
slightly with alkali cellulose but instead form by-products, so efficiency is
poor. Schenck22 and Timell23 achieved etherification by using special
18 E. Heuser, The Chemistry of Cellulose, Wiley, New York, 1944.
17 M. Hagedorn and P. Moller, Cellulosechemie, 12, 29 (1931).
18 J. F. Haskins and D. C. Ellsworth (to E. I. du Pont de Nemours & Co.), U. S.
Patent 2,102,205 (Dec. 14, 1937) ; Chem. Abstracts. 32, 1450 (1938).
19 The Dow Chemical Co., Midland, Mich., unpublished data.
20 E. J. Lorand, Ind. Eng. Chem.t 31, 891 (1939).
21 K. Hess, C. Trogus, W. Eveking, and E. Garthe, Ann., 506, 260 (1933).
22 H. Schenck, Dissertation, Berlin, 1936.
28 T. Timell, Studies on Cellulose Reactions, Esselte Akt., Stockholm, 1950.
DC. DERIVATIVES OP CELLULOSE
swelling agents: Schenck alkylated sodium cupricellulose and Timell
alkylated viscose Ayon swelled by quaternary bases. Cellulose was
methylated with diazomethane by Nierenstein24 and by Reeves and
Thompson.15
Aralkyl halides (such as benzyl chloride) are used to prepare benzyl
cellulose, trityl cellulose, and similar ethers.26 The fine structure of cellu-
lose affects the course26 of such reactions; this will be discussed in connec-
tion with the manufacture of benzyl cellulose.
Triphenylchloromethane (trityl chloride) was found by Helferich and
Koester27 to yield an ether (trityl cellulose) of degree of substitution (D.S.)
1.0, which was split into its components by aqueous hydrochloric acid.
They concluded that only the primary hydroxyls were etherified, and this
premise was later adopted by Sakurada,28 but Hockett and Hudson29
showed that secondary hydroxyls also react. Hearon, Hiatt, and Fordyce80
found that 90% of the primary hydroxyls and a minor part of the secondary
hydroxyls of cellulose were readily tritylated. Honeyman31 found that the
primary hydroxyls of cellulose were tritylated at 58 times the average rate
of the secondary, but that the rate of the primary groups relative to the
secondary rapidly decreased until all hydroxyls became equally reactive.
D.S. 2 was obtained. Thus steric hindrance is a factor in etherification.
Tritylation as a research tool has been succeeded by esterification with
^-toluenesulfonyl chloride (tosyl chloride) (see Sections A and C of this
Chapter IX). The application of tosylation-iodination to cellulose ethers
was reviewed by Timell.23
/fCarboxymethyl cellulose results from the treatment of cellulose with al-
kali and either chloroacetic acid or sodium chloroacetate. Aside from the
evolution of the heat of mixing, which results in loss of degree of polymeriza-
tion (D.P.) if not controlled, it matters little whether the acid, the acid salt,
or the alkali is added to the cellulose first. The usual substitution is D.S.
0.8 or less, but Chowdhury12 and McLaughlin and Herbst32 obtained substi-
24 M. Nierenstein, Ber.t 58B, 2615 (1925) ; Helv. Chim. Acta, 92, 150 (1914).
26 W. Hentrich and R. Kohler (to Procter and Gamble Co.), U. S. Patent 2,284,282
(May 26, 1942); Chem. Abstracts, 36, 6346 (1942).
26 E. J. Lorand and E. A. Georgi, /. Am. Chem. Soc., 59, 1166 (1937).
27 B. Helferich and H. Koester, Ber., STB, 587 (1924).
28 1. Sakurada and T. Kitabatake, /. Soc. Chem. Ind.t Japan, 37B, 604 (1934).
29 R. C. Hockett and C. S. Hudson, /. Am. Chem. Soc., 53, 4456 (1931).
80 W. M. Hearon, G. D. Hiatt, and C. R. Fordyce, J. Am. Chem. Soc., 65, 2449 (1943).
81 J. Honeyman, J. Chem. Soc.9 1947, 168.
" R. R. McLaughlin and J. H. E. Herbst, Can. J. Research, 28B, 731 (1950).
886 CELLULOSE
tutions approaching 2.8 by repeated reagent additions. fi-Chloropropionic
acid was used to prepare carboxyethyl cellulose.19
Alkene oxides and halohydrins react with alkali cellulose to yield hydroxy-
alkyl celluloses. Both ethylene oxide and propylene oxide are used com-
mercially. Morgan38 reported water-soluble hydroxyethyl cellulose of
apparent D.S. 4.1. The glucopyranose residue (i.e., anhydroglucose unit)
contains but three available hydroxyl groups, so much of the apparent sub-
stitution is addition to the hydroxyls of the substituent group. Alkene
oxides readily polymerize34; thus, ethylene oxide can react with the hy-
droxyls of the substituent group at least as readily as with the original cel-
lulose hydroxyls.
Sodium 0-chloroethylsulfonate reacts with alkali cellulose to yield sulfo-
ethyl cellulose, the sulfonic acid analog of carboxyethyl cellulose. Karrer
and coworkers,35 Timell,23'86 and others37 have prepared sulfoethyl cellulose;
its manufacture was attempted in Germany during World War II. Sodium
chloromethylsulfonate does not appear to react with alkali cellulose.19
The difference in reactivity between the methyl- and ethyl-w-haloalkylsul-
fonic acids may be explained on the basis of the tquanticule theory of
Fajans.38
The addition to cellulose of olefins activated by polar substituent groups
is a general method for the preparation of the sulfoethyl,39 cyanoethyl,40-44
88 P. W. Morgan, Ind. Eng. Chem., Anal Ed., 18, 500 (1946).
*4 H. Staudinger and O. Schweitzer, Ber., 62B, 2395 (1929).
35 P. Karrer, H. Koenig, and E. Usteri, Helv. Chim. Acta, 26, 1296 (1943).
36 T. Timell, Svensk Papperstidn., 51, 254 (1948); Swedish Patent 124,025 (Feb. 15,
1949) ; Chem. Abstracts, 43, 9446 (1949).
87 J. B. Dickey and J. G. McNally (to Eastman Kodak Co.), U. S. Patent 2,422,000
(June 10, 1947) ; Chem. Abstracts, 41, 5306 (1947).
88 K. Fajans, Chem. Eng. News, 27, 900 (1949).
89 V. R. Grassie (to Hercules Powder Co.), U. S. Patent 2,580,352 (Dec. 25, 1951);
W. Neugebauer, K. Sponsel, and U. Ostwald (to Kalle & Co., Akt.-Ges.), U. S Patent
2,132,181 (Oct. 4, 1938); Clem. Abstracts, 33, 381 (1939).
40 L. H. Bock and A. L. Houk (to Rohm & Haas), U. S. Patents 2,332,048 and 2,332,-
049 (Oct. 19, 1943); Chem. Abstracts, 38, 1640 (1944); Brit. Patent 562,581 (July 7,
1944) ; Chem. Abstracts, 40, 736 (1946) ; U. S, Patent 2,349,797 (May 30, 1944).
41 British Thomson -Houston Co., Ltd., Brit. Patent 592,352 (Sept. 16, 1947) ; Chem.
Abstracts, 42, 2103 (1948).
42 J. H. MacGregor (to Courtaulds, Ltd.), Brit. Patent 605,357 (July 21, 1948);
Chem. Abstracts. 43, 404 (1949); U. S. Patent 2,482,011 (Sept. 30, 1949); Chem. Ab-
stracts, 44, 1702 (1950); J. Sac. Dyers Colourists, 67, 66, 74 (1951).
48 J. H. MacGregor (to Courtaulds, Ltd.), Brit. Patents 636,020 (Apr. 19, 1950),
636,295 (Apr. 26, 1950); Chem. Abstracts, 44, 6624 (1950).
44 R. C. Houtz (to E. I. du Pont de Nemours & Co.), U. S. Patent 2.375,847 (May 15
1945) ; Chem. Abstracts. 39, 4486 (1945).
IX. DERIVATIVES OF CELLULOSE 887
carboxyethyl,45 carbamylethyl,46 and stdfamylethyl47 ethers of cellulose.
Etherification with acrylonitrile40""48 yields cyanoethyl cellulose which upon
complete hydrolysis yields carboxyethyl cellulose. Low temperatures and
dilute alkali are necessary if the nitrile ether is to be isolated. Other at-
tempts to introduce nitrogen onto the cellulose chain have included etheri-
fication with haloalkyl amines,48 haloalkyl amides,49 aminoalkene oxides,50
and ethylenimine.51
a,w~Dihalogenated etherifying agents react at both ends with cellulose
to yield insoluble, high-melting, cross-linked products52; such agents are
more useful for the surface modification of cellulose than for the preparation
of soluble or plastic products. Fluoroethylene was used to prepare o?-
fluoroethyl cellulose.68
Unsaturated aliphatic ethers have been made with the purpose of obtain-
ing compatibility with synthetic polymers and copolymerizability with
inexpensive monomers and of preparing ethers that can be cast or formed
and then rendered insoluble by baking. The ethers that have been prepared
from the corresponding unsaturated halides include allyl cellulose,19-28'54""69
46 V. R. Grassie (to Hercules Powder Co.), U. S. Patent 2,539,417 (Jan. 30, 1951);
Chem. Abstracts, 45, 4452 (1951).
46 L. H. Bock and A. L. Houk (to Rohm & Haas), U. S. Patent 2,338,681 (Jan. 4,
1944); Chem. Abstracts, 38, 3855 (1944).
47 V. R. Grassie (to Hercules Powder Co.), U. S. Patent 2,580,351 (Dec. 25, 1951).
48Kalle and Co., Akt.-Ges., German Patent 719,241 (May 20, 1938); C. L. P.
Vaughan (to Hercules Powder Co.), U. S. Patent 2,591,748 (Apr. 8, 1952).
49 J. B. Riest and W. H. Delden (to Montclair Research Corp.) , U. S. Patent 2,399,603
(Apr. 30, 1946) ; Chem. Abstracts, 40, 4229 (1946).
50 1. G. Farbenindustrie Akt.-Ges., German Patent 726,199 (Aug. 27, 1942); Chem.
Abstracts, 37, 6458(1943).
51 H. Fink, R. Stahn, and W. K. Bitterfeld (to I. G. Farbenindustrie Akt.-Ges.),
U. S. Patent 2,097,120 (Oct. 26, 1938); Chem. Abstracts, 32, 353 (1938).
52 E. C. Britton and K. G. Harding (to The Dow Chemical Co.), U. S. Patent 2,216,-
095 (Sept. 24, 1940) ; Chem. Abstracts, 35, 893 (1941).
M W. E. Hanford (to E. I. du Pont de Nemours & Co.), U. S. Patent 2,409*274 (Oct.
15, 1946); Chem. Abstracts, 41, 982 (1947).
"C. Dreyfus, French Patent 724,584 (Oct. 15, 1931); Chem. Abstracts, 26, 4950
(1932).
w Gesellschaft fur Chemische Industrie, Basel, Swiss Patent 144,227 (Dec. 15, 1928) ;
Chem. Abstracts, 25, 4124 (1931).
66 Gesellschaft fur Chemische Industrie, Basel, Brit. Patent 342,689 (Nov. 3, 1928);
Chem. Abstracts, 25, 5578 (1931).
" I. Sakurada, Angew. Chem., 42, 549 (1929).
58 R. Haller and A. Heckendorn, Helv. Chim. Acta, 24, 86E (1941).
69 S. N. Danilov and O. P. Koz'mina, /, Gen* Chem, (U. S. S. R.), 18, 1823 (1948);
through Cfaw, Abstracts, 43, 5943 (1949),
888 CELLULOSE
crotyl cellulose,60 and mixed ethers of these and other alkyl halides and
oxides.19 Butadiene monoxide was used to prepare 2-hydroxy-3-butenyl
cellulose.61 Charch62 prepared a cross-vulcanized product of crotyl cellu-
lose and butadiene, and Ushakov and Klimova68 copolymerized allyl hy-
droxyethyl cellulose with sulfur dioxide. The vinylation of cellulose by
acetylene64 is claimed, but positive structural proof is lacking.
Alkylsilyl ethers have been made by reaction of cellulose with alkylhalo-
silanes in pyridine.19'66'66
(b) ALKALI CELLULOSE
Cellulose alone does not react appreciably with etherifying reagents; if
the cellulose is to become reactive, it must be treated with a swelling agent
and a solvating agent. The usual swelling agent is sodium hydroxide;
the usual solvating agent is water. Alkali cellulose preparation consists
of the treatment of cellulose with aqueous sodium hydroxide solution.
(The steeping technology described in Section D of this Chapter IX is
seldom used in cellulose ether manufacture.)
A concentration of 30% NaOH is generally used for the preparation of
alkali-soluble ethers and of water-soluble ethers of low degree of substitu-
tion. Concentrations of from 35 to 76% are used for the preparation of
ethers of higher substitution. At least two moles of sodium hydroxide per
glucopyranose residue are used in the preparation of alkali cellulose for
alkali-soluble ethers; three or more moles of sodium hydroxide per gluco-
pyranose residue are used in the preparation of alkali cellulose for the higher
substituted water-soluble or organosoluble ethers.
The cellulose may be bleached wood pulp or cotton linters. It must be
60 F. C. Hahn (to E. I. du Pont de Nemours & Co.), U. S. Patent 2,082,797 (June 8,
1937); Chem. Abstracts, 31, 5577 (1937).
« D. M. Musser (to Pacific Mills), U. S. Patent 2,455,083 (Nov. 30, 1948); Chem.
Abstracts, 43, 1979(1949). ,
88 W. H. Charch (to E. I. du Pont de Nemours & Co.), U. S. Patent 2.520,711 (Aug.
29, 1950); Chem. Abstracts, 44, 10370 (1950).
"S. N. Ushakov and O. M. Klimova, Zhur. Priklad. Khim.> 25, 46, 191 (1952);
through Chem. Abstracts, 46, 5837, 6826 (1952).
64 A. E. Favorskil, V. I. Ivanov, and Z. I. Kuznetsova, Compt. rend. acad. sci. U. R.
S. S., 32, 630 (1941). V. V. Shtitevski!, N. A. Oblonskaya, and N. I. Nikitin, Zhur.
Priklad. Khim., 24, 1045 (1951) ; through Chem. Abstracts, 46, 4221 (1952).
M H. A. Schuyten, J. W. Weaver, J. D. Reid, and J. F. Jurgens, /. Am. Chem. Soc.,
70, 1919 (1948); H. A. Schuyten, J. W. Weaver, and J. D. Reid (to U. S. Secretary of
Agriculture), U. S. Patent 2,562,955 (Aug. 7, 1951).
"M. J. Hunter (to Dow Corning Corp.), U. S. Patent 2,532,622 (Dec. 5, 1950);
Chem. Abstracts, 45, 1768 (1950).
IX. DERIVATIVES OF CELLULOSE 889
high in alpha-cellulose content, very low in incrustants, free of metals, and
uniformly absorbent of water and of sodium hydroxide solution. It must
also be free of dots or knots of fibers and must not have been overdried.
Uniform penetration of the sodium hydroxide or other swelling agent
throughout the cellulose is the most important factor in cellulose etherifica-
tion. The swelling agent must provide alkalinity at every point in the
cellulose throughout the reaction; uniform swelling of the cellulose aids
the diffusion of the etherifying reagent to the reaction centers. The solvat-
ing agent (water) acts as a solvent and carrier for the sodium hydroxide, as
an agent to swell and solvate the cellulose so that the etherifying reagent
penetrates readily, and as a diluent for the etherifying reagent. There is
almost no etherification in the absence of water.67
The sodium hydroxide concentration that is reached toward the end of
the reaction determines the degree of substitution that is achieved.20
When the reagent reacts with the sodium hydroxide, the final alkali con-
centration depends upon the initial concentration, upon the water formed,
upon the sodium hydroxide added, and upon the sodium hydroxide con-
sumed. Even in the most efficient etherification, the increase in substitu-
tion above D.S. 2.0 stops when the sodium hydroxide concentration falls
to about 30%. The presence of diluents, the kind and amount of agitation,
and the temperature schedule used also affect the D.S. achieved.
The etherification efficiency decreases as the water concentration in-
creases, because hydrolysis of the reagent to by-products increases, and
because the over-all driving force for the etherification of cellulose is some
high power68 of the concentration of the base used as the swelling agent.
General methods for the preparation of alkali cellulose include: (1)
dipping a cellulose sheet in alkali, (2) spraying or mixing a disintegrated
sheet with alkali, (3) squeezing out an alkali-cellulose slurry, and (4) adding
alkali to a slurry of cellulose fibers in an inert diluent. (See Section D of
this Chapter IX.)
(c) OTHER SWELLING AGENTS
The distribution of substituents in cellulose ethers made from alkali
cellulose does not approach closely that calculated by statistical methods;
various swelling agents other than aqueous sodium hydroxide have been
proposed in order to obtain more uniform substitution. These inclu4e
sodium cupricellulose (Normann compound), thallium cupricellulose, or-
ganic bases, liquid ammonia, and other inorganic bases.
67 J. Ch6din and A. Tribot, M6m. services cUm. itat (Paris), 33, 168 (1947).
« H. M. Spurlin, /. Am. Chem. Soc.t 61, 2222 (1939).
890 CELLULOSE
(1) Sodium Cupricellulose
Normann69 first observed the addition compound, sodium cupricellulose,
that forms when cellulose is treated with sodium hydroxide and copper
hydroxide, or when alkali cellulose is treated with a copper salt. Traube70 ~~78
and his fellow workers methylated sodium cupricellulose. The reaction
rate was more rapid and the D.S. of the product was higher than in a con-
trol experiment in which alkali cellulose was used. The solubility of the
ethers from sodium cupricellulose began at very low D.S. values. Products
of D.S. 0.8 were soluble in both cold and hot water. Piwonka73 stated that
the methylation took place only at the 3-hydroxyl of the cellulose, but
Heddle and Percival74 found that the 2-hydroxyl reacted also.
TABLE 15
Substituent Distribution in Methyl Cellulose from Sodium Cupricellulose (Timell23)
Degree of Substitution
0.33
0
98
i
.09
1.39
Substitution
Found Calcd.
Found
Calcd
Found
Calcd.
Found
2-Hydroxyl
0.20 0.18
0.41
0.45
0.40
0.45
0.40
3-Hydroxyl
0.13 0.11
0.07
0.11
0.07
0.10
0.07
6-Hydroxyl
— —
—
—
—
—
0.10
2- and 3-Hydroxyls
— 0 02
0.25
0.21
0.31
0.27
0.34
2-, 3-, and 6-Hy-
droxyls
— —
—
—
—
—
0.05
Traube's products, which were of D.S. 1.0 or less, lacked trisubstituted
units; in this respect they differed from commercial methyl cellulose. It
appeared that the 6-hydroxyl of the cellulose did not react; this has been
confirmed by Timell.23* Traube attributed the good solubility at low D.S.
to lack of trisubstitution and to uniform distribution of the substituents
along the cellulose molecules. Timell found that the substituent groups in
ethers from sodium cupricellulose were distributed along the cellulose mole-
M W. Normann, Chem.-Ztg.t 30, 584 (1906).
« 70 w. Traube and A. Funk, Ber., 69B, 1476 (1936).
71 W. Traube, R. Piwonka, and A. Funk, Ber., 69B, 1483 (1936).
» W. Traube (to "Achetem"), U. S. Patent 2,140,568 (Dec. 20, 1938); Chem. Ab-
stracts, 33, 2709 (1939).
7« R. Piwonka, Ber., 69B, 1965 (1936).
74 W. J. Heddle and E. G. V. Percival, J. Chem. Soc., 1938, 1690.
IX. DERIVATIVES OF CELLULOSE 891
cules in accordance with the laws of probability. This implies that all of
the cellulose was equally available for etherification. Table 15 shows that
typical analyses of methyl cellulose made from sodium cupricellulose agree
well with values calculated from rate constants by Timell,28 The 6-hy-
droxyl of the cellulose did not begin to react with the methylating reagent
until the substitution exceeded D.S. 1.1. Meanwhile the entering methyl
groups were forced to distribute themselves evenly in proportion to the rate
constants because there was a choice between but two positions, rather than
among three.
(2) Thallium Cupricellulose
Traube and Funk70 found that a complex was formed when cellulose was
treated with thallium and copper hydroxides or with salts of these metals
together with sodium hydroxide solution. The use of thallium cupricellu-
lose was soon abandoned in favor of sodium cupricellulose.
(5) Organic Bases
The dispersion of cellulose in quaternary bases of high molecular weight
was first proposed by Lilienfeld.76 Lieser and co workers76"78 showed that
the minimum normality of a quaternary base that was necessary to disperse
cellulose was a straight-line function of the molecular weight of the base.
Bock79 stated that those quaternary ammonium hydroxides which con-
tain un substituted hydrocarbon groups would dissolve cellulose. Strangely
enough, one of the most important factors contributing to the solvent
power of these bases seemed to be their concentration in aqueous solution.
A solution with a concentration of 35 to 50% was usually a good cellulose
solvent almost without regard for the size of the alkyl or aryl groups.
Tetramethylammonium hydroxide was not a useful solvent because of its
low (20%) water solubility. The phenyl quaternary bases were unstable
at room temperature. The benzylammonium hydroxides, such as tri-
methylbenzylammonium hydroxide, were more stable than aliphatic bases,
were readily prepared, and are useful solvents.
Sisson and Saner80 found that the x-ray pattern of swollen cellulose that
75 L. Lilienfeld, U. S. Patent 1,771,462 (July 29, 1930); Chem. Abstracts, 24, 4630
(1930); Brit. Patent 217,166 (June 5, 1923); Chem, Abstracts, 19, 398 (1925).
76 T. Lieser, Ann., 528, 276 (1937).
77 T. Lieser, R. Jaks, and E.-A. Glitscher, Ann., 548, 212 (1941).
7* T. Lieser and E. Leckzyck, Ann.. 522, 56 (1936).
79 L. H. Bock, Ind. Eng. Chem., 29, 985 (1937).
80 W. A. Sisson and W. R. Saner, /. Phys. Chem., 43, 687 (1939).
892 CELLULOSE
had been compounded with an amount of quaternary base insufficient to
cause dispersion indicated equal modification of all of the crystalline cellu-
lose.
Bock81-82 prepared water-soluble methyl cellulose and ethyl cellulose with
the aid of quaternary bases. He obtained water solubility at D.S. 0.6 to
0.7; he assumed that the substituents were more evenly distributed than
in the usual products prepared from alkali cellulose. Compton88 and
Mahoney and Purves84 found that the methoxyl groups in methyl cellulose
that had been prepared in a quaternary base were distributed according to
the laws of probability, just as they are in the methylation of sodium cupri-
cellulose. Johnston85 found that above D.S. 2.0 the methyl cellulose was
precipitated from the quaternary base and was thereafter methylated in
suspension.
Timell23 prepared methyl cellulose, ethyl cellulose, and propyl cellulose
in quaternary bases. He also prepared several branched-chain cellulose
ethers from viscose rayon that was dissolved in such bases. He obtained
water solubility of isopropyl cellulose at D.S. 0.5, but w-propyl cellulose
prepared under similar conditions was not water-soluble. He attributed the
solubility of the isopropyl cellulose to the ability of the branched chain to
hold the cellulose chains apart.
Quaternary ammonium bases are the most useful of the organic
bases, but others have been used for special purposes. Guanidine was re-
ported to be a swelling agent for cellulose by Bock.79 Pyridine was used by
Van Dyke19 and others66'68 as a solvent and base in the preparation of mixed
ethers from etherifying reagents (organosilicon compounds) that are
highly sensitive to moisture. The reaction of pyridine derivatives with
cellulose is important in textile finishing. (Stearamidomethyl)pyridinium
81 D. H. Powers and L. H. Bock (to Rohm & Haas), U. S. Patent 2,009,015 (July 23,
1935); Chem. Abstracts, 29, 6080 (1935).
82 L. H. Bock and A. L. Houk (to Rohm & Haas), U. S. Patent 2,083,554 (June 15,
1937); Chem. Abstracts, 31, 5577 (1937); D. H. Powers, L. H. Bock, and A. L. Houk
(to Rohm & Haas), U. S. Patent 2,087,549 (July 20, 1937); Chem. Abstracts, 31, 6366
(1937) ; L. H. Bock and A. L. Houk (to Rohm & Haas), U. S. Patent 2,084,125 (June
15, 1937) ; Chem. Abstracts, 31, 557 (1937) ; D. H. Powers, L. H. Bock, and A. L. Houk
(to Rohm & Haas), U. S. Patent 2,216,045 (Sept. 24, 1940).
M J. Compton, /. Am. Chem. Soc., 60, 2823 (1938).
84 J. F. Mahoney and C. B. Purves, J. Am. Chem. Soc.t 64, 15 (1942).
85 G. G. Johnston, /. Am. Chem. Soc., 63, 1043 (1941).
IX. DERIVATIVES OF CELLULOSE 893
chloride (Velan P. F. or Zelan A) reacts with cellulose, liberating pyridine
and yielding cellulose stearamidomettiyl ether on the fiber surface.8**87
(4) Liquid Ammonia
The swelling of cellulose by liquid ammonia was reported by Bernardy88
in 1925. Hess,89 Barry, Peterson, and King,90 and Clark and Parker91
applied x-ray studies to this swelling. Clark and Parker found that the
swelling may amount to three times the original fiber diameter.
Scherer and Hussey92'98 found that the ammonia cellulose can quantita-
tively exchange all of its hydroxyl hydrogens for sodium, liberating hydro-
gen and consuming three atoms of metallic sodium per glucopyranose resi-
due:
Rcen(OH)8 + 3 Na LiquM NH> > iWONa), + 'A H, (1)
The sodium celluloses in liquid ammonia were etherified by Peterson and
Barry,94 by Schorigin and Makarowa-Semljanskaja,96 by Scherer and
Gotsch,96 and by Freudenberg.97 High substitution (for example, D.S.
2.85) is obtained under such conditions, but degradation of the cellulose
may occur. The liquid ammonia was said to serve as the solvating agent
in this case, but one cannot be sure that water was completely absent.94-96'98
86 R. J. W. Reynolds, E. E. Walker, and C. S. Woolvin (to Imperial Chemical Indus-
tries, Ltd.), Brit. Patent 466,817 (June 7, 1937) ; Chem. Abstracts, 31, 8195 (1937) ; C. E.
Mullin, Chem. Inds., 47, 404 (1940).
87 H. A. Schuyten, J. W. Weaver, J. C. Frick, Jr., and J. D. Reid, Textile Research J.,
22,424(1952).
88 G. Bernardy, Angew. Chem., 38, 838, 1195 (1925).
89 K. Hess and C. Trogus, Ber., 68B, 1986 (1935) ; K. Hess and J. Gundermann, Ber.t
70B, 1788(1937).
90 A. J. Barry, F. C. Peterson, and A. J. King, /. Am. Chem. Soc., 58, 333 (1936).
91 G. L. Clark and E. A. Parker, /. Phys. Chem.t 41, 777 (1937).
92 P. C. Scherer. Jr., /. Am. Chem. Soc., 53, 4009 (1931).
9* P. C. Scherer, Jr. and R. E. Hussey, /. Am. Chem. Soc., 53, 2344 (1931).
94 F. C. Peterson and A. J. Barry (to The Dow Chemical Co.), U. S. Patent 2,145,273
(Jan. 31, 1939) ; Chem. Abstracts, 33, 3587 (1939) ; U. S. Patent 2,157,083 (May 2, 1939) ;
Chem. Abstracts, 33, 6595 (1939) ; U. S. Patents 2,232,926 and 2,232,927 (Feb. 25, 1941) ;
Chem. Abstracts, 35, 3814 (1941).
95 P. Schorigin and N. N. Makarowa-Semljanskaja, Ber., 69B, 1713 (1936).
96 P. C. Scherer and L. P. Gotsch, Bull. Virginia Polytech. Inst., 32, 11 (1939);
through Chem. Abstracts, 34, 259 (1940).
97 K. Freudenberg and H. Boppel, Ber., 70B, 1542 (1937); K. Freudenberg, E. Plan-
kenhorn, and H. Boppel, Ber., 71B, 2435 (1938).
98 P. C. Scherer, Jr. (to North American Rayon Corp.), U. S. Patents 2,181,919 and
2.181.920 (Dec. 5, 1939) ; Chem. Abstracts, 34, 2172 (1940).
894 CELLULOSE
The presence of a small amount of water could result in the formation of
alkali cellulose in suspension in the ammonia.
The necessity of water as a swelling agent in the absence of liquid am-
monia was shown by Peterson and Barry; ethylation of sodium cellulose
with ethyl chloride yielded D.S. 2.06 in the presence of liquid ammonia,
but only D.S. 0.04 when the liquid ammonia was removed. When water
was added to the system, ethylation was again obtained.
(5) Other Inorganic Bases
The use of inorganic bases other than sodium hydroxide has been de-
scribed." In general, there is no advantage in cost or performance of other
bases over sodium hydroxide.
(d) ACCESSIBILITY OF CELLULOSE TO REAGENTS
The alternative swelling agents that were just discussed (topic c above)
are expensive, and they frequently yield products that are difficult to isolate
or to purify. Alkali cellulose remains the starting material for etherifica-
tion. One must therefore consider how best to prepare and to use alkali
cellulose.
The question of how much of cellulose is readily reactive has been de-
bated (see Chapter IV). Attempts to determine the accessibility of cellu-
lose were summarized by Tarkow,100 by Timell,28 and by Heuser.101 Hy-
drolysis and oxidative methods gave very low results that are not significant
in the present case. The reaction of cellulose with sodium in liquid am-
monia,23 its reaction with formic acid and acetic anhydride,100 and various
physical methods (x-ray diffraction, density, calorimetry, equilibrium mois-
ture regain, sorption isotherms, and deuterium exchange) indicate that both
cotton linters and wood pulp are 35 to 50% accessible. Tarkow stated that
these figures indicate the readily accessible region of cellulose, but they
indicate neither crystallinity nor chemical convertibility.
Heuser101 indicated that the submicroscopic structure of cellulose does
not interfere with the specific reactivity of the cellulose hydroxyls relative
to various reagents; for example, in the heterogeneous system, methyl and
ethyl chlorides preferred the 2- position whereas ethyl sulfate appeared to
react more readily with the 6- position.
Mass estimates of accessibility have not yet satisfactorily assigned the
««H. Dreyfus, U. S. Patent 2,098,335 (Nov. 19, 1937); Chem. Abstracts. 32, 353
(1938); U. S. Patent 2,181.264 (Nov. 28, 1939); Chem. Abstracts, 34, 2172 (1940).
«» H. Tarkow, Tappi, 33 , 595 ( 1950) .
«" E. Heuser, Textile Research /., 20, 828 (1950).^
IX. DERIVATIVES OF CELLULOSE 895
accessible region of cellulose either with respect to the individual fibers or
with respect to the glucopyranose residue; many of the methods proposed
are open to criticism of their basic simplifying assumptions. (See Chapter
IV-B.)
In commercial etherification, concern is not with the inaccessibility of
large portions of the cellulose, but with small numbers of rogue or unreactive
fibers, which can seriously impair the usefulness of a batch of a cellulose
ether. The origin of rogue fibers is not apparent; they may be fibers of
another species, or they may still retain such incrustants as rosin and lignin.
The prevailing trend in cellulose ether manufacture is toward continuous
alkali cellulose manufacture. The time during which the pulp is in contact
with the sodium hydroxide solution varies from a few seconds to a few
minutes, so distribution of the sodium hydroxide solution must be rapid.
This makes necessary a nearly perfect pulp sheet: a sheet uniformly ab-
sorbent of water and sodium hydroxide, uniformly formed, and free of knots
and clots.
Accessibility is but one of the facets of the problems of uniform etherifica-
tion. Others are sodium hydroxide penetration, reagent diffusion, molecu-
lar ratios of reactants, and substituent distribution (topic e below). Uni-
form penetration of the sodium hydroxide throughout the alkali cellulose
is the most important single factor in cellulose etherification. At least two
moles of sodium hydroxide per glucopyranose residue are used as 30%
aqueous solution to make the lower substituted alkali-soluble and water-
soluble ethers. Similarly, at least three moles of sodium hydroxide per
glucopyranose residue are used as 35 to 76% or stronger aqueous solution
to make ethers of high substitution.
Ready diffusion of the etherifying reagent to the reaction centers is the
second most important factor in cellulose etherification. Elevated tempera-
tures aid diffusion. Diluents equalize diffusion and reaction rates and re-
duce steric hindrance. Stepwise addition of reactants is often better than
single addition. The ratios of sodium hydroxide, cellulose, and water to
one another determine the substitution reached and the efficiency of etheri-
fication.
(e) RANDOMNESS OF ETHERIFICATION
The degree of substitution of a cellulose ether and the distribution of the
substituent groups among the available hydroxyls of the cellulose molecule
largely determine the solubility properties and the utility of the ether.
The hypothesis that the distribution of the substituents in a cellulose
ether is random and can be calculated by statistical methods was proposed
896 CELLULOSE
by Spurlin88 (see Chapter IX-A). Distribution was calculated from the
viewpoint that the ratio of the reactivities of the 2-, 3-, and 6-hydroxyls of
the cellulose remains independent of the degree of substitution, however
much the over-all reactivity may vary. Thus, all of the hydroxyls may have
the same chance to react, and the substituents will be distributed among
the available positions according to the laws of probability. Spurlin's
hypothesis has been given exhaustive mathematical and experimental
treatment by Timell.2* Earlier workers contributed less complete data.
Since rate constants depend upon the specific reaction involved, they are
omitted from the following discussion.
Mahoney and Purves84 investigated the substituent distribution in sev-
eral undegraded methyl celluloses and ethyl celluloses. They determined
the number of primary hydroxyl groups present by tosylation-iodination.
Subtraction of the number of primary hydroxyl groups so determined from
the total number of unsubstituted groups in the original alkyl celluloses
left a quantity, H, which was equal to the average sum of the unreacted
secondary hydroxyls.
When alkylation occurred with uniform average density along the cellu-
lose chains (that is, homogeneously), the maximum frequency of completely
unsubstituted glycol units (£2,3) was calculated to be H*/4. When alkyla-
tion occurred in localized regions of the cellulose chains (that is, hetero-
geneously), the probable frequency of glycol units was calculated to be
H/2. The actual occurrence of glycol units was determined by sodium
periodate oxidation. The results obtained by Mahoney and Purves84 are
shown in Table 16. The column (12) Glycol count shows that for the
methyl celluloses the determined 2,3-glycol (£2,3) was equal to the calcu-
lated H/2; the methylations were heterogeneous. For ethylation D the
glycol determined was equal to the calculated H2/4; the etherification in
a quaternary base was homogeneous. Similarly, the low D.S. ethylations
E and F were heterogeneous, but ethylation A to a commercial D.S. value
was quite homogeneous. The uniform substitution of sample D is shown
by its solubility in water at D.S. 0.58.
Honeyman81 treated cotton cellulose with two different concentrations of
sodium hydroxide: 45% and 90%. The resulting alkali celluloses were
reacted with ethyl chloride in a manner similar to commercial practice.
When 45% NaOH was used, the initially heterogeneous reaction became
homogeneous after 3 hrs. When 90% NaOH was used, the initial reaction
was very heterogeneous; after 3 hrs. the reaction followed much the same
course as when the 45% NaOH was used, except that the glycol groups
G*,i persisted at a higher degree of substitution. This means that some por-
IX. DERIVATIVES OF CELLULOSE
M ^
5 o t
d cd £
897
X*
8
I
CO 73
rH r
W
.a
^
I
Q
•4->
§
<+•>
•5
J
b s s b §
+_> +j jij +-> g$
(M ^(M CM 51
flC l£ frj Ji"j frj
0 0 O 0 0 0
rH l>
O
l> rH f^ C6 CO
o o" d o" o" o
00 CO t^ t^» ^O C5
d d d d d d
Tj< O5 rH CD Tj* Cq
t>. -rf t^ l>» CO rH
d d d d d o"
t^ CO rH
cp co t^
rH O rH
00 »0
CO ^^ O^ rH ^j CD
co co c^ ^ GO co
O rH O O O rH
rH rH O> O CO 00
W I »O Cq I CO CO 00
d d d d d d
S o? S S § ^
o rH d d rH cq
1I|S||
S S W W W W
O pQ Q PH W <J
-8
-s hj
i
-8
-9
CO
Q
»
£ d
II
15
&N1
11 §1
Oj OJ _. ^3
OrHCMCO
1-2 §
g fe-1! .
fe|«^
•stwgs
^•^-a-S^
1 1 !^^
•g HCNCO<N
A co'Poo ST
.s -B
§-S £
II §1
«ii
ents
ple.
stitu
ree
bsti
d
898 CELLULOSE
tions of the cellulose did not react beyond the initial alkali-soluble stage
and would not become organosoluble at any degree of substitution.
Timell23 prepared and analyzed cellulose ethers from alkali cellulose, from
sodium cupricellulose, and from cellulose in a quaternary base. His data
are collected in Table 17. Methyl celluloses from alkali cellulose (A and B)
were heterogeneous. Methyl celluloses from sodium cupricellulose (Table
15) had no substitution in the 6- position until above D.S. 1.1. Methyl
cellulose, ethyl cellulose, w-propyl cellulose, and isopropyl cellulose were
substituted homogeneously in quaternary bases.
Carboxymethyl cellulose and sulfoethyl cellulose appeared to approach
homogeneous substitution, although the data were difficult to interpret.
These ethers seemed to be substituted on one or the other of the secondary
hydroxyls of a glucopyranose unit, but not on both. This may have been
due to steric hindrance or to negative polarization of the residual hydroxyl
of a 1,2-glycol pair. Dyer and Arnold102 found that the glycol count of two
different carboxymethyl cellulose samples was intermediate between that
calculated for homogeneous and for heterogeneous reaction, but Ryd-
holm108 concluded that the substitution was essentially homogeneous. The
statistical calculations of Spurlin68 are not applicable to the case in which
substitution of one secondary hydroxyl interferes with that of the other.
The proper formulas for this case are given in Section A of this Chapter
IX. It has been shown104 that the distribution of substituents in carboxy-
methyl cellulose conforms to the revised calculations.
The micropreparation of 2-ethyl cellulose was claimed by Sugihara and
Wolfrom,105 who used copper to render unreactive the 6- position, regener-
ated cellulose (dehydrated azeotropically) to improve accessibility, and
methyl iodide to secure slow diffusion and steric hindrance.
The substituent distribution in hydroxyalkyl celluloses is only partially
known. Table 18 shows the data of Morgan,83 who first published values
of M.S., the moles of ^ethylene oxide consumed that become attached to
the glucopyranose unit and to its substituent chains in hydroxyethyl cellu-
lose.
The reaction of a mole of ethylene oxide with a secondary hydroxyl of a
glucopyranose residue results in the conversion of one secondary hydroxyl
to a primary hydroxyl :
101 E. Dyer and H. E. Arnold, /. Am. Chem. Soc., 74, 2677 (1952).
w« S. Rydholm, Svensk Papperstidn., 53, 561 (1950).
104 T. Timell, Svensk Papperstidn., 55, 649 (1952); T. Timell and H. M. Spurlin,
Svensk Papperstidn., 55, 700 (1952); 56, 311 (1953).
» J. M. Sugihara and M. L. Wolfrom, /. Am. Chem. Soc., 71, 3509 (1949).
IX. DERIVATIVES OF CELLULOSE
899
Ceo
w
<v
W O
£.2
2!
ri
c
s
•*-
JB
CO t> l"» *•* C&
o d o o o CD
o o o o o o
O i-« O O O O
0 0
00
o o
00 -^
" CO
O CD
CM
(N
O CD
O O
o
CD O
1C
O
0 O
CO
CO 00
»O 00
o o o o o o
9 s 85 s s a
CD CD CD O CD CD
TH ^J O O O O
-< pq O P W fa
la.
JH
3
I
e
CD
03
Q
s d
o o - -
111
j Wo
'• 1 ^
, ^^
^Sco
»0t3
CO*-' v «j— gj
, d-agg-a^
^^ 3 "g cr cr'g a;
ITcJco ^ScD ?* g aT c? g 8
-/N-*'>-XSi*/N*^s~/' -S 8 o 8 § 3
^3 o QM o'S
5o>>oo
900
CELLULOSE
CH2CH2O » R,enCHOCH2CH2OH
The number of initial primary hydroxyls remains unchanged:
R0flUCH2OH + CH2CH20 » RoenCH2OCH2CH2OH
(2)
(3)
The number of primary hydroxyls also remains unchanged when chain
growth takes place:
RoeiiCH2OCH2CH2OH + CH2CH2O > lUiiCHjOCF^CHzOCHaCHiOH (4)
TABLE 18
Analyses of Hydroxyethyl Cellulose (Morgan88)
Moles ethylene oxide
used per glucopyranose unit
M.S.
Efficiency,
Solubility
0.25
0.17
68
7% NaOH
Commercial
0.18
__
7% NaOH
Commercial
0.23
_
7% NaOH
0.50
0.25
51
7% NaOH
0.75
0.42
57
Swollen by cold water
1.50
0.73
48
Swollen by many solvents but not
by acetone
Commercial
1.44
—
Soluble in water, nearly soluble in
acetic acid or pyridine
4.05
1.84
46
Soluble in water, acetic acid, or
pyridine
10.0
4.10
41
Soluble in water and in aqueous
ethanol, but not in acetone
Tasker and Purves106 and Cohen and Haas107 published partial analyses
of hydroxyethyl cellulose. The results of Cohen and Haas for four samples
of variant substitution are shown in Table 19.
TABLE 19
Analyses of Hydroxyethyl Cellulose (Cohen and Haas107)
Sample
M.S.
2,3-Glycol,
CM
Primary
hydroxyls
formed
(1 - G*,,)
Primary
hydroxyls
Ethyl ene
oxide
loss on
extended
tosylation
I
0.44
0.71
0.29
__
_
II
0.97
0.57
0.43
_
—
III
1.50
0.48
0.52
1.46
0.84
IV
3.07
0.34
0.66
—
—
» C. W. Tasker and C. B. Purves, /. Am. Chem. Soc., 71, 1023 (1949).
w S. G. Cohen and H. C. Haas, /. Am. Chem. Soc.t 72, 3954 (1950).
DC. DERIVATIVES OF CELLULOSE
901
The data from sample III of Cohen and Haas107 (Table 19) indicate that
one glycol group (£2,3) was used for each new primary hydroxyl that was
formed. It was concluded that sample III appeared to have 50% of its
glucopyranose units substituted in either the 2- or the 3- position, to have
50% of its glucopyranose residues not substituted on the 2- or the 3- posi-
tion (the glycol group £2,3 was still present), and to have no glucopyranose
units substituted on both the 2- and 3- positions.
TABLE 20
Arbitrary Compositions of Hydroxyethyl Cellulose 108'107; *ee «!«> *»'108
Ethylene oxide substitution per mole
Primary
Secondary
Sample
M.S.
D.S.
Mono-
Di-
Mono-
Di-
Tasker and Purves106
0.44
0.37
0.13
0.02
0.17
0.05
Cohen and Haas, lo7 Ia
0.44
0.37
0.06
0.02
0.24
0.05
Cohen and Haas, 107 III0
1.50
0.94
0.10
0.34
0.28
0.22
• See Table 19.
If it is assumed that the length of the polyethylene oxide chains did not
exceed D.P. 2 in samples I and III of Cohen and Haas and in the sample
(similar to I) of Tasker and Purves, the arbitrary compositions shown in
Table 20 may be calculated on the basis of their 3- to 5-hr, tosylation-
halogenation data.23-108
H-4- O CH2 — CHg-j 1
I— — '*
H 0
CHgOH
Fig. 59. Structure of hydroxyethyl cellulose.
J. F. Mahoney and C. B. Purves, /. Am. Chem. Soc.9 64, 9 (1342).
902 CELLULOSE
As the M.S. increased, the polyethylene oxide chains increased in length.
Morgan38 pointed out that hydroxyethyl celluloses, unlike other cellulose
ethers, showed increasing solubility in all media with increasing substitu-
tion. Thus water solubility produced by the wedging apart of the cellulose
chains by hydroxyethoxyl substitution was not lost when the extended
polyethylene oxide chains produced organosolubility.
Hydroxyethyl cellulose is of the structural type shown in Figure 59.
(f) FRACTIONATION OF ETHYL CELLULOSE
The analyses of cellulose ethers that have been presented are average
values for experimental or for commercial products; they do not take into
account the variation in composition of the samples. Since the initial cel-
lulose is a random mixture of molecules of diverse size and its breakdown by
cleavage is likewise random,109 it is not surprising that cellulose ethers are
random mixtures of molecules that differ from one another in both substitu-
tion and size.
The properties that best describe a particular cellulose ether are its degree
of substitution and its viscosity. In practice, the viscosities of solutions of
finite concentration determine commercial utility, but from a theoretical
point of view the limiting or intrinsic viscosity (limit of rjsp/c as c approaches
zero) better describes the average length of a molecule (see Chapter XII-A).
Suitable factors for the conversion of intrinsic viscosity values to numerical
degrees of polymerization for cellulose ethers are not available in the litera-
ture, because closely fractionated samples of uniform D.S. and D.P. have
not been prepared. The determination of intrinsic viscosity has been dis-
cussed in part by Davis and Elliott.110
If a cellulose ether is to be completely described, it must be fractionated
and each fraction must be analyzed both for substitution and for intrinsic
viscosity.
The customary technique of solvent precipitation is particularly applica-
ble to ethyl cellulose. Okamura111 dissolved ethyl cellulose in acetic acid
™ G. Beall and L. J6rgensen, Textile Research /., 21, 203 (1951).
110 W. E. Davis and J. H. Elliott, J. Colloid Sci.t 4, 313 (1949). See also Chapter X-F
of this book.
111 1. Okamura, Cellulosechemie, 14, 135 (1933).
IX. DERIVATIVES OF CELLULOSE 903
and precipitated fractions by adding water. Staudinger and Reinecke112
used dioxane and water. Ushakov and Geller118 used alcohol and gasoline.
Scherer and McNeer114 tried the systems benzene and w-heptane, and ace-
tone and water unsuccessfully. They also used acetic acid-water, but failed
to stabilize their products. They finally used ethyl acetate-acetone as the
solvent and acetone-water as the nonsolvent. Barry19 used chloroform as
the solvent and petroleum ether (Skellysolve, b. p. 100-140°C.) as the non-
solvent. Samsel and Warren19 used chloroform as the solvent and petro-
benzol116 as the nonsolvent. Scherer and lacoviello116 used benzene-metha-
nol as the solvent and heptane-benzene as the nonsolvent.
The course of a fractionation will depend upon the nature of the solvents
and the precipitants used, but in general the material of high molecular
weight will be precipitated first.
If the ethyl cellulose is dissolved in a solvent such as chloroform, and if
to this mixture is added a nonpolar liquid such as petroleum ether, material
of low degree of substitution will be precipitated first; as the solvent mix-
ture is made less active, the degree of substitution of the fractions obtained
will gradually increase. If, on the other hand, the ethyl cellulose is dissolved
in alcohol and water is then added, the first fractions precipitated will be
high in degree of substitution, and the low-substituted material will remain
in solution.
If the determined degree of substitution of a sample is rather high (above
2.5, for example), there can be relatively little low-substituted material
present. If, in addition, the intrinsic viscosity is relatively high, fractiona-
tion will be by chain length rather than by degree of substitution. Thus,
Staudinger and Reinecke112 and Scherer and McNeer114 found a narrow
degree-of-substitution range in the fractionation of high-intrinsic-viscosity
ethyl cellulose of D.S. 2.6.
Scherer and lacoviello116 fractionated ethyl celluloses of D.S. 2.54 and
112 H. Staudinger and F. Reinecke, Ann.. 535, 47 (1938).
11SS. N. Ushakov and I. M. Geller, Plasticheskie Massui Sbornik Statelt 30, (1939);
through P. C. Scherer and R. D. McNeer, Rayon and Synthetic Textiles, 30, No. 8, 56
(1949) ; 31, No. 2, 53 (1950) ; 31, No. 4, 54 (1950).
114 P. C. Scherer and R. D. McNeer, Rayon and Synthetic Textiles, 30, No. 8, 56 (1949) ;
31, No. 2, 53 (1950) ; 31, No. 4, 54 (1950).
116 Petrobenzol is an aliphatic naphtha, manufactured by Anderson Prichard Oil Co.,
and described in H. A. Gardner, Physical and Chemical Examination of Paints, Var-
nishes, Lacquers and Colors, 9th ed., Inst. of Paint & Varnish Research, Washington,
D. C., 1939, p. 570.
116 P. C. Scherer and J. G. lacoviello, Rayon and Synthetic Textiles, 32, No. 11, 47
(1951).
904 CBLLULQSE
2.23. They obtained a high degree of reproducibility of their technique
and report a wide range of D.P. within the samples.
Samsel and Warren19 fractionated ethyl cellulose of D.S. 2.4 and found a
range of 0.1 to 0.2 D.S. in the recovered fractions; these data are shown in
Table 21. The total distribution within the original samples is not de-
TABLE 21
Fractionation of Ethyl Cellulose, D.S. 2.4 (Samsel and Warren")
Sample
batch
Determined
D.S.
Fractions
Intrinsic
viscosity0
Mean
D.S.
Range
Standard
deviation
A
2.42
-2.41
0.21
0.065
1.75
B
2.46
2.43
0.18
0.066
1.7
C
2.41
2.31
0.23
0.083
1.8
D
2.40
2.34
0.18
0.054
1.8
E
2.38
2.34
0.18
0.077
2.0
F
2.38
2.34
0.19
0.076
2.0
G
2.43
2.38
0.18
0.066
2.1
H
2.40
2.36
0.10
0.036
2.11
a Intrinsic viscosity solvent: 70:30 benzene :methanol.
scribed completely by the mean, range, and standard deviation117 of the
fractions, however, because the recovery was only about 90%.
Few fractionations of other cellulose ethers have been reported. Methyl
celluloses were fractionated by Signer and von Tavel118 and by Steele and
Pacsu.119 Timell and Purves120 nitrated alkali-soluble methyl celluloses
by a nondegradative method. The trisubstituted portions of nonuniform
methyl cellulose dissolved in the nitrating mixture. The resultant products
were subjected to fractional precipitation and fractional solution on a
micro scale. The methoxyl degrees of substitution of the fractions varied
inversely as their viscosities. The alkali-soluble methyl celluloses became
more uniform in thislrespect as their degrees of substitution increased.
Alkali-soluble methyl celluloses prepared with methyl chloride121 were
more uniform than similarly substituted methyl celluloses120 prepared with
methyl sulfate.
117 K. A. Brownlee, Industrial Experimentation, 3d American ed., Chemical Publishing
Co., Brooklyn, N. Y., 1949, p. 26.
118 R. Signer and P. von Tavel, Angew. Chem., 50, 902 (1937) ; 51, 535 (1938).
119 R. Steele and E. Pacsu, Textile Research J., 19, 771, 784 (1949).
120 T. Timell and C. B. Purves, Svensk Papperstidn., 54, 303 (1951).
m A. T. Maasberg (to The Dow Chemical Co.), U. S. Patent 2,408,326 (Sept. 24,
1946); Chem. Abstracts. 41, 1104 (1947).
IX. DERIVATIVES OF CELLULOSE 905
3. Properties of Cellulose Ethers
(a) THE SOLUBILITY OF CELLULOSE ETHERS
A given cellulose ether is most soluble in solvents that best match the
ether or its solvated derivative in cohesive energy density.122 When a cellu-
lose ether is prepared, the gradual increase in degree of substitution of the
ether is accompanied by a transition from insoluble cellulose through solu-
bility in the following series of solvent types: aqueous alkali, water, water-
alcohol mixtures, hydrocarbon-alcohol mixtures or equivalent solvents,
and finally aromatic hydrocarbons. The solubility in water and other hy-
droxylic solvents is generally lost as solubility in hydrocarbon solvents is
reached.
The variation of solubility of cellulose ethers with degree of substitution
is shown in Table 22. Ethyl cellulose and, to a lesser extent, methyl cellu-
lose go through the stages outlined. The higher aliphatic ethers, propyl
cellulose and butyl cellulose, are not very soluble in aqueous alkali or in
water at any time, but isopropyl cellulose prepared in a homogeneous man-
ner was found by Timell23 to be water-soluble. Benzyl cellulose made
from fibrous alkali cellulose is not soluble in alkali or in water at any stage.
Sodium carboxymethyl cellulose does not pass beyond the water-alcohol-
soluble stage. The heavy metal salts of carboxymethyl cellulose are solu-
ble only in aqueous alkali. The free acid form, as commonly made by
acidification, is also soluble only in aqueous alkali, but conversion of an
aqueous solution of the sodium salt by means of ion -exchange resins yields
a water dispersion of carboxymethyl cellulose that becomes water-insoluble
when dried. Table 22 likewise shows that ethers prepared homogeneously
(samples B) from sodium cupricellulose or in quaternary bases are soluble
at lower degrees of substitution than are the corresponding ethers prepared
from alkali cellulose (samples A).
Swelling is considered to be solution of the solvent in the cellulose ether;
there is thus little difference between a swollen gel and a solution. An
ether may be swelled by a solvent at degrees of substitution both above and
below the degree of substitution at which it i& soluble in the solvent.
Ethers of a given degree of substitution are more soluble, the lower their
intrinsic viscosity is; degraded cellulose ethers are more soluble than are
undegraded ethers of the same kind.
The solubility of low-substituted ethers in 4 to 10% NaOH solution is
greater near 0°C. than at higher temperatures. Chilling apparently aids
the hydration of the cellulose hydroxyls. Water solubility is held to result
"' H. M. Spurlin, J. Polymer Sci., 3, 714 (1948).
906
CELLULOSE
*R 5J<
,
o |_H
«
A
W
S
*St/;?<
00
CO
iO *H
O r^
*>»
»H
a
TO
co
c
M
**^
Q
ffl
1
i-H
*3
8<
5 «
o
S
|
S Zb co
2 CN CN
c
*>,
j
00
CO a K
o
CN »O CO
§
•M
w
S ^22
§"
•a
O T-H vH
*p
^j
V .I* 8 3S «
04 0»
M*
2
?- • 00 CN Si
2 .
0 0
w
oi
5
O • • •
*o3c^<j
•
en
t/J
I
d d ^H i-J CN
*x
S
O r-i
{j
"3
PQ
g
8
td
bo
a
Q
cr
+3
0
1 «
10 |
CO
1
o
CN £
w
« |
Q
3
CN 4>
^ o
.
S
*cy
S w
CO
Q
d ^
Q
o
1
2 "
oT
0
1
Jo
a
H |
J2
C4
Jj>
^ 00 00
H
§
•8
•S
3
1
JS
tJ
?li
CN J 2 2 » »
Propyl eel
normal-
d
carboxymeti
A
T O CN
O CO O
d d CN
i
in
3
o ?b cb cb P- ob
g
s
03
1
2 T I 1 !
r-H ^ CO i-H Tf CO
•"
1 5 2
1 10 00
i
•8
O O I-H CN CN CN
TH TH
i
•*->
CtJ
i
'S
en
en
*3
>
d
1
1
1
•a % «
1 ll
^1*
ffi rt w 43
111
.1
X
<3 u .3 .S .S
.s .s .s
O ^g
*
Solubilil
I Iff!
o5 bo bo bo
^ o3 bo
00 ^ S
i
-.y d d d
Tfl W KH HH >H
^ d d d
V^ HH I-H H- t
^u 5
1
tH CN CO ^ »O
CN CO ^* >C
1-1 oi co
IX. DERIVATIVES OF CELLULOSE 907
from the wedging apart of the cellulose chains by the substituent groups,
so that the remaining hydroxyls are available for hydration.138 Bulky
substituent groups enable water solubility to be obtained at lower degrees
of substitution than is the case with the less bulky methyl or ethyl groups.
Thus methyl cellulose is water-soluble at D.S. 1.3, ethyl cellulose at D.S.
0.7, isopropyl cellulose at D.S. 0.5, and sodium carboxymethyl cellulose at
D.S. 0.5. Hydroxyethyl cellulose is water-soluble at D.S. 0.8 (see Tables
18 and 20) or M.S. 1.4. It is interesting to note that water solubility occurs
at about the same weight per cent of added substituent in all these cases.
Solubility in organic solvents stems from predominance of the sub-
stituent groups over the remaining hydroxyl groups. The best solvents
most closely resemble the cellulose ethers, and, in general, alcohol-hydro-
carbon mixtures are better solvents than are single solvents. The maximum
over-all solubility is reached at a substitution of less than D.S. 3.0. Thus
ethyl cellulose is most soluble in alcohol-hydrocarbon mixtures at D.S. 2.4
to 2.5 and becomes decreasingly soluble in alcohols as D.S. increases.
Ethyl cellulose of D.S. 2.7 to 2.9 is dispersible in aromatic hydrocarbons.
The proportion of alcohol that must be used in alcohol-hydrocarbon sol-
vent mixtures to obtain minimum viscosity is proportional to the number
of hydroxyls that remain unsubstituted in the cellulose ether. Thus ethyl
cellulose of D.S. 2.2 requires 70 : 30 benzene :methanol to dissolve it, whereas
ethyl cellulose of D.S. 2.6 is soluble in 95:5 benzene : methanol or in 80:20
toluene :ethanol, and ethyl cellulose of D.S. 2.8 is dispersed in benzene or
toluene alone. Better film properties are obtained if the amount of alcohol
slightly exceeds these minimum values.
128 J. F. Haskins and R. W. Maxwell (to E. I. du Pont de Nemours & Co.), U. S. Patent
2,131,733 (Oct. 4, 1938); Chem. Abstracts, 32, 9496 (1938).
124 D. C. Ellsworth and F. C. Hahn (to E. I. du Pont de Nemours & Co.), U. S. Patent
2,157,530 (May 9, 1939); Chem. Abstracts, 33, 6595 (1939); U. S. Patent 2,249,754
(Jan. 22, 1941); Chem. Abstracts, 35, 7190 (1941).
126 E. J. Lorand, Ind. Eng. Chem., 30, 527 (1938).
126 L. Lilienfeld, U. S. Patent 1,589,606 (June 22, 1926); Chem. Abstracts, 20, 3084
(1926).
127 E. Heuser and W. von Neuenstein, Cellulosechemie, 3, 89 (1922).
128 E. Berl and H. Schupp, Cellulosechemie, 10, 41 (1929).
129 H. Staudinger and p. Schweitzer, Ber.t 63B, 2327 (1930).
180 S. N. Ushakov and S. I. Kucherenko, Plasticheskie Massy, 3, 12 (1934).
181 A. W. Schorger (to Burgess Laboratories), U. S. Patent 1,863,208 (June 14, 1932);
Chem. Abstracts, 26, 4174 (1932); U. S. Patent 1,914,172 (June 13, 1933); Chem. Ab-
stracts, 27, 4243 (1933); U. S. Patents 1,941,276, 1,941,277, and 1,941,278 (Dec. 26,
1933); Chem. Abstracts, 28, 1861 (1934).
132 A. W. Schorger and M. J. Shoemaker, Ind. Eng. Chem., 29, 114 (1937).
188 F. D. Farrow and S. M. Neale, /. Textile Inst., 15, T157 (1924).
908 CELLULOSE
(b) THE GELATION OF WATER-SOLUBLE CELLULOSE ETHERS
Homogeneously prepared water-soluble methyl cellulose and water-solu-
ble ethyl cellulose dissolve to give aqueous solutions that can be heated to
near their boiling points without gelation.19 Such nongelling ethers have
been prepared from alkali cellulose, from sodium cupricellulose, and from
reaction in quaternary bases. These preparations are characterized by their
water solubility at lower degrees of substitution than hold for the corre-
sponding heterogeneously prepared ethers. It is believed that their failure
to gel upon heating is due to lack of trisubstitution,134 but it is also possible
that freedom from long unaltered residues of the original cellulose structure
is the important factor. Solutions of heterogeneously prepared ethers gel
from solution upon heating.
The reversible gelation of a methyl cellulose of D.S. 2.22 was studied by
Heymann,186 who called the gelation of this methyl cellulose upon heating
an inverse transformation to distinguish it from the gelation upon cooling
that is observed in solutions of agar-agar and gelatin. The viscosity of a
given solution of methyl cellulose decreased as the temperature of the solu-
tion was increased. Heymann attributed this to a decrease in hydration
of the methyl cellulose that accompanied the rise in temperature. As the
gel point was approached, small increments of temperature produced large
increases in viscosity; maintenance of the solution viscosity required an
increased rate of shear to break down the gels that formed. When the gela-
tion temperature was reached, the application of shearing force could no
longer maintain the solution state, the viscosity rose rapidln| and a large
aggregate of gel separated from the water. When the gelle(£(solution was
cooled, the viscosity returned to its original value by a direct path.
The gelation temperature decreased with increasing concentration of
methyl cellulose of a given viscosity. When the concentration was fixed,
the gel temperature decreased with increasing viscosity of the methyl
cellulose. At low concentrations, or with methyl cellulose of low viscosity,
turbidity was obtained*, rather than actual gelation. This turbidity was
due to gel particles.
The presence of soluble inorganic salts in methyl cellulose solutions gener-
ally lowered the gelation temperature by decreasing the hydration of the
methyl cellulose. High concentrations of salt caused gelation at room tem-
perature. The salt concentration that a solution could tolerate decreased
with increasing methyl cellulose concentration and decreased with increased
viscosity of the methyl cellulose. The effect of salt addition was similar to
"« J. Vacher, Chimie 6f Industrie, 43, 347 (1940).
l« E. Heymann, Trans. Faraday Soc.9 31, 846 (1935) ; 32, 462 (1936).
IX. DERIVATIVES OF CELLULOSE
909
the effect of heating. The salt tolerance depended upon the ions present:
Soluble thiocyanates and iodides raised the gelation temperature; other ions
lowered it in the order phosphate > sulfate > tartrate > acetate > chlo-
ride > nitrite > nitrate. This is the familiar lyotropic series of ions.
Gelation is a very slow process at temperatures just below the gel point,
which indicates that the chains require time to diffuse to the required posi-
tions for association.
(c) THE THERMOPLASTICITY OF CELLULOSE ETHERS
The thermoplastic properties of cellulose ethers at elevated tempera-
tures may be looked upon as an extension of their solution properties, for
10
S 25
iu
i "o
•> 85
O 10
(0
> 25
BENZYL, D.S. 2.05
n-PROPYL, D.S. 2.05
ETHYL, O.S. 2.45
110 120 130 140 ISO 160 170
TEMPERATURE, °C.
Fig. 60. Softening-melting point ranges of benzyl cellulose, w-propyl
cellulose, and ethyl cellulose; viscosity in cps. (The Dow Chemical Co. w).
he degree of substitution at which maximum solubility is obtained is also
hat of the lowest softening or flow temperature of the cellulose ether.
The thermoplasticity of a cellulose ether depends upon the nature of its
ubstituent group, the degree of substitution, and the chain length of the
*her. The softening temperature decreases as the substituent size in-
xeases: Thus methyl cellulose softens at over 190°C., whereas ethyl cellu-
ose of similar substitution softens near 140°C.
Figure 60 shows the effect of the nature of the substituent group upon
he softening-melting point range for some cellulose ethers prepared from
Ikali cellulose and alkyl chlorides by Savage.19 Ethyl cellulose of D.S. 2.42
910 CELLULOSE
was similar in softening temperature to w-propyl cellulose of D.S. 2.05.
These ethers softened at 125° to 135°C.; ethyl cellulose of D.S. 2.05 would
soften only above 170°C. Benzyl cellulose of D.S. 2.05 softened at 105°
to 120°C. The softening points of benzyl cellulose and w-propyl cellulose
approached that of ethyl cellulose as the viscosity increased from 10 to 25
cps. The softening and melting points were determined for sheeting on a
modified Dennis-Shelton136 melting-point bar. Lorand125 reported lower
softening points for benzyl cellulose.
The softening temperature of a cellulose ether of a given degree of sub-
stitution is higher, the higher the intrinsic viscosity of the ether. This is
shown for ethyl cellulose137 in Table 23.
TABLE 23
The Effect of Intrinsic Viscosity upon the Softening Temperature of Ethyl Cellulose
of D.S. 2.45 (Hercules Powder Company137)
Intrinsic
viscosity*
5% Solution*
viscosity, cps.
Softening
temperature, °C.&
0.8-0.9
7-9
151
1.0-1.3
13-20
156
1.6
42.5
161
2.1
93.5
167
2.8
199
171
Intrinsic viscosity and 5% viscosity solvent: 80:20 toluene :ethanol.
b Penetration softening point.137
Small quantities of cellulose ethers dissolve in large amounts of the com-
mon solvents at room temperature or below. The concentration that can
be reached depends upon the chain length of the cellulose ether and upon
the solvent used. As the concentration of the ether is increased, portions
of the ether no longer dissolve well (the solvent becomes saturated with
respect to certain fractions) and the apparent viscosity of the solution is
greatly increased* (The~ratio of the viscosity in such a medium at a high
concentration to the viscosity of, for example, £ 5% solution is termed the
viscosity index; it is a measure of the solvent power on the one hand and
of the uniformity of substitution of the ether on the other hand.) If now
the temperature is raised, the viscosity index is decreased; the amount of
the cellulose ether dissolved may be increased and the amount of the solvent
may be decreased. As the temperature is raised higher, a State of concentra-
tion is ultimately reached that amounts to solution of the solvent in the
l* L. M. Dennis and R. S. Shelton, /. Am. Ckem. Soc.t 52, 3128 (1930),
IVJ Ethyl Cellulose, Hercules Powder Co., Wilmington, Del., 1949.
IX. DERIVATIVES OF CELLULOSE 911
cellulose ether. The cellulose ether swelled by the solvent resembles a
liquid in its properties.
If the product is to be technically useful, it must reassume a certain de-
gree of hardness when its forming by fluid flow or by plastic flow has been
completed. This depends upon applying the right conditions of tempera-
ture and pressure to the swelled cellulose ether, which must have been formu-
lated to give the desired end properties.
The agents that are used to modify a cellulose ether during plastic form-
ing or to impart particular properties to a finished form of the cellulose
ether are called plasticizers. If a permanent effect is desired, a nonvolatile,
stable plasticizer must be used. Cellulose ethers are compatible with a
variety of different plasticizing compounds, but the more effective plasti-
cizers closely resemble in structure the cellulose ether with which they are
used. Thus water-soluble ethers are plasticized by hydroxyl-containing
compounds such as glycols; esters or ethers are suitable for cellulose ethers
of intermediate substitution; nonsolvent oils may be used to plasticize
highly substituted ethers.
Each cellulose ether may be considered to be, in a sense, internally plasti-
cized by its substituent group ; thus benzyl cellulose may be molded without
an added plasticizer. The internal plasticization is greater, the lower the
yield point is. Table 24 shows the yield points of several cellulose ethers.
TABLE 24
Internal Plasticization of Cellulose Ethers Shown by Their Yield Points (The Dow
Chemical Company19)
Cellulose ethers
D.S.
Yield point,
kg./sq. cm.a
Methyl cellulose
2.0
600
Methyl cellulose
2.8
525
Ethyl cellulose
2.3
510
Ethyl cellulose
2.5
475
w-Propyl cellulose
1.5
375
Butyl cellulose
1.8
275
Amyl cellulose
1.8
100
Benzyl cellulose
2.0
390
0 Yield point from load-elongation curves on sheeting tested in Scott tensile strength
tester at 5-kg. load and loading rate of 25.4 mm./min.
The internal plasticization is proportional to the size of the substituent
group. Such substituent groups as hydroxyethyl and hydroxypropyl are
increasingly effective as internal plasticizers as their side chains increase in
length.
912 CELLULOSE
(d) THE MECHANICAL PROPERTIES OF CELLULOSE ETHERS
The mechanical properties of a cellulose ether depend upon the plasti-
cization, whether internal (by substitution) or external (by an added plasti-
cizer), and upon the intrinsic viscosity of the cellulose ether.
The relation of intrinsic viscosity to the mechanical properties of ethyl
cellulose187 is shown in Table 25. The tensile strength, flexibility, and
TABLE 25
The Effect of Intrinsic Viscosity on the Physical Properties of Ethyl Cellulose of D.S.
2.45 (Hercules Powder Company187)
Intrinsic0
viscosity
Solution
viscosity,
5% solution,
cps.°
Sheeting0
thickness, in.
Yield point,
Ib./sq. in."
Tensile
strength,
Ib./sq. in.
Flexibility
by Schopper
double folds
0.85
7.4
0.0031
— .
7960
17
0.93
9.3
0.0030
6500
7670
23
1.05
13.0
0.0030
—
8100
42
1.25
19.7
0.0030
7800
8250
65
1.62
42.5
0.0030
, —
8390
90
2.15
93.5
0.0030
8800
9100
128
2.80
199
0.0030
• —
9240
248
a Intrinsic viscosity, 5% viscosity, and casting solvent: 80:20 toluene : ethanol.
elongation increase with viscosity, provided that fibers and gels are absent.
Both tensile strength and elongation vary with the casting solvent used.
The moisture sorption of a cellulose ether depends upon the number of
free hydroxyl groups and upon the size and the nature of the substituent
group, as shown in Table 26.
TABLE 26
Effect of Substituent Group on the Moisture Sorption of Cellulose Ethers (Lorand116)
Moisture
Cellulose ether
D.S.
Softening*
temperature, °C.
sorption.
% at 72%
relative
humidity
and 19°C.
Ethyl cellulose
2.15
158
3.0
Butyl cellulose
2.28
65
1.7
Amyl cellulose
1.91
45
1.0
• Penetration softening point.187
Figure 61 shows that the moisture sorption of ethyl cellulose decreases
with increasing degree of substitution and increases with increasing relative
humidity, but is less in water at 50°C. than in water at 21 °C.
DERIVATIVES OP CELLULOSE
913
Table 27 shows the moisture vapor transmissions of some cellulose ether
sheets in comparison to ethyl cellulose sheeting of comparable thickness as
a standard, since the moisture vapor transmission values may vary con-
siderably with the thickness of the sheeting.
The values of mechanical properties cited are rather general ones for
typical ethers. It should be remembered that cellulose ethers are mixtures,
both with respect to chain length and with respect to degree of substitution;
here again, if the mechanical properties are to be completely described, the
ether must be fractionated and the properties of the individual fractions
must be determined.114
12
10
o
I 8
§
or
ID
3\
A.
0123
DEGREE OF SUBSTITUTION
Fig. 61. Ethyl cellulose: relation of degree of substitution to moisture
absorption (The Dow Chemical Co.19). Lines: 1, 50% R.H., 21°C.; 2> 70%
R.H., 19°C.; 3t water immersion, 50°C.; 4, water immersion, 21°C.
4. Ethyl Cellulose
The uses of ethyl cellulose are many and varied. It is formulated into
plastics, lacquers, transparent sheeting, melts, varnishes, and adhesives.
Ethyl cellulose is tough; it retains strength and flexibility over an ex-
treme range of temperature. It is useful in both rigid and soft plastics, and
can be fabricated by extrusion, injection molding, compression molding,
drawing, and casting. Ethyl cellulose toughens and hardens most composi-
tions in which it is compatible, it is soluble in many low-cost solvents, and
it is compatible with a wide range of plasticizers and resins. It can be
formulated for many varied uses; it can be tailored for the specific use.
914 CELLULOSE
Ethyl cellulose may be made in three ranges of substitution :
Commercial ethyl cellulose ranges in D.S. from 2.20 to 2.58 (ethoxyl con-
tent from 44.0 to 49.5%) ; it is soluble in common organic solvents and is
thermoplastic.
TABLE 27
Effect of Degree of Substitution of Cellulose Ethers upon Water Vapor Transmission
(Ronda19)
Water vapor transmission,
g./lOO sq. in./24 hr.
Sheeting
D.vS.
Thickness,
in.
Absolute
Relative to
ethyl cellulose,
D.S. 2.4
Methyl cellulose
2.0
0.0025
72
219
Ethyl cellulose
2.4
0 0025
33
100 base
Ethyl cellulose
2.3
0 001
111
219
Ethyl cellulose
2.4
0.001
52
100 base
Ethyl cellulose
2 6
0 001
94
180
Ethyl cellulose
2.4
0.001
52
100 base
Benzyl cellulose
2.0
0.008
3
18
Ethyl cellulose
2.4
0 008
18
100 base
Nearly completely substituted ethyl cellulose ranges in D.S. from 2.60
to 2.80 (ethoxyl content 50.0 to 52.5%); it is soluble in hydrocarbons, but
is not soluble in many of the common oxygenated solvents. It is limited in
compatibility, and suitable plasticizers for its thermoplastic use have not
been found. It is incompatible with the commercial ethyl cellulose of
lower degree of substitution.
Low-substituted ethyl cellulose, ranging in D.S. from 0.8 to 1.7 (ethoxyl
content 19 to 35%), is water-soluble. The control of ethylation to obtain
water solubility is difficult.
Ethyl cellulose is prepared by the etherification of alkali cellulose with
ethyl chloride, followed by the isolation, washing, and drying of the prod-
uct. In this section the manufacture of ethyl cellulose of D.S. 2.20 to 2.58
is discussed in detail; the higher and lower substituted ethyl celluloses are
briefly mentioned.
(a) RAW MATERIALS
The common sources of cellulose for ethylation are bleached and purified
cotton linters and wood pulp. The sheet cellulose, commonly supplied in
roll form, must be uniform in weight, density, water absorbency, and aque-
ous sodium hydroxide absorbency. It must be high in alpha-cellulose con-
DC. DERIVATIVES OF CELLULOSE 915
tent, very low in incrustants, free of metals, free of knots or clots of fibers,
and must not have been overdried. Typical analyses188 of celluloses suit-
able for ethylation are shown in Table 28.
TABLE 28
Typical Celluloses Used in Ethylation (Martin188)
Chemical cotton
Wood pulp
Alpha-cellulose, %
98.7
94.5
Beta-cellulose, %
1.3
3.0
Gamma-cellulose, %
0
2.5
Moisture, %
6
6
Color, G.E. brightness
90
85
Ash, %
0.05
0.06
Iron, p.p.m.
12
12
Calcium, p.p.m.
50
50
Viscosity, TAPPI Standard T 230, 1%
cupriethylenediamine, cps.
40
65
The ethyl chloride used is pure; it should contain less than 20 parts per
million of sulfur and less than 50 parts per million of acetaldehyde. The
sodium hydroxide must contain less than 40 parts per million of iron and
must be very low in alkaline earth metals.
(b) PROCESSES
When alkali cellulose is etherified with ethyl chloride, at least one-half
of the ethyl chloride that is consumed is converted to by-products, chiefly
ethanol and ethyl ether. If other factors are equal, the ratio of ether to
ethanol is greater, the longer the reaction time is.
The ethylation reaction is carried out either with excess sodium hydroxide
and limited ethyl chloride, or with limited sodium hydroxide and excess
ethyl chloride.
Ethylation efficiency is the percentage of the total ethyl chloride con-
sumed that becomes substituent upon the cellulose. Efficiency is related
both to the concentration and the amount of sodium hydroxide present
during the reaction and to efficient by-product recovery and utilization.
Efficiency is reduced by the presence of water or of alcohols. Efficiency is
higher, the more concentrated the sodium hydroxide that is present.
188 A. F. Martin, private communication.
916 CELLULOSE
(c) ALKALI CELLULOSE MANUFACTURE
To prepare alkali cellulose when excess sodium hydroxide is used, the
sheet cellulose is first reduced to fibrous shreds in a Stern shredder.139 The
shredded cellulose is blown into a continuous horizontal conveyor, where it
is moistened by a spray of aqueous sodium hydroxide solution.140 The re-
sultant slurry is loaded to the ethylators without aging or ripening.
When limited sodium hydroxide is used, the alkali cellulose preparation is
much more critical. The cellulose sheet is passed continuously through a
bath of aqueous 50 to 76% NaOH at 55° to 130°C.141-142 The alkali cellu-
lose, which is swelled to maximum size after impregnation, is wiped to
increase its density,142 is aged for a few seconds at a temperature between
its hardening point (15° to 18°C. below the freezing point of the aqueous
sodium hydroxide used) and its degradation point (130°C.), and is cooled.143
The aging time and temperature vary inversely with the intrinsic viscosity
desired in the ethyl cellulose end product.
If the alkali cellulose is not aged, the viscosity may be controlled by the
addition of air to the ethylator.144 Viscosity may also be controlled by the
use of oxygen carriers such as manganese or cobalt compounds, oxidizing
agents,146 or acid gases.146 In essence, the intrinsic viscosity of the product
is controlled by mild hydrolysis146 or by mild oxidation143""145 of the cellu-
lose. If air is used as the oxidant, the amount required (shown in Table
29) is small.
The intrinsic fluidity, [<J>] (that is, the reciprocal of the intrinsic viscos-
189 R. L. Stern (to Hercules Powder Co.), U. S. Patent 2,028,080 (Jan. 14, 1936);
Chent. Abstracts* 30, 1561 (1936) ; A. S. Finlayson (to Hercules Powder Co.), U. S. Patent
2,313,866 (Mar. 16, 1943) ; Chem. Abstracts, 37, 5237 (1943).
140 "Integration of Chemical Plant Facilities,1' Chem. Met. Eng., 52, 129 (Sept., 1945).
i« S. L. Bass (to The Dow Chemical Co.), U. S. Patent 2,143,855 (Jan. 7, 1939);
Chem. Abstracts, 33, 3150 (1939).
141 W. R. Collings, L. DePree, and M. H. Weymouth (to The Dow Chemical Co.),
U. S. Patent 2,143,863 (Jan. 17, 1939); Chem. Abstracts, 33, 3150 (1939); U. S. Patent
2,145,862 (Feb. 7, 1939) ; Chem. Abstracts, 33, 3586 (1939).
148 F. C. Peterson and A. T. Maasberg (to The Dow Chemical Co.), U. S. Patents
2,149,309 and 2,149,310 (Mar. 7, 1939) ; Chem. Abstracts, 33, 4421 (1939).
"4 R. B. Darling (to Hercules Powder Co.), U. S. Patent 2,492,524 (Dec. 27, 1949);
Chem. Abstracts, 44, 2237 (1950).
»« E. D. Klug (to Hercules Powder Co.), U. S. Patent 2,523,377 (Sept. 26, 1950);
Chem. Abstracts, 45, 1344 (1951); E. D. Klug and H. M. Spurlin (to Hercules Powder
Co.), U. S. Patent 2,512,338 (June 20, 1950); Chem. Abstracts, 44, 8656 (1950).
148 R. D. Freeman and M. J. Roberts (to The Dow Chemical Co.), U. S. Patents 2,159,-
375 and 2,159,376 (May 23, 1938) ; Chem. Abstracts, 33, 7108 (1939) ; U. S. Patent 2,159,-
377 (July 19, 1938) ; Chem. Abstracts, 33, 7109 (1939).
DC. DERIVATIVES OF CELLULOSE 917
ity), has an approximately linear relationship to the weight of oxygen con-
sumed. The equation of this line is
[*J « 2.2 + 1.2* (5)
where x is the pounds of air used per 100 pounds oi ceUulose. Tke amount
of air required to control viscosity varies with tke aging temperature, e&sy\
ation temperature, aqueous sodium hydroxide concentration, and otixs
factors.
TABLE 29
Control of Ethyl Cellulose Intrinsic Viscosity (Darling144)
Air/100 Ib. of cellu/ose
Viscosity,
cps.,5%
solution8
Intrinsic
viscosity"
Intrinsic
• fluidity
Cubic feet at
standard
conditions
Pounds
393
3.48
0.29
5.7
0.46
125
2.35
0.42
21.8
1.76
105
2.22
0.45
22.8
1.84
75
1.97
0.51
26.8
2.16
B Viscosity and intrinsic viscosity solvent: 80:20 toluene: ethanol.
(d) ETHYLATION
Ethyl cellulose is prepared by the etherification of alkali cellulose with
ethyl chloride. The reaction is:
Rc.u(OH),.3NaOH + 2 CH8CH2C1 >
Rceii(OH)(OCH2CH3)2 + 2 NaCl + NaOH + 2 H2O (6)
The number of moles that react varies with the degree of substitution and
with the substituent distribution. The reaction is retarded as the water
concentration increases and as the sodium hydroxide concentration de-
creases.
Measurable amounts of carboxyl groups appear in the product as the
result of oxidation of the alkali cellulose or of the ethyl cellulose.
The chief by-product reactions are :
CH»CH2C1 + NaOH > CH8CH2OH + NaCl (7)
the rate of which is proportional to the sodium hydroxide concentration,
and
CH,CH,OH + CH,CH2C1 + NaOH » CH3CH2OCH2CH3 + NaCl + H2O (8)
the extent of which is proportional to the ethanol concentration and to the
reaction time. This last reaction is rapid if alcohol is added147; otherwise
918 CELLULOSE
it is negligible at first but increases with time as alcohol is produced by
reaction 7.
Other side reactions include the oxidation of ethanol to acetaldehyde and
higher aldehydes and the reaction of these aldehydes in the presence of
sodium hydroxide to yield colored resinous bodies.
By-product formation and cellulose degradation are rapid if the primary
ethylation rate is slow.147
Ethylations are carried out in jacketed, agitated, nickel-clad autoclaves.
If ethyl chloride is used alone, the working pressure is about 400 Ib./sq. in.
gage; if a diluent is used, the pressure may be as low as 175 Ib./sq. in. gage.
Thorough agitation is of extreme importance. The reaction is mildly exo-
thermic; the heat liberated is removed by condensation of the solvents on
the shell.
Ethylation methods that are recognized include : single-stage with ethyl
chloride alone144-147-148; and multistage with solid sodium hydroxide addi-
tion,149 either with ethyl chloride alone or with ethyl chloride and a diluent.
These methods may be modified by variations of time, temperature, and
charge.
The alkali cellulose is prepared to conform to the ethylation conditions
that are to be used. The relation of alkali cellulose composition and of
ethylation conditions to efficiency at diverse degrees of substitution was
studied by Swinehart and Maasberg,149 They ethylated alkali cellulose142
that contained from 3.2 to 4.5 moles of sodium hydroxide per glucopyranose
residue and from 2.5 to 3.4 moles of water per glucopyranose residue.
The alkali cellulose was reacted with a large excess of ethyl chloride at 90°
to 120°C. until the D.S. was 1.7 to 2.0. At this point the reaction mass con-
tained 1.8% residual sodium hydroxide; water formation had diluted the
sodium hydroxide to a concentration of 30 to 50%.
Solid sodium hydroxide was then added to bring the sodium hydroxide
concentration back up to 55 to 75%. The reaction was continued until
the desired degree of substitution was obtained; water formation had
again diluted the sodium hydroxide to a concentration of 30 to 50%. The
147 E. J. Lorand (to Hercules Powder Co.), U. S. Patent 2,096,681 (Oct. 19, 1937);
Chem. Abstracts. 32, 353 (1938); U. S. Patent 2,110,526 (Mar. 8, 1938); Chem. Ab-
stracts, 32, 3611 (1938); U. S. Patent 2,130,998 (Sept. 20, 1938); Chem. Abstracts, 32,
9495(1938).
148 W. R. Collings and L. DePree (to The Dow Chemical Co.), U. S. Patent 2,163,869
(June 27, 1939) ; Chem. Abstracts, 33, 8012 (1939).
149 R. W. Swinehart and A. T. Maasberg (to The Dow Chemical Co.), U. S. Patent
2,254,249 (Sept, 2, 1941) ; Chem. Abstracts, 35, 8295 (1941).
IX. DERIVATIVES OF CELLULOSE 919
minimum quantities of sodium hydroxide that are required to obtain cer-
tain degrees of substitution are shown in Table 30.
TABLE 30
Ethylation : Relation of Sodium Hydroxide : Cellulose Ratio to Degree of Substitution
(Swinehart and Maasberg149)
D.S.
Ethoxyl,
%
Minimum NaOH;
: cellulose ratio
Minimum
NaOH
concentration.
%
Weight
Moles
2.2
44
1.1
4.5
48
2.3
46
1.4
5.7
50
2.5
48.5-49
1.7
6.9
56
2.6
50
2.0
8.1
60
2.8
52
2.5
10.1
74
When ethyl chloride is used alone, a multistage reaction with sodium
hydroxide addition is more efficient than a single-stage reaction; a multi-
stage reaction containing such a diluent as benzene is even more efficient.
The relation of these factors to efficiency is shown in Table 31.
TABLE 31
Ethylation: Relation of Method to Efficiency (Swinehart and Maasberg149)
Ethyl chloride : cellulose
weight ratio consumed
Efficiency, %°
D.S.
Ethoxyl,
%
One-
stage
NaOH
Two-
stage
NaOH
Two-
stage
NaOH
in
benzene
One-
stage
NaOH
Two-
stage
NaOH
Two-
stage
NaOH
in
benzene
0.6
15
1.1
0.8
0.45
22
30
53
1.3
30
1.5
1.1
0.9
38
51
62
1.6
35
2.4
1.2
1.1
28
57
62
1.9
40
2.5
1.5
1.35
33
54
60
2,25
45
2.6
2.2
1.6
36
43
58
2.42
47.5
2.9
2.6
1.9
35
39
53
2.47
48
3.1
2.7
2.0
33
38
51
2.62
50
—
—
3.17
—
33
—
0 First-stage efficiency 50 to 60%.
In practice, more than the minimum quantity of sodium hydroxide shown
in Table 31 is used. Since water retards reaction 6, the strength of the
sodium hydroxide that is used to prepare the alkali cellulose must be in-
creased as the desired degree of substitution increases. The minimum
concentration required is shown in the last column of Table 30.
920 CELLULOSE
The consumption of ethyl chloride during the reaction can be followed by
analysis of the reaction mass for salt and sodium hydroxide. When the
desired ethyl chloride consumption is reached, the sample is tested for ap-
proval with respect to degree of substitution and viscosity.
(e) PRECIPITATION
The ethyl cellulose formed amounts to about 8% of the reaction mass;
it must be separated from the mixture of salt, sodium hydroxide, water,
and solvents in which it is dissolved.
The reaction mass, if viscous, is diluted,150 strained, and precipitated150'151
either in an autoclave or in separate equipment. When the volatile solvents
(ethyl chloride, ether, ethanol, and benzene) are flashed off, the ethyl cellu-
lose is left in suspension as dense, porous granules. If excess sodium hy-
droxide was used, the excess is recovered. The solvents are condensed,
fractionated, and re-used. Excess ethanol and ether are either sold or
reconverted to ethyl chloride.
(f) BY-PRODUCT CONVERSION
Effective by-product recovery and conversion is the key to efficient oper-
ation. Ethanol and ethyl ether are reconverted to ethyl chloride by high-
temperature, aqueous-phase hydrochlorination with hydrogen chloride in
the presence of a heavy metal salt catalyst such as zinc chloride.152*158
Ethyl ether is also reconverted to ethyl chloride by high-temperature, gas-
phase hydrochlorination over a solid catalyst.
(g) PURIFICATION
The impure ethyl cellulose is washed with water until free of alkali and
salt. The washing is carried out in one or more agitated wash tanks; the
filter media may be false bottoms in the tanks or they may be separate
filter units. If the granules are coarse they may be ground through a
knife mill to increase their leachability.
The ethyl cellulose is given special purification treatments during wash-
ing to prepare it for its intended end use. Such treatments may include
bleaching with sodium hypochlorite184 or with sodium chlorite, acid treat-
«° W. R. Ceilings (to The Dow Chemical Co.), U. S. Patents 2,121,731 and 2,121,732
(June 21, 1938); Chem. Abstracts, 32, 6461 (1038).
Wl H. M. Spurlin (one-half to Hercules Powder Co.. one-half to The Dow Chemical
Co.), U. S. Patent 2,249,673 (July 15, 1941); Chem. Abstracts, 35, 6791 (1941).
»» H. M. Spurliil (to Hercules Powder Co.), U. S. Patent 2,084,710 (June 22, 1937);
Chem. Abstracts, 31, 6816 (1937).
»» R. P. Carter (to Hercules Powder Co.), U. S. Patent 2,396,639 (Mar. 19, 1946);
Chem. Abstracts, 40, 4076 (1946).
IX. DERIVATIVES OF CELLULOSE 921
ment,165 deashing,168 or fractionation.157'168 Following purification, the
ethyl cellulose is centrifuged and dried to a low moisture content.
(h) STABILITY OF ETHYL CELLULOSE
Ethyl cellulose is stable toward such chemical agents as alkalies, salts,
and water. It is stable in the presence of sunlight or ultraviolet light and
at temperatures above its softening point, provided that it has been properly
prepared and formulated.
The thermal stability of ethyl cellulose is related to the presence of func-
tional groups, such as carbonyl and carboxyl. When the purification of the
ethyl cellulose is completed, the carboxyl groups that were formed by oxida-
tive ring or chain breakage may be left in any degree of neutralization from
a form in which they are completely bound by base metals to a metal-free
form. In the final formulation there must be a suitable balance of acidity
to basicity. If the formulation is too acidic, it will degrade in viscosity and
strength; if the formulation is too basic, it will discolor when subjected to
heat. This effect is of minor concern in uses that do not involve heat,
although Evans and Spurlin156 found that bound metal increases the viscosity
of solutions in nonpolar solvents over that which is found for deashed ethyl
cellulose.
The oxidation of ethyl cellulose was studied by Evans and McBur-
neyi59,i6o ^Q passed oxygen through ethyl cellulose of 0.9 intrinsic viscosity
at 50° to 109°C. with and without irradiation by ultraviolet light. Their
mechanism for the oxidation is based upon hydroperoxide and acetaldehyde
formation. Their work showed the sensitizing action of easily oxidized
material, such as acetaldehyde and oxidized ethyl cellulose itself, and the
effectiveness of such oxidation retardants and inhibitors as diamylphenol,187
diphenylamine, 159 and copper salts. 161 Chamberlain19 in further work found
1M J. McHard and F. C. Peterson (to The Dow Chemical Co.), U. S. Patent 2,238,912
(Apr. 22, 1941) ; Chem. Abstracts, 35, 4951 (1941).
188 A. S. Finlayson (to Hercules Powder Co.), U. S. Patent 2,178,630 (Nov. 27, 1939);
Chem. Abstracts, 34, 1482 (1940).
» E. F. Evans and H. M. Spurlin, /. Am. Chem. Soc., 72, 4750 (1950).
™ H. M. Spurlin (to Hercules Powder Co.), U. S. Patent 2,214,070 (Sept. 10, 1940);
Chem. Abstracts, 35, 893 (1941).
1M J. H. Sharphouse, P. R. Hawtin, John Downing, and W. H. Groombridge (to British
Celanese, Ltd.), Brit. Patent 556,664 (Oct. 15, 1943); Chem. Abstracts, 39, 1992 (1945).
See I. Okamura, Cellulosechemie, 14, 135 (1933).
»• L. F. McBurney, Ind. Eng. Chem., 41, 1256 (1949).
160 E. F. Evans and L. F. McBurney, Ind. Eng. Chem., 41, 1256 (1949).
M1 P. VanWyck (to Hercules Powder Co.), U. S. Patents 2,561,892 and 2,561,893
(July 24, 1951).
922 CELLULOSE
that their mechanism is well supported, but that the rate of oxidation de-
pends largely upon the history of the sample.
The heat stability of ethyl cellulose is determined in practice under the
conditions of its end use. Base flake formulated into molding powder for
plastic end uses is tested by retention at 240°C. in the tunnel of an injec-
tion-molding press or a compression-molding press for an extended time.
The change in intrinsic fluidity is then determined. Chamberlain19 found
that the exclusion of oxygen from the press by nitrogen blanketing does not
bring about great changes in viscosity, but does definitely improve color.
Since there is not sufficient oxygen in a press to produce great changes in
viscosity, it is found that the change in fluidity varies with the history of
the sample, and that both hydrolysis and oxidation may occur.
When properly prepared and formulated, ethyl cellulose will produce
low-color plastics with high retention of strength and durability over a
broad range of temperature.
Ethyl cellulose for outdoor uses is tested by atmospheric exposures in a
suitable climate. Ethyl cellulose having a small intrinsic fluidity rise is
most suitable for these uses, and the working range below excessive fluidity
improves at higher viscosities.
(i) ETHYL CELLULOSE PLASTIC
Ethyl cellulose plastic articles may be fabricated by extrusion or by in-
jection molding. The properties of the plastic may be varied to yield
hard-surfaced extruded shapes, medium-hard molded shapes, or molded
shapes having outstanding low-temperature flexibility. Table 32 sum-
marizes the properties of ethyl cellulose plastics.19 The values given are
conservative, but they exceed the proposed ASTM specifications for ethyl
cellulose molding compounds.
(j) ETHYL CELLULOSE SHEETING
Ethyl cellulose sheeting contains little plasticizer; its properties are
essentially those of the base flake. Table 33 (see p. 924) shows the prop-
erties of ethyl cellulose sheeting.19
(k) ETHYL CELLULOSE BASE FLAKE
The solubility and the thermoplasticity of ethyl cellulose depend upon
its degree of substitution and upon its intrinsic viscosity. The softening
temperature and the melting temperature decrease with increasing degree
of substitution to a minimum at D.S. 2.50; above this substitution these
properties again increase. At a fixed degree of substitution, the softening
IX. DERIVATIVES OF CELLULOSE 923
TABLE 32. Average Properties of Ethyl Cellulose Plastics
(The Dow Chemical Company19)
Test
method
Property A.S.T.M.
Injection
Kxtrusioii
hard
Medium
Low
temperature
1.
Mechanical properties:
Tensile strength, 77 °F.,
Ib./sq. in. D638-49T
7000
5000
3500
Elongation in tension,
77 °F., % D638-49T
6
10-30
10-25
Modulus of elasticity in
tension, 77 °F., Ib./sq. in. D638-49T
3 X 10*
2 X 10*
1.5 X 10*
Impact strength, Izod,
notched,
77°R, ft.-lb./in. D256-47T
2.0
2.5
4.5
0°R, ft.-lb./in.
1.5
1.0
2.0
-40°F., ft.-lb./in.
1.0
0.5
1.2
Hardness, Rockwell D785-48T
R110
R100
R70
2
Thermal properties:
Flow temperature, °C. 1)569-48
155
145
150
Heat distortion, °C , 264
Ib./sq. in D648-45T
70
60
55
Weight loss on heating,
max., % D787-46T
0.2
0.4
2.0
3.
Electrical properties:
Dielectric constant
108 cycles/sec. D150-47T
3.3
3.1
3.4
106 cycles /sec.
3.2
3.0
3.2
108 cycles/sec.
3.0
2.8
2.9
Power factor, %
108 cycles/sec. D150-47T
0.3
1.3
1.0
108 cycles/sec.
2.0
1.6
2.4
4.
Chemical properties:
Chemical resistance —
a
a
a
Water sorption, % gain in
24 hrs. D570-42
1.5
1.2
1.5
Water solubility, max., % D570-42
0.2
0.2
0.4
5.
Fabrication :
Specific gravity D792-48T
1.12
1.10
1.10
Injection molding, flow
grade
H
M
MH
Machine temperature, °F.
—
430
420
Machine pressure, lb./
sq. in. gage
—
15,000
15,000
Extrusion temperature, °F.
440
410
410
Color possibilities
Opaque
Transparent
to opaque
Machining properties
Excellent
Good
Good
a Chemical resistance: weak acids, no effect; strong acids, severe attack; weak alka-
lies, no effect : strong alkalies, slight attack ; organic solvents, widely soluble.
924
CELLULOSE
TABLE 33. Average Properties of Ethyl Cellulose Sheeting
(The Dow Chemical Company19)
Property
Test method
A.S.T.M.
Value
Thickness,
in.
1. Mechanical properties:
Tensile strength, Ib./sq. in.
Elongation, %
Folding endurance, M.I.T. double folds,
75 °F., 50%R.H.
Bursting strength, Ib./sq. in.
Tear strength, Elmendorf, g./O.OOl in.
2. Thermal properties:
Softening temperature, °C.
Melting temperature, °C.
Specific heat, B.t.u./lb.
3. Electrical properties:
Dielectric constant
60 cycles/sec.
108 cycles/sec.
10* cycles/sec.
Power factor, %
60 cycles/sec.
10* cycles/sec.
108 cycles/sec.
Dielectric strength, volts/0.001 in.
4. Chemical properties:
Water sorption, %, 100 °F., 100% R.H.
%, 24-hr, immersion,
80 °F.
Normal moisture content, %, 75 °F.,
50% R.H.
Moisture vapor transmission, g./lOO sq.
in./24 hrs., 100 °F., 95% R.H.
Chemical resistance
5. Fabrication:
Specific gravity
6. Optical properties:
Refractive index, n*i>
Transmission of white light, %
Ultraviolet cut-off, A.
Transmission of infrared (except narrow
absorption band at 10 ~* cm.), %
Fade-Ometer, 200 hrs.
yellowing
embrittlement
D150-44T
D150-44T
D570-42
Modified
General
Foods6
D71-27
8000
20-35
2750
85
97
154
191
0.348
3.2
3.1
3.0
1.2
0.4
2.0
3500
3.3
7.5
1.4-1.7
35
1.15
1.47
88-92
2200
90
almost none
none
0.001
0.001
0.005
0.002
0.002
a Chemical resistance: weak acids, no effect; strong acids, severe attack; weak alka-
lies, no effect; strong alkalies, slight attack; organic solvents, widely soluble.
6 Thwing- Albert Vapometer in modified General Foods Cabinet; cabinet and pro-
cedure modified by The Dow Chemical Co.19
IX. DERIVATIVES OP CELLULOSE
925
and melting temperature increase as the intrinsic viscosity increases. The
relation among these factors is shown in Figure 62; the fine lines represent
the spread in temperature due to variant intrinsic viscosity* 1M
220
210
200
190
180
170
ETHOXYL CONTENT, PERCENT
45 46 47 48
SO
ui
CL
160
ISO
140
130
120
SOFTENING POINT RANGE
2J7
2.24
230 238 2.44 230 233 238 2.60
DEGREE OF SUBSTITUTION
Fig. 62. Ethyl cellulose: relation of softening-melting point range to degree of
substitution and to intrinsic viscosity (The Dow Chemical Co.w).
The solubility of ethyl cellulose varies in the same manner as its thermo-
plasticity; ethyl cellulose of D.S. 2.1 to 2.4 is soluble only in 70:30 ben-
1M Ethocel Handbook. The Dow Chemical Co., Midland, Mich., 1940.
926
CELLULOSE
zenerinethanol, in 60:40 toluene rethanol, in methyl acetate, or in ethylene
dichloride. When the substitution is increased above D.S. 2.4, solubility in
a wide variety of solvents results. Above D.S. 2.55 alcohol solubility is lost
but hydrocarbon dispersibility is retained.137'162 If the substitution is fixed,
the solubility increases as the viscosity is lowered.
150
.9 U> 15 2.0 2.5 &0
INTRINSIC VISCOSITY
Fig. 63. Ethyl cellulose: relation of viscosity of 5% solution (in cps.) at
25°C. to intrinsic viscosity (The Dow Chemical Co.19). Lines: a, D.S. 2.5
in 80:20 toluene :ethanol; bt D.S. 2.3 in 70:30 benzene :methanol.
Two substitution grades of ethyl cellulose satisfy most needs. D.S. 2.24
to D.S. 2.38 (45.0 to 47.0% ethoxyl) is used for plastics and wherever hard-
ness, strength, and heat resistance are factors; D.S. 2.44 to D.S. 2.58 (48.0
to 49.5% ethoxyl) is used where ready solubility, wide compatibility, and
ready softening are factors. Each substitution grade is produced in several
IX. DERIVATIVES OF CELLULOSE 927
viscosity types. A low-viscosity type is used when high concentration in
solution or high flow is a factor; a high-viscosity type is used when strength,
flexibility, or hardness is a factor.
The intrinsic viscosity has been used in this Section of Chapter IX so
that different cellulose ethers may be compared directly. Commerical ethyl
cellulose, however, is sold on the basis of the viscosity of its 5% solution by
weight in a specific solvent. The specific solvents for 5% solution viscosity
measurements are :
Degree of Solvent composition by
substitution Ethoxyl, % volume
2 . 24-2 .38 45 . 0-47 .0 70 : 30 benzene : rnethanol
2 . 44-2 .58 48 . 0-49 .5 80 : 20 toluene : ethanol
Certain specifications require that the viscosity of D.S. 2.24 to 2.38 ethyl
cellulose be reported in terms of 60:40 toluene: ethanol solvents; the ratio
of the viscosity in 60:40 toluene : ethanol to the 70:30 viscosity is 1.34:1.
The intrinsic viscosity of ethyl cellulose in a specific solvent is related to
its 5% viscosity in the same solvent by the empirical equation19:
(5% solution viscosity, cps.) « A (intrinsic viscosity)2'9 (9)
where the intercept constant, A, is:
A D.S. Solvent
(a) 10 6 2.5 80:20 toluene: ethanol
(b) 7.2 2.3 70:30 benzene :methanol
Figure 63 shows the relation of 5% solution viscosity to intrinsic viscosity
for commercial ethyl cellulose.
The viscosity of ethyl cellulose solutions of finite concentration is lower
in good solvents than in poorer solvents. The viscosity is influenced by the
base-exchange state of the end groups.156 When concentrations of 5% or
higher are used, the relation of viscosity to concentration may be expressed
by Philippoff s equation168:
(10)
Kauppi and Bass164 used equation (10) to construct a viscosity-concen-
tration chart (Fig. 64) for ethyl cellulose at finite concentrations. If the
viscosity of an ethyl cellulose is known at one concentration, its viscosity at
any other concentration may be found from Figure 64. Philippoff found
that any means of altering the viscosity (for example, by changing solute,
l" W. Philippoff and K. Hess, Z. physik. Chem., 31B, 237 (1936).
w< T. A. Kauppi and S. L, Bass, Ind, Eng. Chem., 29, 800 (1937).
928
CELLULOSE
solvent, or temperature) changed the intercepts of a line but not its line-
arity.
Fig. 64. Ethyl cellulose: viscosity-concentration chart (Kauppi and Bass164).
(1) FORMULATION OF ETHYL CELLULOSE
The formulation of ethyl cellulose is discussed by the manufacturers.187*162
Aromatic hydrocarbon-alcohol blends are suitable solvents, with modifica-
tion, for most purposes. Plasticizers, usually ethers, esters, or nonsolvent
oils, impart flexibility and flow. Resins impart hardness and adhesion.
Ethyl cellulose is compatible with many resins and with nitrocellulose, but
not with most other high polymers. Ethyl cellulose is compatible with
many waxes to form melts; compatibility is increased by the use of mutu-
ally compatible materials.
IX. DERIVATIVES OF CELLULOSE 929
(m) HIGH-ETHOXYL ETHYL CELLULOSE
Nearly completely substituted ethyl cellulose of D.S. 2.60 to 2.80 (ethoxyl
content 50.0 to 52.5%) has been prepared by the method of Swinehart and
Maasberg.149 The large amounts of sodium hydroxide and salt involved
tend to separate the reaction mass into layers, so that decreased diffusion
lengthens the reaction time. The quantities of reagents involved are shown
in Tables 30 and 31.
Ethyl cellulose of D.S. 2.6 to 2.8 is dispersed in hydrocarbons. Plasticiza-
tion is required to increase strength and to eliminate haze. No plasticizer
has been found to act as a solvent or to lower the melting point. High-
ethoxyl ethyl cellulose is incompatible with ethyl cellulose of lower substitu-
tion. Sheeting of high-ethoxyl ethyl cellulose has a tensile strength of 5000
Ib./sq. in., an elongation of 10 to 30%, a melting point of 240°C., and no
softening point.
(n) LOW-SUBSTITUTED ETHYL CELLULOSE
The very low-substituted ethyl celluloses resemble the more common
water- or alkali-soluble methyl celluloses, carboxymethyl celluloses, and
hydroxyalkyl celluloses in properties. Ethyl cellulose of D.S. 0.3 to 0.5 is
soluble only in 4 to 10% aqueous sodium hydroxide.84'125 Ethyl cellulose
of D.S. 0.7 to 1.3 is soluble in cold water.19'84-125 Such ethers are soluble at
still lower substitution values when the solutions are chilled or when the
cellulose has been degraded.
Alkali-soluble ethyl cellulose is prepared by the treatment of alkali cellu-
lose with ethyl sulfate123-124 or with ethyl chloride.19'126 The residual so-
dium hydroxide is neutralized, and the product is washed free of salt, and
then dried. The alkali-soluble ethers are dissolved in 4 to 10% aqueous
NaOH, cast or formed, and coagulated by acid treatment. The washed
and dried sheeting resembles cellophane.
Water-soluble ethyl cellulose is produced by the ethylation of alkali cel-
lulose with ethyl chloride.19 The alkali cellulose composition and the reac-
tion temperature are controlled to yield the type of solubility and the gela-
tion temperature desired.
(o) MODIFIED ETHYL CELLULOSE
Sonnerskog165 and Jullander166 have described ethyl hydroxyethyl cellu-
loses. At M.S. 0.9 hydroxyethoxyl and D.S. 0.9 ethoxyl, the product is
»» S. Sdnnerskog, Svensk Papperstidn., 48, 413 (1945).
» I. Jullander, Svensk Papperstidn., 55, 197 (1952).
930 CELLULOSE
water-soluble, whereas at M.S. 0.7 and D.S. 1.34 the product is soluble in
methylene chloride-alcohol.
5. Methyl Cellulose
The uses of methyl cellulose stem from its cold-water solubility and from
the toughness of its sheeting. Methyl cellulose thickens water solutions
without precipitation over a pH range from 3 to 12. Methyl cellulose gels
from solution upon heating or upon salt addition; this feature is utilized
in many applications. Methyl cellulose can be cross-linked to insolubility
after forming. The strength and toughness of methyl cellulose films add
strength to adhesives in which methyl cellulose is compounded. The sur-
face-active properties of methyl cellulose aid in the stabilization of latexes
and emulsions. The physiological inertness and the storage stability of
methyl cellulose permit its use in cosmetics, pharmaceuticals, and food
products.
Unplasticized methyl cellulose is heat resistant and not thermoplastic;
when methyl cellulose is plasticized by certain liquids that dissolve it at
high temperatures, the product is thermoplastic.167
Methyl cellulose is made in several ranges of substitution. Commercial
methyl cellulose (D.S. 1.6 to 2.0, methoxyl content 26.5 to 32.6%) is soluble
in cold water; nearly completely substituted methyl cellulose (D.S. 2.4 to
2.8, methoxyl content 38.0 to 43.0%) is soluble in polar organic solvents;
low-substituted methyl cellulose (D.S. 0.1 to. 0.9, methoxyl content 2 to
16%) is soluble in 4 to 10% aqueous sodium hydroxide.
Chemically modified methyl celluloses combine the properties of methyl
cellulose with those of other cellulose ethers.
Methyl cellulose is prepared by the etherification of alkali cellulose with
methyl chloride, followed by the isolation, washing, and drying of the prod-
uct.
(a) RAW MATERIALS
Methyl cellulose is prepared from wood or cotton cellulose suitable for
ethylation. Very high viscosity grades are made from pulp having a 0.5%
viscosity of 40 to 60 cps. (viscosity in 0.5% cupriethylenediamine solution,
TAPPI Standard T 230). Refrigeration-grade methyl chloride and sodium
hydroxide suitable for ethylation are used.
187 G. K. Greminger, R. M. Upright, and L. H. Silvernail, in Protein and Synthetic
Adhesives, Chapter VII, Tappi Monograph Series No. 9t Technical Association of the
Pulp and Paper Industry, New York, N. Y. (1952); L. H. Silvernail (to The Dow
Chemical Co.), U. S. Patent 2,602,755 (July 8, 1952) ; Chem. Abstracts, 46, 989 1(1952)
IX. DERIVATIVES OF CELLULOSE 931
(b) METHYLATION
Methyl cellulose is prepared from alkali cellulose that contains limited
sodium hydroxide.141""143 The aging time varies from a few seconds to a
number of minutes. The time and temperature chosen vary inversely as
the viscosity desired. The reactions are similar to equations 6, 7, and 8.
Methylations are carried out in jacketed, agitated, nickel-clad auto-
claves at a working pressure of about 200 Ib./sq, in. gage. The reaction is
mildly exothermic; the heat liberated is removed by condensation of the
solvents on the shell.
The relation of alkali cellulose composition to methyl cellulose properties
was studied by Maasberg and others.168 An alkali cellulose prepared from
37.8 to 57.2% NaOH and containing a weight ratio of sodium hydroxide to
cellulose of 0.9 to 1.2 and a weight ratio of water to cellulose of 0.9 to 1.5
yielded, on methylation with a slight stoichiometrical excess of methyl
chloride at 50° to 100°C., a cold-water-soluble methyl cellulose of D.S. 1.6
to 2.0 that could be washed with hot water at 85° to 100°C. and dried.
The physical properties of the product were controlled by the relation of
time to temperature and pressure during processing.169
Methylation efficiency varies from 70 to 80% for alkali-soluble methyl
cellulose to 40 to 50% for water-soluble methyl cellulose.
(c) METHYL CELLULOSE BASE FLAKE
Methyl cellulose base flake170 is a white powder of bulk density 0.3 to 0.5
g./ml. It is heat stable up to 225°C., but it browns slightly upon continued
heating above 190°C., and chars without softening at 225° to 230°C. It
is soluble in cold water, but insoluble in hot water or in saturated salt solu-
tions.
(d) METHYL CELLULOSE SOLUTIONS
Methyl cellulose solutions are prepared by wetting the powdery base flake
with water at 75° to 80°C., and then adding the remaining water while
cooling and agitating. Solution clarity is improved by chilling to below
10°C.
Figure 65 shows the relation of solution viscosity to concentration for
methyl celluloses having intrinsic viscosities (in water at 20°C.) of from
*« A. T. Maasberg (to The Dow Chemical Co.), U. S. Patent 2,160,782 (May 30,
1939) ; Chem. Abstracts, 33, 7563 (1939).
Ml R. W. Swinehart and A. T. Maasberg (to The Dow Chemical Co.), U. S. Patents
2,331,864 and 2,331,865 (May 12, 1943) ; Chem. Abstracts, 38, 1693 (1944).
™ Methoccl, The Dow Chemical Co., Midland, Mich., 1949.
932
CELLULOSE
1.4 to 7.3. The nominal 2% solution viscosities are shown in parentheses.
The intrinsic viscosity of methyl cellulose in water is related to its 2%
viscosity in water at 20°C. by the empirical equation19:
(2% viscosity, cps.) » A( intrinsic viscosity)3-6 (11)
10000
5000
CO
Ul
<o
o
(L
H
bl
O
CO
O
O
CO
• 2000
1000
500
A
200
100
10 ie 14
WEIGHT PERCENT
l«
10
02468
CONCENTRATION ,
Fig. 65. Methyl cellulose: viscosity-concentration chart (The Dow Chemical
Co.19). Values of intrinsic viscosity and of viscosity in 2% solution in water at
20 °C. are shown on the curves.
where the intercept constant, A, is 0.34 for granular methyl cellulose and
0.71 for fibrous methyl cellulose. The difference in intercepts between
types is due to the processing of the granular form.189 Figure 66 shows the
IX. DERIVATIVES OF CELLULOSE
933
I 2 3456
INTRINSIC VISCOSITY
Fig. 66. Methyl cellulose: relation of viscosity of 2% water solution at 20°C. to
intrinsic viscosity (The Dow Chemical Co.w).
934
CELLULOSE
3UU0
4000
3000
tooo
1000
0 900
O
OJ
300
•»
CO
III
2 too
o
flL
o «oo
>
CO
8 ••
CO
30
to
10
5
1
V
^S
y{0%
J
^
^
V
^
k
5%
\
V
^^^,
-^i
S_ _j^
^'•^^
\
X
X
4%
^ '
^^^
"^
^v^
*
^^^
X.
J
"^
m*r
10 tO 80 40 *
TEMPERATURE, °C.
70
Fig. 67. Methyl cellulose: relation of solution viscosity and
gel point to temperature and concentration (The Dow Chemical
Co. »). (Intrinsic viscosity, 2»0t )
IX. DERIVATIVES OF CELLULOSE 935
relation of 2% solution viscosity at 20° C. to intrinsic viscosity for commer-
cial methyl celluloses.
When a methyl cellulose solution is heated, the viscosity decreases to a
minimum just below the gelation temperature, then rises rapidly when the
gel point is reached. The relation of solution viscosity to temperature is
shown in Figure 67 for methyl cellulose of intrinsic viscosity 2.0. The gela-
tion temperature decreased as the concentration increased.
The effect of salts167 is similar to that of heating. The effect of sodium
chloride addition to a 2% solution of methyl cellulose of intrinsic viscosity
5.5 is:
Salt concentration, % Gelation temperature, °C.
0 50
1 50
5 44
10 33
Solution of methyl cellulose in water reduces the surface tension of the
liquid-air interface from 72 dynes/cm, for water alone to a value of 50
dynes/cm. The surface tension is independent of viscosity and concentra-
tion.
(e) METHYL CELLULOSE SHEETING
Methyl cellulose sheeting has been cast from water solutions. Such
sheeting can be plasticized by moisture. The properties of methyl cellulose
sheeting are shown in Table 34. The flexibility of methyl cellulose sheeting
TABLE 34
Properties of Methyl Cellulose Sheeting (The Dow Chemical Company170) at 73 °F. and
50% Relative Humidity
Property Value
Specific gravity 1 . 37-1 . 45
Tensile strength, Ib. /sq. in. 8500-1 1 , 400
Elongation, % 10-15
M.I.T. double folds 12,000
Ultraviolet
Resistance Excellent
Transmission, 0.001-in. film, 400 HIM, % 90
290 mM, % 84
210 HIM, % 54
Oil resistance (vegetable, animal, and mineral oils) Impervious
Water content, % 6.5
936 CELLULOSE
is increased by plasticization. Plastitizers and additives include water,
sugars, glycols, polyglycols, phosphates, alcohol amines, and hygroscopic
salts.
(f) COMPATIBILITY
Methyl cellulose can be blended with starches, glues, soaps, dextrins,
and water-dispersible natural gums. Methyl cellulose is compatible with
many water-soluble resins and up to 40% with starches.
(g) INCREASING WATER RESISTANCE
Methyl cellulose can be made insoluble by chemical cross-linking of its
unetherified hydroxyls by bifunctional compounds. Agents that are used
include citric acid, glyoxal,171 dimethylolurea, water-soluble melamine-
formaldehyde resins, quaternary ammonium salts, and water-soluble urea-
formaldehyde resins.
(h) HIGH-METHOXYL METHYL CELLULOSE
Nearly completely substituted methyl cellulose (D.S. 2.4 to 2.8) is soluble
in polar organic solvents and in alcohol-aromatic hydrocarbon mixtures.
This ether may be prepared by the ethyl cellulose process141"144*148'149 or by
modification of the method of Maasberg.168 This ether may be used to
thicken organic solutions.
(i) ALKALI-SOLUBLE METHYL CELLULOSE
Low-substituted methyl cellulose (D.S. 0.1 to 0.9) is soluble in 2 to 10%
aqueous NaOH. Such ethers are prepared by the treatment of alkali cellu-
lose with methyl sulfate128'124 or with methyl chloride.19'121 Maasberg121
prepared alkali cellulose containing a weight ratio of sodium hydroxide to
cellulose of 0.35 to 0.60 by treating cellulose with 27.5 to 45% aqueous
NaOH at 15° to 35°C. This alkali cellulose was reacted with a weight ratio
of from 0.15 to 0.5 methyl chloride to cellulose for from 4 to 10 hrs. at 35°
to 75°C. until the reaction pressure fell to zero. The products were washed
with hot water containing sufficient acid to neutralize the residual sodium
hydroxide and were dried. The products were soluble in 2 to 10% aqueous
NaOH, but not in water. Such products are used as permanent sizing for
cloth or may be formed into sheeting similar to cellophane.
171 A. E. Broderick (to Carbide and Carbon Chemicals Corp.), U. S. Patent 2,329,741
(Sept. 21, 1943) ; Chem. Abstracts, 38, 1112 (1944).
IX. DERIVATIVES OF CELLULOSE 937
(j) CHEMICALLY MODIFIED METHYL CELLULOSES
Methyl cellulose is chemically modified to raise its gelation temperature,
to improve its salt compatibility, to broaden its solubility, and to provide
thermoplasticity,
Hydroxyethyl methyl cellulose was prepared in Germany by the succes-
sive reaction of alkali cellulose with ethylene oxide and with methyl chlo-
ride.172'178 This product was used as a starch substitute in adhesives.
The preparation of carboxymethyl methyl cellulose, with a carboxymeth-
oxyl D.S. of 0.2 to 0.3 and a methoxyl D.S. of 0.7 to 2.1 was described by
Swinehart, Savage, and Kuhlman.174'176 This ether approached the proper-
ties of carboxymethyl cellulose at a pH above 6 and of methyl cellulose at
a pH of 2 to 4. The sodium salt was soluble in water at room temperature
and remained in solution when heated to over 90°C. The salt compatibility
was increased so that a technical product containing by-product salt had
excellent solubility. The acid form gelled from solution at 55°C. ; thus, the
product could be washed with water in the acid form and then converted
to the sodium form.
A hydroxypropyl methyl cellulose,176 with 0.05 to 0.2 hydroxypro-
poxyl M.S. and 1.4 to 2.1 methoxyl D.S., that had a gel point of 70°C. and
corresponding improved salt compatibility, was disclosed in a patent issued
to Schick.177 That this ether is more internally plasticized than is hydroxy-
ethyl methyl cellulose is shown by a tendency toward organosolubility and
thermoplasticity rather than toward increased water solubility.
6. Carboxymethyl Cellulose
Commercially, the term carboxymethyl cellulose is applied to a water-
soluble cellulose ether which is actually the sodium salt of carboxymethyl
cellulose. It is also frequently called cellulose gum, CMC, or sodium cellu-
172 D. Traill and S. Brown, FIAT, Final Report No. 486 (Jan. 10, 1946); through
Library, U. S. Department of Agriculture, Washington, D. C.
173 M. Hagedorn and E. Rossback (to I. G. Farbenindustrie Akt.-Ges.), U. S. Patent
1,877,856 (Sept. 20, 1932); Chem. Abstracts, 27, 413 (1933).
174 R. W. Swinehart, A. B. Savage, and W. D. Kuhlman (to The Dow Chemical Co.),
U. S. Patent 2,476,331 (July 19, 1949); Chem. Abstracts, 43, 9445 (1949).
175 R. W. Swinehart (to The Dow Chemical Co.), U. S. Patent 2,510,153 (June 6,
1950); Chem. Abstracts, 44, 8631 (1950).
™ H. Dreyfus, Brit. Patent 277,721 (Sept. 30, 1927) ; Chem. Zentr., 1928, I, 445;
Brit. Patent 327,157 (Mar. 28, 1930) ; Chem. Abstracts, 24, 5157 (1930) .
177 J. L. Schick (to The Dow Chemical Co.), U. S. Patent 2,538,051 (June 16, 1951);
Chem. Abstracts, 45, 4489 (1951).
938 CELLULOSE
lose glycolate. The uses of sodium carboxymethyl cellulose stem from its
properties as a protective colloid,178 thickening agent, and film-former.
The free acid form, that is, carboxymethyl cellulose, on the other hand,
has found limited application because it is, in general, not water-soluble.
Sodium carboxymethyl cellulose thickens water solutions without pre-
cipitation over a pH range of 4 to 12 and does not gel from water solutions
upon heating. It is not precipitated from solution by dilute solutions of
salts of alkali or alkaline earth metals; it is precipitated from solution by
strong acids, by salts of amphoteric multivalent metals, and by salts of
heavy metals. The protective colloid properties of sodium carboxymethyl
cellulose aid in the building or promoting of detergents and in textile sizing.
The physiological inertness of sodium carboxymethyl cellulose permits its
use in food products.179
Sodium carboxymethyl cellulose is made only in relatively low ranges of
substitution. At D.S. 0.7 to 1.2, it is water-soluble, yields clear solutions
when purified, and is precipitated from solutions only by acidification with
strong acids to pH 1 to 3. Sodium carboxymethyl cellulose of D.S. 0.3 to
0.6 is water-soluble, but is commonly sold in a technical, unpurified form;
it is precipitated from solution by weak acids at pH 3. At D.S. 0.1 to 0.2 it
is soluble in 3 to 10% aqueous NaOH, depending upon the extent of chilling
of the solution, and in 5 to 8% NH4OH.180
(a) RAW MATERIALS
The quality of the raw materials used in carboxymethylation varies with
the process used, with the degree of purification of the product, and with
the intended end use. The cellulose may be finely milled wood pulp,181
purified wood pulp, or purified cotton linters.182 The sodium hydroxide is
of the same quality that is used in other cellulose etherifications. The
sodium chloroacetate or chloroacetic acid used should be free of di- and tri-
chlorinated compounds.
(b) CARBOXYMETHYLATION
Carboxymethylation differs from other cellulose etherifications in that
alkali cellulose need not be prepared separately, although normally it is.
178 C. B. Hollabaugh, L. H. Burt, and A. P. Walsh, Ind. Eng. Chem., 37, 943 (1945).
17» H. A. Shelanski and A. M. Clark, Food Research. 13, No. 1, 29 (1948).
180 Du Pont Sodium CMC, E. I. du Pont de Nemours & Co., Wilmington, Del., 1947.
181 W. F. Waldeck (to Wyandotte Chemicals Corp.), U. S. Patent 2,510,355 (June 6,
1950) ; Chem. A bstracts, 44, 7538 ( 1950) . R. N. Hader, W. F. Waldeck, and F. W. Smith,
Ind. Eng. Chem., 44, 2803 (1952).
*» E. D. Klug and J. Tinsley (tp Hercules Powder Co.), U. S. Patent 2,517,577 (Aug.
8, 1950); Chem. Abstracts, 44, 10318 (1950).
DC. DERIVATIVES OF CELLULOSE 939
The processes used differ widely; they are based upon the experience of the
manufacturers in the preparation of other materials.
The carboxymethylation reaction is:
R»u(OH)8 + ClCH2COONa + NaOH »
RceiiCOHJjOCHzCOONa + NaCl + H,O (12)
By-product sodium glycolate formation occurs according to:
ClCH2COONa + NaOH > HOCH2COONa + NaCl (13)
Technical sodium carboxymethyl cellulose contains the sodium chloride-
glycolate mixture; this is removed from the purified product. It is not
practical to reconvert glycolic acid to chloroacetic acid.
The traditional manufacture of sodium carboxymethyl cellulose1 Ifl2t172
is carried out in a Werner-Pfleiderer type of mixer with toothed sigma-
shaped blades and a cooling jacket. The cellulose may be steeped in sodium
hydroxide (see Section F of this Chapter IX), pressed, and shredded, or the
alkali cellulose may be prepared entirely in the shredder. Schmitz183 pre-
pared alkali cellulose in a slurry and continously removed the alkali cellu-
lose from the slurry with the aid of press and drainage rolls. The sodium
chloroacetate or chloroacetic acid may be shredded into the cellulose before
the sodium hydroxide is added.
Slurry carboxymethylation in the presence of tert-butyl alcohol or iso-
propanol was disclosed by Klug and Tinsley,182 who obtained D.S. 0.88
and a very low fiber content in the presence of these alcohols. Low substi-
tution was obtained in the presence of methanol (D.S. 0.16) or ethanol
(D.S. 0.35).
An unusual continuous carboxymethylation was disclosed by Waldeck.181
Finely milled wood cellulose (40 to 300 mesh) was tumbled in a rotary drum
and sprayed with 49% chloroacetic acid (1.27 weight ratio of acid to pulp).
After thorough tumbling, sodium carbonate (0.34 sodium carbonate : pulp)
was added to neutralize the chloroacetic acid. After thorough mixing, 50%
NaOH solution (0.72 solution : pulp) was slowly sprayed in and the mixture
was again thoroughly tumbled. When it had been dried, the technical
product, D.S. 0.72, contained 1% insoluble material. The efficiency was
67%.
The principle of the alkali cellulose process of Collings and coworkers
was applied to carboxymethylation by Collings, Freeman, and Anthoni-
sen.184 Purified cotton linters sheet was passed continuously through 75%
> »* R. Schmitz (to E. I. du Pont de Nemours & Co.), U. S. Patent 2,392,269 (Jan. 1 .
1946); Chem. Abstracts, 40, 2984 (1946).
184 W. R. Collings, R. D. Freeman, and R. P. Anthonisen (to The Dow Chemical Co.),
U. S. Patent 2,278,612 (Apr. 7, 1942) ; Chem. Abstracts, 36, 5013 (1942).
940 CELLULOSE
chloroacetic acid solution (5 seconds contact time) to obtain a pickup of
0.71 weight ratio of acid: cellulose and through 41% aqueous NaOH (15
seconds contact time) to obtain a pickup of 0.98 weight ratio of sodium
hydroxide : cellulose, and was shredded. A product of D.S. 0.5 to 0.6 re-
sulted. Swinehart and Allen185 used an additional sodium hydroxide treat-
ment to obtain a fiber-free product of D.S. 0.6 to 1.1. Maxwell186 prepared
alkali-soluble ethers by passing the sheet continuously first through sodium
hydroxide solution, and then through the etherifying agent, after which the
excess was pressed out. Rogers, Mueller, and Hembree187 prepared sodium
carboxymethyl cellulose by a different continuous sheet process.
TABLE 35
Efficiency of Carboxymethylation (McLaughlin and Herbst82)
(Mole ratio: sodium hydroxide to cellulose, 1.28:1; sodium chloroacetate to cellulose,
1.02:1)
Mole
ratio,
water: cellulose
Shredding
time,
hr.
Reaction
temperature,
°C.
Reaction
efficiency,
%
D.S.
Water
solubility
0.40
1
10
25
0.31
Particles
0.40
3
10
64
0.80
Particles
0.40
6
10
67
0.84
Particles
0.75
6
25
66
0.83
Particles
1.30
6
25
69
0.86
Clear
1.50
6
25
64
0.80
Clear
3.00
6
25
31
0.39
Fibers
In most of the batch processes it is customary to transfer the reaction
mass to bins, wagons, or tumbling drums for the extended reaction period
after mixing is complete in order to free the expensive mixing equipment
for another batch. The reaction is carried out at 0° to 70°C. The reaction
may be followed by titration of the salt formed and of the sodium hydrox-
ide present.
Efficiency of Carboxymethylation was studied by McLaughlin and
Herbst.82 They found that efficiency was improved by lower reaction tem-
peratures, by decreasing excess sodium hydroxide, and by increasing the
shredding time at low water contents. The optimum water to cellulose
mole ratio was 1.3 to 1. Their data are shown in Table 35.
* ™ R. W. Swinehart and S. R. Allen (to The Dow Chemical Co.), U. S. Patent 2,524,024
(June 26, 1950) ; Chem. Abstracts, 45, 1344 (1951).
t 1M R. W. Maxwell (to E. I. du Pont de Nemours & Co.), U. S. Patent 2,101,263 (Dec.
7, 1937); Chem. Abstracts, 32, 1094 (1938).
187 L. N. Rogers, W. A. Mueller, and E. E. Hembree (to Buckeye Cotton Oil Co.),
U. S. Patent 2,553,725 (May 22, 1951) ; Chem. Abstracts, 45, 8247 (1951).
IX. DERIVATIVES OF CELLULOSE 941
(c) AFTERTREATMENT
The af tertreatment of the carboxymethylation reaction mass varies with
the process. If a technical unpurified product is desired, the wet reaction
mass may be neutralized with sodium bicarbonate, shredded, and sold
wet> 172, iso.iss.isa or the technical product may be dried and sold.180-181 The
technical products are of D.S. 0.3 to 0.7 and of low intrinsic viscosity. Typi-
cal compositions of the technical products are shown in Table 36.190'191
TABLE 36
Composition of Typical Technical Sodium Carboxymethyl Celluloses, D.S. 0.7
(The Wyandotte Chemicals Corp.190 and Hercules Powder Company191)
Composition, weight ratio to
sodium carboxymethyl cellulose
Ingredient • Alto Blfl
Sodium carboxymethyl cellulose
1.0
1.0
Water
0.08
0.05
Sodium chloride
0.26
0.04
Sodium carbonate
0.03
—
Sodium carboxymethyl cellulose (low D.P.) plus
sodium glycolate
0.15
0.02
(d) PURIFICATION
Products of D.S. 0.5 to 1.2 may be fiber free and can be purified to yield
clear solutions. Purification methods involve either the use of alcohol-
water mixtures to extract the salt and the sodium glycolate without solution
of the sodium carboxymethyl cellulose, or conversion to the insoluble acid
form or to an insoluble salt.
Collings, Freeman, and Anthonisen184 neutralized the reaction mass, D.S.
0.6, with hydrochloric acid, dissolved the acid form, precipitated it with
alcohol, and washed it with aqueous alcohol. Klug and Tinsley182 washed
the product of their alcohol slurry process, D.S. 0.88, with 70% aqueous
methanol. Swinehart and Allen185 and Houghtonm used similar methods.
188 Wyandotte Carbose, Wyandotte Chemicals Corp., Wyandotte, Mich., 1952.
188 Hercules CMC Cellulose Gum, Hercules Powder Co., Wilmington, Del., 1949 and
1951.
190 Wyandotte Chemicals Corp., Wyandotte, Mich., unpublished data.
191 Hercules Powder Co., Wilmington, Del.
1M A. A. Houghton (to Imperial Chemical Industries, Ltd.), U, S, Patent 2,513,725
(July 4, 1960) ; Chem. Abstracts, 44, 8656 (1950),
942 CELLULOSE
Freeman and Anthonisen*98 and Houghton194 converted sodium carboxy-
methyl cellulose to the acid form by treatment with strong acid below pH 1 .
A unique method of reducing the amount of acid required to convert to
the acid form was disclosed by Lamborn. 196 Crude product containing 20%
sodium carboxymethyl cellulose (D.S. 0.7), 33% salt, and 47% water was
neutralized to pH 6.5 with sodium bisulfate in a Werner-Pfleiderer mixer.
The crude neutral dough was extruded through orifices in a press to form
strands of diameter 0.038 in. The strands were dried to below 5% moisture
in warm air, and were immersed for 2 hrs. or more in a bath kept at pH 1
with hydrochloric acid. The strands were then washed with water until
free of contaminating salts, dried, and granulated.
Freeman and Roberts196 recovered carboxymethyl cellulose as its alkali-
soluble aluminum salt.
(e) CARBOXYMETHYL CELLULOSE BASE FLAKE
Sodium carboxymethyl cellulose is a light cream to white powder. It
browns upon heating at 180° to 225 °C. and chars upon heating to 210° to
250°C. It is soluble in cold and in warm water. It is highly moisture-
absorbent and may gain its own weight of water at high humidity.197 Bulk
densities range from 0.4 to 0.8 g./ml.
(f) CARBOXYMETHYL CELLULOSE SOLUTIONS
Sodium carboxymethyl cellulose dissolves in efficiently agitated water,
and in aqueous ethanol or aqueous acetone containing over 60% water.
Solution clarity is improved by passing the solutions through a colloid mill
or a homogenizer.
Solutions of sodium carboxymethyl cellulose are thixotropic, and their
viscosities depend upon the rate of shear and other variables, so the vis-
cosity is usually measured under specified conditions with a rotational vis-
cometer such as the Brookfield. Figure 68 shows the relation of solution
viscosity to concentration for sodium carboxymethyl cellulose.
Intrinsic viscosity values can be obtained for solutions of sodium car-
198 R. D. Freeman and R. P. Anthonisen (to The Dow Chemical Co.), U. S. Patent
2,351,258 (June 13, 1944) ; Chem. Abstracts, 38, 5405 (1944).
194 A. A. Houghton and K. J. C. Luckhurst (to Imperial Chemical Industries, Ltd.),
U. S. Patent 2,357,469 (Sept. 5, 1944) ; Chem. Abstracts, 39, 416 (J945).
"« B. T. Lamborn (to Hercules Powder C6.), U. S. Patent 2,513,807 (July 4, 1950) ;
Chem. Abstracts, 44, 8656 (1950) .
198 R. D. Freeman and M. J. Roberts (to The Dow Chemical Co.), U. S. Patents
2,331,858 and 2,331,859 (Oct. 12, 1943) ; Chem. Abstracts, 38, 1641 (1944).
»w C. J. Brown and A. A. Houghton, /. Soc. Chem. Ind. (London), 60, 254T (1941).
IX. DERIVATIVES OF CBLLtfLOSE
boxymethyl cellulose in 5% NaOH solution, in which case Martin's con-
stant varies from 0.14 to 0.16. For example, low-, medium-, and high-
20000
10000
O 2 4 • 8 10
CONCENTRATION, WEIGHT PERCENT
Fig. 68. Sodium carboxymethyl cellulose: viscosity-concentration
chart (Hercules Powder Co.189). Viscosities in 2% solution in water
at 25 °C. are shown in parentheses on the curves.
viscosity types of sodium carboxymethyl cellulose of D.S. 0.8 give intrinsic
944 CELLULOSE
When sodium carboxymethyl cellulose solutions are heated, the viscosity
decrease with increasing temperature is reversible if the maximum tempera-
ture is not over 50°C.; above this temperature a permanent viscosity loss
may occur. iw.m.iw Such viscosity loss is more pronounced in the presence
of alkali, for solutions of alkali-soluble material lose viscosity at room tem-
perature. The relation of solution viscosity to pH is shown in Figure 69.
Carboxymethyl cellulose is a weak acid Colonization = 3 X 10 ~5). The
pH of precipitation of the acid form varies with D.S.; thus D.S. 0.1 to 0.2
precipitates below pH 6; D.S. 0.3 to 0.5 below pH 3; and D.S. 0.7 to 0.9
below pH L The pH of precipitation also varies to some extent with the acid
used. The acid form of carboxymethyl cellulose becomes insoluble if over-
dried, and redissolves only in dilute alkalies.12'199 Chowdhury12 attributed
this behavior to inner lactone formation. A water dispersion of the free
acid can be made by treatment with ion-exchange resins.189
The salt compatibility of sodium carboxymethyl cellulose solutions varies
with the metallic ion present.189 When solutions containing 1% sodium
carboxymethyl cellulose and 5% of various salts were prepared, the results
observed were:
Precipitate
No precipitate Soluble in NaOH Insoluble in NaOH
Ba(N03)2(thixotropic) A12(SO4)3 Cr(NO3)3
CaCl, BaCl2 AgNO3
MgCU SnCl2 FeCl3
MnSO4 FeSO4
Pb(CH8CO2)2
(g) CARBOXYMETHYL CELLULOSE SHEETING
Sodium carboxymethyl cellulose sheeting can be cast from water solu-
tions and is unaffected by most solvents and oils. The physical properties
of such sheeting are markedly dependent on the relative humidity. At
50% relative humidity, the unplasticized film has a tensile strength of
10,000 to 12,000 Ib./sq. in., has 3% elongation, and endures 200 M.I.T.
double folds. At higher humidity, the film becomes weaker and more flex-
ible. The addition of about 20% plasticizer (such as glycerin or ethylene
glycol) will also modify the film properties. Such sheeting has a tensile
strength of 7000 Ib./sq. in., has 15% elongation, and endures 1000 M.I.T.
double folds.
198 Carboxymethocel, The Dow Chemical Co., Midland, Mich., 1945.
1M I. Sakurada, Z. angew. Chem., 42, 640 (1929).
IX. DERIVATIVES OP CELLULOSE
945
(h) INCREASING WATER RESISTANCE
The water resistance of sodium carboxymethyl cellulose sheeting and
coatings can be increased by setting the sheeting to relative insolubility by
salt (for example, alum) treatment or acid treatment, or by cross-linking
with urea-formaldehyde or melamine-formaldehyde resins.
(i) COMPATIBILITY
Sodium carboxymethyl cellulose is compatible in solution with many
compounds, but shows limited compatibility in sheeting. It is compatible
with formamide, hydroxyethyl cellulose, natural gums, pectin, polyvinyl
alcohol, sodium alginate, starch, and urea-formaldehyde resins.
100
V)
<L
<j
8
75
50
> 25
HIGH VISCOSITY
IO II 12 13
pH
Fig. 69. Carboxymethyl cellulose: variation of solution viscosity with pH
(Brown and Houghton197). Viscosities in 1% solution in water at 25°C.
7. Hydroxyethyl Cellulose
Although no sodium hydroxide is consumed directly during the prepara-
tion of hydroxyethyl cellulose from ethylene oxide, sodium hydroxide is
necessary in order to swell the cellulose and to catalyze the reaction.
Ethylene chlorohydrin is also a suitable reagent for hydroxyethylation.
The uses of hydroxyethyl cellulose result from its solubility in cold and
hot water, its salt and solvent compatibility, and its effectiveness as a pro-
tective colloid. Hydroxyethyl cellulose thickens solutions over a wide
range of pH without precipitation.
In hydroxyethyl cellulose preparation, substitution takes place both on
the cellulose hydroxyls (the conventional D.S.) and by chain polymeriza-
946 CELLULOSE
tion on the hydroxyls of previously substituted groups. The average num-
ber of moles of alkene oxide consumed that become attached to a glucopy-
ranose residue in these two ways is termed M.S.107
Hydroxyethyl cellulose is made only in relatively low ranges of substitu-
tion. Hydroxyethyl cellulose of M.S. 0.50 or higher is water-soluble; hy-
droxyethyl cellulose of M.S. 0.05 to 0.4 is soluble in aqueous alkali solutions.
The lower substituted ethers are soluble only upon freezing.
(a) HYDROXYETHYLATION
Much of the previous discussion of carboxymethyl cellulose applies also
to hydroxyethyl cellulose. The quality of the materials used varies with the
degree of purification of the product and with the intended end use. Hy-
droxyethyl cellulose may be prepared by the reaction of alkali cellulose
with ethylene oxide or with ethylene chlorohydrin (2-chloroethanol) ; the
chlorohydrin probably is converted to the oxide, and the oxide then reacts
with the alkali cellulose.
Hydroxyethylation was studied by Morgan33 (see Table 18), who showed
the relation of ethylene oxide consumption to reaction efficiency, solubility,
and M.S., by Tasker and Purves,106 and by Cohen and Haas,107 who pre-
sented partial analyses of hydroxyethyl celluloses (see Tables 19 and 20).
The probable structure of hydroxyethyl cellulose is indicated in Figure 59.
The formation of hydroxyethyl cellulose is shown in equations 2, 3, and 4.
These reactions take place in the presence of sodium hydroxide and water.
Davis200 hydroxyethylated cotton linters in the presence of benzene or
ethyl ether as a carrier. He found that for alkali-soluble products the opti-
mum range of the ratio of sodium hydroxide to cellulose was 0.5 : 1 to 1.0 : 1,
and the optimum range of the ratio of water to cellulose was 1.0 : 1 to 2.0 : 1 .
He found also that for the production of uniformly substituted ethers it
was best to limit the ratio of ethylene oxide to cellulose to approximately
0.25:1. Uniform additional substitution was then obtained by further
treatments with ethylene oxide.200-201
Schorger131'132 and Shoemaker132 prepared low-substituted alkali-soluble
hydroxyethyl cellulose and studied the effect of sodium hydroxide concen-
tration and of freezing upon the solubility of the products. They steeped
cellulose in 30% NaOH and pressed it to a sodium hydroxide : cellulose
ratio of 0.9:1 and a water : cellulose ratio of 1.5:1. They found that the
100 W. E. Davis, Dissertation, New York State College of Forestry, Syracuse, N. Y.,
1941.
Wl F. H. Reichel and R. T. K. Cornwell (to Sylvania Industrial Corp.), U. S. Patent
2,388,764 (Nov. 31, 1945); Chem. Abstracts, 40, 1656 (1946).
DC. DERIVATIVES OF CELLULOSE 947
solubility of such alkali-soluble ethers is greater in 8% NaOH than in
higher or lower concentrations. They obtained optimum filterability when
they froze solutions containing 5 to 6% NaOH.
Reichel and Hindry202 prepared alkali cellulose by the viscose method
(see Section F of this Chapter IX) by treating cellulose with 18% NaOH,
pressing, and shredding. The crumbs were then etherified to M.S. 0.3,
and the excess etherifying agent was used up by the addition of regenerated
cellulose. The product was soluble in 10% NaOH at 0°C.
Hydroxyethyl cellulose may also be made by a slurry method208 and by
a vacuum process.204
(b) AFTERTREATMENT
The aftertreatment of the reaction mass varies with the process and with
the end use. The residual sodium hydroxide must be neutralized in order
to prevent viscosity degradation. Alkali-soluble ethers may be washed
with acid and then with water after neutralization is complete, and dried.
Water-soluble ethers are difficult to purify. Both the alkali-soluble and
the water-soluble ethers may be dried after neutralization and sold as
technical products, or they may be dissolved and sold in solution form.
Solutions may be purified by dialysis.
Kunz205 proposed the addition of material that would form a solid hydrate
in order to take up water and neutralize simultaneously. Crude reaction
mass (sodium hydroxide : cellulose 0.33:1, water : cellulose 1.47:1) was
mixed with phosphoric acid (phosphoric acid : cellulose 0.66 : 1 ; water : cellu-
lose 0.12: 1). The temperature rose to 50° C. during mixing. The product
was dried below 35°C., then raised to 100°C. The product was a mixture
of hydroxyethyl cellulose and hydrated sodium phosphate of pH 8.4.
Aluminum sulfate, esters, alcohols, or ketones may be used to reduce
solubility during the washing of hydroxyethyl cellulose.
(c) HYDROXYETHYL CELLULOSE BASE FLAKE
Water-soluble hydroxyethyl cellulose, M.S. over 0.5, is a white to light
tan powder, soluble in water over a wide temperature range. It slowly de-
202 F. H. Reichel and W. F. Hindry (to Sylvania Industrial Corp.), U. S. Patent 2,172,-
109 (Sept. 5, 1939); Chem. Abstracts, 34, 262 (1940).
803 E. D. Klug and H. G. Tennent (to Hercules Powder Co.), U. S. Patent 2,572,039
(Oct. 23, 1951); Chem. Abstracts, 46, 1256 (1952).
*>< D. R. Erickson, U. S. Patent 2,469,764 (May 10, 1949); Chem. Abstracts, 43, 5592
(1949).
806 W. B. Kunz (to American Viscose Corp.), U. S. Patent 2,488,631 (Nov. 22, 1949) ;
Chem. Abstracts, 44, 1702 (1950).
948 CELLULOSE
composes at temperatures over 100°C., especially in the presence of acids,
alkalies, or salts. It browns at about 180°C. and chars upon continued
heating at over 200°C. Alkali-soluble hydroxyethyl cellulose, like other
alkali-soluble ethers, is a white fibrous material similar to the original cellu-
lose.
(d) HYDROXYETHYL CELLULOSE SOLUTIONS
Water-soluble hydroxyethyl cellulose is soluble in efficiently agitated
water. It is slightly soluble upon heating (up to 1%) in ethylene glycol,
propylene glycol, and glycerol, and (5%) in N-acetylethanolamine.206
Solutions are opalescent in appearance.
Water-soluble hydroxyethyl cellulose is compatible in solution with
starch and starch derivatives, gelatin, natural gums, and sodium carboxy-
methyl cellulose, but is only partly compatible with methyl cellulose and
polyvinyl alcohol. Aqueous solutions tolerate up to equal parts by weight
of the water-soluble alcohols and much larger quantities of polyhydroxy
compounds, carboxylic acids, aldehydes, and amines without gelation. Low
concentrations of water-soluble esters and ketones have a precipitating
effect. Of the common salts, only the sulfates affect solution stability and
only aluminum sulfate has a precipitating effect.
The viscosity of solutions of water-soluble hydroxyethyl cellulose is
lowered by dilute strong acids and by concentrated formic acid. Alkali
causes a slower viscosity decrease.
Alkali-soluble hydroxyethyl cellulose is soluble, depending upon its de-
gree of substitution, in 2 to 10% NaOH, in 10% KOH, and in alkaline
40% urea.207
(e) HYDROXYETHYL CELLULOSE SHEETING
Water-soluble hydroxyethyl cellulose solutions yield clear sheeting that
is heat stable below 100°C. and is soluble in water, but insoluble in most
organic solvents. The flexibility of sheeting is increased by plasticization
with 10 to 50% (based upon the ether) of sorbitol, glycols, liquid polygly-
cols, higher diols, N-acetylethanolamine, or sulfonated castor oil.
Alkali-soluble hydroxyethyl cellulose, like other alkali-soluble ethers,
yields sheeting from aqueous NaOH solutions that can be set by flooding
with water or by neutralization with acid or with salt solutions. Such
sheeting is reported207 to have a tensile strength of 13,000 Ib./sq. in., and an
elongation of 6%, and to withstand 1300 M.I.T. double folds.
206 "Cellosize" Hydroxyethylcellulose, Carbide and Carbon Chemicals Division, Union
Carbide and Carbon Corp., New York, F5339-D.
w W. E. Gloor, B. H. Mahlman, and R, D. Ullrich, Ind. Eng. Chem., 42, 2150 (1950).
IX. DERIVATIVES OF CELLULOSE 949
Alkali-soluble hydroxyethyl cellulose is compatible with gelatin, carboxy-
methyl cellulose, methyl cellulose, and casein.
(f) IMPROVING WATER RESISTANCE
Water resistance (partial in the case of treated, water-soluble hydroxy-
ethyl cellulose) is improved by treatment before drying with glyoxal,171
urea-formaldehyde, or melamineHformaldehyde. Both types are rendered
photosensitive by chromates and by azo dyes and may then be rendered
temporarily insoluble by exposure to ultraviolet light.206'207 The acetal of
glyoxal and hydroxyethyl cellulose hydrolyzes upon standing in water.206
(g) MIXED HYDROXYETHYL ETHERS
The addition of a small amount of hydroxyethyl substitution (0.3 to 0.4
M.S.) to conventional derivatives confers certain unusual properties; thus
the salt precipitation of methyl cellulose and carboxymethyl cellulose is re-
duced, the aliphatic-solvent tolerance of high-substituted ethyl cellulose is
increased, and the acetone solubility of high-substituted cellulose acetate
is increased.207**208
8. Benzyl Cellulose
Benzyl cellulose, which results from the reaction of alkali cellulose with
benzyl chloride, has reached commercial production in Europe, but has not
passed beyond the pilot plant stage in the United States.
Benzyl cellulose is internally plasticized to an extent that it may be used
for coatings, plastics, and lacquers without the addition of plasticizer. It
is soft, but water resistant.
(a) BENZYLATION
Lorand and Georgi26'209 prepared benzyl cellulose with minimum agita-
tion and carried out continuous water removal in order to improve effi-
ciency. They followed the course of the benzylation with successive photo-
graphs (Fig. 70). Benzylation began at active spots upon the fiber surface
(Fig. 70A) and moved inward from one growth layer to the next. Figure
70B shows drops of weak salt-sodium hydroxide solution leaving the water-
, E. D. (to Hercules Powder Co.), Brit. Patent 670,672 (April 23, 1952);
Chem. Abstracts, 46, 8372 (1952); U. S. Patent 2,618,632 (Nov. 18, 1952).
208 M. Hagedorn and P. Moller (to I. G. Farbenindustrie Akt.-Ges.), U. S. Patent
1,994,038 (May 12, 1935); Chem. Abstracts, 29, 3156 (1935); A. E. Broderick (to Car-
bide and Carbon Chemicals Corp.), U. S. Patent 2,330,263 (Sept. 28, 1943); Chem. Ab-
stracts, 38,1365(1944).
»• E. J. Lorand (to Hercules Powder Co.), U. S. Patent 2,056,324 (Oct. 6, 1936);
Chem. Abstracts. 30, 8615 (1936).
950
CELLULOSE
Fig. 70. Benzyl cellulose : stages in benzylation (Lorand and Georgi28). A and B.
fibers at early stages of benzylation (magnification X90); C, benzylated fibers breaking
up as gel sheath dissolves (magnification X90); D and E, sections through benzylated
fiber at an advanced stage of benzylation (magnification X350); F, section through
benzylated fiber showing gelation of outer layers (magnification X350).
IX. DERIVATIVES OF CELLULOSE 951
repellent benzylated areas. Figures TOD and 70E show in section the migra-
tion of weak salt-sodium hydroxide solution from the benzylated outer
layers to the fiber core. Figure 70F shows in section a fiber half benzylated
and swollen, but not dissolved in benzyl chloride. Figure 70C shows
benzylated fibers breaking up as their outer layers dissolve.
The benzylation reaction and the by-product reactions are analogous to
equations 6, 7, and 8.
Gomberg and Buchler,14 who benzylated at low temperatures in the pres-
ence of low sodium hydroxide concentrations, prepared benzyl cellulose that
was only partially soluble. Okada210 prepared benzyl cellulose under reflux ;
he obtained a soluble product from a low sodium hydroxide concentration.
He found that ethers prepared rapidly at high temperature had better
mechanical properties than did products of a slow, low-temperature reac-
tion.
Brandt211 mercerized cellulose in 16 to 18% aqueous sodium hydroxide
and removed the excess sodium hydroxide by strong pressing. His benzyl-
ated products were not completely soluble and his reaction rates decreased
rapidly. He found it necessary to resort to sodium hydroxide stronger than
40% or to continuous water removal, to obtain D.S. 2.0 or higher. His
viscosities were low.
Mienes212 found that the solubility of benzyl cellulose increased in pro-
portion to the squeezing pressure at a given sodium hydroxide to cellulose
ratio. At substitution below D.S. 2.0, the D.S. was said to be a function
of the sodium hydroxide : cellulose ratio; at high substitutions this ratio
was said to be of decreasing influence. He recommended sodium hydroxide
addition213'214 to keep the sodium hydroxide concentration constant. D.S.
2.0 resulted from alkali cellulose prepared from 22% NaOH.
In German benzyl cellulose manufacture,172'215 alkali cellulose (weight
ratio of sodium hydroxide : cellulose, 2.65:1; of water : cellulose, 4.20:1),
prepared in a Werner-Pfleiderer mixer, was reacted with benzyl chloride
210 H. Okada, Cellulosechemie, 12, 11 (1931).
411 K. Brandt, Dissertation, Berlin, 1932; through K. Mienes, CelMoseester und Cellu-
losetither, Chemisch-technischer Verlag Dr. Bodenbender, Berlin-Steglitz, 1934.
212 K. Mienes, Celluloseester und Celluloseather, Chemisch-technischer Verlag Dr.
Bodenbender, Berlin-Steglitz, 1934.
218 G. von Frank and K. Mienes, German Patent 555,930 (Aug. 1, 1932); Chem. Ab-
stracts, 26, 6134(1932).
214 G. von Frank and K, Mienes, German Patent 575,349 (Apr. 27, 1933) ; Chem. Ab-
stracts, 27, 5974(1933). Oft
216 I.G. Farbenindustrie Akt.-Ges., German Patent 492,062 (Feb. 15, 1930); Chem,
Abstracts, 24, 2599 (1930).
952 CELLULOSE
(ratio to cellulose 4.85: 1) in an autoclave at 70° to 110°C. for 8 hrs. with
stirring. The excess sodium hydroxide solution was then drained off for
re-use, and the product was washed with water to remove salt and with
methanol to remove organic by-products, washed again with water, and
dried. Benzyl cellulose, D.S. 2.0 (benzyl cellulose: cellulose 1.9: 1) and by-
products (benzyl alcohol plus benzyl ether .-cellulose 2:1) resulted.
Lorand216 decreased by-product formation by the addition of benzyl
ether to the charge. Mienes212 proposed the use of chlorobenzene as a
solvent for low-substituted benzyl cellulose in the early stages of the reac-
tion; Savage19 used benzene as a diluent; Huber217 used tertiary amines as
dispersants. The patent literature was reviewed by Worden7 and by
Mienes.212
(b) PURIFICATION
Purification by extraction with water and with alcohols has already
been described.172'215 Dorr218 disclosed the addition of such soaps as sodium
oleate to the crude product and precipitation of the mixture as a flocculent,
easily washable powder. Lorand219 dispersed the crude product in -such
solvents as ethers, high water-soluble alcohols, and hydrocarbon-alcohol
mixtures that had a high separation ratio and a low breakup ratio upon
precipitation. Savage19 washed the crude reaction solution by countercur-
rent liquid-liquid extraction with an aqueous alcohol.
(c) BENZYL CELLULOSE BASE FLAKE220
Benzyl cellulose is a white, granular material of specific gravity 1.2.
The melting point of benzyl cellulose varies from 90° to 155°C. depending
upon degree of substitution and viscosity (see Fig. 60). Benzyl cellulose is
soluble in higher aliphatic and cyclic ketones, esters, lower aromatic hydro-
carbons, chlorinated hydrocarbons, and aromatic hydrocarbon-alcohol
mixtures.
216 E. J. Lorand (to Hercules Powder Co.), U. S. Patent 2,001,102 (May 14, 1935);
Chem. Abstracts, 29, 4580 (1935).
217 Eugen Huber (to I. G. Farbenindustrie Akt.-Ges.), U. S. Patent 1,805,365 (May
12, 1931) ; Chem. Abstracts, 25, 3832 (1931).
**» E. Ddrr (to Hercules Powder Co.), U. S. Patent 2,020,934 (Nov. 12, 1935); Chem.
Abstracts, 30, 613 (1936).
111 E. J. Lorand (to Hercules Powder Co.), U. S. Patent 2,056,612 (Oct. 6, 1936);
Chem. Abstracts, 30, 8615 (1936).
110 Benzyl Cellulose, Hercules Powder Co., Wilmington, Del, 1944.
IX. DERIVATIVES OF CELLULOSE
953
(d) BENZYL CELLULOSE SHEETING220
The properties of benzyl cellulose sheeting are shown in Table 37. The
electrical properties of benzyl cellulose are similar to those of ethyl cellu-
lose. The tensile strength of benzyl cellulose is lower than that of other
cellulose derivatives. The light and heat stability are low. The low-tem-
perature flexibility of benzyl cellulose is comparable to that of ethyl cellu-
lose.
TABLE 37
Properties of Benzyl Cellulose Sheeting
(Hercules Powder Company220)
Property
Value, unplasticized
benzyl cellulose,
0.003-in. film
Tensile strength, Ib./sq. in.
Flexibility, 25 °C., M.I.T. double folds
Elongation, %
Moisture permeability, g./sq. m./24 hrs. at 35 °C.
Fade-Ometer
hours to embrittlement
hours to discolor
6700
350
19
82
24
24
(e) BENZYL CELLULOSE PLASTICS220
The properties of unplasticized benzyl cellulose plastics are shown in
Table 38. The sof teness and low moisture absorption of benzyl cellulose are
evident from the table.
TABLE 38
Properties of Benzyl Cellulose Plastics (Hercules Powder Company220)
Benzyl cellulose, D.S. 2.3; viscosity, 70 cps.; ash, 0.11%; stabilized with
1 % diphenyl amine
Test
Method
Extrusion
Injection
Flow temperature, °F.
D569-41T
—
220°
Rockwell M hardness
D229-39
49
46
Charpy impact, ft. -Ib. /inch notch
D256-41T
0.5
1.6
Water absorption, %
D570-40T
0.54
0.44
Heat distortion temperature, °C.
—
47
—
Izod impact, ft.-lb./inch notch
—
1.6
—
Cylinder temperature, °F.
—
~-
385
Nozzle temperature, °F.
—
—
3S5
Mold temperature, °F.
—
—
100
Mold pressure, Ib./sq. in. gage
—
—
9000
• 100 Ib./sq. in.
CELLULOSE
(f) COMPATIBILITY880
Benzyl cellulose is compatible with many resins, with most common
plasticizers, and with polystyrene.19 It is incompatible with other cellulose
esters and ethers and with other vinyl polymers.
9. Addition to Cellulose of Olefins Activated by Polar Substituent Groups
(a) THEORY
The rate of addition of a reagent to an olefinic double bond is influenced221
by the position of the double bond with respect to the rest of the molecule,
by the nature of the other groups present, and by the catalyst and the
solvent that are used. Anionoid reagents do not usually react with the
double bond in olefins, because the double bond cannot act as an electron
acceptor; but, if an activating substituent group is conjugated with the
double bond, anionoid addition may occur. Thus Michael222 showed that
a,/3-unsaturated esters and ketones combined with malonic ester and the
like under the influence of sodium ethoxide. Bruson and Riener223 showed
that, if alkaline catalysts are present, acrylonitrile will add to polyhydric
alcohols according to the mechanism :
* (14)
where X is an anionoid reagent. Hydroxyl groups attract protons, causing
negative ion formation, so that alkaline catalysis is necessary.
Although the earliest literature references to the addition of activated
olefins to cellulose concerned the preparation of sulfoethyl cellulose from
cellulose and vinylsulfonic acid,39 most of the literature describes the use of
acrylonitrile with cellulose to obtain cyanoethyl cellulose, or its hydrolytic
product, carboxyethyl cellulose.
ib) CYANOETHYLATION
Acrylonitrile cyanoethylates the hydroxyls of cellulose in the presence
of alkali. If uniform products are desired, the cellulose should pass into
solution early in the reaction. For this reason viscose is usually used as the
starting material. The technical uses of cyanoethyl cellulose include the
221 w j Hickinbottom, Reactions of Organic Compounds, 2d ed.f Longmans Green,
London, 1948, p. 35.
«» A. Michael, /. prakt. Chem., 35, 349 (1887) ; 37, 496 (1889).
»» H. A. Bruson and T. W. Riener, /. Am. Chem. Soc.t 64, 2850 (1942); 65, 18, 23
(1943); 66,56(1944).
IX. DERIVATIVES OF CELLULOSE 955
preparation of water-soluble yarns, twistless yarns, and the modification of
"brushed rayon."224'226
If strong alkali or long reaction times are used, the cyanoethyl cellulose
will be hydrolyzed to carboxyethyl cellulose. Hydrolysis of cyanoethyl
cellulose with boiling 5 to 10% aqueous NaOH gives carboxyethyl cellulose,
but splitting occurs at the ether linkage and ammonia is liberated. The
cellulosic hydrolysis product contains no nitrogen, but the carboxyl groups
present are equivalent to only about one-fourth of the nitrogen liberated
as ammonia.
The reactions include :
» CHa=CHCN > R^nOCH2CH2CX (15)
Cellulose Acrylonitrile Cyanoethyl cellulose
ReeiiOCH2CH2CN — ^U Rccl,OCH2CH2CONH, (16)
O-Carbamylethyl cellulose
R,eiiOCH2CH2CONH2 H*° > RceiiOCH2CH2COOH + NH, (17)
Carboxyethyl cellulose
RceiiOCHaCH^CN > RceiiOH + CH*=CHCN (18)
In practice it is found that the nitrogen content rises to a maximum and
then falls. Some water-soluble derivatives226 contained less than 10%
carboxyethyl groups, whereas lower or higher substituted ethers contained
none.
The cyanoethylation reaction is very sensitive to temperature
changes.224"227 Hydrolysis is appreciable at temperatures above 15°C. At
a given temperature, the degree of cyanoethylation increases as the mole
ratio of acrylonitrile to cellulose increases. The maximum degree of cyano
ethylation is reached in about 4 hours, and it decreases thereafter.
Increasing the sodium hydroxide concentration increases the rates both
of cyanoethylation and of hydrolysis, so that the net result is a lower degree
of cyanoethylation. The rate of cyanoethylation can be increased by in-
creasing the caustic ratio, if the temperature is kept down to 10°C.
The hydrolytic side reactions of acrylonitrile contribute to inefficiency.
CH2=CHCN + H2O > HOCH2CH2CN (19)
Hydracrylonitrile
HOCH2CH2CN + CHf=CHCN » NCCH2CH2OCH2CH2CN (20)
Bis(/3-cyanoethyl) ether
224 J. H. MacGregor and C. Pugh, J. Soc. Dyers Colourists, 67, 66, 74 (1951).
228 J. A. Somers, British Rayon & Silk /., 26, No. 312, 67 (1950).
«* F. Happey and J. H. MacGregor, Nature, 160, 907 (1947).
887 J. H. MacGregor and C. Pugh, The Reaction of Acrylonitrile with Poly saccharifies,
llth International Congress of Pure and Applied Chemistry, London, 1947, in press.
956 CELLULOSE
Further hydrolysis proceeds in the same manner that was given above for
the cellulose ether.
CH2=CHCN + H2O > CH2=CHCONH2 (21)
Acrylamide
CH2=CHCONH2 + NaOH > CH^CHCOONa + NH3 (22)
Sodium acrylate
(c) ALKALI-SOLUBLE CYANOETHYL CELLULOSE
When the mole ratio of acrylonitrile to cellulose is 0.5:1 to 1.2:1, and
the sodium hydroxide concentration is below 15%, the cyanoethyl celluloses
that are obtained (D.S. 0.2 to 0.3, nitrogen below 2%) are insoluble in
water, but are soluble in dilute aqueous solutions of sodium hydroxide or
of quaternary ammonium bases, and in 25 to 50% solutions of ammonium
thiocyanate or zinc chloride. These ethers are spun into yarns, which take
up 200 to 300% water, but do not dissolve in water.
(d) WATER-SOLUBLE CYANOETHYL CELLULOSE42-228
When the mole ratio of acrylonitrile to cellulose is from 1.5:1 to 4:1,
and the sodium hydroxide concentration is below 10%, the cyanoethyl
celluloses that are produced (D.S. 0.7 to 1.0, nitrogen up to 7%) are soluble
in acid, neutral, or alkaline aqueous solutions. These solutions may be
purified by dialysis, and the ethers may then be precipitated by acetone or
by ethanol. The neutral solutions do not have a gel temperature when
heated. These ethers form transparent sheeting, and are emulsifying agents
and nonionic surface-active agents.
When the mole ratio of acrylonitrile to cellulose exceeds 4:1, a second
series of water-insoluble, alkali-soluble products results. The products are
less soluble in caustic solution as the D.S. is increased, but at D.S. 2.0
they become soluble in 50:50 acetone : water .
TABLE 39
Cyanoethyl Cellulose: Relation of Mole Ratio of Acrylonitrile to Cellulose and Degree
of Substitution (Somers225)
Mole ratio : acrylonitrile : cellulose
Degree of substitution
1.4
1.0
1.5
1.2
1.6
1.4
12.0
1.7 .
Table 39 shows that the substitution is increased rather little by an in-
crease in the ratio of acrylonitrile to cellulose. Table 40 shows the increase
228 J. H. MacGregor and Courtaulds, Ltd., Brit. Patent 588,751 (July 2, 1947) ; Chem,
Abstracts, 41, 6718 (1947); Brit. Patent 636,020 (April 19, 1950); Chem. Abstracts, 44,
6624 (1950).
DC. DERIVATIVES OF CELLULOSE 957
TABLE 40
Cyanoethyl Cellulose: Relation of Sodium Hydroxide Concentration to Degree of
Substitution (Somers)226
Viscose, ripeness, 6;
mole ratio, acrylonitrile: cellulose,
1:1
NaOH, %
Nitrogen, %
Cyanoethyl, D.S.
Carboxyethyl, D.S.
6
1.6
0.16
10 15
1.45 0.76
0.14 0.06
20
0.38
0.02
25
0.15
0.01
Viscose, ripeness, 8;
mole ratio, acrylonitrile -cellulose,
3:1
NaOH, %
Nitrogen, %
Cyanoethyl, D.S.
Carboxyethyl, D.S.
6
0.8
20
0.09
0.12
in hydrolysis that is caused by an increase in sodium hydroxide concentra-
tion.
(e) ORGANOSOLUBLE CYANOETHYL CELLULOSE 40*42'44
Organosoluble cyanoethyl cellulose has been made from alkali cellulose,
from regenerated cellulose, and from alkali-soluble ethers, but not from
viscose. The xanthate group in the latter appears to interfere with the
tris(cyanoethylglucose) structure226 of the trisubstituted ether.
The reaction is cairied out in a short time with weak sodium hydroxide
and a large excess of acrylonitrile. The reaction appears heterogeneous,
when agitation is insufficient or suitable solvents are lacking, but the aver-
age degree of substitution reaches 2.5 to 3.0, and the products are soluble
in acetone, acrylonitrile, dimethylformamide, methyl formate, and p-
ethoxypropionitrile. Fibers are spun and sheeting cast from acetone solu-
tion.
(f) CARBOXYETHYL CELLULOSE
Carboxyethyl cellulose may be prepared from activated olefins by cyano-
ethylation under hydrolytic conditions,40'41'229 by the hydrolysis of cyano-
ethyl cellulose,230 or by the reaction of alkali cellulose with esters of acrylic
acid.45 Either alkali-soluble or water-soluble ethers231 may be prepared.
229 M. M. Cruz (to American Viscose), U. S. Patent 2,577,844 (Dec. 11, 1951) ; Chem.
Abstracts, 46, 2801 (1952).
aw w. M. Hutchinson (to Phillips Petroleum), U. S. Patent 2,481,513 (Sept. 13, 1949) ;
Chem. Abstracts, 44, 316 (1950); U. S. Patent 2,519,249 (August 13, 1950); Chem.
Abstracts, 44, 10319(1950).
*" L. H. Bock and A. L. Houk (to Rohm & Haas), Brit. Patent 562,584 (July 17,
1944) ; Chem. Abstracts, 40, 464 (1946).
958 CELLULOSE
Alkali-soluble carboxyethyl cellulose is an indirect product of cyanoethyl-
ation with 0.2 to 0.5 mole of acrylonitrile per mole of cellulose in the pres-
ence of 15 to 40% NaOH for an extended time. Water-soluble carboxyethyl
cellulose results when at least one mole of acrylonitrile per mole of cellulose
is used.
The addition of acrylate esters to alcohols was studied by Rehberg,232 who
found that high temperatures and long times should be avoided. Esters
of acrylic acid add to cellulose, but the free acid and its salts do not.233
RceiiOH + CH2= CH.COOR _NaOH) R0oiiOCH2CH2COONa + ROH (23]
(g) OTHER ETHERS
A mixed 0-carbamylethyl 0-carboxyethyl cellulose is made by the reac-
tion of acrylamide with cellulose, followed by partial hydrolysis.46
Sulfoethyl cellulose has been prepared from vinylsulfonic acid,39-47 its
salts and esters; from ethionic acid, or its anhydride, carbyl sulfate.37
Sulfamylethyl celluloses soluble in dilute alkali, but precipitated by acid,
were prepared from vinylsulfonamide.234
Vinylamine does not exist, but its cyclic isomer, ethylenimine, leacts
with viscose solutions at 100°C. to give acid-soluble products that corre-
spond in structure to the product to be expected from vinylamine (/3-amino-
ethyl cellulose). Ethylenimine serves as its own catalyst.51 Cellulose
diethylaminoethyl ethers were prepared by Vaughan235 and by Grassie.236
232 C. E. Rehberg, M. B. Dixon, and C. H. Fisher, /. Am. Chem. Soc , 68, 544 (1946)
233 V. R. Grassie (to Hercules Powder Co.), U. S Patent 2,539,417 (Jan. 30, 1951),
Chem. Abstracts, 45, 4552 (1951).
234 V. R. Grassie (to Hercules Powder Co.), U. S. Patent 2,580,351 (Dec. 22, 1951);
Chem. Abstracts, 46, 2802 (1952).
235 C. L. P. Vaughan (to Hercules Powder Co.), U. S. Patent 2,591,748 (April 8, 1952) ;
Chem. Abstracts, 46, 5842 (1952); U. S. Patent 2,623,042 (Dec. 23, 1952); Chem Ah
stracts, 47, 3564 (1953).
286 V. R. Grassie (to Hercules Powder Co.), U. S. Patent 2,623,041 (Dec. 23, 1952)
Chem. Abstracts, 47, 3564 (1953).
F. XANTHATES
EMIL KLINE1
When cellulose is treated with sodium hydroxide solution of mercerizing
strength to form alkali cellulose (sometimes referred to as soda cellulose;
see Section D of this Chapter IX) and this alkali cellulose is then treated
with carbon disulfide, interaction occurs with the formation of a sodium
salt of the cellulose ester of dithiocarbonic acid, more familiarly known as
cellulose xanthate. The formation of this ester of cellulose is broadly
represented as follows :
RoeiiONa + CSa » RcellOCSSNa (1)
This reaction, which was discovered in 1892 by Cross, Bevan, and
Beadle,2 is known as the 'Viscose reaction" since it is the basis of the prepa-
ration of "Viscose" — a solution of the xanthate in dilute sodium hydroxide
solution. (The name "viscose" originates from the viscous nature of the
solution and is a condensation of viscous cellulose.) It is one of the most
interesting reactions of cellulose and, because of low cost, industrially one
of the most important. Unlike most of the other cellulose derivatives,
however, the xanthate is not of interest for itself but only as a means of
solubilizing cellulose, from which solution (i.e., viscose) the cellulose may
be regenerated in almost any desired shape or form. The viscose process
has thus found use in the production of rayon, staple liber, cellophane,
sausage casings, bottle caps, artificial sponges, sizings, and related products,
and has been and still is the leading process in their manufacture. The
industrial importance of the process will be appreciated when it is considered
that in 1950 about 3,350,000,000 Ib. of finished products,3 valued by current
U.S. standards at approximately $1,750,000,000, were produced from vis-
1 The author is indebted to and wishes to acknowledge the extremely valuable as-
sistance of Gertrude M. Weisz in revising this Section.
2 C. F. Cross, E. J. Bevan, and C. Beadle, Ber.t 26, 1090 (1893); C. F. Cross and E.
J. Bevan, Ber., 34, 1513 (1901); C. F. Cross, E. J. Bevan, and C. Beadle, U. S. Patent
520,770 (June 5, 1894); Brit. Patent 8700 (Feb. 6, 1893); /. Soc. Chem. Ind. (London),
12, 516 (1893).
8 For 1950 rayon and staple production and prices see Rayon Organon, 22 (June,
1951).
959
960 CELLULOSE
cose. These products required the preparation of about 20,000,000 tons
of viscose.
Since the changes in cellulose upon reaction with carbon disulfide and the
regeneration of the cellulose from the viscose solution are of interest chiefly
as carried out industrially in the viscose process, the xanthates will be
discussed mainly from this viewpoint.
1. Preparation of Viscose
(a) SMALL-SCALE LABORATORY METHODS
The conventional procedure for the preparation of viscose involves a
series of steps. Alkali cellulose is first produced by steeping cellulose in
sodium hydroxide solution, pressing to remove the excess liquor, shredding,
and aging to the desired viscosity. The alkali cellulose is then xanthated
by reacting with carbon disulfide, and the xanthate is dissolved in dilute
sodium hydroxide solution. The viscose thus produced is filtered, sub-
jected to a vacuum treatment to remove air, and ripened to the desired
point, after which it is ready for use. These operations may be carried out
on a small scale as follows :
Place 100 g. (about 7 sheets, 6 in. by 6 in.) of air-dry, rayon-grade, "low-alpha" sulfite
cellulose sheets (see Table 41) edgewise in a rectangular nickel tank or glass battery jar
(7 in. wide by 7 in. high by 2.5 in. long) in a water bath at 21 °C. Fill the tank slowly
to about 1 in. above the pulp with 18.0% NaOH solution (see Table 42), previously ad-
justed to 21 °C. Allow the pulp to remain in the caustic solution for 1 hr., maintaining
the temperature at 21 °C. db 0.5 °C.
Siphon or pour off the caustic. Stack the steeped sheets carefully on a perforated
nickel plate and cover with another similar plate. Place the plates and the alkali cellu-
lose in a laboratory hydraulic or hand "letter" press and press to a weight of 300 g.
Shred the pressed sheets in a small Werner-Pfleiderer shredder for 2 hrs., maintaining the
temperature at 21 °C. it 0.5 °C., and then place the shredded alkali cellulose in a two-
quart glass fruit jar. Seal and place the jar on mechanical rollers in a water bath at
21 °C. Allow the alkali cellulose crumbs to age for 65 hrs. under these conditions (con-
stant rotation at 21 °C. ± 0.5 °C.).
After aging, add 32 g. of CSj, reseal, and replace the jar on the mechanical rollers in the
water bath. Allow the mass to rotate for 3 hrs. at 21 °C. =fc 0.5 °C. When the reaction
with CS« has been completed, remove the jar from the rolls, evacuate it to remove any
excess CSs, then add 169 g. of 18% sodium hydroxide solution and 764 g. of water. Stir
with a mechanical agitator (nickel or stainless steel) for 2 hrs., until solution is complete,
still maintaining the temperature at 21 °C. db 0.5 °C. This will give about 1265 g. of
viscose containing 7% cellulose and 6% NaOH, with a viscosity of 40 to 50 poises and a
total sulfur content of about 2.0%. After adequate filtration and ripening (discussed
later), the viscose will be comparable to a commercial solution and useful for the prepa-
ration of filaments, films, and the like.
IX. DERIVATIVES OF CELLULOSE 961
As will be clear from the discussion which is to follow, exact specifications covering
viscose composition, alkali cellulose aging time, and viscosity cannot be stated without
identifying in detail the type of cellulose and sodium hydroxide. If the desired viscose
composition and viscosity are not obtained, duplicate runs should be made in which the
alkali cellulose is analyzed for per cent of cellulose and per cent of NaOH, and the alkali
cellulose aging time varied. From the analysis of the alkali cellulose, the correct amounts
of NaOH and water to be added to the xanthate can be calculated. By plotting the
logarithm of the alkali cellulose aging time against the logarithm of the viscose viscosity
for two, three, or more runs, straight-line relationships are obtained from which the exact
time for any given viscosity can be determined.
If cellulose is not available in the form of sheets, viscose can be prepared from bulk
pulp, or from a good grade of absorbent cotton, by carrying out the steeping operation
in a beaker or any suitable vessel equipped with an agitator. The ratio of 18% NaOH
solution to cellulose should be about 25 to 1. After the treatment with NaOH, the excess
liquor may be removed from the bulk pulp contained in a cloth bag by centrifuging or
pressing, and then proceeding as described above.
For the laboratory preparation of small amounts of viscose, a number of
other procedures4'6 may be used provided it is not required to meet com-
mercial standards of composition, viscosity, and other specific properties.
One such procedure,6 based on an emulsion xanthation technique, is :
Mix 3.5 g. of air-dry, rayon-grade pulp with 31 ml. water for 1 hr. in a 150-ml. wide-
mouthed, glass-stoppered bottle. Then add 200 mg. glucose, 3 mg. abietic acid, 50 ml.
22.5% NaOH solution, and finally 3.5 ml. CS2. Stopper the bottle tightly, shake the
contents for 15 min., and then rotate the bottle slowly for 6 hrs. at 20.0 °C. ± 0.1 °C.
4 R. Bartunek, Cellulosechemie, 22, 56 (1944); Das Papier, 2, 442 (1948); G. Jayme
and coworkers, Papier-Fabr., 37, 97, 109 (1939); 38, 93, 101, 113, 277 (1940); Das Papier,
1, 133 (1947); Svensk Papperstidn., 50, 117 (1947); Melliand Textilber., 28, 125 (1947);
K. Hess and coworkers, Kolloid-Z., 98, 148 (1942); Kunstseide u. Zellwolle, 27, 37
(1949); 0. Samuelson and coworkers, Svensk Papperstidn., 51, 331 (1948); 52, 448
(1949); Svensk. Kern. Tid.t 61, 79 (1949); R. Vuori, Svensk Papperstidn., 49, 95 (1946);
Dissertation, Helsinki (1947); H. Haas, Das Papier, 2, 397 (1948); T. Bergek, Norsk
Skogind.t 2, 289 (1948); K. Jung, Kolloid-Z., 98, 192 (1942); 108, 120 (1944); G. Cen-
tola and F. Pancirolli, Ind. carta (Milan), 1, 63, 75 (1947); J. Lobering, Kolloid-Z., 98,
186 (1942); W. Klauditz and coworkers, Cellulosechemie, 22, 20, 121 (1944); W. Schra-
mek, Cellulosechemie, 19, 93 (1941); 20, 38 (1942).
* G. Jayme and coworkers, Zellwolle, Kunstseide, Seide, 48, 47 (1943) ; Cellulosechemie,
21, 73 (1943); Kolloid-Z., 107, 163 (1944); 108, 20 (1944); K. Lauer and W. Mansch,
Zellwolle u. Kunstseide, 2, 138 (1944); T. Kleinert and coworkers, Kolloid-Z. , 108, 137,
144 (1944); Svensk Papperstidn., 51, 541 (1948); L. Skark, Papierfabr.-Wochbl. Papier-
fabr., 75, 146 (1947); Das Papier, 2, 3, 186 (1948); H. Koch, Papier-Fabr., 39, 46
(1941); W. Winkler, Kunstseide u. Zellwolle, 28, 153 (1950).
« G. Jayme and J. Wellm, Kolloid-Z. , 107, 163 (1944).
962 CELLULOSE
(b) LARGE-SCALE MANUFACTURE »
*
(1) Raw Materials
The principal raw materials entering into the manufacture of viscose are
cellulose, sodium hydroxide, and carbon disulfide. Of these, cellulose is the
most important because it also constitutes the end product of the process.
At the present time the only practical sources of cellulose are cotton and
wood. Other possible cellulosic materials8 (agricultural residues including
straw, bagasse, and cornstalks) are satisfactory, but, except in a few coun-
tries, they have been too expensive to refine to the point required by the
process. When cotton is the source of cellulose, it is obtained mainly from
second-cut linters and hull fiber (see Chapter VI -B). Wood cellulose
(from which by far the greatest part of the world's output of viscose prod-
ucts is produced) is derived chiefly from softwoods such as spruce, western
hemlock, and southern pine, although very satisfactory pulps are also being
produced from certain of the hardwoods, including beech, aspen, and gum.9
Actually, the kind of wood is secondary in importance to the methods used
to isolate and purify the cellulose (see Chapter VI -A). For this purpose
the sulfite process has been employed almost exclusively, but recent de-
velopments in the sulfate process, involving prehydrolysis, now make its
use also possible for dissolving pulps.9
The factors which generally determine the type of cellulose to be used in
the manufacture of viscose are the type of product to be produced (rayon,
cellophane, staple, etc.), the quality of product, the price, and alkali cellu-
lose aging facilities. When strength, durability, and toughness are the
prime considerations (such as in the manufacture of high-tenacity rayon,
7 V. Hottenroth, Artificial Silk, Pitman, London, 1928; E. Wheeler, The Manufac-
ture of Artificial Silk, Van Nostrand, New York, 1931; M. H. Avram, The Rayon Indus-
try, Van Nostrand, New York, 1929; O. Faust, Kolloidchem. Tech., No. 2, 124 (1931);
W. Weltzien and K. G6tze, Chemische und Physikalische Technologic der Kunstseiden,
Akadem. Verlagsgesellschaft, Leipzig, 1930; H. G. Bodenbender, Zellwolle, 2d. ed.,
Chemisch-technischer Verlag Dr. Bodenbender, Berlin, 1937; K. Gotze, Chemiefasern
nachdem Viskoseverfahren (Reyon und Zellwolle), Springer, Berlin, 1951; L. H. Smith,
editor, Synthetic Fiber Developments in Germany, Textile Research Institute, New York,
1946; H. W. Rose, Rayon Industry of Japan, Textile Research Institute, New York,
1946; Continuous and Staple Fiber Plants of Germany, PB Rept. 377, 1945.
8 H. Jentgen, Kunstseide u. Zellwolle, 24, 350 (1942); H. Levinstein, Chemistry &
Industry, 1948, 538; M. G. Karnik and D. L. Sen, /. Sci, Ind, Research (India), 7, 351
(1948); 9B, 201 (1950).
9 S. Wang, Rayon and Mettiand Textile Monthly, 15, 227 (1934); J. N. McGovern
and G. K. Dickerman, Paper Trade J., 124, 33 (Jan. 9, 1947); L. L. Leach, Rayon Tex-
tile Monthly, 26, 631 (1945); Chem. Inds., 67, 576 (1950).
IX. DERIVATIVES OF CELLULOSE 963
sausage casings, and bottle caps), pulps high in alpha-cellulose content,
including those from cotton linters, and with uniform chain length, mini-
mum impurities, and a minimum quantity of cellulose with degree of poly-
merization (D.P.) less than about 150, are necessary, or at least are de-
sirable.10 When these requirements are not quite so important, or when
price is a deciding factor, pulps lower in alpha-cellulose content and poorer
in color are used. Although most pulps are produced in what might be
described as a "medium" or "normal" range of viscosity, "low-viscosity"
pulps are also in use ; the latter require only one-half to one-third the aging
time of the normal- viscosity pulps.
The various types of acceptable commercial pulps on the market at the
present time may be divided into five groups according to a combination of
alpha-cellulose content, cuprammonium viscosity, and principal end use.
The normal ranges of the common analytical characteristics of the pulps
in each of these groups are given in Table 41. n
It should be understood that, although analytical characteristics as
shown in Table 41 are of prime importance, a chemical analysis alone of a
pulp is not sufficient to establish its value as a source of cellulose for vis-
cose. As a matter of fact, exact specifications covering the cellulose raw
material cannot be set up since complete information is still lacking to
correlate all of the various characteristics of a pulp with the process and
end product. In addition to having a satisfactory chemical analysis, the
pulp must be sufficiently reactive to NaOH and C$2 to give a good solution
(good-filtering viscose), and it must always be uniform in all respects,
including reactivity, chemical analysis, viscosity, D.P. distribution, color,
and sheet structure.4 In this connection, considerable attention has been
given by a number of investigators to the. influence of various pulp proper-
ties and characteristics, and several test procedures for evaluating pulps
as to their suitability for viscose have been proposed.4'5 Some of these
10 O. P. Golova, Kunstseide, 17, 302 (1935); E. Lindpaintner, Melliand Textilber., 23,
229, 281 (1942); A. LeRollan, Ind. textile, 62, 144 (1945); O. Samuelson, Svensk. Kern.
Tid., 59, 105 (1947); R. Bartunek, Das Papier, 4, 451 (1950); E. A. Tippetts, Tappi, 33,
32 (1950).
11 A. Waller, The Svedberg (Memorial Volume), 1944, 400; A. H. Hooker and co-
workers, U. S. Patent 2,079,120 (May 4, 1937); Chem. Abstracts, 31, 4495 (1937); F. E.
Bartell and H. Cowling, Ind. Eng. Chem., 34, 607 (1942); R. L. Mitchell, Ind. Eng. Chem.
43, 1786 (1951); U. S. Patent 2,542,285 (Feb. 20, 1951); Chem. Abstracts, 45, 4042
(1951). For changes in some of the basic analytical characteristics of pulps due to
various viscose processing operations see: A. Pakschver and coworkers, Trans. Inst.
Chem. Technol. Ivanovo (U. S. S. R.\ 3, 158 (1940); A. Riedemann, Rayon Textile
Monthly, 29, No. 8, 45, No. 9, 82 (1948); R. L. Mitchell, Ind. Eng. Chem.t 41, 2197
(1949).
064
CELLULOSE
SaJ
rH d G
W O I
H^ ^2 »|**
3 8*8
0 0
5^
13 g
d o<
fe
*j
o
— a
II
«
O
pal
O O IO CO
O»OOOrHrH»-(O
pOOlOOOpOrH
grHO^HOOOO
O 00 C^J CO
»o co rfl d d d o" d o
O5 rH <M
I I I I I I I I I
C^I Tf O rH rH
»oooooooo
co
COCOCOOOOOO
O5 rH
I I I I I I ^ I
LO Th O rH
OO»OOOOOO
Tt<CQrH
OOOOCOrHOO
7
1 ' ' ' '
*O O rH
OOOOrHOOO
t^iOCOCOOOOO
00 rH
CO C^| IO
OpppCOrHpp
C<|iOCOl>OOOO>O
?rH rH
I I I I I ^ I I
O O O O rH O O O
eft co co d d d d d
t^
CS|
3s
•2 0)
d a
* a
» o °
fc J2 »S2
las
o aj d
8-1
3-g
d d
E d
o| 8 S
* * TJ •£
d > i ^* bfl
•I 1 1 1
Illl
In addi
alumin
ld be u
IX. DERIVATIVES OF CELLULOSE 965
involve the determination of D.P. distribution, end groups, total hydroly-
zates, and similar properties. Many involve the preparation of viscoses
of special compositions and the examination of these viscoses for particular
characteristics such as filterability, unfiltered residues, and mercerization
resistance. These procedures are undoubtedly useful in establishing ap-
proximate values. However, the only way to reach a final conclusion
regarding the utility of a given type of cellulose is actually to convert it into
commercial-type viscose and into the particular end product (filaments,
films, etc.) desired.12
The principal considerations in connection with the water and caustic
soda used in the viscose process are the impurities which may be present,
mostly metallic salts,11 which affect the viscosity, color, and solubility of
the xanthate and certain characteristics of the final product. It is im-
portant that all metallic and other impurities be uniform and as low as
TABLE 42
Representative Analysis of a 50% Caustic Liquor
Unpublished data compiled from analyses by du Pont Co. and by caustic vendors
Material
%
Material
P.p.m.
Fe
3-6
NaOH
50.0 -50.5
Ca
1-15
Na2SO4
0.001- 0.08
Al
10-20
Na2C03
0.05 - 0.20
Mg
5-15
NaCl
0.05 - 0.20
Mn
0.3-0.5
SiO2
0.005- 0.02
Cu
0.3-0.6
Ni
0.0-0.5
possible in concentration. The water also should be soft, preferably of
zero hardness. In caustic soda, the impurities vary somewhat depending
upon the type (caustic liquor versus solid caustic and ammonia-soda process
versus electrolytic process), although most producers have now so refined
their operations and so reduced all impurities that there is relatively little
variation in chemical analysis from one producer's material to another.
Caustic is usually supplied in the form of 50% liquor, in specially lined
tank cars; some 73% liquor is also made for viscose. Solid caustic, which
was employed almost universally some years ago, still finds some use,
particularly where caustic must be transported over long distances. The
range of impurities usually found in 50% caustic liquor used for viscose
is shown in Table 42.
» C. H. Goldsmith, Rayon and Melliand Textile Monthly, 16, 513 (1935); K. Fabd,
Kunstseidt, 17, 42 (1936).
966 CELLULOSE
The carbon disulfide used for viscose is essentially a chemically pure
product made usually by the direct interaction of carbon and sulfur under
definite conditions. For this reason there is usually no need for concern
over specifications, even though impurities, particularly free sulfur, should
be absent for optimum xanthate solubility.
(2} Preparation of Alkali Cellulose
(a) Steeping. The first step in the viscose process, as indicated above,
is the treatment of the cellulose raw material with sodium hydroxide solu-
tion of mercerizing strength. This treatment, known as the steeping oper-
ation, has two specific and distinct objectives : (1 ) swelling and absorption
of NaOH with the uniform and complete formation of alkali cellulose I,
and (2) the removal of hemicellulose and other impurities from the cellu-
lose.
In conventional industrial operation, the cellulose raw material (pulp), in the form of
sheets, is placed edgewise in a steeping press, that is, a rectangular tank, with a ram at
one end actuated by a screw or hydraulic pressure. The size of the press charge varies
considerably. In the United States, batches of 100 to 600 Ib. are normal, the sheets
varying from about 13 in. by 18 in. to 20 in. by 27 in. In Europe, charges as large as
2000 Ib.18 are said to be in use, with sheet sizes up to 30 in. by 40 in. The press is filled
slowly with caustic solution containing about 18% NaOH and 0.5 to 3.0% hemicellu-
lose; then the reaction is allowed to proceed for 30 to 60 min. at a definite temperature,
the normal range being 15-35 °C. (The reaction being exothermic, a rise in tempera-
ture of 2-3 °C. usually occurs.)
Of the factors that control the steeping operation, the nature of the
pulp,4'6 the time, the temperature, and the concentration of NaOH14 are
all-important. For the formation of alkali cellulose I the concentration
of NaOH must lie in the region in which alkali cellulose I is stable, that is,
about 14 to 20% NaOH. When the operation is carried out in a conven-
tional steeping press, 18% NaOH usually gives about the best results from
the viewpoint of xanthation efficiency and viscose filterability (see Fig.
71). M Deviations from this concentration of the order of ±1 to 2% may
not be particularly harmful if suitable adjustments are made, but neither
are they advantageous except in specific instances. Swelling of the pulp
varies with NaOH concentration; the lower the concentration (to about
18 H. Jentgen, Kunstseide, 18, 408 (1936).
14 A. Breguet and coworkers, M6m. services chim. Mat (Paris), 34, 157 (1948).
15 The data in Figures 71-74, inclusive, are representative of a typical commercial
viscose containing 7.5% cellulose and 6.5% NaOH, made from a 95% alpha-cellulose
pulp, pressed, after steeping, to a press weight ratio of 2.6 to 1.0 and xanthated, unless
otherwise indicated, with 34% CS» based on the cellulose in the alkali cellulose.
DC. DERIVATIVES OF CELLULOSE
967
9%), the higher is the swelling. The NaOH concentration also affects
the composition and the rate of aging of the alkali cellulose; the higher
the concentration of the NaOH, the faster is the aging.
Temperature is important in steeping mainly because of its effect on
swelling, which decreases with increasing temperature. Higher tempera-
tures therefore result, for example, in lower hemicellulose losses, higher
yield, stronger steeped sheets, less slumping in the press, and easier press-
ing. Lower temperatures, on the other hand, by increasing swelling,
result in greater pulp reactivity and in increased absorption of caustic.
From these effects, it will be obvious that steeping temperature also is a
factor in determining viscose viscosity, viscose ripening, and alkali cellulose
composition.
3000-
2 6 10 14 16 22 26 30
<y«»NAOH IN STEEPING LIQUOR
STE£P LIQUOR 2 * :! 2 2
AT 2.6 PRESS
RATIO IN A.C.
Q»
O>
SODIUM HYDROXIDE
Fig. 71. Effect of NaOH content of steep liquor and alkali cellulose (A. C.) on viscose
characteristics16: (A) effect on CS* absorption and distribution; (B) effect on filtera-
bility (plugging value). Courtesy of Research Division, Rayonier Incorporated.
Since mercerization is a relatively rapid reaction, time of steeping is
determined largely by the time necessary to solubilize and remove the hemi-
cellulose and other impurities from the pulp. As much as 15 min. is re-
quired to remove the main portion of hemicelluloses from some pulps in a
conventional steeping press, hence this may be considered about the mini-
mum permissible time. Additional time is desirable, especially with large
sheets, and 30 to 60 min. (in some instances 2 hrs.) is usually employed.
Some of the other variables16 which affect the conventional steeping
operation are type and concentration of hemicellulose, pulp sheet density
(the number of sheets per inch in the press), rate of filling of the steeping
press, pulp sheet formation and structure, and use of wetting agents.
16 Steeping Procedure, Rayonier Incorporated, San Francisco, April, 1941 ; E. Ring-
strom and N. H. Apler, Svensk Papperstidn., 51, 501 (1948).
968 CELLULOSE
The amount of hemicellulose present influences the uniformity of merceriza-
tion, the rate of alkali cellulose aging and xanthation, the solubility of the
xanthate, the rate of viscose ripening, and the characteristics of the final
product. It is desirable that the hemicellulose content of the steeping
solution be kept low. It is not economically feasible, however, to discard
the caustic used in the operation; hence, it is almost universal practice
to return it to the process and to remove the hemicellulose from a portion
of the solution by dialysis, thus maintaining a constant, fairly low con-
centration of hemicellulose in the steeping liquor. Although it has been
proposed to add to the steeping solution various other materials, including
wetting agents (which improve caustic penetration and shredding, thus
giving more uniform xanthation and improved solution), sodium sulfide,
metallic salts, and oxidizing agents (for reducing viscosity),17 alcohols,
proteins, and starch, few, if any, of these have found practical use.
(b) Pressing. Since sodium hydroxide reacts with carbon disulfide to
form useless by-products (see below), it is desirable to remove most, but
not all, of the excess NaOH from the alkali cellulose before this reaction.
For this reason, after steeping, the solution is drained from the tank and
the steeped pulp is pressed to about three times the weight of the original
air-dry pulp. As indicated, the waste solution, containing the hemicellu-
lose removed from the pulp, is dialyzed and used over again.
The ease of pressing is inversely proportional to the degree of swelling
and, hence, is determined by type of pulp, temperature, and concentration
of NaOH. The pressing factor or press-weight ratio (ratio of weight of
pressed alkali cellulose to original air-dry pulp) controls the composition of
the alkali cellulose as well as of the waste caustic solution. With an ordi-
nary rayon-grade pulp (91% alpha-cellulose) and 18% NaOH solution, a
press-weight ratio of 3.0 to 1.0 gives an alkali cellulose containing approxi-
mately 29.8% cellulose, 15.2% NaOH, and 55% H2O. Changing the press-
weight ratio changes the cellulose content, the NaOH content remaining
nearly constant. At a press-weight ratio of 2.5 to 1.0, for example, the
analysis is about 35% cellulose, 15% NaOH, and 50% H2O. For further
details as to the effect of pressing and other variables on the composition
of the alkali cellulose, see Section D of this Chapter IX.
The press-weight ratio also influences shredding, the rate, efficiency, and
uniformity of xanthation, the rate of degradation of the cellulose during
alkali cellulose aging, and the properties of the final product. The lower
17 For example, see A. H. Hooker and coworkers, U. S. Patent 2,079,120 (May 4,
1937) ; Chem. Abstracts, 31, 4495 (1937) ; D. Entwistle and W. R. Weigham, U. S. Patent
2,642,492 (Feb. 20, 1951); Chem. Abstracts, 45, 4454 (1951).
IX. DERIVATIVES OF CELLULOSE 969
the ratio is, the poorer is the shredding and the slower is the alkali cellulose
aging and the rate of carbon disulfide absorption. However, xanthation is
more efficient at the lower ratios, and solubility is improved, provided
shredding is satisfactory. In commercial operation, the pressing factors
employed are in an intermediate range from 2.7-1.0 to 3.2-1.0, which repre-
sents a compromise between poor shredding and low reaction rate on the
one hand and gelatinization of the fibers on the other.
(c) Shredding. After removal of the excess caustic by pressing, the alkali
cellulose is mechanically disintegrated or shredded for 1 to 3 hrs. at 25°
to 35°C. to form a mass of fibers resembling bread crumbs. This serves
to increase the surface of the alkali cellulose and to insure rapid and uni-
form aging and xanthation. To achieve further uniformity and low rost,
two or more steeping press charges may be combined into a single shredder
charge. The operation is conventionally carried out in equipment re-
sembling a dough mixer, that is, a Werner-Pfleiderer shredder, which tears
the sheets between rapidly rotating blades and a stationary saddle.
Optimum shredding is indicated by a maximum of fibrillation without ag-
glomeration or knot formation. Although various tests, such as deter-
mination of particle size and density, have been devised for evaluating
shredding, none are really significant. In addition to the factors already
mentioned, time and temperature18 must be carefully controlled, not only
to secure optimum shredding, but also to control the oxygen absorption
and hence degradation of the cellulose which always occur during this stage.
(d) Aging. After the shredding operation, the alkali cellulose crumbs,
in cylindrical or rectangular steel containers (usually one container per
shredder batch), are placed in a constant-temperature room for 1 to 3
days at 25° to 30°C.; this step in the process is called "aging." In some
cases, however, the aging step has been simplified and/or shortened by the
use of low-viscosity pulps, aging before shredding, high-temperature
shredding, continuous aging, and other means. The crumbs are kept care-
fully covered to prevent reaction with COz from the air and evaporation of
moisture which would be detrimental to xanthation and filterability.
During this step (see Section D of this Chapter IX) further absorption
of oxygen occurs with generation of heat, and the viscosity of the cellulose
is reduced. This reduction in viscosity is essential since the solution vis-
cosities which would otherwise be obtained from present-day, unaged,
commercial celluloses would be considerably higher than desired. The
rate and extent of the viscosity reduction during aging are controlled by
w M. Huzii, J. Soc. Chem. Ind.t Japan, 43, Suppl. binding, 191 (1940).
970 CELLULOSE
temperature and time19; the higher the temperature, the faster is the aging
(the temperature coefficient is high), and the longer the time, the lower is
the viscosity. Other factors influencing aging11-20 include type of cellu-
lose and caustic soda, presence of oxidizing agents or metallic impurities
such as iron and manganese which accelerate the reaction, xanthation con-
ditions, and viscose composition.
(3) Xanthation of Alkali Cellulose
The xanthation of alkali cellulose is the most significant step in the
making of viscose, since here the actual conversion of the cellulose to a
soluble derivative takes place. For this purpose, the aged alkali cellulose
is usually placed in a hexagonal or cylindrical rotating drum or a "wet
churn/'21 and CS2 is slowly admitted as a gas or liquid through a perfo-
rated pipe within the drum. The reaction is permitted to continue for
1 to 3 hrs., in the temperature range 20-35 °C., either at atmospheric pres-
sure or under vacuum. The quantity of CS2 added is usually between
30 and 40%, based on the cellulose in the alkali cellulose. As the reaction
proceeds, the white alkali cellulose gradually becomes yellow and finally
assumes an orange or "carrot" color. The end product usually retains
some of its original crumbly, mealy character although it is somewhat
sticky; its actual color and physical characteristics are determined in part
by the alkali cellulose composition, the CS2 concentration, and the tempera-
ture and time of reaction.
Several of the factors influencing the xanthation reaction have been men-
tioned. It may be added that (1) the nature (e.g., type morphology,
and purification) of the cellulose4-6 affects the rate and degree of xanthation
and the amount of CS2 required to produce good solubility ; (2) the degree
of xanthation increases and solubility improves with increasing concen-
tration of CS2 but both reach a maximum beyond which further amounts
of €82 have little, if any, effect; (3) the uniformity of €82 distribution is
important for good solubility (crowding the xanthation vessel, for example,
affects filtration adversely); (4) the rate of xanthation, which reaches a
maximum and then falls off, is probably independent of the D.P. of the
cellulose but is proportional to the vapor pressure of the €82 and the tem-
perature; (5) the maximum degree of xanthate substitution decreases with
19 A. V. Fitter, /. Soc. Dyers Colour ists, 51, 21 (1935).
» I. Jurisch, Kunstseide u. Zellwolle, 23, 266 (1941) ; O. Samuelson, Cellulosa och Pap-
per 1908-1948, S.P.C.I., pp. 295-325; G. A. SchrSter, Kolloid-Z., 105, 223 (1943).
81 M. Golben, U. S. Patent 2,492,421 (Dec. 27, 1949); Chem. Abstracts, 44, 2240
(1950) ; E. Kline, U. S. Patent 2,513,652 (July 4, 1950) ; Chem. Abstracts. 44, 8658 (1950).
IX. DERIVATIVES OF CELLULOSE 971
increase in temperature, but solubility increases with increase in tempera-
ture in the range 15~35°C., the time and CS* concentration being constant;
if the time is reduced accordingly, solubility decreases above about 30 °C.;
(6) the time of reaction, which may be measured by the time to re-estab-
lish the original xanthation pressure, increases with CS% concentration ; it
should be sufficient for complete absorption of CS2, but the degree of
xanthation does not increase nor does the solubility improve with further
reaction time; (7) the solution viscosity of the cellulose is markedly re-
duced by oxidation during xanthation, hence temperature and presence
of oxygen are important factors; (8) the stability of the xanthate de-
creases with increase in temperature, increase in press-weight ratio, increase
in degree of xanthation, and increase in hemicellulose content; and (9)
xanthation conditions, particularly temperature, CS2 concentration, and
degree of xanthation, influence regeneration and the characteristics of the
final product.4-6-22 For the effect of a number of xanthation conditions on
xanthation rate, CS2 absorption, and solubility (viscose filterability),
see Figures 72 and 73.
At the end of the xanthation reaction the excess CS2 is removed by evacu-
ation, and the mass is dissolved in dilute NaOH solution to form the final
solution — viscose. To achieve greater uniformity, two batches of xan-
thate (four steeping press charges) are often combined at this point to form
a single charge. The amounts of NaOH and water employed in preparing
the final solution depend upon the alkali cellulose analysis and the viscose
composition (% cellulose and % NaOH) desired. The latter varies with
different producers and with the type of end product, although the concen-
trations are usually in the range of 6.0 to 10.0% cellulose and 4.0 to 8.0%
NaOH. For example, with an alkali cellulose containing 29.8% cellulose
and 15.2% NaOH, a press-weight ratio of 3 to 1, 35% CS2, and the com-
bination of four steeping press charges of 150 Ib. of pulp each, the pro-
duction of a viscose solution containing 7% cellulose and 6% NaOH would
require 5490 Ib. of water and 186 Ib. of 100% NaOH, and the total viscose
weight obtained would be about 7650 Ib.
Dissolving of the xanthate is referred to as the "mixing" step from the
fact that it is carried out by simply mixing the xanthate and NaOH solu-
tion in a large cylindrical tank (Vissolver) equipped with an agitator.
The mixing operation normally requires from 2 to 6 hrs., the temperature
being controlled in the range 15~20°C.
22 P. C. Scherer and coworkers, Rayon Textile Monthly, 19, 478, 541 (1938); 20, 24,
81, 498, 577 (1939); 21, 529, 611 (1940); W. Schramek and E. Zehmisch, Kolloid-Bei-
hefte, 48, 93(1938).
972
CELLULOSE
(4) Ripening of Viscose
As initially prepared, viscose is relatively difficult to coagulate and,
hence, is not suitable for most industrial operations. To make it fit for
applications such as spinning and casting, the viscose is transferred (usually
after further blending) to large tanks in another constant-temperature
room where it is stored under controlled conditions of time and tempera-
a co
£9
2hJ 4
XANTHATE SULFUR
IN VISCOSE
SULFUR
ADDED
CS2
10 20 30 40 50 60 70 80
% CS2 USED IN XANTHATION
_1
0 30 60 90 120 150 i60
XANTHATION TIME - MINUTES AT 3<fC
(34% CS2 ON CELL. IN A C.)
:4\V X^^.
80 100 120 140 160 ISO
XANTHATION TIME
MINUTES AT 30° C
60 120 160 240 300
XANTHATION TIME- MINUTES
(34% CS2 ON CELL. IN A.C.)
Fig. 72. Effect of certain xanthation conditions on xanthation rate and sulfur dis-
tribution16: (A and B) effect of (A) €82 'concentration and (B) xanthation time on CSz
absorption and distribution; (C and D) effect of (C) temperature and €82 pressure and
(D) C& concentration on rate-of xanthation. A. C. stands for alkali cellulose. Cour-
tesy of Research Division, Rayonier Incorporated.
ture to permit "ripening/1 During this period various changes occur, the
most important being spontaneous decomposition of the xanthate and an
increase in ease of coagulation of the solution. However, since these
changes continue until coagulation and regeneration of the cellulose are
complete, the viscose cannot be stored indefinitely but must be used within
a few days after preparation.
For any given viscose, the rate at which ripening occurs is governed by
IX. DERIVATIVES OP CELLULOSE
973
temperature; the higher the temperature, the faster is the ripening.
(Ripening is an exothermic process and the temperature coefficient of the
chemical reactions involved is of the order of 2.5-3.0 per 10°C.) The ex-
tent of the ripening (and hence time) is determined (empirically) by the
regenerating conditions to be used and by the use to which the viscose is to
be put, that is, whether it is to be spun into threads, cast into films, or
used for some other purpose. Coagulating methods have been developed
7oy
XANTHATION TIME
Y~IN MINUTES AT 30° C.
150 240 340
30 60 90 120 150 180 210
XANTHATION TIME - MINUTES AT 30» C.
C34«Vo CS2 ON CELL. IN A.C.)
10 20 30 40 50 60 70 80
<yoCS2 USED IN XANTHATION
(BASED ON «*o CELL. IN A.C.)
3-f.oooh
XANTHATION TIME IN MINUTES
USED AT EACH TEMPERATURE ~~\
360 180 120 70 30
10 20 30 40 50
XANTHATION TEMPERATURE-°C ,
(34%CS? ON CELL. IN A.C.)
0 O.I
BARRATTE
0.2 0.3 0.4
LOADING-g./c.c. RATIO
Fig. 73. Effect of certain xanthation conditions on viscose filterability (plugging
value):16 (A) effect of xanthation time; (B) effect of C& concentration; (C) effect of
xanthation temperature; (D) effect of Barratte crowding. A. C. stands for alkali cel-
lulose. Courtesy of Research Division, Rayonier Incorporated.
in recent years which permit the use of relatively "unripe" or "green"
viscose, but in most operations ripening requires 1 to 3 days, the normal
temperature range being 15-25°C.
During this period, various analyses also are carried out, the viscose is
filtered (usually in two or more stages through plate and frame filter presses
dressed with various combinations and types of cotton fabric and cotton
batting), and finally evacuated, to remove air. When the ripening has
974 CELLULOSE
reached the point required for conversion to the particular end product
desired, the viscose is ready for use.
(c) CONTINUOUS AND "QUICK11 PROCESSES
Other methods for preparing viscose, and particularly alkali cellulose,
have come into use, or have been suggested, which are cheaper or which
overcome some of the objections to the multistage operation just described.
These mostly involve continuous methods of operation (except for the
xanthation step, which apparently is not yet being carried out continuously
on an industrial scale) or a reduction in the number of stages. In certain
operations, pulp is treated in bulk form in tanks equipped with suitable
agitators (slurry steeping), and the excess caustic is removed by vacuum
filtration, centrifuging, and/or suitably designed presses or press rolls.
This type of procedure, starting with either sheets or bulk pulp, may also
be carried out continuously. Another continuous-type procedure involves
the passing of pulp in sheets or in roll form on a conveyor through a caustic
solution, followed in some cases by continuous pressing, shredding, and
aging. If the viscose plant is erected adjacent to the pulp plant, the pulp
may be employed without drying. Some of the other variations involve
countercurrent steeping, spraying the pulp with caustic in a shredder,
continuous shredding, continuous and/or quick alkali cellulose aging by use
of elevated temperatures or addition of oxidizing or other agents, combina-
tion of xanthation and mixing in a "xantha tor-mixer, " treatment with
carbon disulfide in the presence of an inert organic solvent or at extremely
low temperatures, and continuous mixing by various disintegrating de-
vices.13'23
88 H. Jentgen, Kunstseide u. Zellwolle, 20, 449 (1938) ; R. Monterray, Rusta-Rayonne ,
13, 517 (1938); Rayon Textile Monthly, 18, 227 (1937); H. Schmidt, Chem. Fabrik., 4,
73, 85, 97 (1931) ; F. Steimmig, U. S. Patent 2,005,811 (June 25, 1935) ; Chem. Abstracts,
29, 5653 (1935); German Patent 604,015 (Oct. 17, 1934); Chem. Abstracts, 29, 926
(1935); French Patent 751,617 (Feb. 27, 1933); Chem .- Abstracts , 28, 893 (1934); H.
von Kohorn zu Kornegg, U. SA Patent 2,218,836 (Oct. 22, 1940); Chem. Abstracts, 35,
1227 (1941) ; G. A. Richter, U. S. Patent 1,955, 092 (Apr. 17, 1934) ; Chem. Abstracts, 28,
4229 (1934); L. Lilienfeld, U. S. Patent 1,658,607 (Feb. 7, 1928); Chem. Abstracts, 22,
1237 (1928); H. Plauson, Brit. Patent 184,533 (April 11, 1911); Chem. Abstracts, 7,
412 (1913); A. J. Hailwood, Brit. Patent 281,117 (Nov. 30, 1926); Chem. Abstracts, 22,
3528 (1928); I. G. Farbenindustrie Akt.-Ges., Brit. Patent 434,540 (Sept. 4, 1035);
E. I. du Pont de Nemours & Co., Brit. Patent 463,056 (Mar. 19, 1937) ; Chem. Abstracts,
31, 6003 (1937); Deutsche Gasgluhlicht-Auer-Gesellschaft m.b.H., German Patent
413,511 (Oct. 1, 1921); Chem. Zentr., 96, II, 368 (1925) ; W. Grotzinger, French Patent
823,836 (Jan. 27, 1938) ; Chem. Abstracts, 32, 6')58 (1938) ; L. H. Smith, editor, Synthetic
Fiber Developments in Germany, Textile Research Institute, New York, 1946; H. W.
Continued on next page.
IX. DERIVATIVES OF CELLULOSE 975
2. Mechanism of Cellulose Xanthate and Viscose Formation
On the basis of the analogy between cellulose and ordinary alcohols, the
reaction between alkali cellulose and CS>2 was first represented in a manner
similar to the reaction between €82 and alcohols in the presence of alkali.
It has been known for a long time that when ethyl alcohol is treated with
C$2 in the presence of NaOH, sodium ethyl xanthate is formed, the reac-
tion being as follows :
C2H6OH + NaOH + €82 > C2H6OCSSNa + H2O (2)
An alcoholate is supposedly first formed by the action of the metal hy-
droxide on the alcohol, and then the metal xanthate is produced by the reac-
tion of CS2 with the alcoholate.24 Similar compounds are formed with other
metal hydroxides and other alcohols, including polyhydric alcohols and
simple sugars.25 If cellulose is substituted for the ethyl alcohol in equation
2, the analogous reaction product is sodium cellulose xanthate26 as shown in
equation 1.
Equation 2, however, may be considered only as a simple, general ex-
pression of the viscose reaction. The actual reaction is of a colloidal na-
Rose, Rayon Industry of Japan, Textile Research Institute, New York, 1946; H. Hoff-
man, Kunstseide u. Zellwolle, 26, 8 (1948) ; H. Von Kohorn, IL S. Patent 2,302,077 (Nov.
17, 1942); Chem. Abstracts, 37, 2193 (1943); R. W. Nash, U. S. Patent 2,338,196 (Jan
4, 1944); Chem. Abstracts, 38, 3840 (1944); O. Kohorn and R. Brandes, Brit. Patent
433,414 (Aug. 14, 1935); D. Entwistle and W. R. Weigham, U. S. Patent 2,542,492
(Feb. 20, 1951) ; Chem. Abstracts, 45, 4454 (1951) ; N. A. Copeland, U. S. Patent 2,355,-
057 (Aug. 8, 1944); Chem. Abstracts, 38, 6561 (1944); J. P. Hollihan, Jr., U. S. Patent
2,355,650 (Aug. 15, 1944); Chem. Abstracts, 39, 192 (1945); W. R. Schmitz, Jr., U. S.
Patents 2,360,984 (Oct. 24, 1944); 2,364,392 (Dec. 5, 1944); Chem. Abstracts, 39, 3666
(1945); W. H. Bradshaw, U. S. Patent 2,452,542 (Nov. 2, 1948); Chem. Abstracts, 43,
2435 (1949); J, Brackett, U. S. Patent 2,480,998 (Sept. 6, 1949); Chem. Abstracts, 44,
1705 ( 1950) ; J. G. Weeldenberg, U. S. Patent 2,499,010 (Feb. 28, 1950) ; Chem. Abstracts,
44, 4678 (1950) ; M. P. Kulp, U. S. Patents 2,510,984 (June 13, 1950) ; Chem. Abstracts,
44, 8658 (1950) ; 2,539,437 (Jan. 30, 1951) ; Chem. Abstracts, 45, 3598 (1951) ; Suddeutsche
Zellwolle A.G., Belgian Patent 450,505 (June, 1943); Chem. Abstracts, 41, 7118 (1947).
24 K. Atsuki and T. Takata, Cellulose Ind. (Tokyo), 16, 21 (1940) ; but see also A. E.
Brodsky and coworkers, /. Chem. Phys.. 11, 342 (1943); I. A. Makolkin, A eta Physico-
chim. ( U. R. S. S.), 17, 319 (1942).
25 M. Ragg, Chem. Ztg., 34, 82 (1910); E. Berl and J. Bitter, Cellulosechemie, 7, 137
(1926); T. Lieser and coworkers, Ann., 495, 235 (1932); 511, 121 (1934); E. Treiber,
Monatsh., 82, 53(1951).
26 The name xanthate is derived from the Greek word "xanthos," meaning yellow,
in view of the yellow color of the cuprous salt of xanthic acid. It should be mentioned
also that xanthic acid is really the ethyl ester of dithiocarbonic acid and therefore the
term cellulose xanthate is a misnomer. The latter compound is not the cellulose ester
of xanthic acid but of unsymmetrical dithiocarbonic acid.
976 CELLULOSE
ture and is far more complicated. It has been the subject of numerous
investigations,2*4'5'22-27 the more important of which will be discussed below.
These have now answered many of the questions concerning the mechanics
and chemistry involved, although a number of the conclusions that have
been reached are even now not universally recognized and further study
of the reaction is still needed for its complete understanding. Progress
in the field has been difficult due to the chemical instability of the xanthate
and even more to the prevalence of questionable conceptions about the
nature of cellulose reactions and the properties of high-polymer solutions.
Unlike many of the other cellulose derivatives such as the trinitrate and
triacetate, the trixanthate does not form readily. In the commercial vis-
cose reaction, for example, the average degree of xanthate substitution is
only of the order of one xanthate group per two anhydroglucose units.
However, higher degrees of substitution and even polyxanthates are pos-
sible under special conditions, and the preparation and study of these higher
xanthates have permitted a better understanding of the viscose reaction.
If cellulose is treated with certain tetraalkylammonium hydroxide solu-
tions,28 it disperses completely instead of merely swelling, as when treated
with sodium hydroxide solution. When these molecular solutions of cellu-
lose react with CS2, compounds are formed which have been shown to be
the salts of cellulose trixanthic acid corresponding to the particular organic
base employed. The reaction depends upon complete dispersion of the
cellulose which in turn depends upon the concentration and molecular
weight of the organic base. With 3 to 4 TV tetraethylammonium hydroxide
at 0°C., for example, the reaction is :
(C6H702(OH)3)n + 3w (C2H5)4NOH + 3n C& »
(C6H702[OCS2N(C2Hs)4]3)« + Bn H2O (3)
The trixanthate formed gives a trixanthogen ([C6H7O6(CS2)3]2)W upon
oxidation with iodine. With tetramethylammonium hydroxide, however,
87 H. Ost, Ann., 382, 340 (1911); R. Wolffenstein and E. Oeser, Ber., 56B, 785
(1923); Kunstseide, 7, 2, 27, 74 (1925); R. Bernhardt, Kunstseide, 8, 173 (1926); E.
Heuser and M. Schuster, Cellulosechemie, 7, 17 (1926); T. Lieser, Ann., 464, 43 (1928);
Cellulosechemie, 10, 156 (1929); E. Geiger, Helv. Chim. Ada, 13, 281 (1930); S. Iwasaki,
/. Soc. Chem. lnd.t Japan, 35, Suppl. binding, 91, 92 (1932); O. Faust, Ber., 62, 2567
(1929); J. Compton, Ind. Eng. Chem., 31, 1250 (1939); H. K. Maeda, /. Soc. Textile
and Cellulose Ind., Japan, 1, 672, 2, 8, 13 (1945); O. Samuelson, Cellulosa och Papper
(1908-1948), S.P.C.L, pp. 295^325.
28 T. Lieser, Chem.-Ztg., 60, 387 (1936); T. Lieser and E. Leckzyck, Ann., 522, 56
(1936) ; T. Lieser, Kolloid-Z., 81, 234 (1937) ; Rohm & Haas, Brit. Patent 439,806 (Dec.
9, 1935).
IX. DERIVATIVES OF CELLULOSE 977
dispersion does not occur and the trixanthate is not obtained. With other
bases and under other conditions of concentration, dixanthates are formed
which upon oxidation with iodine give the dixanthogen ([CeHsCVCS^^n-
If cellulose is allowed to react with metallic sodium in liquid ammonia
and the resulting trisodium derivative is treated with €82 under special
conditions, a product corresponding to the trixanthate is formed.29 Also,
if sodium cellulose xanthate, as produced commercially with 30 to 40%
CS2 based on the cellulose in the alkali cellulose, is partially dissolved
by the addition of NaOH and/or water or completely dissolved to form a
normal viscose solution, and this dispersed xanthate is treated with fur-
ther quantities of CS2 or with a more highly substituted xanthate, the re-
action continues and higher degrees of substitution, that is, polyxanthates,
result. By "after-xanthating" procedures of this type and even by the
direct xanthation of regenerated cellulose (as opposed to native cellulose) or
of alkali cellulose pressed less than normal, products containing up to nearly
two xanthate groups per anhydroglucose unit (7-values30 up to 180) have
been obtained.31'82
Thus, the degree of xanthate substitution is in proportion to the disper-
sion of the reacting cellulose. When the cellulose is completely dissolved,
and presumably all OH groups are made available for reaction, a permutoid
reaction and a stoichiometric compound, that is, a trixanthate, results.
When dispersion is not complete, only the OH groups that are actually
available or accessible undergo reaction.
In commercial xanthation, the reaction occurs between CS2 and swollen,
i.e., fibrous, native cellulose, not dispersed cellulose, and all of the OH groups
are clearly not equally available. That chemical reaction does occur with
conversion of some OH groups to xanthate groups is established by the
high temperature coefficient of the reaction, by spectrochemical analysis,
29 P. C. Scherer and L. P. Gotsch, Bull. Virginia Polytech. Inst., 32, 11 (1939).
30 The 7- value is the number of xanthate groups per 100 anhydroglucose units; thus
a 7-value of 100 is equivalent to 1 mole €82 per C6H10O6, i.e., 46.95% €82 (39.5% S)
on the cellulose or 3.0% xanthate S in viscose containing 7.5% cellulose. See H. Fink,
R. Stahn, and A. Matthes, Angew. Chem., 47, 602 (1934) ; N. V. Nederlandsche Kunstzij-
defabriek, German Patent 421,506 (Jan. 18, 1922); Chem. Zentr., 97, I, 1332 (1926);
L. Lilienfeld, Brit. Patent 212,865 (June 2, 1923); Chem. Abstracts, 18, 2249 (1924).
81 T. Lieser, Ann., 528, 276 (1937) ; W. Schramek and E. Zehmisch, Kolloid-Beihefte,
48, 93 (1938); K. Jung, Kolloid-Z., 108, 120 (1944); K. Lauer and coworkers, Kolloid-
Z., 110, 26 (1945); 119, 151 (1950); K. Hess and coworkers, Naturwissenschaften, 38,
433 (1951).
82 G. Centola, Atti congr. intern, chim., 10th Congr., Rome, 4, 117, 129, 138, 722, 728
(1938).
978 CELLULOSE
by the fact that the reaction curve follows the equation of a unimolecular
reaction, and by other data.22'33 However, the degree of substitution, as
mentioned above, is usually only about 1CS2 : 2CeHioO5. This is equivalent
to 23.5% CS2 or 19.8% S based on the cellulose, which is in line with con-
centrations actually used if it is considered that only about 75% of the total
€82 used reacts with the cellulose. (In Figure 72B, starting with 34%
€82, a maximum xanthate sulfur content of about 1.65% on the viscose or
22% on the cellulose is indicated.) Moreover, soluble xanthates have
been produced with a molar ratio of CS2 : C6H10O5 of only 1 : 2.5 (19% CS2)22
and it has been claimed that 16% CS2 and even less is sufficient.34 But
higher degrees of substitution than the ratio 1:2 are also possible with
fibrous alkali cellulose. Although good solubility does not require it,
CS2 : CeHioOs molar ratios of the order of 1 : 1 are readily produced by in-
creasing the amount of CS2 and time of reaction sufficiently.35 (See Fig.
72A which indicates about 2.9% xanthate sulfur in viscose, or 39% on the
cellulose, starting with 80% CS2.)
This does not necessarily mean that stoichiometric compounds of 1:1
or 1 : 2 ratios are formed. A considerable amount of evidence has been
developed to show that the reaction is topochemical or micellar and that
only the cellulose in the amorphous areas and/or on the surface of the crys-
tallites reacts. Against this, the conclusion has been reached by others that
the reaction takes place also within the crystalline regions and is permutoid
or molecular in nature.
Regardless of the final explanation of the reaction mechanism, the im-
portance of alkali cellulose I in industrial xanthation is generally well
recognized and established. Attempts to produce a soluble xanthate from
fibrous alkali cellulose by using concentrations of NaOH in steeping much
above or below 18% have been unsuccessful.36 However, it is not the con-
33 B. Rassow and W. Aehnelt, Cellulosechemie, 10, 169 (1929); K. Atsuki and T.
Takata, J. Soc. Chem. Ind.t Japan, 43, Suppl. binding, 399 (1940).
84 A. J. Burette, French Patent 430,221 (May 22, 1911); L. Lilienfeld, U. S. Patent
1,658,607 (Feb. 7, 1928); Chem. Abstracts, 22, 1237 (1928); see also E. Berl and J. Bit-
ter, Cellulosechemie, 7, 137 (1926); K. Hess, Die Chemie der Zellulose, Akadem. Ver-
lagsgesellschaft, Leipzig, 1928, p. 326; N. I. Nikitin and T. I. Rudneva, Compt. rend,
acad. sci. U. R. S. S., 28; 240 (1940).
35 G. de Wyss, Ind. En°. Chem., 17, 1044 (1925); O. Faust, Kolloid-Z., 46, 329
(192S) ; Ber., 62, 2572 (1929) ; J. G. Weeldenberg, Chem. Weekblad, 30, 47 (1933) ; P. C.
Scherer and D. W. Miller, Rayon Textile Monthly, 19, 478 (1938) ; H. L. Bredee, KolHl-
Z., 04, 81 (1941); K. Hess, H. Kiessig, and W. Koblitz, Naturwissenschaften, 38, 433
(1951); Z. Elektrochem., 55, 697 (1951).
86 K. Atsuki, Cellulose Ind. (Tokyo), 7, 207 (1931) ; A Lottermoser, Z. angew. Chem.
42, 1151 (1929),
IX. DERIVATIVES OF CELLULOSE 979
centration of the NaOH surrounding the alkali cellulose which determines
the solubility of the xanthate but the composition of the alkali cellulose
crystallites.37 With 18% NaOH, alkali cellulose I is formed and this com-
pound must be considered the basis of the industrial xanthate reaction.
Alkali cellulose II apparently also reacts with carbon disulfide to pro-
duce a soluble xanthate with a CS^CeHnA, ratio of 1:2, but only at ele-
vated temperature (60 °C.). This reaction is possibly even more non-
uniform than that of alkali cellulose I. Likewise, if cellulose is sufficiently
milled or is xanthated by emulsion techniques, it may be converted to a
soluble xanthate by treatment with dilute alkali and €82 without the prior
formation of alkali cellulose I. Although these reactions38 cannot be ig-
nored, they involve conditions quite different from those encountered in
industrial xanthation.
The contention that the reaction is micellar in nature and follows a
heterogeneous course is based in part on the ideas that the ratio of NaOH
to CaHiaOs in alkali cellulose I is 1:2, that the maximum degree of sub-
stitution of fibrous alkali cellulose is of the order of lCS2:2C6Hi0O6, and
that in view of the dispersibility of unxanthated alkali cellulose by cellu-
lose xanthate it is not necessary for all the cellulose to react initially with
CS2 to permit good solution. According to this viewpoint,31*32'37*39 the
reaction of fibrous alkali cellulose with excess CS2 results only in the for-
mation of alkali cellulose IV which is insoluble or, at most, in a degree of
substitution of about 0.8 mole of CS% per anhydroglucose unit. (Higher
degrees of substitution can admittedly be obtained with excess €82 but
only by adding alkali or water, by further xanthation and/or by redistribu-
tion of CS2 groups after dissolving, and such reactions do lead to penetra-
tion of the lattice and a permutoid compound.) Likewise, in commercial
alkali cellulose, in which the molar ratio of NaOH to CeHjoOs is 2:1 (the
usual composition is about 15% NaOH-30% CeHuOb), there is ample Na
in the amorphous areas and on the surface of the crystallites to react with
all the CS2 normally used (about 0.75 mole or 35% CS2 on the cellulose)
and thus to account for a 1 : 2 xanthate. This is true actually whether
the ratio of NaOH to C6Hi0O6 in alkali cellulose I is 1 :2 or 1 : 1. It is also
true even in the formation of a 1 : 1 xanthate, when excess CSz is used, on the
basis that the ratio of NaOH to C6H]0O6 in alkali cellulose I is 1 : 2.
87 W. Schramek and F. Ktittner, Kolloid-Beihefte, 42, 331 (1935).
38 K. Hess and H. Hepp, Melliand Textilber., 29, 305, 343 (1948).
89 J. Frenkel, CellulosecHemie, 9, 26 (1928); P. C. Scherer, Rayon Textile Monthly,
27, 22, 409 (1946); T. Lieser, KottM-Z.f 94, 96 (1941); Bar., 74B, 708 (1941); W.
Schramek. Kolloti-Z., 94, 92 (1941).
980 CELLULOSE
The micellar mode of reaction has also been suggested on the basis
that:31-32-37'39-41
X-ray investigations show that alkali cellulose I is always present in
commercial xanthate; even though the alkali cellulose I diagram dis-
appears on long standing, this does not in itself prove that the reaction is
molecular.
When cellulose xanthate is methylated and the methylated product is
subjected to acetolysis, nearly 50% of unmethylated cellobiose octaacetate
is obtained.
Fractionations of viscose with salt solutions (in the author's experience)
and of a diethylchloracetatnide derivative show the presence of xanthates
of various degrees of substitution, precluding the existence of a definite
chemical compound in the ordinary sense.
Mercerization is a micellar reaction and cuprammonium solutions of
cellulose are micellar.
Molecular and micellar solutions of cellulose have the same viscosity,
and changes in viscosity or D.P. do not necessarily mean changes in crystal-
lite size; it is, therefore, not surprising that the CS2: CeHioOs ratio does not
change with D.P.
Carbon disulfide reacts with fibrous alkali cellulose only in dissolved
form via the free NaOH, first forming dithiocarbonate and the latter then
reacting with the alkali cellulose.
Soluble xanthates can be produced without prior formation of alkali cellu-
lose I (see above).
The initial reaction of CS2 with fibrous alkali cellulose is heterogeneous.
The only plausible explanation for such a reaction is that it is topochemical
or micellar in nature. Since the crystalline and amorphous areas in cellu-
lose exist in the proportion of approximately 1:1, the CS2:C6Hi0Ob ratio
of 1 : 2 simply reflects the ratio of these areas.
Many of the above arguments for a micellar reaction must be considered
of questionable validity in view of the vast amount of contrary data showing
that the reaction is molecular in character.35 The latter viewpoint follows,
for example, from data indicating that the ratio of NaOH to C6Hi0OB in
40 K. Hess and C. Trogus, Cellulosechemie, 13, 84 (1932) ; W. Schramek and coworkers,
Papier-Fabr., 36, Tech.-wiss. Tl., 226 (1938); Z. pkysik. Chem., B50, 298 (1941); Mel-
Hand Textilber., 28, 383 (1947); T. Lieser and coworkers, Kolloid-Z.t 81, 240 (1937);
Ann., 528, 284 (1937); Ann., 548, 204, 212 (1941); K. Jung, Kolloid-Z., 108, 120
(1944); T. Bergek, Norsk Skogind., 2, 289 (1948); P. C. Scherer and R. W. Phillips,
Rayon and Synthetic Textiles, 30, No. 5, 45 (1949).
41 T. Lieser, Ann., 470, 104 (1929); 483, 132 (1930); K. Lauer and coworkers, Kol-
loid-Z., 110, 26 (1945); 119, 151 (1950).
fX. DERIVATIVE'S OF CELLULOSE
alkali cellulose I is 1 : 1 (and not 1 : 2) and that commercial alkali cellulose
(with the usual molar composition of 2NaOH:lC6HioO&) readily forms a
1 : 1 xanthate with excess CS2 and sufficient time. This 1 : 1 reaction must
penetrate the alkali cellulose I crystallites as, it is claimed, there is not
enough Na outside of the crystalline areas. Xanthation must, therefore,
occur also, at least to some extent, within the crystalline areas when less"
CS2 is used since there is nothing fundamental about the usual commercial
1:2 xanthate.
Further evidence of the molecular nature of the reaction and arguments
against a micellar reaction may be briefly summarized :30'32 -85'42
X-ray investigations now show that the alkali cellulose I diagram which
may be present at first in cellulose xanthate, is due to incomplete reaction
and that it disappears completely on standing, even at a low extent of
reaction with CS2.
Fractionation of a diethylchloracetamide derivative by other workers
shows uniform distribution of CSz groups and little variation in degree of
substitution; in a 1:1 xanthate, some glucose residues would have to be
even more highly substituted if the reaction occurs only on the surface.
The CS2:C6HioO6 ratio does not change with D.P.; this is difficult to
explain on the basis of a micellar surface reaction since the area of the sur-
face is not fixed but changes with D.P.
The ratios of the iisp/C values for a series of xanthates of different
D.P.s to the rjSp/C values of the celluloses regenerated therefrom and dis-
solved in cuprammonium solution are constant, which could be so only in a
molecular reaction.
Alkali cellulose treated with sodium sulfide does not produce a soluble
xanthate on reaction with CSz even though it has the same crystal struc-
ture as normal alkali cellulose.
The formation of higher xanthates in solution does not prove that a
fibrous xanthate cannot enter into a molecular reaction as claimed.
In the reaction of alkali cellulose with carbon dioxide, the cellulose
reacts completely and the alkali cellulose I crystallites disappear; they
therefore should not be a hindrance to reaction with CSz in normal xan-
thation.
The methylation results mentioned above do not prove the presence of
unaltered cellulose in the xanthate. They may be explained on the basis
that the CS2 groups are randomly distributed with some glucose units carry-
48 G. Centola, Ann. chim. applicata, 31, 525 (1941); H. Staudinger and coworkers,
Ber., 71B, 1995 (1938); /. prakt. Chem., 156, 261 (1940); O. Samuelson, Svensk Kern.
Tid., 58, 285 (1946).
982 CELLULOSE
ing more than one xanthate group and some none. Likewise, the dissolv-
ing of unxanthated cellulose in viscose does not mean that the cellulose
dissolves as such or prove that the original viscose contained unaltered
cellulose. It is due simply to the well-established redistribution of CS2
groups that occurs after xanthation.
Even though the reaction is nonhomogeneous, it can still be intramicellar
or molecular.
With regard to the position of the reactive OH group or groups, the
idea that the primary OH group in the 6- position is most capable of under-
going the xanthation reaction was rejected some time ago in favor of the
secondary OH group in the 2- position.41 More recent observations
appear to confirm the reactivity of the 2- position in that it reacts first,
but only up to a degree of substitution of about lCS2:2C6Hi0O6.
After this, the 3- position reacts. Here again, the evidence is not too con-
clusive. It seems more likely that xanthation does not occur uniformly
on any specific OH group but that all three OH groups react and that the
CS2 is randomly distributed.43 There is some evidence in this connection
that although the secondary hydroxyls can and do react initially, the final
equilibrium favors the primary hydroxyls.44
Most of the evidence developed to date has thus been contradictory and
variously interpreted, and it is still not clear exactly what happens in
xanthation. Further clarification of the reaction is desirable, including
more conclusive data as to the formula for alkali cellulose I and the relative
amounts of amorphous and crystalline areas in alkali cellulose. Until
then, probably only the trixanthate will be accepted universally as stoichio-
metric.
For the present, it would appear that the ' 'primary" xanthate reaction
in the viscose process involves chemical reaction, but the ratios of 1CS2>
2CeHi0O6 usually found in industrial practice and of 1CS2: ICeHioOs found
as the upper limit in the xanthation of fibrous alkali cellulose do not signify
stoichiometric compounds and do not mean that there is necessarily a
regular arrangement of the substituent groups along the cellulose chains.
For example, the 1 : 2 product most certainly does not have precisely every
second anhydroglucose unit substituted in, say, the 2- position. The real
significance of the 1 : 2 ratio in commercial practice lies only in the fact that
this is enough to insure solubility. (The same substitution is required
for xanthate as, for example, for carboxymethyl and carboxyethyl sub-
48 T. Lieser, Ann., 522, 58 (1936); Papier-Fabr., 36, Tech.-wiss. TL, 272 (1938);
Kolloid-Z.t 94, 96 (1941).
44 A. Matthes, Faserforsch. u. Textiltech., 4, 127 (1952).
IX. DERIVATIVES OF CELLULOSE 983
stitution. Even nonacid groups such as hydroxyethyl and ethyl give prod-
ucts soluble in 6% NaOH at the 0.5 substitution level.)
In xanthation, initially at least, even with excess CSa, the xanthate groups
are probably distributed at random according to some form of statistical
tendency and depending also upon the morphology of the fiber, with the
more accessible anhydroglucose units containing perhaps even more than
one group. The reaction also involves swelling and physical dispersion
of the cellulose, and the resulting compound is probably a mixture of cellu-
lose molecules of varying chain length, substituted to varying degrees.
As was shown in Section A of this Chapter IX, however, the initial
distribution of substituents is not the same as the final distribution.
Continuous further reaction and redistribution of CS2 groups occur after
initial xanthation so that, on sufficiently long standing or with excess CSj,
the alkali cellulose I crystallites disappear. If the xanthate is dissolved
before the reaction is complete, reaction continues in solution and eventu-
ally the product is probably molecular in character with the C£>2 groups
distributed more or less equally and with most of the crystallites dispersed.
Although uniformity of reaction will be approached after a sufficient period
of equilibration, the redistribution cannot go to completion with the for-
mation of a perfectly uniform product because of the gradual loss of xan-
thate groups which takes place during viscose ripening (see below) .
In addition to the main reaction between alkali cellulose and carbon di-
sulfide, side reactions also occur during xanthation. These are of two
types: (1) secondary reactions of the xanthate itself and (2) direct re-
action between CS2 and free NaOH. The secondary reactions of the
xanthate are brought about by its instability and the reversible nature of
the reaction with CS2. The primary reaction (see equation 1) should thus
be written as an equilibrium reaction which is forced to the right by in-
creasing the concentration of CS2. The following secondary reactions of the
xanthate also have been suggested :
+ 2 NaOH - > Na2CO2S + NaSH + Cellulose (4)
+ NaSH - » Na2CS3 + Cellulose (5)
The thiocarbonate may further react with NaOH forming carbonate and
hydrosulfide:
Na2C02S + NaOH - > Na2CO3 + NaSH (6)
The other side reaction, between CSa and NaOH, may be expressed in the
standard way:
3 CS* + 6 NaOH - » 2 Na*CS3 + Na2CO8 + 3 H2O (7)
984 CELLULOSE
Regardless of the exact nature of these side reactions,45 the secondary
products, which cannot be avoided industrially, consist of carbonate, tri-
thiocarbonate, and probably hydrosulfide and sulfide. The rate of for-
mation of these products is fairly high at the beginning of xanthation but
then gradually slows down. They are more stable than the primary xan-
thate and are responsible for the orange or carrot color of commercial xan-
thate and viscose. (Pure cellulose xanthate is practically colorless.) In
the formation of the xanthate, an equilibrium is undoubtedly reached be-
tween it and the secondary products mentioned, although, because of the
complicated nature of the mixture, it is difficult to determine satisfactorily
the predominating equilibrium. Under normal industrial conditions,
about 75% of the €82 used reacts with the cellulose and 25% appears as
by-products in the freshly prepared viscose. (This excludes a small (5-
10%) proportion of the CS2 added which either does not react or is lost
mechanically.) Thus, a fresh commercial viscose containing 7% cellulose
and 2.0% total sulfur (equivalent to 34% CS2 based on the cellulose)
normally shows the presence of about 1.5% xanthate sulfur.
Besides the conversion of OH groups to xanthate groups, colloidal changes
occur during xanthation which are characterized by an extraordinary
swelling (during which the form of the fiber is partially destroyed), by
increased solubility, and by a drop in viscosity. The latter is due to deg-
radation of the cellulose, that is, a reduction in chain length, the extent
of the degradation being dependent upon the time, temperature, and
amount of oxygen present.
The solution of cellulose xanthate in water or dilute sodium hydroxide
solution may be considered simply as a continuation of the swelling process
which started with the formation of alkali cellulose. From a colloidal
standpoint it is essentially a peptization of a hydrophilic colloid. How-
ever, the xanthation reaction apparently continues, probably intramicel-
larly in part, during and after solution, between unchanged alkali cellu-
lose and CS2 and/or by redistribution of xanthate groups.31-32'46 Disper-
sion of the xanthate is, therefore, a gradual process and continues for some
time after the dissolving step. On this basis it may be said that only when
solution is complete does true or final xanthate formation occur. It has
45 M. Ragg, Chem.-Ztg., 32, 630, 654, 677, 730 (1908); 34, 82 (1910); H. Leuchs,
Kunstseide, 7, 286 (1925) ; G. Kita and R. Tomihisa, Cellulose Ind. (Tokyo), 2, 26 (1926) ;
Cettulosechemie, 10, 134 (1929).
48 I. Sakurada and R. Inoue, J. Soc. Chem. Ind., Japan, 35, Suppl. binding, 127
(1932); O. Kratky and coworkers, Kolloid-Z., 98, 301 (1941); W. Schramek and E.
Zcbmisch, Kolloid-Beihefie, 48, 93 (1938).
DC, DERIVATIVES OF CELLULOSE 985
also been suggested that association takes place during solution and that
NaOH or H2O is combined chemically with conversion of the C=S group
to HO— C— SNa or HO— C— SH. X-ray data, however, do not indicate
that any new chemical compound is produced.37-47
The actual degree of dispersion of the xanthate after dissolving in dilute
sodium hydroxide solution, i.e., the nature and structure of the xanthate
solution, has been the subject of the same controversy as the primary
xanthate reaction itself (see above) and is still open to question. On the
basis that the reaction between alkali cellulose and CSz is micellar, the
xanthate solution may also be considered to be made up of micellar or poly-
molecular particles. This viewpoint is in line also with the analogous vis-
cosity behavior of viscose and micellar soap solutions, the idea that only a
low degree of xanthate substitution is possible or necessary for solution in
fibrous xanthates, the formation of water-soluble cellulose by dialyzing
the product obtained by xanthating cellulose swollen in quaternary am-
monium bases and the fact that only alkali cellulose I or cellulose hydrate
(depending upon the conditions of coagulation) have been found in the
solution by x-ray examination (even though the x-ray diagram of alkali
cellulose I disappears on standing) . It means that solution overcomes the
fiber bond but does not necessarily destroy the alkali cellulose I crystal-
lites 32,39-41,48
As in the case of the initial xanthate reaction, considerable evidence is
available, on the other hand, indicating that the above ideas are not valid
and that the solution is of a molecular nature, i.e., polymer analogous
reactions, redistribution of €82 groups, the fact that with excess CS2 the
xanthate reaction occurs intramicellarly, that on sufficiently long standing
the alkali cellulose I lines in viscose disappear from the x-ray diagram, that
dispersion cannot occur without previously overcoming the cohesive forces
between the cellulose chains, etc.32'35'42-49
47 B. Rassow and W. Aehnelt, Cellulosechemie, 10, 169 (1929). See also P. Herrent
and G. Jnoff, /. Polymer Sci.t 3, 834 (1948).
48 W. Schramek and coworkers, Papier- Fabr., 36, Tech.-wiss. Tl., 226 (1938) ; Kolloid-
Z., 94, 92 (1941); Z. physik. Chem., B50, 298 (1941); Melliand Textilber., 28, 383
(1947); T. Lieser and coworkers, Cellulosechemie, 18, 121 (1940); Z. physik. Chem.,
B74, 708 (1941); Ann., 548, 195, 212 (1941); Chem.-Ztg., 67, 197 (1943); Kolloid-Z.,
94, 96 (1941); 98, 142 (1942); 108, 125 (1944); P. C. Scherer, Rayon Textile Monthly,
27, 74, 409 (1946); K. Lauer and coworkers, Kolloid-Z., 112, 112 (1949); R. Vuori,
Dissertation, Helsinki (1947).
49 S. M. Lipatov and N. A. Krotova, Melliand Textilber., 15, 553 (1934) ; H. Staud-
inger and F. Zapf, J. prakt. Chem., 156,261 (1940); G. Centola, Boll. sci. facolta chim.
ind., Bologna, 1941, 7-12; O. Samuelson, Svensk Papperstidn., 48, 517 (1945); F. Gart-
ner and O. Samuelson, Svensk Papperstidn., 53, 635 (1950).
986 CELLULOSE
Other suggestions concerning the nature of the solution include the ideas
that the particles present are aggregates of molecular chains, lattice-like
particles and characteristic of so-called ' 'fringe micellae" and that they are
"cellulose-chain mixed bodies."60
Summarizing the various viewpoints, it appears that dispersion of the
xanthate and the structure of the solution vary according to circumstances.
Commercial viscose, in which the cellulose concentration is high, is probably
a mixture of some micellar (polymolecular) particles and molecular ag-
gregates, the solution being the result of an equilibrium following redis-
tribution of CS2 groups after xanthation and the dissolving of the un-
xanthated portions of the alkali cellulose, and aggregation of any individual
molecules that may be formed. The cellulose is, or at least need be, only
partly xanthated in such solutions since solubility is achieved by dispersion
of the unxanthated portions of the cellulose chains by the highly hydro-
philic xanthate groups. It does not seem likely under these conditions that
dispersion to a wholly molecular condition occurs. Although many indi-
vidual molecules may be present in a fresh solution, association probably
occurs almost at once. In xanthate solutions in which the cellulose con-
centrations are sufficiently low and/or in which the cellulose is more highly
substituted (CSo : C6Hi0O6 ratios of 1:1 or more and certainly for the tri-
xanthate) the dispersion would be expected to be much more complete
than in commercial viscoses and under these conditions the solutions are
more likely to be, and probably are, almost entirely molecular in nature.
In any case, the dispersion or solubility is influenced by practically every
variable in the process, including type of cellulose, steeping, pressing, shred-
ding, degree of xanthate substitution, mixing conditions, and viscose com-
position.
In addition to its effect on the dispersion of the cellulose, the composition
of the viscose in cellulose and free NaOH also affects the rate of ripening,
the viscosity, the regenerating characteristics (spinning, casting, etc.), and
the properties of the regenerated product. Other factors being constant,
rate of ripening, viscosity, and reciprocal filterability increase with increase
in cellulose concentration or with decrease in free NaOH concentration.
Spinning and film-casting properties depend on the ratio of cellulose to
caustic as well as upon the absolute concentrations, and a proper balance
must be maintained between these constituents and the composition of the
coagulating medium used. In comparison, for example, with cellulose
* H. L. Bredee, Kolloid-Z.t 94, 81 (1941) ; J. J. Stdckly, Kolloid-Z., 105, 190 (1943) ;
M. Takei, Kolloid-Z.t 106, 30 (1944); W. P. Conner and P. I. Donnelly, Ind. Eng.
Chem., 43, 1136(1951).
IX. DERIVATIVES OF CELLULOSE 987
acetate, the optimum concentration of cellulose in viscose is relatively low,
being in the range of 6 to 10%. Attempts to employ more highly concen-
trated solutions have been unsuccessful.
3. Characteristics and Reactions of Viscose
Viscose is a sirupy, oily, orange- or carrot-colored solution, with a char-
acteristic odor which is due, in part, to hydrogen sulfide and polysulfides.
It possesses many of the characteristics of a typical polyelectrolyte, such
as high and anomalous viscosity, relative stability toward electrolytes,
syneresis, weak electric charge, flow birefringence, and relatively great
conductivity, and it undergoes various chemical and colloidal reactions.
(a) VISCOSE VISCOSITY
The viscosity of viscose depends upon several factors, the more important
of which are the D.P. of the cellulose (determined by the type of cellulose,
alkali cellulose aging conditions, etc.), the cellulose content, the degree of
dispersion (alkali content and other factors influencing dispersion), tem-
perature, and degree of ripening. The actual viscosity used in practice
varies with different producers and with the nature of the product to be
made. In addition to the requirements of the mechanics involved in
transferring the solution through pipe lines and in filtration, evacuation,
spinning, casting, and similar operations, there is an optimum viscosity
For every combination of production factors (e.g., viscose composition,
coagulating bath, and speed of extrusion). From the standpoint of the
strength and durability of the regenerated product, the use of a relatively
iiigh viscosity (high and uniform cellulose chain length) would seem de-
sirable, but to date both mechanical and chemical difficulties have made it
impossible to achieve optimum conditions in this respect. As a result,
commercial viscoses range from 30 to 60 poises in viscosity, the weight-
iverage D.P. of the regenerated cellulose varying from about 250 to 600.51
The dependence of viscosity on the D.P. of the cellulose is well known
md need not be discussed here. The relationship of the viscosity to the
concentration of dilute viscose solutions is probably analogous to that of
Dther cellulosic solutions (see Chapter X, Sections E and F). Whether
my of the equations so far proposed relating viscosity, concentration, and
D.P. give more than approximate results with industrial viscose solu-
:ions containing 6-10% cellulose remains to be more clearly established.
In this connection it should be mentioned that the intrinsic viscosity [77] or
61 H. Schwartz and H. A. Wannow, Kolloid-Z.t 97, 193 (1941); 99, 190 (1942).
988 CELLULOSE
Km constant of viscose apparently varies depending upon the degree of
xanthate substitution, being higher with lower substitution.62
Viscosity is also dependent upon degree of dispersion, and hence the
alkali content, of the viscose.63 Although radical reduction in the viscosity
of a viscose, after preparation, is not possible except by degrading the cellu-
lose (indicating that the cellulose in viscose is partially in the form of a
molecular dispersion), slight but important changes in viscosity can be pro-
duced by altering the NaOH concentration. Increasing the NaOH con-
tent from 4 to 6%, for example, decreases the viscosity about 35%, The
degree of dispersion, and hence stability, of viscose is greatest at a concen-
tration of about 9% free alkali and at this concentration the viscosity is at a
minimum (i.e., the viscosity increases both above and below about 9%
NaOH). (For a further discussion of the effect of NaOH concentration
on the viscosity of viscose as well as of the behavior of cellulose xanthate
as a polyelectrolyte, see Chapter X-C. Since cellulose itself shows its
maximum solubility in caustic at a NaOH concentration of about 9% the
minimum viscosity at this point undoubtedly reflects the solubilizing in-
fluence of the NaOH on unsubstituted portions of the chains and on por-
tions of the sample that were incompletely xanthated.)
Although the characteristic viscosity change during viscose ripening
(see below) does not involve the D.P. of the cellulose, viscose viscosity
may be affected by the presence of air.44'64 This is due perhaps to oxida-
tion of the by-products but there is also some evidence that it is possible
actually to reduce the D.P. of the cellulose somewhat by atmospheric oxy-
gen, if contact is intimate enough, with some reduction in viscosity.
Like other cellulose and high-polymer solutions, and as might be ex-
pected of a polyelectrolyte, viscose exhibits pronounced structural vis-
cosity.65 That is, in addition to the factors already discussed, its viscosity
is related to its gel structure and is a function of rate of flow, decreasing
markedly as the velocity gradient or rate of shear increases. This effect
62 H. Staudinger and cwvorkers, Ber., 7lB, 1995 (1938); /. prakt. Chem., 156, 261
(1940); W. Philippoff and H. E. Kruger, Kolloid-Z., 88, 215 (1939); G. Jayme and J.
Wellm, Kolloid-Z., 107, 163 (1944) ; M. Takei, Kolloid-Z., 106, 30 (1944).
68 H. Staudinger and F. Zapf, /. prakt. Chem., 156, 261 (1940) ; E. Heuser and H. Y.
Charbonnier, Ind. Eng. Chem., 33, 402 (1941) ; T. Bergek, Norsk Skogind., 2, 289 (1948) ;
C. W. Tait and coworkers, /. Polymer Sci., 7, 261 (1951).
64 A. Lottermoser and F. Schwarz, Z. angew. Chem., 43, 18 (1930); Kolloid-Beihefte,
42, 419 (1935) ; O. Samuelson, Svensk Papperstidn., 47, 597 (1944).
w W. Philippoff and coworkers, Cellulosechemie, 17, 57 (1936); Kolloid-Z. , 88, 215
(1939); H. Staudinger and F. Zapf, /. prakt. Chem., 156, 261 (1940); H. L. Bred6e and
J. de Booys, Kolloid-Z.t 96, 24 (1941) ; H. Erbring, Kolloid-Z.t 108, 152 (1944) ; A. Lude,
Rec. trav. chim., 68, 1030 (1949).
IX. DERIVATIVES OF CELLULOSE
is negligible in the measurement of viscosity as carried out industrially since
the shearing stresses involved here are relatively low and under these con-
ditions the solution is Newtonian. However, structural viscosity is a
factor in the passage of viscose through pipe lines and pumps and particu-
larly in the spinning of rayon where the solution passes through spinneret
holes at pressures usually exceeding 25 Ib./sq. in. Under the latter condi-
tions the viscosity may be only Vio to about l/?& of that determined by, say,
a ball-fall test. Due to deviations from laminar flow, the viscosity reduc-
tion in spinning is probably not actually as great as might be expected
from the pressures involved. In any case, if it were not for this great
reduction in viscosity, the spinning pressures required would be many
times higher than those normally used. Also, because of the anomalous
flow characteristics of viscose, different viscosities caused by alkali cellu-
lose aging tend to be equalized in spinning.
In this connection, the flow curve of viscose (shearing force versus veloc-
ity gradient) is said to show two points of inflection, suggesting the presence
in viscose of two types of particles, primary molecules or micelles and second-
ary aggregates. This, however, does not appear to be in line with the ob-
servation that only one optical relaxation time has been observed in a well-
dissolved solution. It should also be noted that the drop in viscosity
which occurs at high shearing stresses (high flows, agitation, or other me-
chanical action) and which is due to an alteration of the gel structure is
reversible. The original viscosity is restored on further standing, and this
process can be repeated (suggesting further that viscose may also be thixo-
tropic).66-57
(b) VISCOSE RIPENING
One of the most important characteristics of viscose is its instability.
From the previous discussion it will be clear that cellulose xanthate is un-
stable, beginning to decompose immediately after formation, and that
viscose is a complicated mixture of cellulose xanthate, NaOH, CS2, and
compounds formed by the interaction of these materials (discussed above).
During the ripening of viscose, spontaneous decomposition of the xanthate
continues and all of these materials undergo further reaction. These
changes are both chemical and colloidal in nature.
66 T. Bergek and T. Ouchterlony, Svensk Papperstidn., 49, 470 (1946); but see also
P. Herrent and coworkers, Research (London), 2, 486 (1949) and A. Lude, Rec. trav.
chim., 68, 1030 (1949).
57 R. Signer and W. Meyer, Helv. Chim. Acta, 28, 328 (1945); V. E. Gonsalves, Proc.
Intern. Congr. Rheology, 1948, 2, 239 (1949); W. P. Conner and P. I. Donnelly, Ind.
Eng. Chem.. 43, 1136 (1951).
990 CELLULOSE
(1) Chemical Changes
The most important chemical change is the decomposition of the cellulose
xanthate itself which results in the gradual splitting off of CS2 and regenera-
tion of the cellulose. As might be expected in the case of a substance
made up of a strong base and a weak acid, the reaction is predominantly a
hydrolysis reaction, 27'44-47'58 with the formation of free cellulose xanthic
acid and NaOH until an equilibrium is reached according to :
[Sodium cellulose xanthate] fH2Q] __ „ ,~v
[Xanthic acid] [NaOH] ( '
Since the free xanthic acid is also unstable, liberating C$2, the reaction
is forced to the right, and more and more xanthate is decomposed, until
finally the regeneration of the cellulose is complete. The reactions may be
represented by the following equations :
RoeiiOCSSNa -f H2O > Roe,,OCSSH + NaOH (9)
RceiiOCSSH -f H2O > HOCSSH + Cellulose (10)
HOCSSH > CS2 -f H2O (11)
The same end products result by saponification, which also occurs to a
slight extent, being more noticeable as the caustic content of the viscose
increases :
RcetiOCSSNa + H2O » HOCSSNa -f Cellulose (12)
HOCSSNa > CS2 + NaOH (13)
This decomposition of the xanthate during ripening, as represented by
the changes in xanthate sulfur content, is shown graphically in Figure 74A.
In practice, the hydrolysis is not allowed to proceed to completion. Thus,
a 7% cellulose viscose for use in rayon, containing initially approximately
1.5% xanthate sulfur (2.0% total sulfur), is normally ripened until the
xanthate sulfur content reaches 0.9- to 1.1%. This represents a change in
the average degree of~esterification from about IC&^CeHuOs to 1CS2:
SCeHioOs. For many purposes, the viscose is used even before the xan-
thate sulfur has reached this range.
The redistribution of CS$ groups which occurs during ripening has al-
ready been mentioned and need not be further discussed. Besides this
and the decomposition of the xanthate, other chemical changes occur dur-
58 B. Rassow and M. Wadewitz, /. prakt. Chem., 106, 266 (1923); R. Bernhardt,
Kunstseide, 7, 193 (1925) ; 8, 257, 314 (1926) ; J. Frenkel, Cellulosechemie, 9, 25 (1928) ;
W. Klauditz, Papier-Fabr.t 37, Tech.-wiss. XL, 251 (1939); W. Vieweg, Papier- Fabr.,
37, Tech.-wiss. Tl., 269 (1939).
IX. DERIVATIVES OF CELLULOSE
991
ing ripening, involving the secondary products. The formation of sodium
trithiocarbonate and carbonate continues by reaction between NaOH and
the C£>2 liberated from the free xanthic acid. Other by-products may in-
clude sodium sulfide, polysulfide, and thiosulfate. Whether sulfide orig-
inates as an intermediate in the formation of Na2CSs or as the result
VISCOSE RIPENING TIME
HOURS AT 20° C.
50 100 150 200 250
VISCOSE RIPENING TIME
HOURS AT 20° C.
Fig. 74. Changes during viscose ripening:16 (A) in salt index and xanthate sulfur con-
tent; (B) in viscosity. Courtesy of Research Division, Rayonier Incorporated
of the decomposition of the trithiocarbonate is not certain but its presence
in viscOvSe is reasonably well established. These side reactions have been
variously represented, in addition to equation 7 as follows:27-45-47'59
5 CS2 + 12 XaOH —
CS2 4- Xa2S —
Xa2CSa 4- 3 H2O —
NaaCSs 4- 2 H2O —
H2C& —
NasCSs 4- 3 XaOH --
CS2 4- 2 NaHS -
CS2 4- H2O
Na2S + 2 Xa2CO3 -f 3 Na2CS3
Na2CSs
Na2CO8 4- 3 H2S \
H2CS«4- 2 XaOH [
H2S + CS2 J
» 3 NaHS 4- Xa2CO8
> Xa2CSa 4- H2S
4- COS
6 H2O
(14)
(15)
(16)
(17)
(18)
(19)
The chemical changes may then be summarized by saying : In xanthation,
sodium cellulose xanthate and by-product salts are formed side by side,
59 B. Rassow and K. Schwarze, Papier- Fabr., 28, Tech.-wiss. XL, 746 (1930); H.
Lotze, Kunstseide, 15, 194 (1933); C. L. Moore, Silk and Rayon, 8, 505, 563 (1934);
O. Samuelson, Cellulosa och Papper 1908-1948 S.P.C.I., pp. 295-325; R. S. Neumann
and coworkers, Cettulosechemie, 17, 16 (1936). For spectrochemical changes during
ripening see K. Atsuki and T. Takata, J. Soc. Chem. Ind.t Japan. 43, 402B (1940);
B. Rassow and W. Aehnelt, Cdluhsechemie, 10, 169 (1929); P. Herrent and G. Jnoff,
/. Polymer Set. , 3, 487 ( 1948) .
992 CELLULOSE
with free €82 and NaOH remaining. After solution, new NaOH is added.
Even though xanthate formation takes place more rapidly than the other re-
actions the xanthate, being very unstable, is also decomposed faster than the
secondary products. Xanthate sulfur content and free NaOH, therefore,
decrease markedly, while the salt content increases. Several equilibriums
are established until all the CS2 has been converted and then a "salting
out" effect sets in, due to the various salts present, which results eventually
in precipitation of the cellulose as cellulose hydrate.60
(2) Colloidal Changes
As a hydrophilic colloid, viscose also undergoes significant changes in its
colloidal properties during ripening. It has even been suggested that the
colloidal processes predominate at first and induce the chemical changes,61
although this seems doubtful.
One of the most peculiar colloidal changes during viscose ripening is that
involving the viscosity of the solution, which at first drops rather rapidly,
then passes through a minimum point, and finally rises slowly until coagu-
lation is complete. This course of the viscosity change during the ripening
of a representative commercial viscose is shown in Figure 74B, but it should
be understood that the shape of the curve may differ somewhat from that
shown depending upon viscose composition, temperature, CS2 concentra-
tion, dissolving time, and other factors. The change in viscosity is not
due to any change in the degree of polymerization of the cellulose. Al-
though some degradation of the cellulose occurs in every other step of the
process, little, if any, occurs during ripening.62 The initial drop in vis-
cosity has been explained on the basis that solution of the xanthate is a
slow process and requires time or that final xanthate formation does not
take place until solution occurs. Even though the xanthate apparently
60 There is some evidence indicating that conversion of alkali cellulose I to cellulose
hydrate occurs during ripening by way of alkali cellulose IV; see W. Schramek, Kolloid-
Beiheftc, 42, 331 (1935). It should be added also that viscose may absorb atmospheric
oxygen during ripening with the formation of the disulfide, RcellOCSS — SSCORcell,
and some thiosulfate; see S. N. Danilov and coworkers, /. Gen. Chem. ( U. S. S. R.), 19,
826 (1949) ; also, References 44 and 54.
81 K. Atsuki, /. Faculty Eng.t Tokyo Imp. Univ.. 17, 135 (1927); Cellulosechcmie, 9,
106 (1928).
61 E. Heuser and M. Schuster, Cellulosechemie, 7, 17 (1926); S. Rogowin and M
Schlachover, Cellulosechemie, 14, 17, 40 (1933); A. Lottennoser and F. Wultsch, Kol
loid-Z.t 83, 180 (1938); H. Staudinger and coworkers, Papier- Fabr.t 36, 557 (1938);
Ber.. 71B, 1995 (1938) ; J. prakt. Chem., 156, 261 (1940) ; J. Compton, Ind. Eng. Chem.,
31,1250(1939).
IX. DERIVATIVES OF CELLULOSE 993
dissolves during the mixing operation, the dispersion is "coarse" and in-
complete. Further solution, including penetration of the alkali cellulose
crystallites, continues for 24 to 48 hrs., due either to continued progress
of the C£>2 reaction or redistribution of the CS2 groups, or both32'39 (which
may explain the hump in the first part of the salt index curve in Figure
74A), until the disintegration is complete, the viscosity decreasing as the
dispersion approaches an optimum. In some cases the viscosity may actu-
ally rise slightly for a short time before the drop just described. This rise
has been attributed to a further increase in the degree of hydration of the
cellulose, and the initial decrease in viscosity may conceivably be simply
the result of dehydration of the dispersed particles, that is, a decrease in the
amount of water held by the internal dispersed phase due to osmosis.61-68
The subsequent increase in viscosity is due to the opposite effect, that is,
a decrease in solubility. As hydrolysis proceeds, causing the ratio of
combined sulfur to cellulose to become constantly smaller, the fraction of
the chain that becomes unsubstituted and hence insoluble becomes con-
stantly greater, and the degree of dispersion of the cellulose decreases due
to association and aggregation, until visible coagulation sets in. The
increase in viscosity which takes place is thus the manifestation of the
decreasing solubility and increasing degree of cross-linking of the cellulose
molecules, and is not the result of change of molecular weight. Additional
evidence of the nature of these changes is shown by the formation of gel
structure, changes in flow birefringence, and other characteristics during
ripening,67'64 and by the fact that the original viscosity may be restored
(in fact, the whole ripening process may be reversed) by adding CS2 to
viscose at any stage of the process. This does not mean that no chemical
reactions are involved since it is clear that the changes in degree of substitu-
tion which occur are responsible, at least in part, for the viscosity changes.
Perhaps the most significant colloidal change in viscose during ripening
is the change in its coagulation properties. As initially prepared, viscose is
relatively stable and difficult to coagulate. Paralleling the chemical
changes mentioned above, however, the solution coagulates spontaneously
by virtue of the constantly decreasing solubility of the dispersed phase and
increasing salt formation. Since the — CSSNa group is the solubilizing
factor in the xanthate molecule (the cellulose itself is insoluble), as the
68 C. J. J. Fox, J. Soc. Chem. Ind., 49, 83T (1930); see also J. J. Stockly, Kolloid-Z.
105, 190 (1943) ; T. Bergek and T. Ouchterlony, Svensk Papperstidn., 49, 470 (1946).
64 E. Berl and A. Lange, Cellulosechemie, 7, 145 (1926) ; J. J. Stockly, Kolloid-Z., 105,
190 (1943); G. Centola, Boll. sci. facolti chim. ind., Bologna, 1941, 7-12; M. Takei,
Kolloid-Z., 106, 30 (1944).
994 CELLULOSE
number of — CSSNa groups per anhydroglucose unit decreases, the solu-
bility of the material decreases. If ripening is allowed to proceed uninter-
ruptedly and all the — CSSNa groups are split off, it becomes completely
insoluble, and the solution is converted completely and spontaneously to a
gel of hydrated cellulose. The rate of this change and the actual time re-
quired for gelation will vary considerably, depending upon such factors
as temperature, type and D.P. of the original cellulose, viscose composition,
oxygen and CSz content; i.e., the higher the D.P., the oxygen and cellulose
contents, and temperature are, the faster is the ripening; the higher the
NaOH and CS2 contents are, the slower is the ripening.
As spontaneous coagulation proceeds during ripening, the hydrophilic
character of the solution decreases and it becomes more and more hydro-
phobic in nature. This change is manifested by increasing instability
and ease of coagulation with various agents such as acetic acid, alcohols,
and inorganic salts. Being the salt of a stronger acid, the xanthate is not
(readily) decomposed by monocarboxylic acids of the fatty series such as
formic, acetic, and lactic acids, or by CC>2, SO2, and other weak acids.
(These acids do react with the free NaOH and decompose the by-product
sodium salts with the formation of the sodium salts of the acid used and
liberation of €82, H2S, and CCV) Moreover, these acids do not usually
coagulate freshly prepared viscose. As ripening proceeds, however, a
point is reached where addition of an acid such as acetic also causes coagu-
lation, and, as ripening continues, lesser and lesser amounts of acid are re-
quired for coagulation.
Methyl and ethyl alcohols, and alkali and ammonium salts, also coagulate
viscose, precipitating the xanthate unchanged. This behavior is entirely
analogous to that of other polyelectrolytes such as carboxymethyl cellulose
and proteins. With inorganic salts, coagulation is based on a "salting-
out" effect (dehydration of the dispersed phase) which follows an initial
neutralization of the negatively charged xanthate particles by the first
addition of the electrolyte. Whereas large amounts of alcohol and salts
are required to coagulate the viscose when freshly prepared, lesser and lesser
amounts are necessary as ripening progresses and the hydrophobic character
of the solution increases. This change is shown by the salt index curve
in Figure 74A. As initially prepared, about an 8.0% NaCl solution is
required for coagulation of a drop of this particular viscose. After 45
hrs., a 4.0% salt solution suffices, and after about 75 hrs. only a 2.0%
solution is required. When 10% NH^Cl is added to the same viscose
solution, 27.2 cc., 10 cc., and 6.4 cc. are needed for coagulation after the
respective ripening times. Tests based on these coagulation properties
are described below.
IX. DERIVATIVES OF CELLULOSE 995
Among other characteristics of viscose, to which reference has
been made, is flow birefringence. Even at low rates of shear, viscose
shows marked optical double refraction, and studies of the magnitude of the
birefringence, of relaxation times, and related properties, permit certain
conclusions regarding the structure of the solution and the changes in the
structure during viscose ripening and spinning.67 As might be expected,
flow birefringence and optical relaxation time increase during ripening.'
It may be added that it has been found in certain viscoses that the curves
for relative viscosity and the precipitation potential of silver xanthate as
functions of ripening time both pass through a minimum at the same
degree of ripening. This suggests that there may be an "optimum struc-
ture" for obtaining optimum mechanical properties of the end products at
this particular degree of ripening but just what this structure is has not yet
been established.66
Other changes during ripening include syneresis, increased turbidity,
increased particle size, decreased dissociation, decreased conductivity, and
changes in light absorption, Tyndall effect, Brownian movement, and vol-
(5) Control of Ripening
Since the spinning and casting properties of viscose as well as the quality
and characteristics (tenacity, elongation, luster of filaments, clarity of
films, dyeing, softness, etc.) of the end product (rayon, cellophane) are
determined in part by the degree of ripening, the normal instability of
viscose is a major source of concern in its industrial utilization. For this
reason, precautions are taken to maintain constant conditions (constant
salt index) and to inhibit or retard ripening. This is usually accomplished
mechanically by such measures as the use of relatively low viscose tempera-
tures (15~20°C.) during ripening, refrigeration and stepwise reduction in
size of supply lines, and blending of batches (including the recirculation
of viscose around the spinning machines and feed tanks). Chemical meth-
66 P. Herrent and coworkers, Research (London), 2, 486 (1949). See also W. Schra-
mek, Melliand Textilber., 28, 420 (1947) for other changes in structure during ripening.
66 T. Mukoyama, Kolloid-Z.t 41, 62 (1927); 42, 79, 180, 353 (1927); S. M. Lipatov,
Kolloid-Z.t 49, 441 (1929); R. Bernhardt, Melliand Textilber., 7, 55, 318 (1926); T.
Sugita, Cellulose Ind. (Tokyo), 8, 3, 166 (1932); R. O. Herzog, Kolloid-Z., 35, 193
(1924); Schwedler, Dissertation, Leipzig, 1927; B. Rassow and W. Aehnelt, Cellulose-
chemie, 10, 169 (1929); K. Atsuki and T. Takata, J. Soc. Chem. Ind., Japan, 43, 402B
(1940); G. Centola, Boll. sci. facolta chirn. ind., Bologna, 1941, 7-12; P. C. Scherer,
Rayon Textile Monthly, 26, 69, 1 17 (1945).
996 CELLULOSE
ods may also be used to retard ripening. It has been mentioned that the
CSg and free alkali concentration influence ripening; the higher the content
of NaOH (up to about 9%) and the higher the CS2 are, the more stable is
the viscose and the slower is the ripening. Other materials which have been
proposed for addition to viscose to retard ripening include sodium sulfite,
sodium cyanide, arsenites, certain urea and other arnino derivatives, alkyl
xanthates, phenols, calcium acetate, gallic acid, acrylonitrile, pyridine, and
others.67'68 With the exception of sodium sulfite, however, none of these
addition agents appears to be used in practice.
Consideration has been given also to methods for accelerating ripening in
order to reach the desired degree of esterification more rapidly and thus to
reduce or even eliminate the ripening step. It is obvious that this can be
done in several ways, such as by raising the ripening temperature, by add-
ing electrolytes, and by using low concentrations of NaOH and CS2.69
The addition to viscose of hemicellulose, poly alcohols like glycerol, ether,
hydrogen peroxide, sulfide, polysulfides, air, and other materials44'70 also
accelerates ripening. Although the effect of some of these agents may be
colloidal in nature, the action of most of the above-mentioned ripening ac-
celerators and inhibitors is due to an actual change in rate of chemical de-
composition of the xanthate.68
67 C. A. Ernst, U. S. Patent 863,793 (Aug 20, 1907) ; Chem. Abstracts, 2, 478 (1908) ;
R. W. Maxwell, U. S. Patent 2,011,227 (Aug. 13, 1935); Chem. Abstracts, 29, 6758
(1935); R. Linkmeyer and H. Hoyermann, German Patent 312,392 (Nov. 17, 1917);
Chem. Zentr., 90, IV, 1017 (1919) ; Soc. Lorch and Hamrn, French Patent 728,682 (Dec.
21, 1931) ; Chem. Abstracts, 26, 6137 (1932) ; E. B. Castillo, Analesfis. y quim. (Madrid),
43,60(1947).
68 For the action of acrylonitrile see J. P. Hollihan and S. A. Moss, Jr., Ind. Eng.
Chem., 39, 222 (1947); J. H. MacGregor and C. Pugh, J. Soc. Dyers Colovrists, 64, 71
(1948).
69 C. A. Ernst, U. S. Patents 896,715 (Aug. 25, 1908); 863,793 (Aug. 20, 1907);
Chem. Abstracts, 2, 478 (1908); Vereinigte Kunstseide-Fabriken Akt.-Ges., Brit. Patent
17,502 (Aug. 8, 1902); Soci£te Fran^aise de la Viscose, Brit. Patent 8179 (1907); Chem.
Abstracts, 2, 1762 (1908) ; Soc. Anon. Soie de St. Chamond, Brit. Patents 1436 (Aug. 10,
1910); 24,291 (Dec. 18, 1914); Chem. Abstracts, 10, 1600 (1916); Deutsche Zellstoff-
Textilwerke, German Patents 339,050 (Oct. 12, 1918) ; Chem. Zentr., 92, IV, 669 (1921) ;
342,641 (Oct. 30, 1919); Chem. Zentr., 93, II, 48 (1922); W. Mendel, German Patent
566,691 (Aug. 30, 1930) ; Chem. Abstracts, 27, 2578 (1933) ; A. J. Burette, French Patent
430,221 (May 22, 1911); J. A. Calhoun, Jr., and F. C. Wedler, U. S. Patent 2,558,037
(June 26, 1951) ; Chem. Abstracts, 45, 7791 (1951).
70 R. L. Cairncross and G. H. Goodell, U. S. Patent 1,814,543 (July 14, 1931); T.
Mukoyama, Kolloid-Z., 42, 180 (1927) ; J. Sauvy, Ind. textile, 63, 143 (1946) ; O. Samuel-
son, Cellulosa och Papper 1908-1948, S.P.C.I., pp. 295-325.
IX. DERIVATIVES OF CELLULOSE 997
(c) PURIFICATION OF VISCOSE
Various suggestions have been made for the purification of viscose and the
preparation of cellulose xanthate in a stable, dry form free from the usual
by-product salts. Viscose can be coagulated with alcohol, salts, or a weak
acid, or a combination of these agents, and the by-products can be removed
by washing the precipitated xanthate with fresh precipitating solution..
The washed xanthate may then be redissolved in caustic solution or dried
if desired. Such procedures are of interest in studying the composition
of the xanthate and viscose, since purified, dried xanthate is reasonably
stable. Although a number of purification methods have been described
and patented,27-30-71 they are expensive, and they are superfluous so far as
the major technical applications of viscose are concerned. They are, there-
fore, not used.
(d) ADDITIONS TO VISCOSE
Besides the materials mentioned for controlling ripening, literally
hundreds of others have been suggested or patented for addition to viscose
for various purposes. These include solid and liquid, inorganic and organic
compounds of almost every conceivable type, and they are added for al-
most every conceivable purpose, including improving or modifying the
luster, dyeing, color, strength, elongation, softness, and other character-
istics of the end product, and the clarity, color, surface tension, viscosity,
ripening rate, and spinning characteristics of the viscose. In actual prac-
tice only a limited number of these materials are in common use and they
are added mainly for securing low luster (such as titanium oxide and mineral
71 L. Lilienfcld, U. S. Patent 980,648 (Jan. 3, 1911); Chem. Abstracts, 5, 1188 (1911);
Brit. Patent 14,339 (June 15, 1914) ; Chem. Abstracts, 9, 3359 (1915) ; A. Bernstein, U. S.
Patent 1,121,605 (Dec. 22, 1914) ; Chem. Abstracts, 9, 377 (1915) ; G. A. Richter and P.
C. Scherer, U. S. Patent 1,880,041 (Sept. 27, 1932); Chem. Abstracts, 27, 414 (1933);
H. B. Dykstra, U. S. Patent 2,072,738 (Mar. 2, 1937) ; Chem. Abstracts, 31, 3275 (1937);
Vereinigte Kunstseide-Fabriken Akt.-Ges., Brit. Patent 8742 (1908); H. Lyncke, Brit.
Patent 8023 (1908); Chem. Abstracts, 3, 714 (1909); Viscose Syndicate Ltd., German
Patent 133,144 (Mar. 31, 1901); Chem. Zentr., 73, II, 411 (1902); Soci6t6 Fran^aise de
la Viscose, German Patent 187,369 (Aug. 13, 1904); Chem. Abstracts, 2, 730 (1908);
Continentale Viscose Compagnie, German Patent 209,161 (Oct. 20, 1903); Chem.
Abstracts, 3, 2223 (1909); J. P. Bemberg Akt.-Ges., German Patent 197,086 (Mar. 29,
1907); Chem. Abstracts, 2, 2301 (1908); F. Becker, German Patent 234,861 (Aug. 16,
1910); Chem. Abstracts, 5, 3157 (1911); Deutsche Gasgliihlicht-Auer-Ges., German
Patent 408,822 (Apr. 29, 1922); Chem. Zentr., 96, I, 1471 (1925); R. Linkmeyer and
H. Hoyermann, German Patent 312,392 (Nov. 17, 1917); Chem. Zentr., 90, IV, 1017
(1919) ; F. B. Cramer, U. S. Patent 2,369,718 (Feb. 20, 1945) ; Chem. Abstracts, 39, 3668
(1945) ; K. Atsuki and T. Takata, /. Soc. Chem. Ind., Japan, 43, 394B (1940).
998 CELLULOSE
oil), for dispersing delustering agents, for controlling milkiness and spin-
neret incrustations, and for sponge and bottle cap (film) manufacture.
Very few of the other suggested additions to viscose find any industrial
application.72
(e) ANALYSIS OF VISCOSE
The analysis of viscose usually includes the determination of the amount
of cellulose, total alkali, total sulfur, viscosity, filterability, xanthate sulfur
(degree of ripeness or esterification by chemical tests), and degree of ripeness
by coagulation tests.
The cellulose content of viscose may be determined readily by regenera-
tion in the form of a film with salt solution or a mixture of salt and mineral
acid. Total alkali is obtained by titration with sulfuric acid. Analysis
for total sulfur is made by oxidizing the viscose with hypochlorite, hypo-
bromite, or a mixture of hydrogen peroxide and nitric and perchloric acids,
and estimating it as barium sulfate. Total sulfur may also be determined
volumetrically by treating the viscose with sodium zincate and titrating
the resulting zinc sulfide with iodine.73
Determination of apparent viscosity is carried out industrially by either
the falling-ball or flow method. Although the viscometers used are of the
standard types, specifications regarding size of balls, tube and capillary
diameters, and other dimensions vary considerably throughout the industry.
It should be understood that in view of the anomalous viscosity of viscose,
72 A few of the more recent patents covering additions to viscose are: J. S. Creadick,
U. S. Patent 2,307,760 (Jan. 12, 1943) ; Chem. Abstracts, 37, 3605 (1943) ; L. Ubbelohde,
U. S. Patent 2,322,981 (June 29, 1943) ; Chem. Abstracts, 38, 253 (1944) ; P. H. Schlosscr,
U. S. Patents 2,331,935-6 (Oct. 19, 1943); 2,362,217 (Nov. 7, 1944); Chem. Abstracts,
39, 2650 (1945); 2,373,712 (Apr. 17, 1945); Chem. Abstracts, 39, 4223 (1945); 2,392,103
(Jan. 1, 1946); Chem. Abstracts, 40, 2984 (1946); 2,393,817 (Jan. 29, 1946); Chem. Ab-
stracts. 40, 2305 (1946); R. C. Smith, U. S. Patent 2,334,358 (Nov. 16, 1943); Chem.
Abstracts, 38, 2820 (1944); J. W. Hill, U. S. Patent 2,335,592 (Nov. 30, 1943); Chem
Abstracts, 38, 3144 (1944); J. T. Marsh, U. S. Patent 2,337,398 (Dec. 21, 1943); Chem.
Abstracts, 38, 3475 (1944); R. S. Bley, U. S. Patent 2,341,509 (Feb. 15, 1944); Chem.
Abstracts, 38, 4443 (1944); T. Koch, U. S. Patent 2,345,345 (Mar. 28, 1944); Chem.
Abstracts, 38, 4443 (1944); J. Nelles, U. S. Patent 2,356,079 (August 15, 1944); Chem.
Abstracts, 39, 199 (1945); J. E. Kirby, U. S. Patent 2,371,052 (Mar. 6, 1945); Chem.
Abstracts, 39, 5086 (1945); O. W. Boies, U. S. Patent 2,379,783 (July 3, 1945); Chem.
Abstracts, 39, 4224 (1945); H. Cowling, U. S. Patent 2,397,338 (Mar. 26, 1946); Chem.
Abstracts, 40, 3262 (1946) ; N. L. Cox, U. S. Patents 2,535,044-5 (Dec. 26, 1950) ; Chem.
Abstracts, 45, 2669, 2670 (1951); 2,536,014 (Dec. 26, 1950); Chem. Abstracts, 45, 2207
(1951). See also L. Thoria, /. Indian Chem. Soc., Ind. & News Ed., 11, 63 (1948).
78 H. L. Barthilemy and L. Williams, Ind. Enz. Chem., Anal. Ed., 17, 624 (1954).
IX. DERIVATIVES OF CELLULOSE 999
flow curves relating velocity of flow to pressure are required to obtain really
complete information regarding the flow characteristics of any viscose.
To establish the "xanthate" viscosity of a pulp (the viscose viscosity that
will be obtained as the result of a given set of process conditions), it is
necessary to convert the pulp to viscose under carefully controlled condi-
tions. The small-scale laboratory procedures for the preparation of viscose
described at the beginning of this section are satisfactory for this purpose.
By using the appropriate formula, data obtained by these procedures may
also be employed to determine, at least approximately, the cellulose D.P.74
(1} Degree of Ripeness and Degree of Esterification
Both chemical and colloidal methods are employed for determining the
degree of ripeness or degree of esterification of viscose. Of the chemical
methods, probably the best is the procedure based on the reaction of the
xanthate with diethylchloroacetamide which converts it into an insoluble,
stable derivative30-75 whose composition is:
( C«H702 1 OH ] , [O- CSvS— CHr-CQ— N( C2H5)2] ,_, ) „
After precipitation of this compound, it is filtered off, and the nitrogen
is determined by the Kjeldahl method. Since every nitrogen atom corre-
sponds to one xanthate group in the original sample, this procedure gives
the degree of esterification of the cellulose, and the results may be converted
to xanthate sulfur.
The oldest chemical test for determining xanthate sulfur and degree of
ripening is based on the reaction between cellulose xanthate and iodine
which results in the formation of a so-called disulfide :
2 Rccl,OCSS\a + L> » R,.eHOCSS— SSCORceii + 2 Nal (20)
Because of the interference of the sulfur-containing by-products in vis-
cose, this method is not as straightforward as the one in which diethylchloro-
acetamide is used, and a number of variations in procedure have been sug-
gested.2-22'76 However, the method is used extensively and, with suitable
74 G. Jayme and coworkers, Kol!oid-Z., 107, 163 (1944); 108, 20 (1944); Melliand
Textilber., 27, 155(1946).
™ H. Fink, Angew. Chem., 47, 429 (1934).
76 J. d'Ans and A. Jager, Kunstseide, 8, 17, 43, 57, 82, 110 (1926) ; Cellulosechemie, 16,
22 (1935); H. Jentgen, Laboratoriumsbuch filr die Kuntseide- und Ersatzfaserstoff-Indus-
trie, W. Knapp, Halle (Salle), 1923, p. 55; O. Faust, E. Graumann, and E. Fischer,
Cdlulosechemie, 7, 165 (1926); R. Bernhardt, Kunstseide, 8, 164 (1926); J. Eggert, Die
Herstellung und Bearbeitung der Viscose unter bes. Berucks. d. Kunstseidefabrikation, 2d
ed., J. Springer, Berlin, 1931 ; G. Kita, Kunstseide, 8, 221 (1926) ; G. de Wyss, Ind. Eng.
Continued on next page.
1000 CELLULOSE
precautions, reproducible results can be obtained which agree with the di-
ethylchloroacetamide method. Perhaps the best and simplest procedure77
is to remove the by-product sulfur compounds by treating the viscose with
acetic acid in the presence of calcium carbonate which acts as a buffer,
blowing with air (or, better, oxygen-free nitrogen), and then titrating with
standard iodine solution. The change in xanthate sulfur content during
ripening of a representative commercial viscose is shown in Figure 74A.
The by-product sulfur content of viscose may be estimated by calcula-
tion as the difference between the total sulfur and xanthate sulfur contents.
It may also be determined directly77 by absorbing the CS2 and H2vS, which
are expelled as described in the above method for xanthate sulfur, in alco-
holic NaOH and cadmium acetate solutions, respectively, and titrating
with iodine.
Colloidal methods for determining the degree of esterification or ripeness
of viscose are empirical in nature and measure the ease of coagulation
rather than the amount of any chemical compound. They are more prac-
tical than chemical methods and they are also simpler. Two such methods
are used extensively in industry as regular control methods. They in-
volve the coagulation of viscose with sodium chloride and ammonium
chloride solutions,76 >7S and are based on the fact that with increasing age
viscose may be coagulated by constantly smaller amounts of electrolytes.
1. The Salt Point or Salt Index Method gives the concentration of sodium chloride
solution which isjtitt sufficient to coagulate a definite, small quantity of viscose which
is allowed to fall into it dropwise. There are several modifications of this test. One pro-
cedure is to allow one drop of viscose to fall from the end of a small glass rod (3/io in.
diameter) into a 250-cc Erlenmeyer flask containing 40 cc. of salt solution of known
concentration. The solution is immediately shaken mechanically for a definite period,
and the salting-out effect is noted. If the drop of viscose dissolves, the salt solution is too
dilute. If it coagulates as a heavy precipitate, the solution is too concentrated. At the
correct concentration of NaCl, the drop forms two or three freely suspended, ''comma-
shaped" particles, readily seen by the naked eye.
Chem., 17, 1043 (1925); H. Lotze, Kunstseide, 16, 290 (1934); E. Berl and H. Dillenius,
Cellulosechemie, 13, 1 (1932); K Jung, Kollmd-Z., 108, 120 (1944); K. Kriiger, Kunst-
seide u. Zellwolle, 25, 370 (1947).
77 W. H. Fock, Kunstseide, 17, 117 (1935). For potentiornetric titration methods for
determining the by-product compounds and the degree of ripening, see R. S. Neumann
and coworkers, Cellulosechemie, 17, 16 (1936) ; P. Herrent and G. Jnoff, 7. Polymer Sci.,
3,487,834(1948).
78 K. Ziegler and W. Schafer, Cellulosechemie , 15, 89 (1934); V. Hottenroth, Chem.-
Ztg., 39, 119 (1915); T. Mukoyama, Kolluid-Z., 43, 349 (1927); X. Herthe, Ind. textile,
59, 287 (1942); E. B. Castillo, Anales fis. y quim. (Madrid), 42, 1019 (1946); 43, 60
(1947) ; F. Genert, Kunstseide u. Zellwolle, 23, 80 (1941).
IX. DERIVATIVES OF CELLULOSE 1001
2 The Ammonium Chloride or Hottenroth Index Method gives the volume in cc. of a
10% XH4C1 solution which is necessary to coagulate the viscose under certain conditions.
Viscose (20 g.) is diluted with 30 cc. of water and the solution titrated with 10% NH4C1
solution, with rapid stirring, until coagulation just sets in.
It will he noted that in the salt index method a small quantity of viscose is added to
a large volume of coagulating agent. The concentration and nature of the salt solution
are therefore riot appreciably affected. In the ammonium chloride method, the viscose
solution is in excess. Moreover, the latter method depends upon the conversion of the
ammonium chloride, first, by the free NaOH in the viscose forming sodium chloride and
free ammonia, and, second, by the combined sodium, forming ammonium cellulose
xanthate and more sodium chloride. The ammonium chloride method is, therefore,
more sensitive than the salt index method and more dependent upon the composition of
the viscose. The change in salt index during ripening of a representative commercial
viscose is shown in Figure 74A. It will be observed that there is a direct relationship
between the salt index and xanthate sulfur curves, and hence between the colloidal and
chemical method^ for determining ripeness.
Other methods (including chemical, spectrochemical, physical, and col-
loidal) have been suggested for analyzing and examining viscose to estab-
lish its quality, the presence of insoluble matter, air, gel structure, pulp
reactivity, and other properties. Among these may be mentioned filter-
ability (which will be discussed below), microscopic examination (with
dark-field illumination, under polarized light or otherwise) for gels, fibers,
filtration residues, and the like, flow birefringence, and ultraviolet absorp-
tion.4i5>57>79 With the exception of filterability and some microscopic tests
these methods are not used extensively in industry.
(2) Viscose Filterability
The filterability of viscose is the simplest and perhaps the best measure
of itsquality from the point of view of xanthate and cellulose solubility, of the
efficiency of the various viscose processing steps, and of the presence of in-
soluble matter. As mentioned above, viscose is usually subjected to several
filtration steps before it can be used industrially. The first of these filtra-
tions is by far the most important, both from a technical and economic
standpoint, since it removes most of any undissolved fibers, gels, xanthate,
or other impurities. Industrially, therefore, the efficiency of the first
filtration step, that is, the amount of viscose which can be passed through
79 E. Berl and H. Dillenius, Cellulusechemie, 13, 1 (1932); M. Numa, Kunstseide, 9,
597 (1927) ; C. L. Moore, Silk and Rayon, 8, 563 (1934) ; E. Kiihnel, Kunstseide u. Zell-
wolle, 21, 369, 394 (1939); I. Jurisch, Kunstseide u. Zellwolle, 22, 346 (1940); 23, 5
(1941); K. Atsuki and T Takata, J. Soc. Chem. Ind., Japan, 43, 394B (1940); E. B.
Castillo, Ion., 6, 306 (1946); F. Castellani, Chimica c industria (Milan), 28, 6 (1946);
E. Schauenstein and E. Treiber, Melliand Textilber., 32, 43 (1951).
1002 CELLULOSE
these filters before plugging, is a direct reflection of the viscose quality.
A naly tically, viscose filterability can be determined in the laboratory by a
number of procedures which are useful not only for plant control purposes
but for predicting the behavior of raw materials, for determining the effect
of other variables in viscose manufacture and for determining the reactivity
of pulps to NaOH and/or CS2. All of these procedures involve the prepa-
ration of viscose under a standard set of conditions and passing it through a
suitable standard filtering apparatus under fixed conditions of pressure,
temperature, time, and filter medium. Depending upon the type of data
obtained, filterability may be reported in various terms. Among these
may be mentioned80 :
(1) The amount of viscose filtered in a given time, such as 60 min.
(2) The actual plugging value- -the amount of viscose actually required
to plug the filter by carrying the test to complete plugging.
(3) The calculated plugging value : (a) the amount of viscose obtained
by determining the rate of filtration at intervals such as 5, 10, 15, and 30
min., plotting the rates against the corresponding total amounts filtered,
drawing the best straight line through the points, and extrapolating to
/erorate; or, (b) plugging value:
2(p* + Pl)
2 - jyp,
(4) Plugging constant :
(Ku) = 100,000 X ~
Jr\ -f- JT2
where PI = grams of viscose filtered from 0 to 20 min. and P2 = grams of
viscose filtered from 20 to 60 min.
Most of the factors affecting filtration have already been mentioned. It
is beyond the scope of this Section to discuss the subject in any detail ex-
cept to say that bad^ filtration is usually the result of slime, gels, fibers, and
other incompletely dissolved fiber components, the causes of which are
varied and numerous. The type (e.g., morphology, D.P., and native
fiber structure) of pulp, impurities (organic and inorganic) in raw materials,
80 L. H. Smith, editor, Synthetic Fiber Developments in Germany, Textile Research
Institute, New York, 1946, p. 195. For theoretical considerations involved, including
criticism of some of these formulas see: P. H. Hermans and H. L. Bredee, Rec. trav.
chim., 54, 680 (1935); T. Bergek and T. Ouchterlony, Svensk Papperstidn., 49, 470
(1946); P. H. Teunissen, Svensk Papperstidn., 51, 497 (1948); H. L. Vosters, Svensk
Papperstidn., 53, 29, 613, 771 (1950) ; V. E. Gonsalves, Rec. trav. chim., 69, 873 (1950) ;
A. Matthes, Ghent. Tech. (Berlin), 3, 13 (1951).
IX. DERIVATIVES OF CELLULOSE 1003
carbon disulfide concentration, viscose composition, and every step in the
processing of viscose are all vital in achieving good solubility and a good
filtering solution.4'5-80'81 (For the effect of some viscose processing factors
on filterability, see Figures 71 and 73.)
(f) COAGULATION AND REGENERATION
In view of the fact that viscose is of interest solely as an intermediate
product from which the dissolved cellulose may be recovered in some de-
sired physical form, its most important reaction is that by which the re-
generation of the cellulose is brought about rapidly. The spontaneous de-
composition of the xanthate during ripening obviously occurs too slowly
to be of use industrially for this purpose, and the reaction with neutral
salts results only in coagulation, with the xanthate remaining unchanged.
Mineral acids and acid salts, however, decompose the xanthate directly,.
causing both coagulation and regeneration of the cellulose.82 At the same
time, the free alkali is neutralized and the by-product salts are decomposed
with liberation of CS2, H2S, CO2, and free sulfur. The main reactions, as
occurring with sulfuric acid, may be represented as follows:
2 Rcc,,OCSvSNa + H2SO, > 2 Rrf.,,OH -f- Na2SO4 + 2 CS2 (21)
2 NaOH + H2SO4 > Na2SO4 -f 2 H2O (22)
Na2CS3 + H2SO4 > Na2SO4 -f H2S + CS2 (23)
Na2CO3 + H2SO4 > Na2SO4 + CO2 + H2O (24)
Na2S + H2S04 > Na2SO4 + H2S (25)
If sodium sulfite is present in the viscose, liberation of H2S is prevented
and colloidal sulfur formed instead :
Na2SOs + H2S04 > Na2SO4 + H2O + vSO2 (26)
SO2 + 2H2S > 2 H2O + 3S (27)
With few exceptions, these reactions, combined with the coagulation
effect of neutral salts, are the basis of all industrial uses of viscose. The
viscose is either extruded into a bath containing mineral acid plus one or
more salts causing immediate and simultaneous coagulation and regenera-
81 A. Marschall, Kunstseide u. Zellwolle, 24, 188 (1942) ; O. Samuelson, Sven$k Papper-
stidn., 48, 517 (1945) ; 52, 448, 465 (1949) ; Svensk Kern. Tid., 61, 227 (1949) ; T. Klein-
ert and V. Mossmer, Osterr. Chem.-Ztg., 51, 29 (1950); H. A. Wannow, Reyon, Synthe-
tica, Zellwolle, 29, 135 (1950).
82 Cellulose xanthate may also be decomposed by elect rophoresis, the regenerated
cellulose depositing at the anode. See P. Karrer and T. Lieser, Cellulosechemie, 7, 3
(1926).
1004 CELLULOSE
tion, or it is partially coagulated and regenerated in an acid-salt bath
followed by completion of the reaction by a second acid-salt bath or other
treatment, or it is first just coagulated in a solution of one or more salts
followed by regeneration in a second bath containing mineral acid.83 The
first two "one-bath" and "two-bath" systems are used most widely for the
manufacture of filaments, whereas the latter "two-bath" system is em-
ployed mostly in the production of film structures.81 Regardless of the
regenerating methods employed, the cellulose is obtained as hydra ted or
mercerized cellulose, the characteristics of which have already been de-
scribed in Chapter I V-D.
(1) Technical Applications7
In the production of filaments (rayon arid staple fiber), the viscose is
forced under pressure through very fine holes contained in cup-shaped
nozzles (spinnerets) immersed in a coagulating bath. The emerging viscose
filaments are coagulated immediately, and the coagulated fibers from each
spinneret (each spinneret contains a number of holes) are combined into a
main thread which is simultaneously drawn through the bath, stretched,
and collected on a suitable take-up device. The size of the main thread,
in the case of continuous filament rayon, varies from about 40 to 2200
denier, the number of filaments per thread varying from about 14 to 9(>().
In the production of staple fiber, the number of holes per spinneret is much
greater, in some cases several thousand, and the size of the main thread,
per spinneret, may be as high as 10,000 denier or even higher.
Three methods are used in rayon production for collecting the freshly
spun threads. They may be (1) wound on a rotating spool or bobbin, in
the "spool" or "bobbin" process; or (2) collected in the form of a "cake"
by centrifugal force in a revolving bucket, in the "bucket" or "pot" spin-
ning process85; or (3) they may be wound on a specially designed rotating
reel or other thread storage device which receives and then advances the
thread automatically and continuously, to a series of other similar reels
83 For example, see F. L. Durr, Rayon Textile Monthly, 27, 126, 196, 246, 300, 357,
414 (1946).
84 Dry spinning methods have also been suggested but they have not been applied
industrially. See H. Hoffmann, Papier-Fabr., 39, 14 (1941); R. O. Herzog and H.
Hoffmann, U. S. Patent 2,036,752 (Apr. 7, 1936); Chem. Abstracts, 30, 3645 (1936);
J. L. Costa, U. S. Patent 2,317,152 (Apr. 20, 1943); Chem. Abstracts, 37, 5867 (1943);
L. Paulus, Swedish Patent 116,135 (Apr. 2, 1946). .
85 C. P. Walker, Rayon and Synthetic Textiles, 31, No. 12, 34 (1950) ; Silk and Rayon,
21, 1386(1947); A. J. Hall, Fibres, 9, 19(1948).
IX. DERIVATIVES OF CELLULOSE 1005
(as in the Industrial Rayon Corporation86 "continuous" spinning and puri-
fication process) or along a single pair of canted rollers (as in the recently
developed Nelson, American Viscose Corporation, and Kohorn processes87).
After spinning, and when coagulation of the viscose is complete (regenera-
tion of the cellulose may or may not be complete), the yarn is subjected to a
number of operations, including stretching, washing, desulfuring, bleach-
ing, finishing or oiling, drying, twisting, and winding to suitable packages
(e.g., skeins, cones, tubes). These operations vary considerably through-
out the industry, although in the "spool" and "pot" spinning processes they
are all discontinuous. Each spool or cake, representing 0.50 to 2.0 Ib.
of rayon, is treated individually as such, and the whole operation, from
spinning to the final drying of the yarn, requires from one to six days.
In the Industrial Rayon Corporation "continuous" process, all of these
operations except the final coning step are continuous; each thread is
carried forward individually over a series of thread-advancing reels from
the moment of extrusion until it is completely processed, dried, twisted,
and wound on bobbins, and all of the operations are accomplished in about
() min. (see Fig. 75). So far as is known, the other "continuous" processes
86 T. R Olive, Chem 6r Met. Eng., 45, 668 (1938) ; Silk and Rayon, 20, 1298 (1946) ;
21, 1386 (1947); J. V. and S. L. Sherman, The New Fibers, Van Nostrand, New York,
1946, pp. 222-5.
87 Rayon Textile Monthly, 28, 633 (1947); A. J. Hall, Fibres, 9, 62, 107 (1948); S. W.
Barker, U. S. Patents 2,525,760 and 2,526,110 (Oct. 17, 1950); Chem. Zentr., 122, II,
2825 (1951); R. W. Stanley, U. S. Patents 2,516,157 (July 25, 1950); 2,532,465 (Dec.
5, 1950); 2,582,293 (Jan. 15, 1952); J. W. Coleman and coworkers, U. S. Patent 2,536,-
093 (Jan. 2, 1951) ; /. Textile Inst., 42, A286 (1951) ; H. J. McDermott and J. W. Ped-
low, U. S. Patent 2,536,094 (Jan. 2, 1951); / Textile Inst., 42, A286 (1951); Rayon
and Synthetic Textiles, 32, No. 11, 9 (1951) ; H. Yon Kohorn, Australian Patent Applica-
tion No. 29,713/49, filed Sept. 7, 1949.
Many other schemes for the continuous spinning and purification of rayon have been
proposed but none of them appear to be in use industrially. See: F. L. Durr, Rayon
Textile Monthly, 28, 468 (1947); S. W. Barker and R. Alleston, /. Textile Inst., 39, PI
(1948) ; P. W. Frisk, Rayon and Synthetic Textiles, 30, No. 9, 49 (1949) ; A. J. L. Moritz,
U. S. Patents 2,302,792 (Nov. 24, 1942); Chem. Abstracts, 37, 2577 (1943); 2,346,696
(Apr. 18, 1944) ; Chem. Abstracts, 38, 6094 (1944) ; A. L. Ewing, U. S. Patents 2,317,747
(Apr. 27, 1943); Chem. Abstracts, 37, 6140 (1943); 2,435,430 (Feb. 3, 1948); J. H.
Truesdail, U. S. Patents 2,333,278-9 (Nov. 2, 1943); Chem. Abstracts, 38, 2510 (1944);
W. E. Neff, U. S. Patent 2,340,611 (Feb. 1, 1944); C. F. Gram, U. S. Patent 2,319,812
(May 25, 1943) ; Chem. Abstracts, 37, 6459 (1943) ; H. O. Naumann, U. S. Patent 2,476,-
757 (July 19, 1949); G. A. M. Heim, U. S. Patent 2,334,325 (Nov. 16, 1943); Chem.
Abstr.f 38, 2820 (1944); F. A. J. van Hall, U. S. Patent 2,539,980 (Jan. 30, 1951); /.
Textile Inst.. 42, A287 (1951) ; R. Levison, U. S. Patents 2,566,455-6-7 (Sept. 4, 1951) ;
H. A. Kuljian, U. S. Patents 2,495,936 (Jan. 31, 1950); 2,501,776 (Mar. 28, 1950);
2,504,703 (Apr. 18, 1950) ; and many others.
1006
CELLULOSE
are restricted mainly to the production of rayon for tires and no desulfuring,
bleaching, or twisting steps are included.
Excess coagulating bath is first removed from the freshly spun yarn by
SPINNING
Fig. 75. A continuous spinning and processing machine. This view
shows how rayon is spun, purified, finished, dried, and twisted in syn-
chronized operation. Courtesy of Industrial Rayon Corporation.
washing with water, sometimes in the presence of small amounts of an
alkaline agent such as ammonia or sodium bicarbonate. (Carbon disul-
file recovery is sometimes carried out at this point.) In the ' 'spool" and
4 'pot" spinning processes, the yarn may be dried after this washing stepfc
IX. DERIVATIVES OF CELLULOSE 1007
or further treated directly in the "gel" state (as is, or after reeling to skeins).
Sulfur, which is always present, probably in both free and combined forms,
is then removed by treatment with a dilute (0.5-1.5%) solution of sodium
sulfide, ammonium sulfide, sodium hydroxide, or sodium carbonate.88
Other desulfuring agents have also been proposed.89 When bleaching is
required, it follows the desulfuring operation and is usually carried out with
sodium hypochlorite solution90 containing about 0.05% available chlorine.
After desulfuring and bleaching, the yarn may receive other treatments such
as finishing, sizing, and tinting, after which it is dried. The latter opera-
tion, which is extremely important in establishing the properties of the
final product, may be carried out with or without tension, depending upon
whether or not the yarn is to possess any residual shrinkage. In any case,
the tension during drying must be uniform to minimize any residual shrink-
age differences.
The manufacture of rayon staple resembles that of continuous filament
rayon except that the main threads are much larger, as already pointed out,
and the spinning and purification operations are nearly always continuous.
Two main processes are in use, depending upon the cutting operation:
(1) The filaments may be cut into staple immediately after spinning or
washing and the remaining operations carried out on the cut fibers or (2)
the thread or rope from a number of spinnerets may be combined and proc-
essed in rope form, after which it is cut to the desired length and then
dried.91
88 L. A. Paley, U. S. Patent 1,779,103 (Oct. 21, 1930); Chem. Abstracts, 25, 205
<1931); H. H Parker, U. S Patent 1,931,266 (Oct. 17, 1933); Chem. Abstracts, 28, 321
(1934); H. B. Kline, U. S. Patent 1,932,789 (Oct. 31, 1933); Chem. Abstracts, 28, 659
<1934).
89 P. C. Scherer, Ind. Eng. Chem., 25, 1319 (1933); A. D. Conley and E. C. Stillwell,
U. S. Patent 1,371,300 (Mar. 15, 1921); Chem. Abstracts, 15, 1812 (1921); E. K. Glad-
ding and T S. Sharpe, U. S. Patent 1,655,097 (Jan. 3, 1928); Chem. Abstracts, 22, 1050
<1928); A. Hartmann and J. Uytenbogaart, U. S. Patent 2,194,470 (Mar. 26, 1940);
Chem. Abstracts, 34, 4924 (1940); I. G. Farbenindustrie Akt.-Ges., Brit. Patent 279,437
<Oct. 14, 1927); Chem. Abstracts, 22, 2847 (1928); A. E. Stein, Brit. Patents 428,955
(May 22, 1935); 429,165 (May 24, 1935); J. G. Evans, Brit. Patent 464,116 (Apr. 9,
1937); Chem. Abstracts, 31, 6463 (1937); Herminghaus & Co., French Patent 655,729
(June 14, 1928); Chem. Abstracts, 23, 4072 (1929).
90 See also J. S. Fonda and G. W. Filson, U. S. Patent 2,064,300 (Dec. 15, 1936);
Chem. Abstracts, 31, 889 (1937); R. O. Denyes, U. S. Patent 2,479,605 (Aug. 23, 1949);
Chem. Abstracts, 44, 840 (1950); J. W. Jacokes, U. S. Patent 2,488,667 (Nov. 22, 1949);
Chem. Abstracts, 44, 7068 (1950).
91 For example, see: O. Von Kohorn, U. S. Patent 2,308,576 (Jan. 19, 1943) ; Chem.
Abstracts, 37, 3952 (1943); L. E. Lovett, U. S. Patent 2,315,265 (Mar. 30, 1943).
1008 CELLULOSE
In the production of cellophane, viscose is forced under pressure (cast)
through a slit in a suitable hopper which is immersed in a coagulating bath.
As the viscose emerges from the slit, it is coagulated at once in the form
of a thin, wide sheet or film. This is led by means of suitable rollers through
a series of baths in which, after regeneration of the cellulose is complete,
the film is washed, desulfured, bleached, etc., and finally dried. The en-
tire operation is continuous, the machine being somewhat similar to, al-
though much smaller than, a paper machine.92
Sausage casings, bands, and similar structures are made by extruding
viscose through an annular slot immersed in a suitable coagulating bath.
The viscose is coagulated in the form of a tube, which is processed on a con-
tinuous machine through the various purification and drying steps.93
Bottle caps, bands, and the like may be formed by covering suitable man-
drels with a film of viscose by dipping, and then immersing them in a coagu-
lating medium. The viscose is thus regenerated as a film in the form of the
particular mandrel used, after which the structure is removed from the
mandrel and subjected to the necessary purification operations. Other
cellulose structures are made from viscose in an analogous manner.94
(2) Coagulating Baths and the Spinning Operation
Since viscose is used most widely in the production of rayon and staple
fiber, its coagulation and regeneration are of concern mostly from the view-
point of filament formation, that is, spinning.96 In this connection, the
composition of the coagulating bath (the importance of raw materials,
viscose composition, viscosity, degree of ripening, and other factors have
already been mentioned) is a prime consideration. The bath serves several
functions. It must coagulate the cellulose dispersed in the viscose and
92 J. E. Brandenberger, U. S. Patents 1,548,864 (Aug. 11, 1925); Ghent. Abstracts, 19,
3018 (1925) ; 1,601,289 (Sept. 28, 1926) ; Chem. Abstracts, 20, 3814 (1926).
»3 W. P. Cohoe, U. Ss Patent 1,163,740 (Dec. 14, 1915).
94 For example, see: G. Pum and A. Glaessner, U. S. Patent 1,142,619 (June 8,
1915); Chem. Abstracts, 9, 1978 (1915); L. Mostny, U. S. Patent 1,611,056 (Dec 14,
1926) ; Chem. Abstracts, 21, 321 (1927).
96 H. Erbring, Kolhid-Beihefle, 44, 171 (1936) ; R. Klaus, Kunstseide, 15, 9 (1933) ; 16,
148 (1934); A. Wehrung, Cellulosechentie, 11, 170 (1930); X. Matsumoto, /. Soc.
Chem. Ind.t Japan, 41, B380 (1938); R. Inoue, J. Soc. Chem. Ind., Japan, 41, B334,
B357(1938); 42, B18 (1939) ; F. H. Miiller, Physik. Z., 42, 123 (1941); P.H.Hermans,
Physics and Chemistry of Cellulose Fibres, Elsevier, The Netherlands, 1949; /. Polymer
Sci., 1, 389, 393 (1946); 2> 632 (1947); 3, 1 (1948); R. Gaebel, Kunstseide u. Zellwolle,
27, 153 (1949); J. Miiller, Kunstseide u. ZcllwoHc, 28, 385 (1950); H. L. Vosters, Svensk
Papperstidn., 53, 59 (1950).
IX. DERIVATIVES OF CELLULOSE 1009
decompose the remaining xanthate, at least in part; it must also neutralize
the free alkali, decompose the by-product salts, and be compatible with
the by-products of the reaction. Moreover, it must coagulate and regener-
ate the viscose as a homogeneous plastic mass, at a suitable speed so that
the coagulated filaments may be readily and continuously led through the
bath, stretched to a considerable extent if desired, and collected, all without
interruption. These results have been achieved most satisfactorily by
combining a reagent which actively decomposes the xanthate (dilute
mineral acid) with an agent which simply coagulates (a salt).96 The so-
called Miiller bath,97 consisting of dilute sulfuric acid combined with a
soluble sulfate salt, represents such a combination, and baths of this general
type, although considerably modified, are in general use throughout the
industry. The composition used is normally in the range of 7 to 139o
H2SO4 and 13 to 25% Na2SO4. MgSO4 or (NH4)2SO498 is sometimes used
in place of all or part of the Na2SO4. In addition, most coagulating baths
today contain small amounts (0.5-2.0%) of ZnS04" and, in some cises,
also organic materials like glucose100 and wetting or dispersing agents.101
% For some early bath compositions see: C. F. Cross, E. J. Bevan, and C. Beadle,
U vS. Patent 604,206 (May 17, 1898); C H. Steam, U. S. Patents 622,087 (Mar. 28,
1899); 716,778 (Dec. 23, 1902); 725,016 (Apr. 7, 1903); Brit. Patent 2,529 (1902).
97 F. Steimmig, Kunstseide, 12, 242 (1930); M. Miiller, U. S. Patent 836,452 (Nov.
20, 1906); Vereinigte GlanzstofT-Fabriken Akt.-Ges., German Patents 187,947 (Apr.
2, 1905); 287,955 (Feb. 15, 1912); Chcm. Abstracts, 10,2145(1916).
98 T. H. Verhave, U. S Patent 1,280,338 (Oct. 1, 1918); Chem. Abstracts, 12, 2695
(1918) ; J. C. Hartogs, German Patent 324,433 (Mar 1, 1914) ; Chem. Zentr., 91, IV, 486
(1920)
99 S. S. Napper, U. S. Patent 1,045,731 (Nov. 26, 1912); Chem. Abstracts, 7, 706
(1913); F. C. Niederhauser and H. B. Kline, U. S. Patent 1,661,574 (Mar. 6, 1928);
Chem. Abstracts, 22, 1483 (1928); J. J. Stockly and A. Brotz, U. S. Patent 2,015,201
(Sept 24, 1935) ; Chem. Abstracts, 29, 7657 (1935) ; Vereinigte GlanzstofT-Fabriken Akt.-
Ges., German Patent 260,479 (Sept. 16, 1911) ; Chem. Abstracts, 7, 3236 (1913).
100 L. P. Wilson, U. S. Patent 970,589 (Sept. 20, 1910); Chem. Abstracts, 4, 3298
(1910); E. Bronnert, U. S. Patent 1,426,953 (Aug. 22, 1922); Chem. Abstracts, 16, 3763
(1922); Vereinigte Glanzstoff-Fabriken Akt.-Ges., German Patent 240,846 (Sept. 26,
1908) ; Chem. Abstracts, 6, 2169 (1912).
101 J. J. Polak and J. G. Weeldenburg, U. S. Patent 2,125,031 (July 26, 1938); Chem.
Abstracts, 32, 7725 (1938); L. Rose, U. S. Patent 2,302,589 (Nov. 17, 1942); Chem.
Abstracts, 37, 2590 (1943); R. S. Bley, U. S. Patents 2,310,207-8 (Feb. 9, 1943); Chem.
Abstracts, 37, 4247, 4248 (1943); 2,345,570 (Apr. 4, 1944); Chem. Abstracts, 38, 4444
(1944); 2,348,203 (May 9, 1944) ; Chem. Abstracts, 39, 1291 (1945) ; B. W. Collins, U. S.
Patents 2,359,749-50 (Oct. 10, 1944) ; Chem. Abstracts, 39, 3933, 1052 (1945) ; 2,519,227
(Aug. 15, 1950) ; Chem. Abstracts, 44, 1111 1 (195:)) ; H. B. Kline, U. S. Patents 2,394,519
(Feb. 5, 1946); Chem. Abstracts, 40, 2628 (1946); 2,422,021 (June 10, 1947); Chem.
Abstracts, 41, 5721 (1947); IX E. Drew, U. S. Patent 2,-'}6(),405 (Oct. 17, 1944); Chem.
Continued on next page.
1010 CELLULOSE
Of the hundreds of reagents which have been studied for spinning baths,
perhaps the most important are those in the class of inorganic salts. The
coagulating effect of sodium and ammonium salts has been mentioned.
Heavy-metal salts also react with and coagulate viscose to give the corre-
sponding metal xanthates.2'102 Since yarn properties (such as strength,
elongation, dye affinity, luster, softness, and cross section) vary with the
nature and concentration of the salt, many of them have become of definite
interest as spinning bath constituents.
The action of salts is due in part to changes in pH,43-103 to changes in the
speed of regeneration (which is reduced by the addition of salts104), and
to the hydration of the ions. Most important, however, is the dehydrating,
salting-out action which is common to all salts. In this connection the
valence law of colloid chemistry and the Hofmeister series apply,106 but
only in part, in view of the hydrophilic character of viscose and the chemi-
cal reactions (mentioned above) which occur simultaneously. Thus,
ammonium salts have a greater coagulating power than sodium salts,
whereas the coagulating effect of divalent Mg++ is of the same order as
monovalent Na4", and heavy-metal salts like Zn + +, Fe++, and Ni++ are
several hundred times more effective than either Na+ or Mg~f~f".106 The
retardative influence on regeneration shown by these cations is in about
Abstracts, 39, 1538 (1945); A. Cresswell, U. S. Patent 2,442,331 (June 1, 1948); Chem.
Abstracts, 42, 6114 (1948); T. A. H. Blaas, U. S. Patent 2,451,148 (Oct. 12, 1948);
Chem. Abstracts, 43, 850 (1949) ; S. A. Moss, Jr., U. S. Patent 2,489,310 (Nov. 29, 1949) ;
Chem. Abstracts, 44, 7062 (1950) ; K. R. Brown, U. S. Patent 2,495,833 (Jan. 31, 1950) ;
Chem. Abstracts, 44, 10320 (1950).
102 Many of these heavy-metal xanthates are highly colored compounds. For
example, the lead salt is carmine red; the copper salt, chocolate brown; the iron salt,
brownish red; the nickel salt, cherry red; the zinc salt, white; the mercury salt, yellow;
the cobalt salt, brownish black; the bismuth salt, red-brown; the cadmium salt, yellow;
the silver salt, rose-brown; the antimony salt, red-yellow. Since solutions of pure
xanthate do not give such intensely colored precipitates and since the compounds can
also be obtained from the reaction products of C&> and NaOH, they are probably only
in part cellulose xanthate salts. See H. Seidel, Mitt. Tech. Gewerb.-Mus. Wien, 10, 35
(1900); R. Wolffenstein and E. Oeser, Kunstseide, 7, 29 (1925); T. Lieser, Cellulosc-
chemie, 10, 156(1929).
103 K. Tanemura, Cellulose Ind. (Tokyo), 11, 12, 100 (1935).
104 C. L. Moore, Silk and Rayon, 9, 19 (1935) ; S. Hase, /. Soc. Chem. Ind., Japan, 35,
Suppl. binding, 367 (1932).
106 E. Berl and H. Dillenius, Cellulosechemie, 13, 1 (1932); K. Leuchs, Chem.-Ztg., 47,
801 (1923).
106 L. Mirlas, Cellulosechemie, 16, 37 (1935); M. Horio, Textile Research J .. 20, 373
(1950); D. Vermaas and J. J. Hermans, Rec. trav. chim., 67, 983 (1948).
DC. DERIVATIVES OF CELLULOSE, 1011
the same order as their coagulating power,107 and, when two or more of
them are combined, their activity is usually greater than the sum of the
individual salts.
These characteristics explain the use of various salts in baths and the
special interest in zinc and other polyvalent cations (with zinc sulfate, for
example, a relatively stable film containing zinc cellulose xanthate is,
formed on the surface of the filaments), particularly in stretch-spinning
processes to permit the application of tension, to increase strength, and to
produce fine filament yarns. MgSC>4, being more soluble than Na^CU,
permits higher salt concentrations in the bath without crystallization diffi-
culties, whereas ZnSO4, even in the small concentrations mentioned above,
deepens dyeing and improves strength, softness, and luster. Further very
significant effects along these and other lines are obtained by higher con-
centrations (2.5-30.0%) of ZnSO4,108 and other salts, including the sulfates109
of aluminum, chromium, iron, arsenic, nickel, cobalt, and manganese,
have been proposed for various specific purposes. Sodium bisulfite, so-
dium and ammonium phosphates, sodium borate, sodium thiosulfate,
sodium and ammonium bicarbonates, sodium silicate, sodium salts of
fatty acids, sodium benzene sulfonate, salts of certain organic bases, and
others110 have also been suggested for use.
107 V. Duchesnoy, Russa. 9, 641 (1934); A. Pakshver and coworkers, Trans. Inst.
Chem. Technol Ivanovo ( U. S. S. R.), 1940, No. 3, 178.
™ W. P. Dreaper, U. S. Patent 1,626,454 (Apr. 26, 1927); Chem. Abstracts, 21, 2063
(1927); Brit. Patent 239,254 (May 2, 1924); Chem. Abstracts. 20, 2079 (1926); R. Pic-
ard, U. S. Patent 1,831,030 (Nov. 10, 1931); Chem. Abstracts. 26, 844 (1932); I. P.
Davis, U. S. Patent 2,114,915 (Apr. 19, 1938); Chem. Abstracts. 32, 4784 (1938); J. H.
Givens, H. W. Biddulph, and L. Rose, U. S. Patent 2,192,074 (Feb. 27, 1940); Chem.
Abstracts. 34, 4569 (1940) ; N. L. Cox, U. S. Patents 2,535,044-5 (Dec. 26, 1950) ; Chem.
Abstracts. 45, 2669, 2670 (1951).
»» C. H. Steam, U. S. Patent 725,016 (Apr. 7, 1903) ; J. C. Hartogs, U. S. Patent
1,573,062 (Feb. 16, 1926); Chem. Abstracts. 20, 1328 (1926); I. P. Davis, U. S. Patent
2,114,915 (Apr. 19, 1938); Chem. Abstracts. 32, 4784 (1938); C. H. Stearn and C. F.
Topham, Brit. Patent 16,604 (July 28, 1903); /. Soc. Chem. Ind.. 23, 784 (1904); H.
Kizu and K. Kadowaki, /. Soc. Chem. Ind.. Japan. 38, Suppl. binding, 195 (1935);
J. J. StSckly, U. S. Patents 2,315,559-60 (Apr. 6, 1943); Chem. Abstracts. 37, 5590
(1943); N. L. Cox, U. S. Patents 2,347,883-4 (May 2, 1944); Chem. Abstracts. 39, 192
(1945); 2,364,273 (Dec. 5, 1944); Chem. Abstracts. 39, 3668 (1945); E. B. Castillo,
Rev. acad. cienc. exactas.fis.-quim. y nat. Zaragoza. [2], 3, No. 1, 33 (1948).
»°C. N. Waite, U. S. Patents 759,332 (May 10, 1904); 816,404 (Mar. 27, 1906);
C. A. Ernst, U. S. Patents 798,027 (Aug. 22, 1905) ; 792,888 (June 20, 1905) ; E. Bron-
nert, U. S. Patents 1,102,237 (July 7, 1914) ; Chem. Abstracts. 8, 2947 (1914) ; 1,374,718
(Apr. 12, 1921); Chem. Abstracts. 15, 2730 (1921); 1,376,672 (May 3, 1921); Chem.
Abstracts. 15, 2730 (1921) ; 1,387,882 (Aug. 16, 1921) ; Chem. Abstracts. 15, 4063 (1921) ;
Continued on next page.
1012 CELLULOSE
The nature of the acid is important also.111 Although sulfuric acid is
perhaps the only acid used commercially, others112 have been proposed,
such as acetic and other fatty acids, hydrochloric, nitric, benzenesulfonic,
phosphoric, and arsenic acids.
Many organic compounds118 have been investigated as constituents of
spinning baths, but few are actually used. For the most part, organic
additions are of interest in influencing thread formation apart from coagu-
lation and regeneration. The outstanding materials of this type are glu-
cose and certain wetting or dispersing agents. Glucose is used rather
widely in concentrations of 2 to 10%, mainly to prevent crystallization.
It also suppresses the action of the acid and the oxidation of H2S, and in-
fluences the continuity of spinning, softness, and other yarn properties.100
Various wetting and dispersing agents are used in the bath (in relatively
low concentrations) in some plants to prevent or reduce the formation of
deposits in the spinneret holes.101 These deposits, representing insoluble
by-products of the reaction between viscose and spinning bath, adversely
affect the quality of the thread and the continuity of spinning. They
may be controlled under ordinary conditions by proper attention to various
spinning factors, but the production of very fine filament yarns and the
use of higher spinning speeds and of certain bath compositions to permit
high stretching have aggravated this condition in recent years, creating
interest in new and novel methods of correction. 114
C. P. Cross, U. S. Patent 1,538,689 (May 19, 1925); Chem. Abstracts, 19, 2136 (1925);
H. Chavassieu, U. S. Patent 2,034.711 (Mar. 24, 1936) ; Chem. Abstracts, 30, 3256 (1936) ;
S. Peessarer, Brit. Patent 16,583 (Aug. 15, 1905) ; W. F. Underwood, U. S. Patent 2,413,-
123 (Dec. 24, 1946).
111 Y. Kami and M. Nozaki, Cellulose Ind. (Tokyo), 5, 117 (1929); O. Faust, Ber.,
626,2567(1929).
112 C. A. Ernst, U. S. Patent 792,888 (June 20, 1905) ; C. H. Steam, U. S. Patent 725,-
016 (Apr. 7, 1903) ; L. P. Wilson, U. S. Patent 970,589 (Sept. 20, 1910) ; Chem. Abstracts,
4, 3298 (1910); E. Bronnert, U. S. Patents 1,102,237 (July 7, 1914); Chem. Abstracts,
8, 2947 (1914); 1,376,672 (May 3, 1921); Chem. Abstracts, 15, 2730 (1921); 1,464,805
(Aug. 14, 1923) ; Chem. Abstracts, 17, 3259 (1923) ; J. C. Hartogs, U. S. Patent 1,534,382
(Apr. 21, 1925); Chem. Abstracts, 19, 1781 (1925); L. Lilienfeld, U. S. Patent 1,881,740
(Oct. 11, 1932); Chem. Abstracts, 27, 601 (1933).
113 C. A. Ernst, U. S. Patent 792,888 (June 20, 1905) ; M. T. Callimachi, U. S. Patent
1,449,380 (Mar. 27, 1923); Chem. Abstracts, 17, 1888 (1923); F. C. Niederhauser and
A. E. Sunderland, U. S. Patent 1,625,562 (Apr. 19, 1927); Chem. Abstracts, 21, 2063
(1927).
114 R. Soukup, U. S. Patent 2,324,437 (July 13, 1943) ; Chem. Abstracts, 38, 253 (1944) ;
I. F. Walker, U. S. Patent 2,364,407 (Dec. 5, 1944); Chem. Abstracts, 39, 4223 (1945);
G. H. White, U. S. Patent 2,394,957 (Feb. 12, 1946); Chem. Abstracts, 40, 2627
(1946); A. Cresswell, U. S. Patent 2,515,697 (July 18, 1950); Chem. Abstracts, 44,9674
(1950) ; A. Wehrung, Kunstseide u. Zellwolle, 28, 84 (1950).
IX. DERIVATIVES OF CELLULOSE 1013
Although coagulation and regeneration usually occur simultaneously,
they do not occur at the same rate. The absolute concentrations and the
ratio of acid and salts 104-115 determine which of these reactions predominates
and, in turn, the shrinkage characteristics of the gel and the quality and
characteristics of the final yarn. In this connection, the presence of vari-
ous materials in small amounts as impurities, such as traces of metals, may
have a profound effect on the spinning operation. The cross section of the
filaments, which is indicative of the initial shrinkage and of many yarn
properties, also varies characteristically with bath (and viscose) com-
position.116 ''High-swelling" baths, such as neutral salt solutions or con-
centrated sulfuric acid alone, give a circular cross section, with a relatively
smooth outline and little, if any, "skin." "Low-swelling" conditions, pro-
duced, for example, by acid baths with high salt concentrations and high
degrees of xanthate substitution, cause the initial formation of a semiperme-
able surface film or "skin" ; the lower the swelling is, the thicker is the skin.
As coagulation and regeneration progress, an exosmosis of water takes
place from the interior to the exterior of the filament, causing a decrease in
volume and, hence, a shrinkage of the surface film, and giving an irregular,
crenulated cross section. Although considerable information has been de-
veloped117 as to the actual mechanism of "skin" formation as well as to the
relationship and orientation of the "skin" and "core" and the irregularity
of cross sections of viscose filaments, these phenomena are still not com-
pletely understood.
Besides the chemical composition of the bath, a number of physical fac-
tors play important roles in the spinning operation. These include bath
temperature (the rates of the reactions involved increase with temperature) ,
bath travel, spinneret composition, spinning speed, spinning tension, and
stretch. All of these factors118 as well as viscose and bath composition are
interdependent, so that it is nearly always impossible to change any one
116 P. Martin, Rusta-Rayonne, 11, 275 (1936); E. Bronnert, U. S. Patent 1,393,197
(Oct. 11, 1921); Chem. Abstracts, 16, 837 (1922); and other Bronnert patents; P.
Thivet, Rusta-Rayonne, 11, 401 (1936).
116 E. Bronnert, /. Soc. Dyers Colourists, 38, 153 (1922) ; R. O. Herzog, Leipzig Monat-
schr. Textile-Ind., 41, 352 (1926); A. Jager, Kunstseide, 13, 325 (1931); P. A. Koch,
Klepzig's Textil-Z., 40, 17, 284 (1937).
117 J. M. Preston, /. Soc. Chem. Ind.t 50, 199T (1931); J. Textile Inst., 40, T327
(1949); F. F. Morehead and W. A. Sisson, Textile Research J., 15, 443 (1945); W.
Schramek and E. Zehmisch, Kolloid-Beihefte, 48, 93 (1938); K. Wuhrmann, Helv. Chim.
Acta, 28, 666(1945).
IIB Y. Mitugi and coworkers, /. Soc. Chem. Ind., Japan, 46, 944 (1943) ; R. Prince and
J. Seiberlich, /. Phys. Chem., 50, 222 (1946).
1014 CELLULOSE
of them without also making compensating changes in others. Bath
temperature is usually controlled in the range of 45 to 55°C. where one-
bath systems are employed, and the bath travel may vary from about 10 in.
to several hundred inches. Spinnerets are made mostly of precious met-
als, high-platinum alloys being preferred.119 Spinning speed varies con-
siderably, depending upon such factors as type of spinning machine, fila-
ment size, and bath composition. Most operations today employ speeds
in the range of 3000 to 4000 in. per minute, although considerable atten-
tion is being given to faster operation.120 Speeds up to 10,000 in. per min-
ute and even higher are possible, although they involve radical changes
from present practice.
The tension-stretch relationships during spinning are vitally important
from the viewpoint of fiber properties. In the viscose, the cellulose par-
ticles are dispersed at random, although it may be that the chains are more
or less straightened and parallel to each other. In the spinning operation,
alignment of the particles (orientation) occurs as a basic process accom-
panying or following coagulation, regeneration, and the initial gel shrink-
age; the degree of orientation determines in part the strength, elongation,
dyeing, and other characteristics of the fiber. Some orientation of the
particles on the dutside of the filaments is brought about by surface friction
as the viscose passes through the spinneret holes. The major orienting
effect, however, is accomplished by a stretch imparted to the filaments due
to the fact that the velocity of collection of the filaments is always greater
than the extrusion.
The total stretch for any given yarn is governed by the spinneret hole
diameter, which varies with different producers and with the filament size
from 0.0020 in. for the finest filaments to about 0.0060 in. for the coarsest.
The effect of the stretch depends upon the stage at which it is applied. In
the early methods of spinning "without tension " in which the thread passes
directly from spinneret to collecting device, the stretch is imparted almost
exclusively dose tor the spinneret where the filament is still in a "fluid"
state. This gives little orientation and the fibers are characterized by low
»• For example, see R. V. Williams and E. R. McKee, U. S. Patent 2,135,611 (Nov.
8, 1938) ; Chem. Abstracts. 33, 1496 (1939) ; H. Whitehead, U. S. Patent 2,334,890 (Nov.
23, 1943) ; Chem. Abstracts, 38, 2820 (1944). For spinneret hole design see H. J. Jones,
U. S. Patent 2,341,555 (Feb. 15, 1944) ; Chem. Abstracts, 38, 4444 (1944).
l»H. Jentgen, Kunstseide, 19, 261 (1937); R. Soukup, U. S. Patents 2,307,863-4
(Jan. 12, 1943); Chem. Abstracts. 37, 3606 (1943); F. R. Millhiser, U. S. Patent 2,440,-
057 (Apr. 20, 1948); Chem. Abstracts, 42, 4364 (1948); N. Drisch and R. Prion, U. S.
Patent 2,511,699 (June 13, 1950); Chem. Zentr., 122, 1, 544 (1951); J. W. Pedlow and
coworkers, U. S. Patent 2,510,135 (June 6, 1950).
IX. DERIVATIVES OF CELLULOSE 1015
tenacity and high elongation. Present methods, for the most part, in-
volve so-called "tension spinning" or "stretch spinning" practices121 in
which at least part of the total stretch occurs relatively farther from the
spinneret, that is, later in the coagulation-regeneration cycle. This is
accomplished by interposing various mechanical "braking" devices122
(such as fixed guides, freely rotating guides, and positively driven rollers)
between the spinneret and the final collecting device. Although these
stretching procedures are most commonly used in conjunction with baths
of the Miiller type, highest breaking strengths are obtained when they
are combined with special bath compositions such as strong (50-86%)
sulfuric acid123 (which imparts a simultaneous coagulating, swelling, and
plasticizing effect) or with Miiller-type baths followed by secondary baths
or treatments, usually at higher temperatures.124 By thus delaying the
121 O. Faust, Kunstseide, 14, 362 (1932).
11S C. A. Ernst, U. S. Patent 808,148 (Dec. 26, 1905); W. Harrison, U. S. Patent
1,930,803 (Oct. 17, 1933); Chem. Abstracts, 28, 321 (1934); H. Pfannenstiel and H.
Meyer, U. S. Patent 1,933,999 (Nov. 7, 1933); Chem. Abstracts, 28, 640 (1934); F. H
Griffin, U. S. Patent 1,950,922 (Mar. 13, 1934) ; Chem. Abstracts, 28, 3586 (1934) ; H. A.
Schrenk, U. S. Patent 1,968,912 (Aug. 7, 1934); Chem. Abstracts, 28, 6310 (1934);
W. H. Bradshaw, U. S. Patent 2,012,984 (Sept. 3, 1935); Chem. Abstracts, 29, 7075
(1935); W. H. Bradshaw and G. P. Hoff, U. S. Patents 2,083,251-2 (June 8, 1937);
Chem. Abstracts, 31, 5579 (1937) ; B. Borzykowski, Brit. Patent 149,295 (Aug. 31, 1917) ;
A. J. L. Moritz, U. S. Patent 2,302,971 (Nov. 24, 1942); Chem. Abstracts, 37, 2577
(1943) ; J. W. Coleman and coworkers, U. S. Patent 2,536,093 (Jan. 2, 1951) ; /. Textile
Inst., 42, A286 (1951).
188 These baths have apparently not been commercially satisfactory. See L. Lilien-
feld, U. S. Patent 1,683,199 (Sept. 4, 1928) ; Chem. Abstracts, 22, 3990 (1928) ; and other
patents of Lilienfeld.
124 H. C. Stuhlmann, U. S. Patent 1,901,007 (Mar. 14, 1933); Chem. Abstracts, 27,
3075 (1933); A. Bernstein, U. S. Patent 1,996,989 (Apr. 9, 1935); Chem. Abstracts, 29,
3513 (1935); I. P. Davis, U. S. Patent 2,114,915 (Apr. 19, 1938); Chem. Abstracts, 32,
4784 (1938); J. H. Givens, H. W. Biddulph, and L. Rose, U. S. Patent 2,192,074 (Feb.
27, 1940); Chem. Abstracts, 34, 4569 (1940); A. Hartmann and J. Uytenbogaart, U. S.
Patent 2,194,470 (Mar. 26, 1940); Chem. Abstracts, 34, 4924 (1940); Zellstofffabrik
Waldhof and A. Bernstein, Brit. Patent 335,605 (June 27, 1929); Chem. Abstracts, 25,
1672 (1931); Chatillon (Soc. anon, italiana per la seta artificiale), Brit. Patent 370,943
(Jan. 13, 1931); Chem. Abstracts, 27, 3088 (1933); S. Riko, Y. Akizuki, and Y. Kikuti,
J. Soc. Chem. Ind., Japan, 39, Suppl. binding, 31 (1936) ; I. P. Davis, U. S. Patent 2,312,-
152 (Feb. 23, 1943); Chem. Abstracts, 37, 4899 (1943); H. Fink, U. S. Patent 2,327,516
(Aug. 24, 1943) ; Chem. Abstracts, 38, 865 (1944) ; G. I. Thurmond, U. S. Patents 2,328,-
307 (Aug. 31, 1943); Chem. Abstracts, 38, 865 (1944); 2,369,190-1 (Feb. 13, 1945);
Chem. Abstracts, 39, 5488 (1945) ; A. S. Brown, U. S. Patent 2,433,733 (Dec. 30, 1947) ;
Chem. Abstracts, 42, 2114 (1948) ; E. A. Tippetts, U. S. Patent 2,439,829 (Apr. 20, 1948) ;
H. W. Swank, U. S. Patent 2,440,226 (Apr. 20, 1948); G. M. A. Kayser, U. S. Patent
2,452,130 (Oct. 26, 1948); Chem. Abstracts, 43, 1990 (1949); F. R. Millhiser, U. S.
Continued on next page.
1016 CELLULOSE
stretching, orientation of the interior of the filaments as well as of the sur-
face occurs, and, hence, higher tenacity results. (The higher tenacity is
obtained, however, at the expense of elongation.) Whereas the older spin-
ning methods, "without tension," give breaking strengths of the order
of 1.6 g. per denier (31,000 Ib. per sq. in.), "stretch spinning*' methods are
now common which give 2.5 to 4.25 g. per denier, and, experimentally,
strengths as high as 6.0 g. per denier (117,000 Ib. per sq. in.) have been
obtained.
Space does not permit further discussion of the spinning operation (fila-
ment formation^ or of the extrusion of viscose for the production of films
and the like. Much of the investigational work on the subject is covered
by References 84-124, inclusive, but it should be made clear that the proc-
esses involved are extremely complicated and still require considerable
study for their complete elucidation.
After completion of the spinning operation (i.e., coagulation, regenera-
tion, and orientation), there is little if any change in the basic structure or
characteristics of the gel filaments during the steps of washing, desulfuring,
or bleaching. In the drying operation, however, the final shrinkage of the
gel structure takes place. Drying conditions (such as temperature, rate,
uniformity, and amount of tension) therefore are extremely important in
determining the final structure and hence the properties (dyeing, tensile
strength, elongation, residual shrinkage, etc.) of the final product.
(g) FURTHER REACTIONS OF VISCOSE
Viscose undergoes several other reactions not mentioned in the previous
discussion. For example, xanthic esters may be formed according to the
general reaction :
RceiiOCSSNa + X—R » Roel,OCSSR + NaX (28)
in which R is an organic radical and X the negative ion of an inorganic acid.
Thus methyl iodide forms methyl cellulose xanthate,125 and diethyl sulfate
forms ethyl cellulose xanthate. Both of these compounds are soluble in
alkali and in certain organic solvents, and may be spun or formed into plas-
tic masses.
Patent 2,453,332 (Nov. 9, 1948); J. A. Calhoun, Jr., U. S. Patent 2,484,012 (Oct. 11,
1949); Chem. Abstracts, 44, 1714 (1950); M. Horio and S. Nagata, /. Soc. Chem. Ind,,
Japan, 46, Suppl. binding, 155B (1943); W. Schramek, Melliand Textilber., 28, 420
(1947).
126 L. Lilienfeld, U. S. Patent 1,680,224 (Aug. 7, 1928); Chem. Abstracts, 22, 3777
(1928); Brit. Patent 252,654 (Oct. 18, 1926) ; Chem. Abstracts, 21, 2384 (1927) ; German
Patent 519,138 (May 30, 1926); Chem. Abstracts, 25, 2847 (1931).
IX. DERIVATIVES OF CELLULOSE 1017
When cellulose xanthate or viscose is allowed to react with halogen de-
rivatives of polyvalent alcohols like a-dichlorohydrin, or with halogen
derivatives of anhydrides of polyvalent alcohols such as epichlorohydrin,
or with trithiocarbonic acid esters of polyvalent alcohols, one or more of
the OH groups of the cellulose is replaced126 and compounds useful for
spinning into filaments and the like are produced.
With halogenated fatty acids, salts, and esters, viscose forms cellulose
xantho fatty acid derivatives.30'127 With monochloroacetic acid, for ex-
ample, the reaction is :
RceiiOCSSNa + ClCH2COONa > RcenOCSSCH2COONa + NaCl (29)
Sodium salt of cellulose
xanthoacetic acid
The xantho fatty acid derivative thus formed may react further with am-
monia, or with organic amines, such as aniline, I28 as follows :
RceiiOCSSCK2COONa + C6H5NH2 >
Rc0iiOCSNHC6H5 + CH2(SH)COONa (30)
Cellulose Sodium
phenylthiourethan thioglycolate
Cellulose
xanthoanilide
If the salt of the above thiourethan is now treated with an ester of an in-
organic acid, such as ethyl iodide, the corresponding ethyl ester of cellulose
phenylthiourethan results.129
Cellulose xanthate, itself, also reacts with organic amines130 forming
N-substituted thiourethans of the type :
RcoiiOCSNHR
126 L. Lilienfeld, U. S. Patents 1,938,032 (Dec. 5, 1933); Chem. Abstracts, 28, 1188
(1934); 2,004,876 (June 11, 1935); Chem. Abstracts, 29, 5285 (1935); 2,021,862-4
(Nov. 19, 1935); Chem. Abstracts, 30, 614 (1936); see also L. Lilienfeld, Brit. Patents
335,906 (Mar. 25, 1929); Chem. Abstracts, 25, 1995 (1931); 335,993-4 (Mar. 25, 1929);
356,286 (Mar. 10, 1930) ; Chem. Abstracts, 26, 5756 (1932).
127 T. Nakashima, J. Soc. Chem. Ind., Japan, 31, Suppl. binding, 31 (1928) ; L. Lilien-
feld, U. S. Patent 1,642,587 (Sept. 13, 1927); Chem. Abstracts, 21, 3742 (1927); Brit.
Patents 231,800 (Nov. 23, 1925); Chem. Abstracts, 19, 3592 (1925); 341,843 (June 20,
1929); Chem. Abstracts, 25, 4125 (1931).
128 L. Lilienfeld, U. S. Patents 1,674,401 (June 19, 1928); Chem. Abstracts, 22, 2839
(1928); 1,674,405 (June 19, 1928); Chem. Abstracts, 22, 2840 (1928); 1,906,910 (May
2, 1933) ; Chem. Abstracts, 27, 3605 (1933) ; Brit. Patent 231,802 (May 22, 1924).
129 L. Lilienfeld, U. S. Patent 1,674,402 (June 19, 1928); Chem. Abstracts, 22, 2839
(1928).
180 L. Lilienfeld, U. S. Patent 1,881,741 (Oct. 11, 1932); Chem. Abstracts, 27, 601
(1933).
1018 CELLULOSE
If the cellulose xanthoacetic acid, mentioned above, is dissolved in methyl-
aniline,181 a salt, cellulose xanthoacetic acid amine, is formed:
RCeiiOCSSCHaCOOH.HN(CH,)C6H6
Xanthates of hydroxyalkyl derivatives of cellulose are obtained from the
reaction of cellulose with a cyclic ether of a polyhydric alcohol and then
treating the resultant compound with carbon disulfide and alkali.132
Diazomethane converts the xanthate to monomethyl dicellulose:
[CeH702(OH)2.6(OCH8)o.5]n
Diazonium compounds183 also react with cellulose xanthate according to the
equation :
RoeiiOCSSNa + C1N==NR > RceiiOCSSN=NR + NaCl >
RceiiOCSSR + N2 (31)
in which R represents an aromatic nucleus such as Cells.
On treatment with acrylonitrile, cellulose xanthate (in viscose) is readily
converted to cellulose cyanoethyl ether which, in the presence of the NaOH
in the viscose, then hydrolyzes with the formation of the sodium salt of
cellulose carboxyethyl ether :
RceiiOCSSNa + CH2=CHCN > RCeiiOCH2CH2CN >
RceiiOCH2CH2COONa (32)
The cyanoethyl ether may be isolated as such, or the reaction mixture may
be used directly for spinning modified rayons. 134
The reaction of viscose with iodine has been described. The light-
yellow disulfide formed here is insoluble and quite stable, although it is
converted back to the normal cellulose xanthate by sodium amalgam.41
Still other xanthates of cellulose, and other reactions of ordinary cellulose
xanthate, as well as mixtures of cellulose xanthate and cellulose ethers, are
possible.185
181 L. Lilienfeld, U. S. Patent 1,890,393 (Dec. 6, 1932) ; Chem. Abstracts, 27, 1718
(1933).
"» L. Lilienfeld, U. S? Patents 1,910,440 (May 23, 1933); Chem. Abstracts, 27, 4080
(1933); 2,306,451 (Dec. 29, 1942); Chem. Abstracts, 37, 3270 (1943).
188 J. H. Helberger, German Patent 562,180 (Oct. 1, 1931); Chem. Abstracts, 27, 841
(1933).
184 J. P. Hollihan and S. A. Moss, Jr., Ind. Eng. Chem., 39, 929 (1947); J. A. Somers,
Brit. Rayon and Silk J., 26, No. 312, 62 (1950).
185 L. Lilienfeld, U. S. Patents 2,051,051 (Aug. 18, 1936); Chem. Abstracts, 30, 6943
(1936); 2,052,478 (Aug. 25, 1936); Chem. Abstracts, 30, 7341 (1936); 2,100,010 (Nov.
23, 1937); Chem. Abstracts, 32, 779 (1938); 2,163,607 (June 27, 1939); Chem. Abstracts,
33, 8913 (1939); 2,169,207 (Aug. 8, 1939); Chem. Abstracts, 33, 9641 (1939); 2,176,085
(Oct. 17, 1939); Chem. Abstracts, 34, 1172 (1940); 2,176,799 (Oct. 17, 1939); Chem.
Abstracts, 34, 1173 (1940).
G. DEGRADATION OF CELLULOSE DERIVATIVES
L. F. McBURNEY
The susceptibility of cellulose derivatives to degradative processes is a
reflection both of the chemical nature of the cellulose chain molecule and of
that of the substituents along the chain. The extent to which each of these
factors contributes to the total effect is dependent upon the type and
degree of substitution of the individual polymer. For convenience of
discussion it is desirable to group the derivatives into two broad classifica-
tions.
A satisfactory segregation of the common cellulose derivatives can b£
based on their solubility either in aqueous or in nonaqueous media, as is
apparent from a consideration of their solubility properties. The water-
soluble derivatives, in the main, are much less completely substituted than
are the organosoluble types. As a result, the stability of the water-
soluble types is less dependent upon the chemical nature of the substituent
and is more nearly comparable to that of the parent cellulose. The stability
of the organosoluble types, on the other hand, is influenced primarily by
the chemical nature of the substituent and to a much lesser extent by the
cellulose backbone. Both classes, however, are subject to a solvent inter-
action which also plays a part in stability. This effect is more pronounced
with the organosoluble types since a wide variety of solvents or plasticizers
are employed in their use as coatings or plastics.
Both water-soluble and organosoluble cellulose derivatives will undergo
degradation by hydrolytic, oxidative, and microbiological processes. The
relative ease with which this deterioration occurs will vary widely, not
only between the two classes of compounds but also between different mem-
bers of the same class; for example, although both organosoluble ethyl
cellulose and cellulose acetate are susceptible to oxidative degradation by
atmospheric oxygen, cellulose acetate will be perfectly stable at tempera-
tures where ethyl cellulose shows a rapid breakdown. The deterioration of
cellulose derivatives usually manifests itself by a loss in viscosity with a
corresponding loss of tensile strength, the formation of color, and in many
cases marked changes in solubility properties.
The stability of the water-soluble cellulosics will be discussed first in this
1019
1020 CELLULOSE
section with their susceptibility to microbiological, hydrolytic, and oxida-
tive degradation being considered in that order. These derivatives have
achieved commercial importance comparatively recently and, as a result,
information dealing with their susceptibility to deterioration is very in-
complete. The order of presentation has been chosen since it represents
both the importance and level of knowledge of the degradative process in-
volved. The water-soluble cellulosics which have reached commercial
importance today are all ethers of degrees of substitution of two or less;
thus this discussion will serve as a brief introduction to the subject of ether
stability in general.
Solvent-soluble cellulosics, in direct contrast to their water-soluble coun-
terparts, have been utilized for years, and a very considerable literature
has been accumulated in reference to their stability. With these more
highly substituted derivatives, oxidation by molecular oxygen is by far
the most important single degradative reaction for both organic ethers and
esters, with hydrolytic breakdown occupying a secondary role, and thermal
degradation being important only in the case of the inorganic ester, cellu-
lose nitrate. The portion of this discussion devoted to organic-soluble
cellulosics will, therefore, emphasize and attempt to interpret primarily
the processes involved in the oxidative degradation reaction and less atten-
tion will be paid to the other causes of instability with the exception of the
thermal breakdown of cellulose nitrate.
1. Water-Soluble Cellulose Derivatives
The water-soluble cellulose derivatives, such as the sodium salt of car-
boxymethyl cellulose, are finding an increasing commercial utilization in
the food, pharmaceutical, textile, and adhesive industries, where hydro-
philic colloids with suspending, thickening, stabilizing, and film-forming
properties are required. These materials are quite stable to storage in the
dry state; in solution, however, they will undergo deterioration. The
major source of this degradation has been shown to be the attack by a
number of genera of bacteria and fungi of widespread occurrence in nature.
Similar deterioration can result from hydrolytic or oxidative attack.
(a) MICROBIOLOGICAL DEGRADATION
The growth of microorganisms in solutions of these water-soluble de-
rivatives produces a variety of undesirable properties; namely, lique-
faction, cloudiness, discoloration, and odor.1 It is believed that the en-
1 G. G. Freeman, A J. Baillie, and C. A. Machines, Chemistry & Industry, 1948, 279.
IX. DERIVATIVES OF CELLULOSE
1021
zyme which hydrolyzes the 1,4-glucosidic linkage in carboxymethyl cellu-
lose is identical with that responsible for similar attack in cellulose (see
Chapter III-C-5) . The ease of attack by these microorganisms appears to
be independent of degree of polymerization (D.P.) but to be quite sensitive
to the degree of substitution (D.S.) of the material.2
The effect of degree of substitution on ease of enzymatic hydrolysis 'of
carboxymethyl cellulose is illustrated in Figure 76 from which it is apparent
0.4 0.6 0.8 1.0 1.2
DEGREE OF SUBSTITUTION
Fig. 76. Effect of degree of substitution of sodium carboxy-
methyl cellulose on extent of enzymatic hydrolysis to glucose (Reese,
Siu, and Levinson2). Samples were hydrolyzed with filtrates of
Aspergillus fumigatus cultures for 2 hrs. at 50 °C.
that those samples containing one or more carboxymethyl groups per an-
hydroglucose unit are iquite resistant to attack, whereas those containing
unsubstituted units in the cellulose chains will be readily degraded. Evi-
dence of this type is responsible for the conclusion that the degrading
enzyme is the same as that which brings about cellulose degradation and is
capable of utilizing only the unsubstituted glucose residues. Commercial
carboxymethyl celluloses have degrees of substitution below one, and thus
their solutions are susceptible to such attack.
Two general methods are available for control of microbiological deg-
radation: One is the use of heat sterilization; the other is the use of bac-
teriostats. The effectiveness of the latter reagents (with cellulose deriva-
tives) can be seen frem Table 43 where samples of sodium carboxymethyl
cellulose treated with w-cresol and monochloro-3,5-xylenol are compared
2 E. T. Reese, R. G. H. Siu, and H. S, Levinson, /. Bacterial,, 50, 486 (1950).
1022
CELLULOSE
with an untreated sample. The data in Table 43 illustrate the fact that
microbiological degradation can be readily controlled and need not be a
detrimental factor in solution use of water-soluble cellulosics.
TABLE 43
Bacteriostatic Action of w-Cresol and Monochloro-3,5-xylenol in 2% Solutions of
Sodium Carboxymethyl Cellulose in Aqueous Sulfate Medium (Freeman, Baillie, and
Machines1)
No
bacteriostat
w-Cresol
Monochloro-3,5-
xylenol
0.1%
0.2%
0.1%
0 2%
Viscosity at 20°C., centi-
stokes
Initially
67
60
72
72
68
After 10-day incubation
8
52
60
60
56
Bacterial count per ml.
Initially
500
100
100
100
100
After 2 days
250,000
-—
—
_
—
After 3 days
50,000,000
100
100
100
100
After 7 days
50,000,000
80,000
80,000
100
100
After 10 days
200,000,000
100,000
100,000
4,000
1000
pH after 10 days
6.6
8.3
9.0
7.2
6.8
Final appearance of solutions
H2S present
No
No
No
No
FeS precip-
change
change
change
change
itated
(b) HYDROLYTIC DEGRADATION
The sensitivity of water-soluble cellulose derivatives to acid hydrolysis
is closely analogous to that of cellulose itself (see Chapter III-C). It
differs only in that physical structure, which is the controlling factor in
cellulose hydrolysis, has little effect on the water-soluble derivatives since
they are reasonably well dispersed in solution and have lost the charac-
teristic fiber structure. Reduction in degree of polymerization is fre-
quently accompanfed by a reduction in degree of substitution.8'4 In
most cases the chain-cleavage reaction is much more rapid than is the loss
of substituent from the ether because of the greater reactivity of the glu-
cosidic (acetal) type of bond in comparison to the normal ether bond.
The parallelism which exists between the acid hydrolysis of cellulose and
that of its water-soluble derivatives has been clearly demonstrated by
Gibbons6 in his studies on methyl cellulose. This work established that
8 A. Pacault, M. Bouttemy, and O. Tanaevsky, BulL $oc, chim. France^ 1952, 345.
4 T. E. Timell, Svensk Papperstidn., 56, 245 (1953).
5 G. C. Gibbons, /. Textile Inst., 43, T25 (1952).
IX. DERIVATIVES OF CELLULOSE
1023
the activation energy of the hydrolysis was independent of the degree of
substitution and the concentration of the hydrolyzing acid and had a value
of 30,150 cal. The activation energies reported for the hydrolysis of
cellulose solutions in concentrated acids are in the same range, being
29,800 cal. in 51% sulfuric acid6 and 27,260 and 29,600 cal. in phosohoric
acid.6'7
6.5
6/i
N
V
. 0
7 E
N
\i
.81 N H(
f% OCM3
:i
7 0
1.81 N
28.8J I
)CHfS
\
M ' *
M
0
Jfl ,
\
\
o. o
80
:
.19 N 1
18.8$ OC
C\^\
*3 ^
^
9c
\
\
^
0.0030
0. 0032
0. 0034
Fig. 77. Arrhenius plots for hydrolysis of methyl celluloses in HC1 (Gibbons*).
The slopes of the straight lines obtained by plotting the logarithm of the ex-
perimentally determined reaction rate constant k against the reciprocal of the
absolute temperature T all give an activation energy E of 30,150 cal. for the
hydrolysis reaction.
The data in Figure 77 show the constancy of the Arrhenius plot obtained
from two methyl celluloses of different D.S. and show as well the effect of
different acid concentrations. The uniformity of the slopes of the curves
shows the same reaction to be occurring in all cases. These same data
illustrate the dependence of the rate of reaction on the degree of substitu-
tion, from which it can be concluded that the rate of hydrolysis decreases as
the degree of substitution is increased.
« K. Freudenberg, W. Kuhn, W. Diirr, F. Bolz, and G. Steinbrunn, Ber., 63B, 1510
(1930) ; K. Freudenberg and G. Blomqvist, Ber., 68B, 2070 (1935).
7 G. V. Schulz and H. J. L6hmann, /. prakt. Chern., 157, 238 (1941).
1024
CELLULOSE
De-etherification of cellulose ethers can be accomplished by treatment
with hot concentrated mineral acids. More dilute acids or treatment with
concentrated acids at room temperature have generally been considered
only to attack the glucosidic chain bonds and to leave the substituent
ether linkage untouched. Pacault, Bouttemy, and Tanaevsky8 have
reported that sulfuric acid solutions in concentration above 5 N will bring
about the de-etherification of carboxymethyl cellulose at room tempera-
ture as shown in Figure 78. The loss of ether groups is accompanied
by a very considerable change in viscosity as shown in Figure 79.
iS 0.6
2
£ 0.5
H
£ °'4
w 0.3
8"
„ „
Q 0.0
1 Hr.
5 Hrs. .
* A 24 Hrs.
A 55 Hrs.
20
25
5 10 15
ACID NORMALITY
Fig. 78. De-etherification of sodium carboxymethyl cellulose (original D.S. =
0.55) on contact with aqueous sulfuric acid solutions at room temperature for
various lengths of time (Pacault, Bouttemy, and Tanaevsky8).
Timell4 has reinvestigated this de-etherification phenomena because of
its important implications which would invalidate many conclusions re-
lating to the substituent distributions in water-soluble cellulose derivatives.
It is the conclusion from Timell's work that the measured change in D.S.
is the result of excessive destruction of the amorphous regions of the cellu-
lose fibers, which are also more highly substituted with carboxymethyl
groups. These areas are rendered soluble and are lost during the washing
process. Thus the undissolved residue will have a reduced D.S. but the
loss is the result of chain cleavage and not of de-etherification.
The acid lability of the cellulose chain is an inherent property of its
chemical structure, and there is no magic inhibitor or retarder which can be
added to overcome its breakdown. The chain cleavage of the water-
soluble derivatives can be minimized by careful control of the pH of the
solution, either by maintaining the solutions slightly on the alkaline side or
IX. DERIVATIVES OF CELLULOSE
1025
by the use of buffer salts to control the acidity at or near the neutral point.
Storage of acidic aqueous solutions of cellulosics will invariably result in a
loss in viscosity which will increase with the acidity of the solution.
(c) OXIDATIVE DEGRADATION
The end uses of water-soluble derivatives, up to the present time, are
such that behavior of these derivatives toward conditions favorable to
oxidative breakdown has been relatively unimportant. The literature, as a
result, contains no information bearing on their so-called weathering stabil-
ity which includes the effects of elevated temperatures, high alkali concen-
tration, air, and ultraviolet light, conditions which normally lead to oxida-
tive instability in high polymers. Under the influence of these reagents,
Steurer and Mertens8 have shown the organosoluble methyl celluloses to
1.5
JJ 1.0
1-4
>
o
t-t
to
-0.5
w
1 Hr.
5 Hr«.
10 15 20
ACID NORMALITY
25
Fig. 79. Decrease in specific viscosity of sodium carboxyraethyl cellulose
on contact with aqueous sulfuric acid solutions at room temperature for various
lengths of time (Pacault, Bouttemy, and Tanaevsky3).
degrade, and similar effects have been found with the higher substituted
ethyl celluloses, which will be discussed under the organosoluble derivatives.
Cellulose itself is unstable to air and alkali, as shown previously (see Chap-
ter III-C). These facts lead to the conclusion that the water-soluble de-
rivatives also would be labile. The ease of this reaction and extent to
which it will occur remain at present as a relatively unexplored field of
cellulose chemistry which, as new uses for the water-soluble gums are found,
will undoubtedly become of major importance.
8 E. Steurer and H.-W. Mertens, Ber., 74B, 790 (1941).
1026 CELLULOSE
2. Organosoluble Cellulose Derivatives
The industrial utilization of organosoluble cellulose derivatives as pro-
tective coatings, films, plastics, and fibers requires them to be resistant to a
wide spectrum of degradative processes. The most critical specification
which they are called upon to meet is the ability to withstand outdoor
weathering conditions. Under such exposure the cellulose derivatives are
subjected to varying degrees of heat, light, moisture, and oxygen, all of
which are capable of bringing about severe deterioration. The result of
these degradative environmental processes can be lumped together under
the general heading of oxidative instability. Since the various derivatives
differ in their general susceptibility to these factors, they will be discussed
individually.
(a) CELLULOSE ETHERS
The principal organosoluble cellulose ether, from the commercial view-
point, is ethyl cellulose, the derivative which has therefore received the
most intensive investigation. Benzyl cellulose was utilized in Europe
for a time but is no longer of commercial importance. The discussion of the
oxidative degradation of ethyl cellulose will apply equally well to the benzyl
ether; the latter, however, is much more sensitive to oxidation.
(1) Oxidative Degradation
Ethyl cellulose has been shown to undergo considerable embrittlement
upon exposure to oxygen at high temperature.9'10 Axilrod and Kline11
reported that, after three months of outdoor exposure, clear plastic sheets
became opaque and brittle. Reports of this nature resulted in a large
volume of empirical investigation directed toward finding antioxidants or
inhibitors to prevent such deterioration. 12~~17 The effectiveness of such
9 E. Berl and G. Rueff, Cellulosechemie, 14, 44 (1933).
10 H. Staudinger, HrStock, and K. F. Daemisch, Melliand Textilber., 22, 620 (1941);
Chem. Abstracts, 37, 6889 (1943).
11 B. M. Axilrod and G. M. Kline, /. Research Natl. Bur. Standards, 19, 367 (1937).
12 S. L. Bass, L. A. Rauner, and P. H. Lipke, Jr. (to The Dow Chemical Co.), U. S.
Patent 2,^83,361 (Aug. 21, 1943) ; Chem. Abstracts, 39, 5487 (1945).
18 G. M. Kline, Soc. of Plastics Ind. Conference, Los Angeles, Feb. 23, 1943.
14 W. Koch (to Hercules Powder Co.), U. S. Patent 2,389,370 (Nov. 20, 1945) ; Chem.
Abstracts, 40, 1352 (1946); U. S. Patent 2,333,577 (Nov. 2, 1943); Chem. Abstracts, 38,
2489(1944).
15 J. H. Sharphouse and J. Downing (to British Celanese Ltd.), Brit. Patent 578,286
(June 21, 1946) ; Chem. Abstracts, 41, 3295 (1947) ; Brit. Patent 580,359 (Sept. 4, 1946) ;
Chem. Abstracts, 41, 3295 (1947).
IX. DERIVATIVES OF CELLULOSE 1027
materials in preventing breakdown of the plastic was measured in terms of
the percentage of viscosity retention obtained under a given set of con-
ditions of heat and availability of air or oxygen. This work was successful
in permitting formulations to be described17 which would reduce the sensi-
tivity of the product to oxidative deterioration. During these investiga-
tions, several inconsistencies appeared which pointed to the need for a thor-
ough knowledge of the basic mechanism involved in the oxidation.
The failure of the empirical method to provide a satisfactory answer to
the problem can be attributed to three factors. First, viscosity retention
was used as the criterion of stability. This is important from commercial
aspects but since viscosity changes are a result of the degradation reaction,
they are not a direct measure of the reaction causing the breakdown and
cannot be relied upon to tell the complete story. Decomposition could
occur, for example, so that cross-linkage would take place and the viscosity
would either remain unchanged or in the extreme case could actually show
an increase. This effect has actually been observed, and viscosity reten-
tions of 118% or more are reported in the literature.17 Second, an examina-
tion of the structure of the polymer, with its high concentration of ethoxyl
linkages, would lead to the immediate assumption that oxidative degrada-
tion is analogous to the oxidation of simple ethers. From the abundant
literature relative to oxidation of monomeric ethers18"21 it can be concluded
that the oxygen availability during exposure will greatly influence the ulti-
mate viscosity achieved. Third, it is conceivable that the products of
decomposition will also influence the rate of oxidation. The ultimate
viscosity would then depend upon whether or not such materials could
escape from contact with the cellulose derivative.
In order to establish the mechanism of the reaction, McBurney and
Evans22-28 measured the rate of oxygen absorption by ethyl cellulose under
controlled conditions of temperature, pressure, and oxygen concentration.
The initial experiments in this study showed the rate of oxygen absorption
16 J. S. Tinsley (to Hercules Powder Co.), U. S. Patent 2,337,508 (Dec. 21, 1943);
Chem. Abstracts, 38, 3473 (1944); U. S. Patent 2,375,708 (May 8, 1945); Chem. Ab-
stracts, 39, 4521 (1945).
17 B. Berry and W. Koch, Modern Plastics. 25, 154 (Oct., 1947).
» A. M. Clover, /. Am. Chem. Soc., 44, 1107 (1922).
11 N. A. Milas, /. Am. Chem. Soc., 52, 739 (1930) ; 53, 221 (1931).
10 C. Moureu and C. Dufraisse, Chem. Revs., 3, 113 (1926). /. Soc. Chem. Ind., 47,
819,848(1928).
11 H. Wieland and A. Wingler, Ann., 431, 317 (1923).
aa L. F. McBurney, Ind. Eng. Chem.t 41, 1251 (1949).
M E. F. Evans and L. F. McBurney, Ind. Eng. Chem.t 41, 1256 (1949).
1028
CELLULOSE
by ethyl cellulose to be a function of time as well as temperature. This
observation was verified by oxidizing a series of samples under identical
conditions; the samples, however, differed in both chain length and
ethoxyl content. The induction periods for these samples were found to
vary as did the initial rates of absorption. Once the steady state had been
reached, the rates of oxygen absorption were quite similar in all cases as
shown in Figure 80. The steady-state oxidation, therefore, is not a func-
24
28
0 4 8 12 16 20
TIME IN HOURS
Fig. 80. Rate of oxygen absorption by various types of ethyl cellulose at 90 °C.
(McBurney22). Samples were dried, ground films of ethyl cellulose. Ethoxyl
content: curves 1 and 2, 48%; curve 3, 44%; curve 4, 44.5%; curve 5, 49.2%.
Viscosity, 5% solution: curves 1 and 2, 1000 cps.; curve 3, 50 cps.; curve 4, 100
cps. ; curve 5, 100 cps.
tion of either the degree of substitution or the chain length of the polymer.
Some explanation is required, however, for the variance observed in the
induction periods of the different samples.
In reactions of this type an induction period may be an indication either
of the presence of inhibitors, such as transition state metals, which are de-
stroyed as the reaction proceeds, or of a truly autocatalytic process. With
ethyl cellulose, it might be assumed that reaction would be initiated on an
aldehyde group at chain ends. Such a process would follow a reaction
Of CtfLLtTLOSFJ
1029
course as illustrated below :
R«0CHO + 02
[RecCO-1 + [-H02j
» [RecCO(00-)]
(1)
[RecCO] + 02 - » [RecCO(00-)] (2)
[RecCO(00-)] + R«CH - > ReoCO(OOH) + [Rec-] (3)
where Rec stands for the ethyl cellulose residue exclusive of the atoms shown.
u 40 60 120 160 200 240 280 320 360 400
OXYGEN ABSORBED IN Ml Ltl ATOMS PER AHH YOROGLUCOSE UNIT
Fig. 81. Chemical changes accompanying oxygen oxidation of ethyl cellulose at 90 °C.
(McBurney22). Dry, ground samples of ethyl cellulose were oxidized to various levels
of oxygen absorption and the functional group content was measured.
If this interpretation is correct, the length of the induction period would
be a function of the concentration of aldehyde groups as well as of the de-
gree to which the oxidation may have progressed during storage or prior
handling of the ethyl cellulose sample. It is reasonable to assume, conse-
quently, that different samples will show varying induction periods. The
correctness of this hypothesis was established for the ethyl cellulose oxida-
tion by showing that the length of the induction period could be extended
1030
CELLULOSE
by removal of aldehyde end groups and shortened by the addition of free-
radical sources. This evidence indicated that oxidation by molecular
oxygen was autocatalytic and that it proceeded by means of a peroxide
mechanism.
Further substantiation was obtained by measuring the hydroperoxide
development during degradation as well as the formation of carboxyl
groups and the effect of oxidation on the ethoxyl content. These data are
illustrated in Figure 81 where it can be observed that peroxide content
shows a rapid initial increase and reaches a maximal value. At higher
1.00
40 80 120 l«0 200 240 200 320 360 4OO
OXYGEN ABSORBED IN MIUIATOMS PER ANHYDROGLUCOSE UNIT
Fig. 82. Heat degradation of ethyl cellulose at 90 °C. as shown by changes
in intrinsic viscosity and intrinsic fluidity (McBurney22). Samples were dried,
ground films of ethyl cellulose.
degrees of absorption the measured values show a decrease in magnitude.
Concurrently the carboxyls show an initial rapid increase which then
becomes slower although giving no evidence of reaching a maximum value.
The ethoxyl content, on the other hand, shows a short interval of no
change followed by a rapid and then a diminishing rate of decrease.
These findings are of major importance in assigning an over-all mecha-
nism to ethyl cellulose peroxidation. The character of the peroxide for-
mation curve is indicative of the controlling step in the oxidation — that of
peroxide formation by oxygen attack, followed by peroxide decomposition.
DC. DERIVATIVES OF CELLULOSE 1031
In all probability the peroxide decomposition is initially thermal in nature,
since no salts, acids, or bases were present. As reaction progresses, it is
conceivable that further breakdown is the resultant of thermal and acid
catalysis, the latter arising from the carboxyls generated during the process.
In simultaneous reactions of this type, a point is reached at which the rate
of decomposition is so great that the measured peroxide content shows a
decrease, although the rate of oxygen absorption by the sample has shown
no change.
The initial rapid carboxyl development may be a reflection of the in-
duction mechanism previously postulated, in which a peracid is formed from
the aldehyde end groups present in the molecule. The propagation of the
reaction does not require end groups alone, and for that reason the rate of
carboxyl development shows a decrease and approaches a steady state
with oxygen absorption.
This conclusion is further substantiated by the behavior of the ethoxyl
content. The data show clearly that extensive oxidation is accompanied
by a considerable loss in ethoxyl. It is justified, therefore, to assume that a
major portion of the attack must be on the ethoxyl groups present in the
molecule. This reaction can be pictured thus :
[H-H
H— C— OCHCH8 > I H— O-O- | + CH8CHO + [-OH] (4)
O
4
H
The occurrence of this cleavage has been substantiated by isolating acetal-
dehyde in appreciable quantities from the oxidized ethyl cellulose. The
nature of the ethoxyl loss curve also lends further support to the proposed
mechanism in that during the initiation stage of the reaction the ethoxyl
content does not change. Such a behavior would be required since the
initial oxygen attack has been postulated to be on an end group and/or
on an impurity not on the ethoxyl groups.
The structural inhomogeneities in cellulose derivatives make it of pri-
mary importance from mechanistic considerations to establish that the oxy-
gen absorption rates are a true measure of the reaction kinetics and not
simply a measure of rate of oxygen diffusion into the polymer. Measure-
ment of the activation energy of the reaction in the temperature range of
69-90°C. gave a value of 25,000 cal. A value of this magnitude was con-
sidered to be stjfficiently high to indicate that diffusion was not the con-
1032,
CELLULOSE
trolling factor. The random nature of the oxidation received further
support from the intrinsic viscosity and fluidity changes which occur during
oxidation. These data are shown in Figure 82, as is also the linear relation-
ship between the intrinsic fluidity and oxygen absorption in confirmation
of the random nature of the oxidation (see Chapter III-C) .
The discussion of the oxidative stability of ethyl cellulose thus far has
been concerned with two major factors only: heat and air, or oxygen. The
third important variable in outdoor weathering is the effect of light on the
320
0 40 80 120 160 200 240 280 320 360 400
TIME IN HOURS
Fig. 83. Oxidation of ethyl cellulose at 50 °C. with and without ultra-
violet (U.V.) light (Evans and McBurney28). Samples were dried,
ground films of ethyl cellulose (48.2% ethoxyl content; viscosity of 5%
solution in 8:2 toluene :2B alcohol, 14 cps.).
stability of cellulose ethers. Ultraviolet radiation has been found to be a
catalyst for the molecular oxygen oxidation of ethyl cellulose.23 The ex-
tent to which the rate of oxygen absorption is accelerated is shown in Figure
83. With illumination, the induction period is greatly reduced over the
dark reaction, and the rate of reaction at the steady state is also markedly
increased. This figure also serves to illustrate another important consider-
ation, namely, the lack of a post-irradiation effect. The rapid decrease in
oxygen absorption rate upon turning off the illuminating source is evidence
that, in this system, no post-irradiation effect is present. The develop-
IX. DERIVATIVES OF CELLULOSE
1033
ment of peroxides and carboxyl groups and the loss of ethoxyl are quite
similar with or without ultraviolet as was the change in viscosity and
fluidity.
The general similarity of data for the samples oxidized with and without
ultraviolet light indicates that there is no essential difference in the mech-
anism of oxidation. The only striking dissimilarity lies in the greatly
accelerated rate of oxygen absorption in the presence of ultraviolet. This
conclusion is similar to that expressed by Milas19 for similar autoxidation
of monomeric ethers. The ultraviolet behaves entirely as a catalyst as is
to be expected if the reaction is chain propagative in character.
From the facts which have been discussed and also by analogy between
this and similar reactions,24 it is possible to deduce a reasonable mech-
anism for the oxidation of ethyl cellulose as follows :
Initiation
RecCHO + O2
[RecCO •] +
[RecCO(OO-)] + H— C— OCH2CH3
cCO •] -f [-OOH]
[lUCO(OO-)]
(5)
(6)
R,,CO(OOH) +
H— C— OC
L I
Propagation
H— C— OCHCHs -f O-.
OCHCH3
OCHCHt
(7)
(8)
O
H— C— OCHCH3
. I !
O
I
O
•M-
-OCHoCH*
f !
H— c— oc:
L i •
I
— C— OCHCH3 -f- I H— C—OCHCHs I (9)
! ! L I ' J
O
O
I
H
24 W. A. Waters, Ann. Repts. on Progress Chem. (Chem. Soc. London). 42, 130 (1945).
1034
Decomposition
OCHCH3
I J
O
CELLULOSE
:— OCHCHs -M-OH]
O
(10)
H— C— O- + CH,CHO
L I J
(11)
RH
•OH]
-> H—C—OH + [R-l (RH - acetaldehyde (12)
| or ethyl cellulose)
C=O + H2O
! ¥ H— C— OR (cross-linkage)
rearrangement
chain scission
(13)
(14)
(15)
On the basis of this mechanism, the initiation of the reaction occurs on
chain ends or aldehyde groups. Propagation of the oxidation does not
require those end groups; rather it occurs on the ethoxyls along the chain.
Chain termination is the resultant of the complex decomposition reaction
of the celluloseoxy free radical. Reactions 12, 13, and 14 are hypothetical
in that it is not possible to isolate materials corresponding to those indi-
cated in the reaction scheme. Reaction 14, which would lead to the pos-
sibility of cross-linkage, has qualitative support in the reports of viscosity
retentions in excess of 100% under certain conditions.17
The effectiveness~of antioxidant can be seen by reference to Figure 84
in which the monobenzyl ether of hydroquinone has been employed as a
retarder. The data in the graph show that the inhibitory effect is a func-
tion of concentration for the very low concentration and does not increase
proportionally with increasing amounts.
In summary, then, it is possible to draw the following conclusions rela-
tive to the mechanism of ethyl cellulose oxidation :
1. Oxidation proceeds by means of a peroxide-catalyzed chain-propa-
gated mechanism.
2. The reaction appears to be initiated on the end groups or aldehydes
IX. DERIVATIVES OF CELLULOSE
1035
present either initially in the chemical cellulose or developed during the
ethylation reaction.
3. Chain propagation is accomplished by reaction on the ethoxyl
groups, substituted along the polymer chain, to form hydroperoxides.
4. Formation of carboxyl groups, loss of ethoxyl, and polymer-chain
scission are a result of the decomposition of ether hydroperoxides by ther-
mal-catalyzed and/or acid-catalyzed mechanisms.
140
100
HO RCTARDEft
SO
60
40
0.026* RETARDER
0.331 RETARDER
20 24 28
TIME IN HOURS
Fig. 84. Retardation of ethyl cellulose oxidation at 90°C. by hydroquinone mono-
benzyl ether (McBurney22). Retarder was incorporated in the ethyl cellulose samples
prior to drying and grinding.
5. Oxidation is subject to positive catalysis by free-radical sources or
ultraviolet light, and can be inhibited or retarded by typical free-radical
chain terminators.
(b) CELLULOSE ORGANIC ESTERS
The organic esters of cellulose are the most stable of the cellulose deriva-
tives under normal conditions of aging. Samples have been reported26
to have retained their initial properties after storage for 25 years. At ele-
28 L. C16ment and C. Riviere, Congr. Mm. ind.t Compt. rend. 18me Congr., Nancy,
1938, 703; Chem. Abstracts, 33, 6042 (1939).
1036 CELLULOSE
vated temperatures, cellulose acetate, for example, may suffer degrada-
tion,25""29 but it is many orders of magnitude more stable, under comparable
conditions, than cellulose nitrate or cellulose ethers. Prolonged exposure of
cellulose ester compositions to sunlight or to ultraviolet light brings about
rather severe deterioration and produces a highly colored material with
greatly reduced mechanical strength.30""35 At relatively high temperatures,
for example 160°C., cellulose acetate can be oxidized in a manner compa-
rable to the cellulose ethers with a corresponding loss in desirable proper-
ties.86
(1) Oxidative Degradation
Cellulose acetate, in contrast to materials such as alkali cellulose, ethyl
cellulose, or benzyl cellulose, is quite resistant to oxidative decomposition,
even at fairly high temperature; for example, ethyl cellulose will oxidize
rapidly in an oxygen atmosphere at 90°C., whereas under similar conditions
cellulose acetate is quite stable. If, however, the temperature is suffi-
ciently high, for example 160°C., cellulose acetate will also oxidize as shown
by a loss in strength, increase in color, and solubility changes.
The mechanism of the oxidation of this cellulose ester is rather obscure,
in contrast to ethyl cellulose or similar ethers where the mode of oxygen
attack can readily be understood. It is not possible to apply a similar
mechanism to cellulose acetate oxidation. One of the major products of
the acetate decomposition is carbon monoxide; hence, in order to follow
the rate of oxygen absorption in closed systems, it is necessary to provide a
means for the constant removal of this volatile decomposition product.
It is also necessary that other materials such as water, carbon dioxide, and
acetic acid be removed, because they too can influence the shape of the
28 J. R. Hill and C. G. Weber, J. Research Nail. Bur. Standards, 17, 871 (1936).
27 P. Krais, Leipzig. Monatschr. Textil-Ind., 43, 257 (1928); Chem. Abstracts, 22,
4828 (1928).
28 C. J. Staud, Paint, Oil and Chem. Rev., 89, No. 18, 8 (1930).
29 W. Wehr, Kolloid-Z., 88, 185 (1939). .
80 V. A. Karfunkel and D. A. Fedorov, J. Applied Chem. ( U. S, S. R.), 13, 1375 (1940) ;
Chem. Abstracts, 35, 2715 (1941).
31 M. DeBuccor, Papeterie, 63, 49 (1941).
82 R. E. Montonna and C. C. Winding, Ind. Eng. Chem., 35, 782 (1943).
88 S. Oguri, M. Takei, and N. Fujita, J. Soc. Chem. Ind., Japan, 42, Suppl. binding,
54(1939).
84 T. S. Lawton, Jr., and H. K. Nason, Ind. Eng. Chem.t 36, 1128 (1944).
86 L. W A. Meyer and W. M. Gearhart, Ind. Eng. Chem., 37, 232 (1945).
86 E. F. Evans and L. F. McBurney, Ind. Eng. Chem., 41, 1260 (1949).
IX. DERIVATIVES OF CELLULOSE
1037
oxidation curves. When such precautions are taken, the curves of oxygen
absorption versus time have the general characteristics shown in Figure 85.
Curves 2, 3, and 4 were obtained under conditions of complete volatile
by-product removal, whereas for curve 1 the carbon monoxide was not
removed. This will serve to indicate the major effect such a volatile
360
320
OXIDATION FLASK FLUSHED
WITH OXYGEN
70
8O
90
100
40 SO 60
TIME IN HOURS
Fig. 85. Effect of carbon monoxide on oxidation of cellulose acetate at 160 °C.
(Evans and McBurney36). Curve 1 : film-type granular cellulose acetate sample (55.5-
56.5% combined acetic acid content), carbon monoxide not removed; Curve 2: same
sample, carbon monoxide removed; Curves 3 and 4: plastics-type, granular cellulose
acetate sample (52.5-53.5% combined acetic acid content), carbon monoxide removed.
product will have on the apparent oxidation rate. In general, these curves
are quite typical of an autocatalytic free-radical type of process. They
show an induction period followed by a linear zero-order reaction which
is characteristic of such processes.
The absorption of oxygen by cellulose acetate is accompanied by ex-
tensive chemical and physical changes in the molecule. In Figure 86 is
1038
CELLULOSE
illustrated the production of volatile products by this reaction. The upper
curve is a composite of all volatiles produced (that is, water, carbon dioxide,
and acetic acid) exclusive of the carbon monoxide. The development of
the latter is shown by the lower curve. Whereas the production of vola-
tiles other than carbon monoxide appears to be decreasing in rate with
degree of oxidation, the carbon monoxide evolution gives evidence of being
autocataly tic ; it is constantly increasing with degree of oxidation.
The source of this carbon monoxide is not definitely known. Since
a-ketocarboxylic acids87 are known to give carbon monoxide upon pyrolysis,
WATER, CARBON DIOXIDE,
ACETIC ACID
80 120
OXYGEN ABSORBED
160 2OO 240 200 320 360
IN Ml LLl ATOMS PER AHHYDROQLUCOSE UNIT
Fig. 86. Formation of volatile products during cellulose acetate oxidation at
160°C. (Evans and McBurney36). Sample was granular cellulose acetate
with 60% combined acetic acid content.
however, it is conceivable that such a structure is produced within the cellu-
lose acetate molecule by oxidation, and the production of the carbon monox-
ide is then a secondary pyrolytic reaction of the oxidation.
From commercial aspects, freedom from color is an important factor in
plastics applications. Oxidation of cellulose acetate has been found to be
deleterious in this respect. From Figure 87 it can be seen that color for-
mation during oxidation is a linear function of degree of oxidation over the
87 C. D. Kurd, The Pyrolysis of Carbon Compounds, Chemical Catalog Co.. New York,
1929, pp. 556-9.
IX. DERIVATIVES OF CELLULOSE
1039
>ntire range studied. Figure 87 also shows the changes in combined acetic
icid content which occur for the same period of oxidation. These com-
bined acetic results are somewhat in doubt. The values were obtained by
;wo methods (see Chapter XII-B-2) — the saponification method and the
nethod of Cramer, Gardner, and Purves38 — and both methods gave simi-
ar results; that is, a trend toward higher combined acetic acid content as a
•esult of oxidation. The need for this cross-checking was caused by the
Dossibility of the formation of cellulosic carboxyls which would consume
40 60 120 160 200 240 26O 320 960 4OO 44O
OXYGEN ABSORBED IN MIUIATOMS PER AMHYOROQIUCOSE UNIT
Fig. 87. Color and combined acetic acid content of cellulose acetate oxidized at
160°C. (Evans and McBurney36). Sample was granular cellulose acetate with 60%
combined acetic acid content.
dkali in the saponification procedure and result in high values for combined
icetic acid.
The viscosity of the cellulose acetate was lowered by oxidation as shown
n Figure 88. In this plot the intrinsic viscosity and the change in intrinsic
luidity are shown. Here again from the commercial utility viewpoint,
the change in viscosity is of primary importance; from the viewpoint of the
nechanistn of oxidation, however, the nature of the fluidity change curve is
nore revealing. In this case the change in intrinsic fluidity is linear up
to an oxygen absorption of 240 milliatoms per anhydroglucose unit, after
88 F. B. Cramer, T. S. Gardner, and C. B. Purves, Ind . Eng. Chem., Anal. Ed., 15, 319
;i943)<
1040
CELLULOSE
which the value begins to decrease. Since the intrinsic fluidity and chain
breaks are related, it is apparent that the reaction has shifted from a normal
random reaction to a more specific type of attack. The oxygen attack
may, for example, have shifted to chain ends; consequently, the number of
chain breaks for a given oxygen absorption would be greatly reduced.
Since the mechanism of this oxidation is not known, such possibilities are
purely speculation.
80 120 160 200 240 200 320 360 40O 440
OXYGEN ABSORBED IN MILLIATOMS PER ANN YDR06LUCOSE UNIT
Fig. 88. Degradation of cellulose acetate on oxidation at 160°C. (Evans and Mc-
Burney36). Sample was granular cellulose acetate with 60% combined acetic acid
content. The intrinsic fluidity change, A [<£], is the difference between the reciprocal of
the intrinsic viscosity of the oxidized sample and that of the original cellulose acetate.
The autoxidation of cellulose acetate can be catalyzed by the presence
of small amounts of acidic materials such as sulfuric acid. This is illus-
trated in Figure 89. Curves 5 and 6, for example, show the effect of the
addition of 0.001% and 0.01% of sulfuric acid to the cellulose acetate sample
whose basic oxidation rate is illustrated by curve 3. It is to be noted that
these curves show a considerably different initial behavior from those in
Figure 85; an immediate absorption of oxygen is followed by a period of
relative inhibition, after which the typical free-radical absorption becomes
dominant. In the case illustrated in Figure 89, the acetate had been cast
as a film in order to permit the uniform introduction of trace amounts of
catalysts. Even though these films were dried rigorously, it is apparent
that some solvent has been tenaciously retained and has changed the initial
absorption rates. This can be appreciated by a comparison of curves 1 and
3 in Figure 89. Curve 1 is a cellulose acetate flake as obtained from a plant
IX. DERIVATIVES OF CELLULOSE
1041
batch, and curve 3 is the same material dissolved in acetone, cast as a film,
dried and ground, then oxidized.
The data discussed thus far have emphasized the role of oxygen in the
thermal degradation of cellulose esters. It is not to be implied, however,
360
30 40 50
TIME IN HOURS
Fig. 89. Effect of acids and retained solvents on oxidation of cellulose acetate films at
160 °C. (Evans and McBurney86). Curve 1: flake or granular commercial plastics-
grade cellulose acetate with 52.1% combined acetic acid content; Curve 2: same sample
as curve 1, but dissolved in acetone, dried in vacuum oven at 105°C. for 24 hr., ground,
and then oxidized ; Curve 3 : same as curve 2, but dried in vacuum desiccator at room
temperature; Curve 4 : same as curve 3, but with 0.1% of a previously oxidized sample
added; Curves 5 and 6: 0.001 and 0.01%, respectively, of H2SO4, based on cellulose
acetate, added to cellulose acetate film.
that oxygen is the sole contributor to the thermal instability of these poly-
mers. The presence of another nonoxidizing mechanism is apparent from
the data in Table 44 from the work of DeCroes and Tamblyn.89 In this
89 G. C. DeCroes and J. W. Tamblyn, Modern Plastics, 29, 127 (April, 1952).
1042 CELLULOSE
instance the cellulose acetate butyrate was used as an unplasticized powder
to avoid any complications due to solvent or plasticizer interaction. The
values in Table 44 indicate that in the absence of oxygen this ester will
decompose thermally to a very considerable degree; in an atmosphere of
oxygen, however, the degradation will be greatly increased.
TABLE 44
Oxidative and Thermal Degradation of Cellulose Acetate Butyrate (DeCroes and
Tamblyii89)
Degrada-
tion in
Apparent
molecular
acetyl
weight,
Conditions of exposure
Atmosphere
content, %
Color
%
Unexposed
—
34.7
White
0
Exposed 24 hrs.
at 150°C.
N2
36.6
Brown
64
02
37.2
Brown
81
Exposed 24 hrs.
at 180°C.
N2
37.2
Black
48
0,
37.2
Black
90
In actual practice, cellulose esters are always used in combination with
plasticizers, which may exert a marked influence on the stability of the
resulting plastic. Some of the more commonly used materials show a re-
markable instability to oxygen and thus may actually behave as catalysts
for the oxidation of the polymer. The oxidizability of some of these com-
pounds is shown in Table 45. It is interesting to note that those plasti-
cizers, such as dimethyl phthalate and triphenyl phosphate, which are
TABLE 45
Oxidizability of Plasticizers (DeCroes and Tamblyn39)
Conditions: 25 ml. (or 25 g.) of plasticizer shaken under oxygen for 3 hr. at 150°C.
Plasticizer
Apparent
oxygen
absorbed, ml.
Acid
produced ,
milliequiv.
Peroxide
produced,
milliequiv.
Dimethyl phthalate
0
0
0
Diethyl phthalate
26
2.0
1.3
Dibutyl phthalate
52
2.3
0.5
Bis(2-ethylhexyl) phthalate
42
1.5
0.7
Dibutyl sebacate
67
1.7
1.6
Tripropionin
20
1.8
1.0
Triphenyl phosphate
0
0
0
inert to oxygen do not contain in their molecules either methylene or
methylidene groups, which are known to be readily susceptible to oxidative
IX. DERIVATIVES OF CELLULOSE 1043
attack.40'41 The ability of sensitive plasticizers to induce oxidation in cellu-
lose acetate butyrate is illustrated in Table 46. These illustrations will
suffice to show how the inherent oxidative sensitivity of organ osoluble
TABLE 46
Induced Oxidative Degradation of
Cellulose Acetate Butyrate (DeCroes and Tamblyn89)
Conditions: 1 g. of ester in 25 ml. (or 25 g.) of plasticizer heated under oxygen for 3 hrs.
at 150°C.
Plasticizer
Apparent oxygen
absorbed, ml.
Degradation in
molecular
weight, %
None
0
0
Dimethyl phthalate
0
11
Triphenyl phosphate
0
1.5
Diethyl phthalate
29
86
Dibutyl phthalate
32
69
Bis(2-ethylhexyl) phthalate
84
76
Dibutyl sebacate
66
71
Tripropionin
39
80
cellulose derivatives can be compounded and magnified by the plasticizer
system with which they are formulated.
Stability can be complicated even more by the presence of substances
which will act as catalysts for the reaction. Included among these com-
pounds are a number of heavy metal soaps such as cupric, ferric, nickelous,
chromic, and silver stearate as well as pigments such as titanium dioxide.
The deleterious effects of plasticizers and catalysts fortunately can be
greatly reduced by the use of antioxidants or inhibitors such as 2,6-di-fer/-
butyl-£-cresol, J\f-ter/-butylaniline, Af-phenylglycine, and 2- [TV-ethyl-TV- (p-
nitro-£-phenylazoanilino)]ethanol. The importance of proper choice of
stabilizer and plasticizer systems in compounding cellulose ester plastics
cannot be overemphasized.
Ultraviolet light plays as important a role in the degradation of cellulose
esters exposed to outdoor weathering as it does with the cellulose ethers.
Some breakdown is experienced by the esters in an inert atmosphere, but it
is quite minor in comparison to the effect obtained with air or oxygen pres-
ent. The magnification of the ultraviolet instability of cellulose acetate
butyrate by the oxygen in air can be seen from the data in Table 47. Plasti-
cizers will have an effect in the ultraviolet-catalyzed oxidation similar to
<° R. Criegee, H. Pilz, and H. Flygare, Ber.t 72B, 1799 (1939).
41 P. George and A. D. Walsh, Trans. Faraday Soc.9 42, 94 (1946).
1044 CELLULOSE
that shown in the heat-catalyzed reaction. With ultraviolet, a further com-
plication is introduced in that there are considerable variations in the ultra-
violet absorptive capacity of the various plasticizers used in formulation.
Meyer and Gearhart85 have reported that aromatic plasticizers even con-
tribute some protection against breakdown by virtue of their increased
absorption over that of the aliphatic plasticizers. Organic inhibitors such
as phenyl salicylate are remarkably effective in stabilizing cellulose esters
toward ultraviolet-induced oxidation.
TABLE 47
Breakdown of Cellulose Acetate Butyrate in
Sealed Quartz Tubes Exposed to Ultraviolet Light
(DeCroes and Tamblyn39)
Exposure, 800 hrs. Degradation in
in Weather-Ometer molecular weight, %
Unprotected 44
Sealed in air 44
Sealed in nitrogen 12
Egerton42 has investigated the photochemical degradation of cellulose
acetate rayon yarns dyed with dispersed and soluble acetate dyes. The
effect of such substances on the stability of the yarn is quite minor, in
contrast to the corresponding instability observed with dyed cotton (see
Fig. 33, Chapter III-C-4). In several cases the loss in strength of the dyed
yarn is less than that of the undyed acetate rayon exposed under identical
conditions. The extent to which the degradation does occur is largely
dependent upon the relative humidity of the surrounding atmosphere;
in most cases loss in strength is negligible at 0% relative humidity. The
maximum deterioration is usually observed at a relative humidity of 100%
as shown in Table 48.
There does not appear to be any good correlation between the color
shade and its sensitizing action on the acetate rayon, a behavior which is
distinctly different from that observed with dyed cotton (see Fig. 33,
Chapter III-C-4). These data can be interpreted to indicate that the
breakdown is the result of an oxidation process in which hydrogen peroxide
is formed and then actually causes the oxidative degradation. The rela-
tive inertness of cellulose acetate toward hydrogen peroxide explains the
minor effects observed on its tensile strength, and the effect of high humidity
is understandable on the basis of its being required for the production of
hydrogen peroxide.
42 G. S. Egerton, Am. Dyestuff Reptr., 38, 608 (1949).
IX. DERIVATIVES OF CELLULOSE 1045
The outdoor weathering stability of cellulose acetate can be summarized
as follows:
1. Cellulose acetate will undergo a heat-catalyzed and light-catalyzed
oxidation which will result in an increase in color, a loss in tensile strength,
and the production of volatile by-products such as CO, CO2, water, and
acetic acid.
TABLE 48
Loss in Tensile Strength of
Solacel-Dyed Cellulose Acetate Yarns (Egerton42)
Conditions: Exposed to sunlight for 4 months (June-Sept.)
Loss in tensile strength, %
Solacel Dye
0%
relative humidity
100%
relative humidity
None
14
21
Fast Yellow G
0
1
Fast Orange 2GK
0
3
Fast Scarlet B
1
3
Fast Crimson B
0
5
Fast Red 5BG
2
3
Violet B
7
10
Violet R
0
5
Fast Green 2G
0
5
Fast Blue 2B
6
4
Navy Blue G
0
7
2. Cellulose acetate will also undergo a strictly thermal decomposition
with deleterious effects comparable to those above.
3. Plasticizers, heavy metal soaps, pigments such as titanium dioxide,
and acids such as bound sulf uric acid may act as positive catalysts for the
reaction.
4. The oxidative breakdown can be inhibited by the proper choice of
antioxidants and plasticizer systems.
5. The exact mechanism of the reaction is not known with any degree
of certainty.
(c) CELLULOSE INORGANIC ESTERS
Cellulose nitrate (also called nitrocellulose) is the only inorganic cellu-
lose ester of commercial importance. It is unique among the industrially
significant high polymers because of its ability to decompose exothermically
without the participation of oxygen. In spite of this, it is satisfactorily
stable und&r ordinary conditions if high temperatures are avoided, alkaline
1046
CELLULOSE
or even slightly basic materials are not used in its formulation, and there is
reasonable protection from ultraviolet light.
Solutions of cellulose nitrate are known to undergo an aging reaction
which results in a lowered viscosity (see Section B of this Chapter IX).
Both Wehr43 and Campbell and Johnson44 have reviewed the literature
relative to this phenomenon and have drawn the following general con-
clusions :
"(1) Cellulose nitrate solutions, on standing in the dark, show a decrease
in viscosity. (2) The decrease in viscosity is accelerated by standing in
strong light. (3) The decrease in relative viscosity is more marked the
1.50<
25
200 225
50 75 100 125 ,150 175
ETXPOSURE TIME IN HOURS
Fig. 90. Degradation of cellulose acetate and cellulose nitrate in the
presence of ultraviolet light at 60 °C. (Lawton and Nason84). Samples
of cellulose acetate (38.0% acetyl content) and cellulose nitrate (11.10%
N) cast as 0.005-in. films were exposed to ultraviolet light in alternate
atmospheres of nitrogen and oxygen.
more concentrated the solution. (4) The decrease in viscosity is more
marked in the early than in the late stages of the aging process. (5) The
decrease in viscosity has been observed with solutions in acetone and in
many other solvents.1'
(1) Oxidative Degradation
The breakdown of cellulose nitrate in film form under ultraviolet irradia-
tion is greatly accelerated by the presence of oxygen. This can best be
« W. Wehr, Kolloid-Z., 88, 185, 290 (1939),
44 H. Campbell and P. Johnson, J. Polymer Sd.. 5, 443 (1960).
IX. DERIVATIVES OF CELLULOSE 1047
illustrated by examining the data in Figure 90 taken from the work of
Lawton and Nason.34 In this example, cellulose acetate and cellulose
nitrate films are compared under the same conditions. The greatly in-
creased rate of viscosity change in the atmosphere of oxygen is readily
apparent.
Inorganic peroxides also have a catalytic effect on cellulose nitrate deg-
radation similar to that of oxygen. The common oxidation and poly-
merization inhibitors of the neutral type, such as hydroquinone, are effec-
tive in retarding the rate of decomposition.
The action of oxygen in these ultraviolet-catalyzed degradations may be
secondary. The denitration reaction produces NOa and HNOa as well as
organic reducing materials. The latter convert the NO2 to NO. In all
probability, the oxygen acts merely as a reagent for the reconversion of
NO to NOa and HNOs. The latter compounds will cause further degrada-
tion of the cellulose nitrate and liberation of nitrogen oxides, instigating
an autocatalytic process. It is more likely, therefore, that oxygen cannot
be considered as a primary factor in cellulose nitrate deterioration, but
merely as a secondary contributor to the over-all mechanisms.
(2) Thermal Degradation
The chief source of degradation of cellulose nitrate is its extreme thermal
sensitivity. Nitric esters are very similar to organic peroxides in the chem-
istry of their decomposition. In both cases, there is a weak link which
can be broken with an activation energy of approximately 40,000 cal.
This analogy is illustrated by comparing the decomposition of ter/-butyl
peroxide (equations 17 and 18) and cellulose nitrate (equations 19-21).
H,Cv /O - Ov xCH8 THsCv X>- 1
)C( ;C( - > 2 )C( (17)
H8CX \CH3 H8CX \CH3
C=0 + [-CH,] (18)
Acetone is produced by the decomposition, and the free methyl radical
disappears by recombining to form ethane or by further reaction with other
materials present in the reaction system. Extending these observations to
the case of nitric esters leads to the following reaction sequence45:
K-i
-A
H— C— ONOj
H
— C
I
L. Phillios. Nature. 160. 753 C1947).
H— C— ONOf
+ NO, (19)
1048 CELLULOSE
H— C— O-
— ONO2
H— C=O + H— C— ONO2 (20)
I L I J
H— C— ONO2 > H— C=0 + NO, (21)
L I J I
The organic free radical formed in reaction 19 has the same general struc-
ture as that obtained from ter/-butyl peroxide and is many times more
reactive than the NO2 also formed. The reactivity is such that all these
processes occur simultaneously to bring about carbon-carbon bond cleav-
age and the production of aldehydes or ketones. It is possible that the
free radicals will dehydrogenate other organic matter to produce alcohols,
or they may add to unsaturated compounds.
The cleavage reaction results in a rapid lowering of chain length, and the
addition reaction can result in cross-linking with subsequent insolubiliza-
tion in some systems. This spontaneous decomposition is not amenable
to inhibition, and will progress in the case of cellulose nitrate at the rate
of about 1% per hour at 135°C. Reduction in temperature is effective
in slowing down the decomposition so that at room temperature degrada-
tion is negligible. As in the case of hydroperoxides, many substances are
capable of accelerating the rate of degradation of cellulose nitrate. These
include easily reducible materials such as aldehydes, as well as acids and
bases.
Lucas and Hammett46 have studied the base-catalyzed decomposition
of benzyl nitrate and have shown that two first-order reactions are occur-
ring:
C«H6CH2ONO2 IOH] > C6H6CH2OH + [NO3] - (22)
C«H6CH2ON02 f°H] ) C6H5CHO + [NO2] - + H2O (23)
It is an interesting speculation to extrapolate these findings to the case of
cellulose nitrate, and from them to develop a possible mechanism for the
base-catalyzed degradation reaction. Rewriting these two equations for
cellulose nitrate leads to the following:
H ONOt H OH
I I
-cH/9 9\?
[OH] I/ OH H\l
* C\H /c + 2INO»]~ C24)
H C 0 °~
CH2ONOj CHaOH
46 G..R. Lucas and L. P. Hammett, /. Am. Chem. Soc.t 64, 1928 (1942).
IX. DERIVATIVES OF CELLULOSE 1049
ONO» H 0
-C. H _ -0-./4 \H (26)
H Q [OH1 * Cv gH \j + 2[NOJ- + 2H.O
I I
CH2ON02 CHO
Reaction 24 is the simple ester hydrolysis reaction and would be expected
to be nondegrading, since the acid formed would be neutralized by the hy-
drolyzing base. Reaction 25, on the other hand, could result in the pro-
duction of an aldehyde group at the 6- position, a ketone group at the 2- or
3- position, or both. Such a compound would then be extremely sensitive
to alkaline cleavage according to the mechanism discussed in Chapter
III-C-3. Chain scission would occur as shown :
OH 0 O OH
I®"©
®CHO ®CHO
A. Ketone at carbon atom 2 B. Ketone at carbon atom 3
In the case A where an aldehyde group is present at position 6 and a ke-
tonic group at 2, they are both in the beta configuration to the 4-glucosidic
bond and would doubly enhance the alkali sensitivity of that glucosidic
link. In the other case, B, the ketonic group is beta to the 1-glucosidic
linkage and the aldehyde is beta to the 4-glucosidic bond; thus, both chain
bonds are made alkali labile and will cleave so that the oxidized glucose
unit is effectively chopped out of the chain. It is a distinct possibility
that the mechanism of attack by the hydroxyl ion in these hydrolyses in-
volves the hydrogen atom on the carbon atom attached to the nitric ester
group. This attack would quite likely result in activating the /3-glucosidic
linkage toward alkaline cleavage without first requiring the conversion
of the ester group to either aldehyde or ketone. In any event the actual
cleavage mechanism would be the same, since it is the direct effect of beta
activation by electronegative groups. The stabilization of cellulose nitrate
has been exhaustively investigated, and a wide variety of compounds have
been proposed for this purpose. Included in such a list are substances
1050 CELLULOSE
such as chalk, sodium silicate, or sodium carbonate, 47~BO ammonia,61
petroleum jelly,52"*58 glucosides57 containing unsaturates, phenanthrene,56
tartaric and other polybasic acids,58 amyl alcohol,59'60 dimethylaniline,60
dicyandiamide,61 carbazole,62""68 diphenylbenzidine,64 and triphenylamine.60
The effectiveness of most of these compounds depends upon their ability
to consume NC>2 and HNOs by nitration reactions, forming inactive prod-
ucts. It should be noted that these stabilizers do not slow down the initial
decomposition reaction of the nitrate; they simply prevent the occurrence
of a self-accelerating reaction. Stabilizers of this type would not be ex-
pected to be effective unless the products of decomposition were N(>2 or
HNOs. It would be anticipated, therefore, that the presence or absence of
air or oxygen would have a marked influence on the effectiveness of such
substances, because of the necessity for oxidizing the volatile decomposition
products to an absorbable state. Weak acids are added to neutralize
small quantities of basic substances which might be present and which
would bring about the alkaline cleavage discussed previously.
Commercial cellulose nitrate is commonly prepared by nitrating cellulose
with mixed nitric and sulfuric acids, and as a result the initial product con-
tains some bound sulfuric acid as its half -ester (see Section B of this Chapter
IX). When present in the final product this group can hydrolyze off,
47 F. A. Abel, Proc. Roy. Soc. (London). 15, 417 (1867).
« lenk, Centralblitt, Neue Folge, 9, 906 (1864).
49 Lenk, Centralblatt, Neue Folge, 11, 570 (1866).
60 T. H. Pelouze and Maurey, Compt. rend., 59, 363 (1864).
61 R. E. Reeves and J. E. Giddens, Ind. Eng. Chem., 39, 1303, 1306 (1947).
62 L. Monti, D. Dinelli, and F. Buni, Gazz. chim. ital, 63, 713 (1933) ; Chem. Abstracts
28, 3903 (1934).
68 M. Tonegutti, Z. ges. Schiess- u. Sprengstoffw., 21, 127 (1926); Chem. Abstracts,
21, 1185(1927).
54 M. Tonegutti, Ann. chim. applicata, 22, 620 (1932); Chem. Abstracts, 27, 1176
(1933).
85 M. Tonegutti, Z. ge's. Schiess- u. Sprengstoffw., 32, 300 (1937); Chem. Abstracts,
32, 1099(1938).
68 M. Tonegutti and E. Brandimarte, Atti V congr. nazl. chim. pura applicata Rome,
1935, Pt. II, 916 (1936) ; Chem. Abstracts, 31, 7649 (1937).
67 R. Poggi, Ann. chim. applicata, 21, 500 (1931) ; Chem. Abstracts, 26, 1788 (1932).
68 C. Krauz and A. Majrich, Chem. Obzor, 8, 213 (1933); Chem. Abstracts, 28, 4907
(1934).
69 E. Berger, Bull. soc. chim. France, [4], 11, 1049 (1912).
60 M. Marqueyrol, Mem. poudres, 23, 128, 158 (1928).
•l J. Walter, Z. angew. Chem.t 24, 62 (1911).
« C. E. Reese, U. S. Patent 1,358,653 (Nov. 9, 1921).
" R. Dalbert, Mint, poudres, 28, 147 (1938).
" A. Douillet and ft. Ficheroulle, Mim. poudres, 27, 105 (1937).
IX. DERIVATIVES OP CELLULOSE 1051
giving sulfuric acid and initiating the autocatalytic decomposition out-
lined previously. One of the major problems of cellulose nitrate manu-
facture is therefore to eliminate even the slightest trace of bound sulfuric
acid and to wash it out of the fiber. If the half-ester is not completely
removed, some stabilization may be obtained by exactly neutralizing that
which is retained. Extreme care should be taken when using such an
expedient since an excess of base will activate the alkaline instability of the
polymer.
The stability characteristics of cellulose nitrate can be summarized by the
following:
1. Thermal degradation of cellulose nitrate is a spontaneous reaction
whose rate is temperature dependent and which cannot be prevented by
antioxidants.
2. The products of decomposition, NO2 and HNO3, unless removed, will
initiate a further autocatalytic phase of the decomposition.
3. The formation of carbonyl groups on the cellulose chain as a result
of thermal denitration sensitizes the chain to alkaline cleavage.
4. Oxygen plays only a very minor part in the degradation of cellulose
nitrate in contrast to its importance with the other cellulose derivatives.
5. Cellulose nitrate is subject to photochemical decomposition which is
accelerated in air or oxygen.
(d) HYDROLYTIC DEGRADATION
Degradation of organosoluble cellulose derivatives by hydrolytic reac-
tion can be attributed to acid cleavage of the glucosidic links in the polymer
molecule, with the exception of the alkaline scission of cellulose nitrate
discussed under the previous topic. The chain cleavage is accompanied by
de-esterification with the esters and by de-etherification with the ethers.
This reaction is of considerable importance in the manufacture of secondary
cellulose acetates, where it is customary to utilize an acid-catalyzed homo-
geneous de-esterification reaction to convert the acetone-insoluble triester
to an acetone-soluble derivative of lower D.S.
Under homogeneous conditions, depolymerization can occur either by
acetolysis or by hydrolysis as illustrated in the accompanying reaction
scheme:
C
!
Hv /C C\ XX /° C\ ,H
c c c c H2°[H]4
\0/\ /\ / \rv'' CHsCOOH
u Xo — cx
HH
1052 CELLULOSE
v^
TT XC Cv JOOCCHl TJH /O Cv TT
*V \Aou + \y \y
' \ H/ v^/ v-
In a homogeneous system containing both acetic acid and water, deg-
radation will occur as a result of the medium effect (acetic acid hydrolysis)
and can be greatly accelerated by the presence of strong acids such as sul-
furic acid.65 The reaction appears to obey the kinetic laws derived by af
Ekenstam66 for the random degradation of polymers and the rate of re-
action is shown to be temperature dependent, increasing with temperature
as would be anticipated. The reaction rates obtained by Hiller66 are re-
corded in Table 49. From these data it is apparent that sulf uric acid ac-
TABLE 49
Reaction Rate Constants for Degradation of Cellulose Acetate (Hiller85)
Temperature,
°C. Reaction rate constant X 10«, hr."1
Uncatalyzed reaction
84.4
0.77
94.2
2.1
104.4
8.1
115.1
20
Sulfuric acid-catalyzed reaction
46.0
22
54.8
64
65.7
177
celerates the degradation to a marked degree.
It is also of importance to compare the relative degradation tendency of
cellulose acetate with its rate of deacetylation. Table 50 shows such a
correlation from which several pertinent conclusions can be drawn.
The degradation occurring during this reaction has a larger activation
energy67 than does the deacetylation reaction. The degradative reaction
is thus more temperature dependent than is the de-esterification; also,
increases in reaction temperature will increase the extent of degradation
« L. A. Hiller, Jr., /. Polymer Sci., 10, 385 (1953).
w A. af Ekenstam, Ber., 69B, 549, 553 (1936).
67 S. Glasstone, K. J. Laidler, and H. Eyring, The Theory of Rate Processes, McGraw-
Hill, New York, 1941.
IX. DERIVATIVES OF CELLULOSE 1053
obtained for a given change in degree of substitution. It is also interesting
to observe that in the uncatalyzed reaction, the activation energy for de-
acetylating primary hydroxyls is much less than that required for secondary
hydroxyls or for degradation. With sulfuric acid catalysis, however, the
energy requirements for the hydrolysis of the secondary hydroxyls have
been greatly reduced and the reaction appears to go faster than that of the
primary hydroxyls.
TABLE 50
Energetics of Cellulose Acetate Deacetylation and Degradation (Hiller65)
Deacetylation
Thermodynamic function Primary Secondary
Eyring notation" groups groups Degradation
Uncatalyzed reaction
Experimental energy of activation,
Etxpn., kcal. mole-1 14.3 db 1.4 20.6 db 0.4 29,6 =fc 1.0
Heat of activation, A#4= (25° C.),
kcal. mole-1 13.7 20.0 29.0
Entropy of activation AS^ (25°
C.), entropy units mole"1 —39 .2 -24 . 1 - 12 . 6
Free energy of activation A Fiji
(25°C.), kcaL mole"1 25.4 27.2 32.8
Sulfuric acid-catalyzed reaction
Experimental energy of activa-
tion, Eexpti., kcal. mole-1 16.1 ± 5.8 8.5 db 0.2 22.7 db 2.0
Heat of activation, Affiji (25°C.),
kcal. mole-1 15.5 7.9 22.1
Entropy of activation, AS* (25°
C.), entropy units mole-1 -29.6 -54.9 -17.7
Free energy of activation AFifc
(25°C.), kcal. mole-1 24.3 24.3 27.4
The loss in viscosity which organosoluble cellulose derivatives may under-
go under acidic conditions is the direct consequence of the acid lability
of the /3-glucosidic bond joining the anhydroglucose units in the cellulose
chain. It would be anticipated, therefore, that organosoluble esters and
ethers of cellulose would be equally susceptible to acid degradation if they
were exposed under conditions of a homogeneous reaction, where all of the
/3-glucosidic bonds would be equally available for attack. Experimental
investigation has shown the above conclusion to be an oversimplification
since it has neglected the possible role of the substituents in modifying the
1054
CELLULOSE
rate of cellulose chain cleavage under identical homogeneous reaction
conditions.
Change in intrinsic viscosity versus time, as shown in Figure 91, was
compared under identical acetolysis conditions for samples of ethyl cellu-
lose, cellulose acetate, and cellulose nitrate.68 It is apparent from the
curves in Figure 91 that ethyl cellulose is the least stable and cellulose ace-
tate the most stable, with cellulose nitrate occupying an intermediate
2.5
Ka.
O
O
CO
t-t
>
u
Cellulose Nitrate (D.S. •» 2.9)
, Celluloee Acetate (D.S. - 2. 7)
•O— O
Ethyl CelluloscftD. S.» 2.6)
1000 2000
TIME IN MINUTES
3000
Fig. 91. Rates of acetolysis at 25 °C. for ethyl cellulose, pfellulose acetate,
and cellulose nitrate ^(Lincoln, Reid, and McBurney*8). Acetolysis was
carried out in an anhydrous medium, 95:5 acetic acid: acetic anhydride
which was 0.78 N with respect to HC1 and 0.105 N with respect to LiCl.
position. The extreme difference in the rate of loss of intrinsic viscosity
between ethyl cellulose and cellulose acetate is a striking example of how
substituents can modify the basic characteristics of the cellulose molecule.
Cellulose nitrate shows a somewhat ambiguous behavior which can be
attributed to the complex nature of the denitration reaction.
Some attention in the literature has been directed toward the hetero-
61 D. C. Lincoln, A. R. Reid, and L. F. McBurney, unpublished results.
IX. DERIVATIVES OF CELLULOSE 1055
geneous hydrolysis "of cellulose acetate69""71 and methyl cellulose.72 The
reaction in the case of methyl cellulose is very similar to the corresponding
cellulose hydrolysis in that an initial rapid reaction is succeeded by a slow
one, and the degradation appears to be tending toward a "leveling off"
D.P. The conditions used with cellulose acetate resulted in deacetylation
which was accompanied by a negligible amount of depolymerization.
This discussion serves to point up a serious gap in our knowledge of
cellulose derivative behavior, namely, the development of a sound theory
to explain the substituent effect on the hydrolytic sensitivity of such com-
pounds. It would be of great help to understand the effect of substituent
size, polarity, extent of substitution, and uniformity of substitution on the
activation energy and frequency factor in these cases.
8* F. S. Sherman and I. O. Gol'dman, /. Applied Chem. ( U. S. S.R.), 25 (English trans.),
87 (1952) ; Chem. Abstracts , 46, 5836 (1952).
70 E. Elod and A. Schrodt, Z. angew. Chem., 44, 933 (1931).
71 1. Sakurada and T. Morita, /. Soc. Chem. Ind.t Japan, 41, Suppl. binding, 385
(1938).
71 R. Steele and E. Pacsu, Textile Research J., 19, 771 (1949).