Full text of "The metallurgy of iron and steel"

See other formats

THE METALLURGY OF IRON AND STEEL

Published by the

McGraw-Hill Book. Company

New "Yor-k.

Succ arsons to theBookDeparbnents of the)

McGraw Publishing Company Hill Publishing" Company

Publishers of Books for

Elec trlcal Worl d The Engineering and Mining Journal

Engineering Record Power and The Engineer

Electric Railway Journal American Machinist

Metallurgical and CKemical Engineering

'vaw JIT >T JT >T \T>Tir\T ir ir>TJinTff>Tffff>T ir snrirjirffff JtrimTiMT if >T>Tir if if iMT

THE METALLURGY OF

IRON AND STEEL

BY
BRADLEY STOUGHTON, Pn.B., B.S.

FIRST EDITION — FIFTH IMPRESSION

(With numerous revisions)

McGRAW-HILL BOOK COMPANY

239 WEST 39TH STREET, NEW YORK

6 BOUVERIE STREET, LONDON, E.G.

1908

-ft

ENTERED AT STATIONERS' HALL, LONDON, ENGLAND

TO
PROFESSOR HENRY MARION HOWE, A.M., LL.H,

BESSEMER MEDALLIST, KNIGHT OF THE ORDER OF

ST. STANISLAS, ETC., ETC.,

PRACTITIONER, INVESTIGATOR, AUTHOR, INTERPRETER,

EDUCATOR AND PHILOSOPHER, WHOM THE WORLD

OF SCIENCE DELIGHTS TO HONOR,

THIS VOLUME IS
AFFECTIONATELY DEDICATED

222483

PREFACE

THE purpose of this book is to serve as a text-book, not only
for college work, but for civil, mechanical, electrical, metallurgical,
mining engineers and architects, and for those engaged in work allied
to engineering or metallurgy. America now produces almost as
much iron and steel as the rest of the world together, although
less than eighteen years ago she held second rank in this industry.
It seems fitting that the record of this progress should be brought
together into one volume covering every branch of the art of ex-
tracting the metal from its ores and of altering its adaptable and
ever-varying nature to serve the many requirements of civilized
life.

I take pleasure in acknowledging here, with sincere thanks
the assistance of many who have aided in the make-up of the
volume, and especially The Adams Co. (Figs. 204-7), American
Electric Furnace Co. (Figs. 303-4), American Sheet & Tinplate Co.
(Figs. 9, 11, 23, and 79), Bethlehem Steel Co. (Figs. 128-9), Brown
Specialty Machinery Co. (Fig. 218), Connersville Blower Co. (Figs.
230-1), Crocker- Wheeler Co. (Figs. 155-6, 167), Francis G. Hall
Esq. (Figs. 188-91, 196), Holland Linseed Oil Co. (Figs. 197-
200), Chas. W. Hunt, Esq., Secretary, American Society of Civil
Engineers (Fig. 284), Professor James F. Kemp (Fig. 8), Mack-
intosh, Hemphill & Co. (Figs. 136-7, 150, 152), Morgan Construc-
tion Co. (Figs. Ill, 175, 179, 180), National Tube Co. (Fig. 171),
S. Obermayer Co. (Figs. 194-5, 201, 203, 223-5, 270), J. W.
Paxson Co. (Figs. 226-8), Henry E. Pridmore (Figs. 208-12),
John A. Rathbone (Figs. 213-6, 219-20), each of whom have kindly
loaned electrotypes. And of Dr. H. C. Boynton (for Fig. 289), the
Brown Hoisting Machinery Co. (Fig. 14), Buffalo Furnace Works
(Fig. 266), H. H. Campbell, Esq. (Figs. 115-6), Professor William
Campbell (Fig. 290 and those on page 186), Carnegie Steel Co.
(Figs. 1 , 80, 160, 163-4, 172, 177) , W. M. Carr, Esq. (Fig. 108) , Central
Iron & Steel Co. (Figs. 44, 46-7, 49, 51), Crucible Steel Company

Vi PREFACE

of America (Fig. 59), Fiske & Robinson (Fig. 27), The Foundry
(Fig. 117-8, 202), Harbison-Walker Refractories Co. (Figs. 100,
181, 229), Joseph Hartshorne, Esq. (Figs. 41, 43, 45, 48, 50), Pro-
fessor Henry M. Howe (Figs. 18, 21, 36-7, 60, 67-70, 120-2, 125-6,
254, 283), Lackawanna Steel Co. (Figs. 13, 20, 65, 73-4), Marion
Steam Shovel Co. (Fig. 10), Mesta Machine Co. (Figs. 17, 138,
143-4, 153, 187), Morgan Engineering Co. (Figs. 127, 147, 159),
Professor A. H. Sexton (Fig. 56), Wm. Swindell & Bros. (Figs.
113-4), United Coke & Gas Co. (Figs. 2-6), United Engineering &
Foundry Co. (Figs. 78, 81, 135, 141, 146, 148-9, 154, 170, 264),
Wellman-Seaver-Morgan Co. (Figs. 92-3, 102-4, 107, 112), Whiting
Foundry Equipment Co. (Figs. 271-3). And of O. S. Doolittle,
Esq., for information upon the paint given on page 433, Frank E.
Hall, Esq., for the analyses in Table XVIII, and W. J. Keep, Esq.,
for the figures in Table XXVI.

But especially I am indebted to the following gentlemen, each
of whom has read a section of the book and made suggestions for
its revision which have been very valuable to me: Messrs. W.
Arthur Bostwick, Stanley G. Flagg, Jr., Alfred E. Hammer, Joseph
Hartshorne, J. E. Johnson, Jr., Carleton S. Koch, Frank N. Speller,
Herbert L. Sutton, and Hugh P. Tiemann.

BRADLEY STOUGHTON.
January 20, 1908.

TABLE OF CONTENTS

CHAPTER I. INTRODUCTION — IRON AND CARBON . . . .

Definitions, 6. General text-books, reference books and
periodicals on the metallurgy of iron and steel, 8.

CHAPTER II. THE MANUFACTURE OF PIG IRON ....

Blast-furnace fuels and fluxes, 11. Varieties and distribu-
tion of iron ores, 14. United States deposits and transpor-
tation, 16. Handling raw material at a modern furnace, 22.
The blast furnace and accessories, 24. Smelting practice and
products, 30. Calculating a blast-furnace charge, 46.

CHAPTER III. THE PURIFICATION OF PIG IRON ...

Comparison of purification processes. 59. Distinguishing
between the different products, 66. Miscellaneous purifica-
tion processes, 67. General reference books on steel, 72.

CHAPTER IV. THE MANUFACTURE OF WROUGHT IRON AND CRUCI-
BLE STEEL ... . . . ...

The manufacture of wrought iron, 74. The carburization
of wrought iron, 85. References on the manufacture of
iron, 93.

CHAPTER V. THE BESSEMER PROCESS .

References on the Bessemer process, 125.

CHAPTER VI. THE OPEN-HEARTH OR SIEMENS-MARTIN PROCESS

Open-hearth plant, 127. Open-hearth furnace, 132. Basic
open-hearth practice, 145. Acid open-hearth practice, 153.
Special open-hearth processes, 155. Open-hearth fuels, 160.

CHAPTER VII. DEFECTS IN INGOTS AND OTHER CASTINGS .
References on defects in ingots, 183.

CHAPTER VIII. THE MECHANICAL TREATMENT OF STEEL

The forging of metals, 187. The reduction of metals in
rolls, 193. Parts of rolling mills, 200. Rolling-mill practice,
215. Wire drawing, 224. Pressing, 227. Comparison of
mechanical methods, 229. Heating furnaces, 229. Refer-
ences on mechanical treatment, 235.

CHAPTER IX. IRON AND STEEL FOUNDING . .

The making of molds, 237. Design of patterns, 260. Cupola
melting of iron for castings, 262. Comparative cupola prac-
tice, 279. Other melting furnaces, 284. Melting steel for
castings, 286. References on foundry practice, 291.

vii

PAGES

3-10

11-50

51-73

74-94

95-126
127-172

173-184
185-235

236-291

Vlll

TABLE OF CONTENTS

PAGES

CHAPTER X. THE SOLUTION THEORY OF IRON AND STEEL . . 292-315

The freezing of alloys of lead and tin, 295. The freezing of
iron and steel, 304. The solid solution of iron and carbon, 308.
The complete Roberts- Austen, Roozeboom diagram, 312. Ref-
erences, 315.

CHAPTER XI. THE CONSTITUTION OF STEEL 316-332

The micro-constituents of steel, 316. The strength of steel,
324. Hardness and brittleness of steel, 328. Electric con-
ductivity of steel, 329. Magnetic properties of steel, 330. Ref-
erences on the constitution of steel, 332.

CHAPTER XII. THE CONSTITUTION OF CAST IRON .... 333-355

The effect of carbon on cast iron, 337. The effect of silicon,
sulphur, phosphorus, and manganese on pig iron, 341. The
properties of cast iron, 345. References on the constitution
of cast iron, 355.

CHAPTER XIII. MALLEABLE CAST IRON 356-369

References on malleable cast iron, 369.

CHAPTER XIV. THE HEAT TREATMENT OF STEEL . . . . 370-395

Improper heating of steel, 370. Hardening of steel, 382.
The constituents of hardened and tempered steels, 389. Ref-
erences on the heat treatment of steel, 395.

CHAPTER XV. ALLOY STEELS . . . . . . . , . 396-421

Nickel steels, 398. Manganese steel, 405. Chrome steel,
407. Self-hardening and high-speed tool steels, 408. Silicon
steels, 413. Vanadium steels, 414. Titanium steels, 419.
References on alloy steels, 419.

CHAPTER XVI. THE CORROSION OF IRON AND STEEL . . . 422-436

The cause and operation of corrosion, 422. Preservative
coatings for iron and steel, 429. References on corrosion, 436.

CHAPTER XVII. THE ELECTRO-METALLURGY OF IRON AND STEEL 437-447

Electro-thermic ore smelting, 438. Electro-thermic manu-
facture of steel, 442. Electrolytic refining of iron, 446. Ref-
erences on the electro-metallurgy of iron and steel, 447.

CHAPTER XVIII. THE METALLOGRAPHY OF IRON AND STEEL . 448-457

Preparation of samples for microscopic examination, 449.
Developing the structure for examination, 452. Microscope
and accessories, 454. Magroscopic metallography, 454. Ref-
erences on the metallography of iron and steel, 457.

CHAPTER XIX. CHEMISTRY AND PHYSICS INTRODUCTORY TO

METALLURGY . . . . . . . . ^ . . 458-488

Oxygen, 462. Thermo-chemistry, 464. Chemical equations,
466. Hydrogen, 469. Elements, compounds, and radicals,
471. Chemical reactions and compounds, 475. Chemical
solutions, 480. Some principles of physics, 482. Physical
.properties of metals, 484.

THE METALLURGY OF IRON AND STEEL

/•*.: :'*:?ie.'4**-^ifeo£.ikAST FURNACE AND STOVES.

•i r1

INTRODUCTION— IRON AND CARBON

This chapter is written for those students of iron and steel, — whether
they be students engaged at some university, or in an engineering
or metallurgical profession, — who have previously completed a course
in chemistry and physics. For the benefit of those who have not
had a technical education, Chapter XIX has been especially pre-
pared, and it is hoped that, if they will read that chapter before be-
ginning elsewhere in the book, all the subjects discussed in these
pages will be readily intelligible to them.

The Ferrous Metals. — Iron and steel together form the largest
manufactured product in the world, and each of them enters
into every branch of industry and is a necessary factor in every
phase of our modern civilization. Cast iron, because of the ease
with which it can be melted, is produced in final form in almost
every city in the United States, and only slightly less widely in
other civilized countries. The manufacture of steel is more
centralized, for economical reasons, but is several times as great
as cast iron in volume. Wrought iron is lesser in amount than
either of the others, but has its own importance and uses. These
three products, — cast iron, steel, and wrought iron, — together
comprise the whole of the so-called " ferrous group of metals" — that
is, the group which we classify together under the name of "iron
and steel." They have two characteristics in common: First,
that iron is present in all to the extent of at least 92 per cent.;
and second, that carbon is their next most important ingredient,
and regulates and controls their chief qualities. Their manufac-
ture represents nearly 15 per cent, of all the world's manufacturing
wealth, and is far greater than any other like industry. (See
Table I.)

Cast Iron. — Cast iron is impure, weak, and must be brought
to its desired size and form by melting and casting in a mold. A

4C, i,\^ t^THB;MEp^LLtf$GY OF IRON AND STEEL

typical example would contain about 94 per cent, iron, 4 per cent,
carbon, and 2 per cent, of other ingredients or impurities.

Pig iron is a raw form of cast iron, and malleable cast iron is
a semipurified form.

Steel. — Steel is purer than cast iron, much stronger, and may
be produced in the desired size and form either by melting and
casting in a mold or by forging at a red heat. It usually contains
about 98 per cent, or more of iron, and, in different samples,
from 1.50 per cent, down to almost no carbon, together with small
amounts of other ingredients or impurities.

Wrought Iron. — Wrought iron is almost the same as the
very low-carbon steels, except that it is never produced by melting
and casting in a mold, but is always forged to the desired size
and form. It usually contains less than 0.12 per cent, of carbon.
Its chief distinction from the low-carbon steels is that it is made
by a process which finishes it in a pasty, instead of in a liquid form,
and leaves about 1 or 2 per cent, of slag mechanically disseminated
through it.

Iron. — Iron as such, — by which I mean pure iron,— does not
exist as an article of commerce, but appears in service and in the
market only in the form of cast iron, steel or wrought iron, — that is,
when contaminated with carbon and other impurities. Some
of these impurities are present because they cannot cheaply be
gotten rid of, and others, because, like carbon for example, they
benefit the metal by giving it strength or some other desirable
property. Pure iron is a white metal and one of the chemical
elements. It is with one exception the commonest and most
abundant metal in the earth, and almost all rocks contain it in
greater or less degree, from which we extract it if it is large enough
in amount to pay for working. It practically never occurs in
nature in the form of a metal, but is always united with oxygen
to form either a blackish, brownish, reddish or yellowish substance.
Indeed, if it should occur in metallic form it would very soon be-
come oxidized by the action of air and moisture.

It is the abundance of iron in the earth which is the chief
cause of its cheapness, and therefore one reason why it is used
more than any other manufactured material. The other reason
is the ease with which we can confer upon it at will some of the
qualities most useful to man, of which the most valuable is prob-
ably its unequaled strength, and the most wonderful its magnet-

INTRODUCTION— IRON AND CARBON

1 ^2 C^ 00 CO £*» C^ 00 CO C^ *-O 00 ^-^ C^ *-^ to ^O ^D *"•* C^l CO 0s! ^^ C^ 00 O^ ^D ^H CO »~H
O C^ CV> ?i rH rH 00 00 CM CO O> Cil> O 00 O OS 1C CO O CQ uo OOCO iO t^St^ «

»o os Q o co os oo *» »s co oO o» oo os c«,eo O& •* ^<

r- iC^COCOCOCS^OCOr- irHU-2<MCiOO";frHCOiOI>

r-GC co rH oo o

COOOCiOOOOOOOOOC^O

corHx-o OOOOOOOOOOG

'— i^fCiOOOOOOOO

r- 1 OOCO'*lOOO5»O>-tiCOOOOLO'-OiOiOl>O>O'^l(M(MCOT-i(MCO»O
CO COCOOCOCV>COCOCOCvlCOCOC1OCOPOCOCOCO-^t|LOCOCOcOCOCOcOCO

00 O f-i
CO (M 00 <M
r-ut

•OOI> ••* CS (M (N u

. 00 *O (N C^ <N CO "* t— i (N GO O O -— i (N <M C5 CO iO •* CO t^ 1-1 O

<N <N C^l <M <N CO CO

CO CO CO Tt< rH

O lO CO CO Tfi CO

00—<

COCO CO O»OOOt^COt^
tO r-t »-OOO(Mi— ir- i

t^-TM 10 TftocOCOJMOt^i— iOcOCOOOO
COCO O OOCO^COi— (l>t^OOlO(MCO(MCO

lOLOO
I>*iCOOO

OO'OOOOi-H
-^OOOOCs-l'— irH

l> 1> I> 00

CS O i-H I> rH (

O i>- O t»

lOcOCOCO

O t* OS 53 CO »O O 00 CO t* CO IN.

Or^GOO'*
i—icOOOI>i— (

LO CO O C5 rH O CO

C^OOCOCO^COrHCDCCCOCOCi^OI^OCOt^C^t^rHOOCviC^OSf^rH
t^OOOOOOCOiOiOCOt^OOOOOOOCO^iOiOt^CO-^OOOJOCO

O (N 00

^H TH CO

o oo CS o oo oo

00 CO O O rH rH

OOOO^OOCO'— iQOrHOOrH

OS >O <M

rH (N (M (M CO CO CO COCO"#Tt< •* "* "tf "* ^ *& *O CO CO 1> l> I> t^ 00 O O O (N

iOCO-^OOrHOOT^(NC500CO»OI>

-_
-— iCOC<l>-OCO<MOO<M>-OCO<NrHOOCO

r-crH -^ o «5 oo o co o i> cs

O OS (N CO t^ CO (M O5 CO O

CO'— <t>-OI>»OiOO5'!t1'— i

O 00

i— i r -
"* CO

OOOOOOOOCO>OOOCO
OO Oi r^ ^O 00 CO rH Ci 1s* ^O O ^H

00 O CO CO CO

^ o oo ^o o

•^ T^ CS <M rH

CO r^rH
LO O— i

COOOOOOOOOOOOOOOGOOCCiOOCiOOCiC^OCiOO
CCGOOOOCOOOOOOOCOOOCOOOOOOOOOOOCOOOOOOOOCiOi

6 THE METALLURGY OF IRON AND STEEL

ism, in which it is not even approached by any other substance.
What these two properties alone mean in modern structural and
electrical engineering can scarcely be estimated.

Carbon. — Carbon is also a chemical element and familiar
to everyone; graphite, lamp-black, charcoal, and diamond are
the various allotropic forms in which it appears. It is a common
substance and present in every form of organic matter, while its
oxides, — carbon monoxide, CO; and carbon dioxide, C02 — are
well known gases. Its chemical affinity for iron is very great;
iron practically always contains some amount, and, if it is de-
sired to remove it entirely, the last traces are eliminated only with
extreme difficulty.

Iron and Carbon. — Carbon has the peculiarity of conferring
on iron great strength, which, strange to say, it does not itself
possess, and also hardness, which it possesses only in its diamond
allotropic form. At the same time it takes away from the iron
a part of its ductility, malleability, magnetism and electric con-
ductivity. So important is the influence of carbon in regulating
and controlling the characteristics of the ferrous metals, that they
are individually and collectively classified according to the amount
and condition of the carbon in them. The potent effect of carbon
must be constantly borne in mind when we come to describe the
manufacture of iron and steel and to discuss the methods of
regulating the carbon.

DEFINITIONS

The following definitions are selected from the report of March
31, 1906, of the Committee on the Uniform Nomenclature of Iron
and Steel of the International Association for Testing Materials,
with slight changes :

Cast Iron. — Generically, iron containing so much carbon or
its equivalent that it is not malleable at any temperature. Spe-
cifically, cast iron in the form of castings other than pigs, or
remelted cast iron suitable for casting into such castings, as dis-
tinguished from pig iron, i. e., cast iron in pigs.

The committee recommends drawing the line between cast
iron and steel at 2.20 per cent, carbon for the reason that this
appears from the results of Carpenter and Keeling to be the
critical percentage of carbon corresponding to the point "a" in
the diagrams of Roberts-Austen and Roozeboom. (See page 314.)

INTRODUCTION— IRON AND CARBON 7

Pig Iron. — Cast iron which has been cast into pigs direct
from the blast furnace. This name is also applied to molten
cast iron which is about to be so cast into pigs, or is in a condition
in which it could readily be cast into pigs before it has ever been
cast into any other form.

Gray Pig Iron and Gray Cast Iron. — Pig iron and cast iron
in the fracture of which the iron itself is nearly or quite concealed
by graphite, so that the fracture has the gray color of graphite.

White Pig Iron and White Cast Iron. — Pig iron and cast iron
in the fracture of which little or no graphite is visible, so that
their, fracture is silvery and white.

Mottled Pig Iron and Mottled Cast Iron. — Pig iron and cast
iron, the structure of which is mottled, with white parts in which
no graphite is seen, and gray parts in which graphite is seen.

Malleable Cast Iron. — Iron which when first made is cast in
the condition of cast iron, and is made malleable by subsequent
treatment without fusion.

Malleable Iron. — The same as wrought iron. A name used
in Great Britain, but not in the United States, except carelessly
as meaning "Malleable cast iron."

Steel. — Iron which is malleable at least in some one range
of temperature, and in addition is either (a) cast into an initially
malleable mass; or (6) is capable of hardening greatly by sudden
cooling; or (c) is both so cast and so capable of hardening.

Wrought Iron. — Slag-bearing, malleable iron, which does not
harden materially when suddenly cooled.

In the definition of steel the first sentence ("is malleable at
least in some one range of temperature") distinguishes steel
from cast iron and pig iron; the second sentence ("is cast into an
initially malleable mass") distinguishes it from malleable cast
iron, and the third sentence ("is capable of hardening greatly by
sudden cooling") distinguishes it from wrought iron. At the
best, however, the definition of steel is in a shockingly bad con-
dition, and has been brought to it by a series of events which shows
the carelessness of the buying public and the greed of men who
will appropriate the name for their product that will bring them
the best price without regard to whether the name really fits or
not.1

1 See page 173 of reference No. 1, at the end of this chapter, and page 6
of No. 2.

8 THE METALLURGY OF IRON AND STEEL

GENERAL TEXT-BOOKS, REFERENCE BOOKS AND PERIODICALS ON
THE METALLURGY OF IRON AND STEEL

(For further references, see the end of Chapter III.}

1. Prof. H. M. Howe. "Iron, Steel, and Other Alloys," 1903.

Published by Sauveur & Whiting, Boston, Mass. This
book contains three chapters upon the "Manufacture of
Iron and Steel " and ten chapters upon its constitution and
properties, especially from the standpoint of metallography.
Upon this latter subject it is without an equal and, like
the same author's larger work, bids fair to remain the
standard authority for many years to come.

2. H. H. Campbell. "The Manufacture and Properties of

Iron and Steel." Fourth edition. New York and London.
1907. This is a great reference book by one of the best
of the practical American metallurgical engineers. It is
undoubtedly the best reference book upon the manufacture
of iron and steel, but is not intended especially for beginners
or those without technical education.

3. James M. Swank. " Directory to the Iron and Steel Works

of the United States." Embracing the Blast Furnaces,
Rolling Mills, Steel Works, Forges, and Bloomaries in
Every State and Territory. Prepared and published by
The American Iron and Steel Association. Philadelphia.
The first edition of this book appeared in 1873, and the
seventeenth edition in 1907. The data given are very com-
plete and are classified for convenient reference.

4. James M. Swank. " History of the Manufacture of Iron in

All Ages, and Particularly in the United States from 1585
to 1892." Philadelphia.

5. "Ryland's Colliery,- Iron, Steel, Tin-Plate, Engineering and

Allied Trades' Directory (For Great Britain only) with
Brands and Trade Marks." 1906. Published by Eagland
& Co., Ltd., London.

6. Andrew Alexander Blair. "The Chemical Analysis of Iron."

A Complete Account of all the best known Methods for the
Analysis of Iron, Steel, Pig Iron, Iron Ore, Limestone,
Slag, Clay, Sard, Coal, Coke, ard Furnace and Producer-
Gases. Sixth edition. Philadelphia and London. 1906.

INTRODUCTION— IRON AND CARBON 9

7. J. 0. Arnold. "Steel Works Analysis." London and New

York. 1895.

8. The Journal of the Iron and Steel Institute. Published in

London. Vol. i, 1869; vol. Ixxiii, 1907. This periodical
appears twice a year and contains not only many original
articles of very great value but also an almost complete
collection of abstracts of the literature of iron and steel
that is published anywhere, classified under headings for
convenient reference. Anyone beginning the study of any
branch of iron and steel metallurgy should commence with
this journal^ as soon as the text-books have been con-
sulted.

9. Stahl und Eisen. Published in Duesseldorf. Vol. i, 1881;

vol. xxvii, 1907. This is the best German periodical on
iron and steel, and contains not only many valuable original
articles and abstracts but also translations. I have found
it particularly useful in this latter connection, because of
its translations of many articles from the Swedish.

10. Revue de Metallurgie. Published in Paris. Vol. i, 1904;

vol. iv, 1907. This is a very valuable periodical for those
who read French, not only for its original articles but also
for its abstracts. Upon the more scientific side of metal-
lurgy, that is to say the properties and constitution of
iron and steel, alloy steels, etc., it is without an equal.

11. The Mineral Industry. Its statistics, technical and trade.

Published in New York. Vol. i, 1892; vol. xvi, 1907. This
contains a review every year of the technology and trade
of each of the metals listed alphabetically, as well as the
statistics of production, price, etc. The articles usually
include a review of the progress of the metallurgy during
the year.

12. The Iron Age. Published in New York. Vol. i, 1869;

vol. Ixxix, 1907. This is the oldest and largest of the
American iron and steel technical magazines, and deals not
only with the scientific and technical side of the subject,
but also acts as a sort of a weekly newspaper upon the con-
dition of the iron trade and recent happenings of interest.

13. Transactions of the American Institute of Mining Engineers.

Published in New York. Vol. i, 1871; vol. xxxviii, 1907.
The American Institute of Mining Engineers is the leading

10 THE METALLURGY OF IRON AND STEEL

aggregation of both mining engineers and metallurgists in
America. These transactions contain many original articles
of value.

14. James M. Swank. Annual Statistical Report of the Secretary

of the American Iron and Steel Association, containing
detailed statistics of the American and foreign iron trade.
Published in Philadelphia.

15. Iron Trade Review. Published in Cleveland, Ohio. Although

this magazine aims to deal principally with iron trade
conditions, it contains also a great many technical articles
of importance.

16. Metallurgie. Published in Halle am See. Vol. i, 1904; vol.

iv, 1907. This German magazine contains a great many
original articles and abstracts.

17. Annales des Mines. Published in Paris. Vol. i, 1816. Tenth

series. Vol. xi, 1907.

18. Revue Universelle des Mines, de la Metallurgie des Travaux

Publics, des Sciences et des Arts Appliques a r Industrie.
Vol. i, 1857; Fourth series, vol. xvii, 1907.

19. Osterreichische Zeitschrift fur Berg- und Huttenwesen. Pub-

lished in Vienna. Vol. i, 1853; vol. Iv, 1907.

THE MANUFACTURE OF PIG IRON

WHATEVER material we are to manufacture — cast iron,
wrought iron, or steel — or for whatever purpose the metal is to
be used, the first step in the operation is smelting iron ore in a
blast furnace with fuel and flux, and obtaining cast iron or pig
iron, terms used synonymously in the United States.1 The pig
iron thus produced is an impure grade of iron, containing usually 3
to 4 per cent, of carbon, up to 4 per cent, of silicon, up to 1 per cent,
of manganese, and a few hundredths of 1 per cent, each of sulphur
and phosphorus.2 The amount of pig iron made exceeds that of
any other product manufactured by man.

BLAST-FURNACE FUELS AND FLUXES

Fuels are impure forms of carbon. By their union with oxygen
they furnish heat:

C + O =CO (generates 29,160 calories).3
CO + O =COa ( " 68,040 " ).

C+O2=CO2( " 97,200 " ).

The temperatures necessary for smelting are obtained in this
way. They also act as the chemical agents to separate the iron
from the oxygen with which it is combined in ores:

Fe2O3+3 C =3 CO+2 Fe (absorbs 108,120 calories).3
Fe3O4+4C=4CO+3Fe ( " 154,160 " ).

The carbon contained in the pig iron is also dissolved from the
fuel, directly or indirectly.

1More strictly speaking, 'pig iron' applies to the virgin product of the
blast furnace, and ' cast iron ' designates pig iron that has been cast into
molds of some final and useful shape, usually after a remelting.

2 In foundry and basic pig irons, the impurities are higher than this.

3 All heat effects of chemical reactions are given in calories per molecular
weight in grams throughout the book.

12 THE METALLURGY OF IRON AND STEEL

Charcoal. — The purer the carbon the better it serves the pur-
poses mentioned. For this reason charcoal, which has the least
amount of objectionable impurities, was once the great metallurgi-
cal fuel. Even to-day blast furnaces use charcoal for the produc-
tion of pure pig relatively free from sulphur. Except in favored
localities, charcoal is costly. Furthermore, its weakness permits
it .to crush easily; so charcoal furnaces are restricted to small sizes,
and 'charcoal iron' is higher in price.

Anthracite. — Anthracite is purer than coke, but its denseness
makes it offer a large resistance to the blast in furnaces having
much height.

Coke. — Coke is the great blast-furnace fuel, and a near-by
supply of this material makes Pittsburg, Chicago, Alabama, and
Colorado the great smelting centers that they are. In the United

— STANDARD AMERICAN BEEHIVE OVEN.

Charge, 5 net tons coal. Coking time, 72 hours. The gases
distilled from the coal burn in the dome of the oven and thus heat
the coal to produce more distillation.

States practically 94 per cent, of the pig iron is made with coke
as fuel, 5 per cent, with anthracite as fuel, and 1 per cent, with
charcoal.

A bituminous coking coaK contains about 30 per cent, by
weight of volatile matter. When this coal is heated, or ' coked/
the volatile matter is driven off, leaving a porous, spongy mass of
a silvery gray color and good strength. This is coke, and an

THE MANUFACTURE OF PIG IRON

analysis of a specimen from the famous Connellsville region near
Pittsburg is: Volatile matter =0.67 per cent.; fixed carbon =87.05
percent.; ash = 10. 60 per cent.; sulphur =0.74 per cent.; phos-
phorus = 0.016 per cent.

Fluxes. — The ash of fuels will not melt readily. By adding
the correct amount of lime to them they are transformed into a
fusible mass, which remains in a liquid form in the furnace and is
easily removed by opening a hole in the side. This fusible ma-

FIG. 3. — 50 OTTO-HOFFMANN BY-PRODUCT, OR RETORT, COKE OVENS.

terial is known as slag or cinder, and the added lime is known as
flux. The flux is usually added in the form of limestone (CaCO3),
but the heat in the upper layers of the furnace drives off the car-
bonic acid, leaving lime (CaO).

The gangue of our iron ores consists usually of silica, alumina,
etc., and, like the fuel ash, requires the addition of the correct
amount of limestone flux to make it into a fusible slag. In the
Pittsburg district we charge about 1200 Ib. of limestone, 2200 Ib.
of coke, and 4000 Ib. of ore for every long ton of pig iron made.
The amount of each is increasing, however, from time to time, as
the higher-grade ores are becoming exhausted and there are more
impurities to be fluxed and melted.

THE METALLURGY OF IRON AND STEEL

VAKIETIES AND DISTRIBUTION OF IRON ORES

The iron ores used for smelting consist of chemical compounds
of iron and oxygen containing more or less water, either in the

FIG. 4. — HORIZONTAL SECTION THROUGH A RETORT OF A
BY-PRODUCT OVEN.

The gases from the coal burn in flues on the side of the retort which contains the coal.

form of moisture or chemically combined as water of crystalliza-
tion.

Hematite (Fe2O3) . — The best known of these ores is hematite,
containing when pure 70 per cent, of iron. The red or brown

FIG. 5. — STRUCTURE OF BEEHIVE COKE.

hematites are the richer varieties (Lake Superior deposits, contain-
ing, in some cases, as much as 68 per cent, of iron), while the

THE MANUFACTURE OF PIG IRON

hydrated hematites, or limonites, usually contain a good deal of
water of crystallization and are consequently poorer in iron, not
often yielding much more than 50 per cent. iron.

Oolitic hematite is a variety that exists in the form of spherical
grains or nodules. It is important because it sometimes contains
limestone and is, therefore, valuable not
only for the iron but for the fluxing quality
of the lime. The Minette ore of Lothringen
(formerly Lorraine), Luxemburg, and France
is an enormous deposit of this oolitic hema-
tite, running from 30 to 35 per cent, iron and
giving a pig iron containing about 2 per
cent, of phosphorus. This ore is the basis
of the iron industry of Germany, France,
and Belgium, and, upon judicious mixing of
varieties, when necessary, is self-fluxing.

Magnetite (Fe3O4). — Magnetite contains,
when pure, enough iron (72.4 per cent.) to
attract the magnet. In the United States it
is often mixed with other impurities, such as
silica, titanium, and phosphorus, so much so
as to render the ore either too poor in iron
to be smelted profitably, or too high in phos-
phorus to make good steel, or so high in
titanium as to interfere with the blast-furnace
smelting by producing sticky slags which are
not easily handled.

The magnetite ores of Sweden are, how-
ever, the purest ores that exist in large
quantities anywhere, and form one of the
sources of the Swedish iron and steel, which
is famous all over the world for its purity,
that is, for its freedom from the objection-
able elements sulphur and phosphorus. It
is these Swedish products which supply the steel industry of
Sheffield with pure material for its tool steel and cutlery.

Siderite (FeCOs). — Another variety of iron ore is the so-
called 'spathic' iron ore, or siderite, which is, however, without
any importance in the United States. This forms the famous
* clay ironstone ' of the Cleveland district in England. It is poor

FIG. 6. — CROSS-SEC-
TION OF RETORT.
Structure of By-product
Coke.

THE METALLURGY OF IRON AND STEEL

in iron and is therefore no longer smelted in any quantity in the
United States in competition with the rich hematites. This ore is

FIG. 7. — BATTERIES OF BEEHIVE OVENS.

almost always calcined before smelting to expel the carbonic acid,
in order to save the blast furnace the extra work of this expul-
sion in its upper levels.

UNITED STATES DEPOSITS AND TRANSPORTATION

In the United States ores of iron are very widely distributed,
as will be seen by reference to the map on page 17, the black spots
on which represent notable deposits. The smelting of ore also
shows a wide distribution. Blast furnaces are in operation in
twenty-two states, including Washington, Minnesota, New York,
and Massachusetts on the north; Colorado, Texas, and Alabama on
the west and south. The great pig iron centers are: (1) The dis-
trict that includes Western Pennsylvania and Ohio, which pro-
duces more than one-half of the pig iron of the country; (2) Illinois,
and (3) Alabama.

It is not to be supposed that all the deposits marked on the
map are extensively worked for their iron. The rich hematite
deposits of the Lake Superior district furnish annually about
40,000,000 tons, which yield more than three-quarters of the pig-
iron production of the country. The only other districts which
produce more than 1,000,000 tons a year are in the states of Alabama
and New York. Most of the other deposits are mined only for local

THE MANUFACTURE OF PIG IRON

treatment. In addition, a total of nearly 1,000,000 tons of ore per
year are imported from Cuba, Spain, and other foreign countries,

100 .Longitude West 90 from Greenwich 75-^

A N A

FIG. 8. — FROM KEMP'S "ORE DEPOSITS OF THE UNITED STATES."

principally for smelting by blast furnaces on or near the Atlantic
coast.

Ore Transportation. — The peculiarity of the Lake Superior
deposits is that almost none of the ore is smelted locally, but is
transported a distance of 800 miles or more in order to bring it to

FIG. 9. — A LAKE SUPERIOR ORE MINE.

the coke. Thus, South Chicago, Western Pennsylvania, and Ohio
receive the bulk of the ore shipped from the Lake Superior mines.
Since the amount of coke used in the blast furnaces is only about
one-half the weight of the ore, it might seem uneconomical to carry

THE MANUFACTURE OF PIG IRON

the latter to the former. But coke is bulky in proportion to its
weight; furthermore, it suffers a good deal of waste in transporta-
tion in consequence of its friability and of the fact that so much
of it is broken down into pieces less than an inch in diameter
(technically known as ' breeze ') which is not suitable for charging
into the blast furnace. The ore, on the other hand, may be handled
by the cheapest and most rapid labor-saving devices. Indeed, in

FIG. 11. — LOADING AN ORE BOAT.

many cases, the ore is never touched by shovels in the hands of
man, but is mined, charged, and discharged in units of several tons
each, and often by means of gravity.

The mining and transportation of this great amount of ma-
terial is in itself a mighty industry, every advance in which
has contributed in no small share to the increasing volume and
importance of the iron, steel, and other industries of the United
States.

Mining. — Some of the Lake Superior deposits lie near the
surface and are therefore cheaply mined. This is especially true
of the soft, earthy deposits of the Mesabi range, which are some-
times worked in great open cuts, the ore being loaded upon cars by
mammoth steam shovels, or sometimes by the caving method, the
ore falling by gravity into cars situated in underground tunnels.
The massive, or rock, ores are more costly to extract, and the
utmost skill of American blast-furnace men has been exercised to

THE METALLURGY OF IRON AND STEEL

employ as large a portion of the earthy ores as possible without
choking up the furnace.

Transportation. — On reaching the shore of the lake the ore
train is run out above a long line of ore bins supported on a wharf
extending over the deep water of Lake Superior. Alongside of
this wharf the ore boats, capable of taking a load of 10,000 or
13,000 tons of ore, are docked. The hatches of these great boats

FIG. 12. — BROWN-HOIST APPARATUS UNLOADING AN ORE BOAT.

are placed such a distance apart that the hinged ore chutes of the
bins may be swung down and, when the gates are opened, the ore
allowed to flow directly into the hold of the vessel. In a few
minutes the vessel has received her full cargo and is ready to start
on its long journey down the chain of inland lakes.

Sometimes in long strings of three or four, in tow together,
sometimes singly or in pairs, the boats travel from one end of Lake
Superior to the other and come to the great canal of Sault Ste.
Marie. This canal deserves a passing mention because of the
enormous benefit which its construction has conferred upon the"
iron industry. It is two miles long, and for about three months of
the year is closed to navigation by the ice; nevertheless, the total
tonnage of the cargoes passing through it annually, by far the
greater part of which consists of iron ore, amounts to nearly
55,000,000, a volume three times as great as that borne by the
next greatest canal, namely, the Suez Canal, which forms the

THE MANUFACTURE OF PIG IRON

great water highway from Europe and the Mediterranean to the
East.

After passing through the one lock of the 'Soo' canal, the
stream of ore divides into two parts. One part turns to the west-
ward and supplies the great blast furnaces of Chicago and Mil-
waukee ; but much the larger portion travels down Lake Huron and
Lake St. Clair and is discharged at some one of the many great un-
loading points on the southern shore of Lake Erie, where it is either
smelted near by or loaded on railroad cars for transportation to
Pittsburg, Youngstown, or one of the other great blast-furnace
centers.

Unloading. — The unloading of boats is accomplished with
almost as great celerity as the loading, and by means of mechanical
unloading machinery a steamer containing as much as 10,300 tons

FIG. 13. — HULETT ELECTRIC UNLOADER.

of ore has been completely discharged in 4 hours and 30 minutes.
Nor is any time wasted in coaling the vessel for a second journey up
the lakes and back. Great machines pick up whole railroad cars
of fuel and empty them bodily into the chute which connects with
the bunkers of the vessel, many of the ore steamers being so con-
structed that this wholesale loading of coal can go on at the same
time that ore is being discharged.

22 THE METALLURGY OF IRON AND STEEL

HANDLING RAW MATERIAL AT A MODERN FURNACE

Behind the blast furnace are situated two long rows of storage
bins, one of which is shown in elevation in Fig. 15. These bins
are filled by bottom-dumping railroad cars which bring the ore to
the furnaces, or by mechanical apparatus from the great piles of
ore stored conveniently near. Between and under these two rows
of bins runs a track on which little trains of ore larries are trans-
ferred back and forth, being first filled with a weighed amount of
ore, limestone, or fuel, and then switched into a position from
which they can deposit their contents into the loading skip of the
blast furnace.

Loading the Furnace. — The next step in the handling of the
raw material is to bring the ore, together with the necessary fuel
and flux, into the mouth of the huge furnace that is to convert it
into pig iron. In one of the big modern American furnaces, work-

FIG. 14. — ORE-HANDLING MECHANISM AT BLAST FURNACE.

ing at top speed, the amount of material which must be dumped
into the top during 24 hours will frequently exceed 2000 tons, and
the charging must go on for 365 days a year with never a delay of
more than a few hours at a time.

In the modern type of furnace this loading is accomplished
altogether by mechanism operated and controlled from the ground
level, and no men are required to work at the top of the furnace.
In Fig. 15 is a section of such a furnace showing one method of
loading, — a double, inclined skipway extending above the top of
4he furnace. One skip is seen discharging its load of ore, or fuel

THE MANUFACTURE OF PIG IRON

and flux, into the hopper, while the second skip is at the bottom of
the incline ready to be loaded with its charge.

Double Bell and Hopper. — The upper hopper of the furnace is
closed at the bottom by an iron cone, known as a 'bell/ This bell

FIG. 15. — CROSS-SECTION OF BLAST FURNACE AND SKIP HOIST. ~

is pressed up against the bottom of the hopper by the lever of the
counterweight, as shown, but may be lowered by operating the

24 THE METALLURGY OF IRON AND STEEL

cylinder Af, to allow the charge to fall into the true hopper, /, of
the blast furnace. In this way the true hopper of the furnace is
progressively filled with ore, flux, and fuel. This hopper, /, is also
closed at the bottom by a similar bell, A. The lowering of this
bell is also controlled by mechanism operated from the ground
level. At intervals this operation is effected and the contents of
the hopper allowed to fall in an annular stream, distributing itself
in a regular layer on top of the material already in the furnace and
reaching to within a few feet of the bottom of the bell. As the
upper bell, B, is now held up against the bottom of the upper
hopper, there is never a direct opening from the interior of the
blast furnace to the outer air, so that the escape of gas, resulting
formerly in the long flame rising from the top of the blast furnace
whenever material Was dropped into the interior, no longer occurs
at our modern plants.

This is not the only means of handling the raw material for the
blast furnace. Several varieties of mechanism are extensively
used, but the description given heretofore will serve to illustrate
the general principles of labor-saving mechanisms in connection
with charging the blast furnace. It will be seen that the ore is
transferred from one receptacle to another by means of gravity
wherever possible.

THE BLAST FURNACE AND ACCESSORIES

The blast furnace itself consists of a tall cylindrical stack lined
with an acid (silicious) refractory fire-brick, the general form and
dimensions being shown in Fig. 15. The hearth or crucible is the
straight portion occupying the lower 8 ft. of the furnace. Above
that extends the widening portion, called the bosh, which reaches
to that portion in the furnace having the greatest diameter. The
stack extends throughout the remainder of the furnace, from the
bosh to the throat. The brickwork of the hearth is cooled by
causing water to trickle over the outside surface.

Tuyeres. — Through the lining of the furnace, just at the top
of the hearth, extend the tuyeres — 8 to 16 pipes having an in-
ternal diameter of 4 to 7 in., through which hot blast is driven to
burn the coke and furnish the heat for the smelting operation.
The ' tuyere notches/ or openings through which the tuyere pipes
_ enter, as well as the tuyeres themselves, are surrounded by hollow

THE MANUFACTURE OF PIG IRON

bronze rings set in the brickwork, through which cold water is con-
stantly flowing to protect them from being melted off at the inner
ends. The number and size of the tuyeres are in proportion to the
diameter of the hearth, the volume and pressure of the blast, etc.,
the blast being given sufficient velocity to carry it, distributed as
evenly as possible, to the very center of the furnace.

Discharge Holes. — On the side of the furnace, and 30 to 40 in.
below the level of the tuyeres, the 'cinder notch' or ' monkey' is
situated. This is protected by a water-cooled casting, and the
hole is closed by chilling the iron in it with an iron plug.

In the front, or breast, at the very bottom level of the crucible,
is the iron tap-hole, from which all the liquid contents of the

Eye Sight

FIG. 16. — PARTS OF A BLAST FURNACE TUYERE.

furnace can be completely drained. This is a large hole in the
brickwork, and is closed with several balls of clay.

Bosh. — The hottest part of the furnace is near the tuyeres
and a few feet above them. In order to protect the brickwork of
the bosh from this heat, a number of hollow wedge-shaped cast-
ings are placed therein, through which cold water circulates. The
brickwork is furthermore protected by a deposition of a layer of
carbon, similar to lampblack, on its internal surface, covered by a
layer of a sort of slag, replacing part of the brickwork. This
deposition of carbon comes about through the reaction of the

THE METALLURGY OF IRON AND STEEL

furnace operation itself, in the following manner: For the correct
conduct of the smelting operation, and especially for the carrying
off of the sulphur in the slag, it is necessary that a very powerful
reducing influence must exist; this reducing influence is produced

FIG. 17. — A BATTERY OF BLAST FURNACE BLOWING ENGINES.

THE MANUFACTURE OF PIG IRON 27

by an excess of coke, and one of its results is the precipitation of
finely divided carbon on the internal walls of the furnace. It is
this thin layer of slag and carbon which is most effective in pro-
tecting the acid lining of the furnace from the corrosive action of
the basic slag.

Hot Blast. — The air for smelting is driven into the furnace by
immense blowing engines ranging up to 2500 H.P. each, and capable
of compressing 50,000 to 65,000 cu. ft. (=4875 Ib.1) of free air per
minute to a pressure of 15 to 30 Ib. per sq. in., which is about what
one furnace requires. It actually requires about 4 to 5 tons of air
for each ton of iron produced in the furnace. After leaving the
engines and before coming to the furnace, the air is heated to a
temperature of 425 to 650° C. (800 to 1200° F.), by being made to
pass through the hot-blast stoves.

Hot-blast Stoves. — Each furnace is connected with four stoves.
These are cylindrical tanks of steel about 110 ft. high and 22 ft. in
diameter, containing two fire-brick chambers. One of these cham-
bers is open, and the other is filled with a number of small flues
(see Fig. 18) . Gas and air are received in the bottom of the open
chamber, B, in which they burn and rise. They then pass down-
ward through the several flues in the annular chamber surrounding
B, and escape at the bottom to the chimney as waste products. In
passing through the stove they give up the greater part of their
heat to the brickwork. After this phase is ended, the stove is ready
to heat the blast.

The blast from the blowing engine enters at the bottom of the
flues, E, passes up through the outer chamber, and down through B
to the furnace. • In this passage it takes up the heat left in the
brickwork by the burning gas and air. Sometimes there are three
passes, instead of two as described. In a blast-furnace plant one
stove is heating the blast while the other three are simultaneously
in the preparation stage, burning gas and air. By changing once
an hour a pretty regular blast-temperature is maintained. The gas
used for the heating is the waste gas from the blast furnace itself,
which amounts to about 90,000 cu. ft. per minute at a temperature
of 235° C. (450° F.) , and has a calorific power of about 85 to 95 B. T.
U . per cubic foot. The latent and available heat of this gas is equiva-
lent to approximately 50 per cent, of that of the fuel charged into

1 At 70° F. and atmospheric pressure, each 1000 cu. ft. of air weighs 75 Ib.

FIG. 18. — HOT-BLAST STOVE. From Howe, "Iron, Steel and other Alloys."

Solid arrows show passage of air that is heating. Broken arrows show passage of burning
gas. This is but one of several types of stove.

FIG. 19.

30 THE METALLURGY OF IRON AND STEEL

the furnace. Only about 30,000 cu. ft., or one-third of this gas, is
needed for keeping the stoves hot, and the remaining two-thirds is
used to produce power.

Power from Waste Gas. — The waste gas comes down the
downcomer T, Fig. 19, settles out dirt in the dust-catcher W , and
is then led to the stoves or power-producer. This gas varies in
composition, but will average about 61 per cent, nitrogen, 10 to
17 per cent. C02, and 22 to 27 per cent. CO. The latter can be
burned with air to produce heat.

CO + O-COa (generates 68,040 calories).

If burned under boilers, the available gas will generate enough
power to operate the blowing engines, hoisting mechanism, and
other machinery used in connection with the furnace. At several
plants the gas available for power is cleaned carefully and utilized
in gas engines, whereby much more power is obtained, the excess
being usually converted into electricity and transmitted to more
distant points.

SMELTING PRACTICE AND PRODUCTS

The furnace is filled with alternate layers of fuel, flux, and ore,
down to the top of the smelting zone. The exact location of this
zone will be dependent upon the volume and pressure of blast, size
of furnace, character of slag made, etc., but will extend from the
level of the tuyeres to a few feet above them, or about to the top of
the bosh. It will require perhaps 15 hours for the material to de-
scend from the top of the furnace to the smelting zone. During
this descent, it is upheld partly by the resistance of the upward-
rushing column of hot gases,1 partly by its friction on the walls of
the furnace, and partly by the loose column of coke which extends
through the smelting zone and to the bottom of the furnace, and
which alone resists melting in the intense heat of this zone. The
blast, entering the furnace through the tuyeres, consists of 23 per
cent, by weight of oxygen and 77 per cent, by weight of nitrogen,
together with varying amounts of water vapor from moisture in
the air (see page 38). The nitrogen is practically inert chemically
and performs no -function other than that of absorbing heat in the

1 There is a great fall in the pressure of the blast between the tuyeres
and the throat, which represents the work done by the air in helping to support
-the stock.

w I

o 8

5 &

B 1

a I

OQ '

g a

o I

§ 5

o g

5 .£

^ -s

S 1

« -

THE METALLURGY OF IRON AND STEEL

smelting zone and giving it out at higher levels. The oxygen
attacks all the coke in the smelting zone and as much of it below
the level of the tuyeres as is not covered by accumulations of iron

Fusion Level

Molten Slag
Molten Iron

Legend ; — Lumps of Coke

Lumps of Iron Ore ®

Lumps of Lime O

Drops of Slag _._d

Drops of Iron I

Layer of Molten Slag

Layer of Molten Iron

FIG. 21. From Howe, "Iron, Steel, and other Alloys."

and slag in the hearth, producing a large volume of carbon mon-
oxide gas (CO) and a temperature of about 1510° C. (2750° F.).1

1 It is of minor importance whether the CO gas is formed directly or as a
result of the two following reactions:

C +O2=CO3
CO2 + C =2CO

THE MANUFACTURE OF PIG IRON 33

The CO and nitrogen pass up between the particles of solid ma-
terial, to which they give up the greater part of their heat. The
former also performs certain chemical reactions, and thus in both
ways the rising column of gases prepares the charge for its final
reduction in the smelting zone.

Chemical Reactions in the Upper Levels. — As soon as the iron
ore enters the top of the furnace, two reactions begin to take place
between it and the gases :

(1) 2Fe2O3+8CO=7CO2+4Fe+C;

(2) 2 Fe2O3+ CO =2 FeO+Fe2Os+CO2;

and this continues with increasing rapidity as the material becomes
hotter. The carbon formed by reaction No. 1 deposits in a form
similar to that of lampblack on the outside and in the interstices of
the ore. This reaction, however, is opposed by two reactions with
carbon dioxide gas :

(3) Fe+CO2=FeO+CO;

(4) C+CO2=2CO.

Reaction No. 3 begins at a temperature of about 300° C. (575°
F.), which is met with about 3 or 4 ft. below the top level of the
stock; and No. 4 begins at about 535° C. (1000° F.), or 20 ft. below
the stock line. Reaction No. 4 is so rapid that the deposition of
carbon ceases at a temperature of 590° C. (1100° F.). All the way
down the ore is constantly losing a proportion of its oxygen to the
gases. At higher temperatures than 590° C., FeO is stable, and
practically all of the Fe2Os (or Fe304 if magnetite is being smelted)
has been reduced:

(5) Fe3O4+CO =3 FeO+CO2.

The reaction between iron oxide and solid carbon begins at 400°
C. (750° F.).

(6) Fe2O3 + 3 C =2 Fe + 3 CO.

At 700° C. (1300° F.) solid carbon begins to reduce even FeO:

(7) FeO+C=Fe+CO.

Practically all the iron is reduced to a spongy metallic form by
the time the temperature of 800° C. (1475° F.) is reached. This is
about 45 ft. from the stock line and less than 30 ft. above the
tuj^eres. At 800° C. the limestone begins to be decomposed by
the heat, and only CaO comes to the smelting zone :

(8) CaC03=CaO+COa.

34 THE METALLURGY OF IRON AND STEEL

The foregoing facts are summarized in Fig. 22, which is adapted
from H. H. Campbell, with certain changes.1 It is not supposed
that these figures are exactly correct for the different levels, and it
is probable that they change from day to day and from furnace to
furnace, but a general idea may be obtained from this sketch. It
will be seen that the upper 15 or 20 ft. of the stock is a region of
Fe2O3 and Fe3O4, gradually being converted to FeO by CO gas,
and forming quantities of C02 gas. If these reactions were the
only ones, the top gases would contain no CO and would have no
calorific power, but reaction No. 1 produces both metallic iron and
carbon, both of which reduce C02 and waste much energy, as far as
the blast furnace is concerned:

Reaction No. 3, Fe+CO2 =FeO+CO, absorbs 2340 calories,
but wastes 68,040 calories.

Reaction No. 4, C+CO2 =2 CO, absorbs 38,880 calories.

From 20 to 35 ft. below the stock line is the region of FeO,
gradually being converted to metallic iron sponge by carbon. On
the lower level of this zone the limestone loses its C02, which joins
the other furnace gases. From 35 ft. down to the smelting zone is
the region of metallic iron. This spongy iron is impregnated with
deposited carbon which probably to some extent soaks into it and
dissolves, in a manner like in nature but not in degree to the way
ink soaks into blotting-paper. This carburization of the iron
reduces its melting-point and causes it to become liquid at a higher
point above the tuyeres than it otherwise would.

On reaching the smelting zone the iron melts and trickles
quickly down over the column of coke, from which it completes its
saturation with carbon. At a corresponding point the lime unites
with the coke ash and impurities in the iron ore, forming a fusible
slag which also trickles down and collects in the hearth. It is
during this transit that the different impurities are reduced by the
carbon, and the extent of this reduction determines the character-
istics of the pig iron, for in this operation, as in all smelting,
reduced elements are dissolved by the metal, while those in the
oxidized form are dissolved by the slag. Only one exception
occurs, namely, that iron will dissolve its own sulphide (FeS) and,
to a less extent, that of manganese (MnS), but not that of other
metals, as, for instance, CaS.

1 See pp. 54 and 62 of Book No. 2, page 8.

THE MANUFACTURE OF PIG IRON

Chemical Reactions in the Smelting Zone. — There is always a
large amount of silica present in the coke ash, and some of this is
reduced according to the reaction :

(9) SiO2+2C=Si+2CO.

The extent of this reaction will depend on the length of time the
iron takes to drop through the smelting zone, the relative intensity
of the reducing influence, and the avidity with which the slag takes

Stock

75-0 450° F

Line

(1)

(2)

575° (3)

65 10 770° F

750° W

55 20 1090°F

1025° (4)

1100°

1300° (7)

45 30 1410° F

•«- $

35 40 1730°F

1830° (4)
p

25-50 3050°F

c»

L5 60 2370° F

/
Smelting j

Tuyeres -

5 -70 Feet
^.SftT.

Zone (J

(1) 2Fe2O3 + SCO = 7CO2 4- 4Fe +C (begins)

(2) 2Fe2O3+ CO=2FeO + CO3 + Fe2O3 (begins)

Fe + COa = FeO + CO (begins)
Fe9O3 + 3C = 2Fe + SCO

C + COQ = 2CO (rapid)
Deposition of carbon ceases

FeO + C = Fe + CO (begins)

FeO + C = Fe 4- CO (complete)
= CaO + C0.2

i) C + C0a = 2CO (prevails)
CO2 cannot exist below this level

(9) SiO., + 2C = Si + 2CO

(10) FeS + CaO + C = CaS + Fe + CO

(11) MnO.,+ 2G = Mn + 20O
P2O5 + 5C = 2P + 5CO

FIG. 22. — DIAGRAM SHOWING CHEMICAL ACTION IN BLAST FURNACE.

up silica. A slag with a high melting-point will trickle sluggishly
through the smelting zone and cause the iron to do the same, to
some extent, thus giving it more chance to take up silicon. A
higher temperature in the smelting zone, which increases dispro-
portionately the avidity of carbon for oxygen, will promote re-
action No. 9. We can produce this higher temperature by supply-

36 THE METALLURGY OF IRON AND STEEL

ing hotter blast. A larger proportion of coke to burden1 will
further promote this reaction, because this not only increases the
amount of the reducing agent, but also raises the temperature and,
therefore, the chemical activity of this agent. Thus the coke has
both a physical and a chemical influence in increasing the intensity
of the reduction in the smelting zone. A basic slag, because of its
avidity for silica, will oppose reaction No. 9; it is one of the princi-
pal means of making low-silicon pig iron. This is in spite of the
fact that the basic slags are sluggish, and therefore trickle slowly
through the smelting zone, thus exposing the silica longer to
reducing influence, and also increasing the temperature of the
materials in this zone (1) by causing them to pass through it more
slowly and absorb more heat, and (2) by reducing the level of the
smelting zone nearer to the tuyeres, which confines the intense
temperature to the smaller area, or, in other words, diminishes the
passage of heat upward.

Sulphur comes into the furnace chiefly in the coke. It is partly
in the form of iron sulphide (FeS) , and partly in the form of iron
pyrites (FeS2) , which loses one atom of sulphur near the top of the
stock and becomes FeS, which will dissolve in the iron unless con-
verted to sulphide of calcium (CaS). This is brought about, ac-
cording to the explanation of Professor Howe, by the following
reaction:

(10) FeS+CaO+C=CaS+Fe+CO.

CaS passes into the slag, and the odor of sulphur is very strong
when the slag is running from the furnace. It is evident from
reaction No. 10 that intense reduction, which increases the silicon
in the iron, has the contrary effect on the sulphur, and this explains
the common observation that iron high in silicon is liable to be low
in sulphur. Indeed, this relation is so constant that it is almost a
rule. There are two exceptions, however: (1) Increasing the
proportion of coke has a doubly strong influence in putting silicon
in the iron. As regards sulphur, on the other hand, it has a self-
contradictory effect; by increasing the amount of sulphur in the
charge it tends to increase it in the iron, which is partly or wholly
counteracted by its effect in reaction No. 10. (2) A basic slag may
hold silicon from the iron, and it also holds sulphur from the iron by
dissolving CaS more readily. In other respects the conditions

1 The burden is the amount of material that the coke has to melt. We
Tighten the burden by increasing the amount of coke, and vice versa.

THE MANUFACTURE OF PIG IRON

TABLE II.— COMPOSITION OF BLAST-FURNACE SLAGS
From H. H. Campbell. P. 50 of No. 2.

SLAG

IRON

REMARKS

3
4
5
6
7
8
9
10
11
12
13
14

It)
17
18
11)
20
21
22
23
24
25
26

SiO2

A1202

CaO

MgO

FeO

Total
not in-
clud-
ing S

tr.
tr.
.01
.06
.11
.05
.03
.02
.02
.02
tr.
.02
.03
.03
.07
.020
.029
.028
.032
.017
!040
.095
.101
.048
.038
.034

tr.
.025
.020
.027

.07
.03
.063
.040

33.10
32.27
24.26
32.68
32.28
34.50
34.98
34.70
33.68
29.86
28.95
30.62
32.55
30.08
31.46
36.08
37.19
36.86
32.06
33.57
35.38
36.35
33.70
35.11
35.10
35.84

Avera

33.21
34.84
31.77
35.55
A.vera
33.15
30.73
34.75
35.35

14.9240.76
14.57 41.02
11.5340.25
13.50 43.28
9.38 46.95
7.94 46.47
12.05 41.33
11.4441.27
11.9345.96
12.0445.20
12.0449.30
10.4749.13
11.1347.16
11.4446.36
11.5044.85
12.8541.69
12.65 35.47
10.7442.46
11.9742.46
10.6544.11
11.7638.19
10.2140.10
12.5638.12
14.21 22.48
14.75 27.95
14.34 32.71
ges for hot i

13.67(40.68
11.75|41.30
11.9845.58
12.0540.52
»es for modt
10.27 45.57
11.3247.36
11.3040.12
14.43 29.69

9.67
10.30
13.28
9.44
9.52
10.47
9.62
9.96
6.69
11.41
8.46
7.49
6.61
8.76
10.41
7.25
11.32
6.62
10.25
8.55
12.32
10.95
11.60
22.38
22.28
17.46
urnac
11.08
9.79
9.05
8.86
jrate (
9.81
8.35
10.86
20.71

98.45
98.16
98.32
98.90
98.13
99.38
97.98
97.37
98.26
98.51
98.75
97.71
97.45
96.64
98.22
98.41
97.53
97.31
97.37
97.69
98.53
98.60
98.30
100.12
100.08
100.35

98.64
97.68
98.38
97.66
•nace :
98.80
97.751
98.29
100.18!

3.37

3.18
4.81
1.25
0.70
0.69
2.60
2.32
1.27
1.27
0.57
0.26
0.15
0.58
0.20
2.15
1.92
1.50
1.59
0.94
1.13
0.66
0.50
1.37
1.85
1.60

3.79
2.46
1.27
1.79

0.88
0.35
0.81
1.61

Cuban ore, He

it (

{( {
Wz
] C°,(
Spanish ore, P

1 Lake ore
1 and part
| anthracite
J^coal;
mostly
Connells-
ville coke

1 Lake ore
• and Con- -
J nellsville
coke.

Cuban ore.
Spanish ore.

ie ((

Lake ore.

Cuban ore.
Spanish ore.
Lake ore.

>t furnace

irm '

>1 <

i t
tot furnace

)0l

Hot furnace
Fairly hot

Normal
u

Cool

I "
[Av. of 8 wks
Av. of 7 wks
Av. of 7 wks

. .

..
0.54
0.90
0.63
0.63
0.81
0.90
0.99
0.32

1.62
1.70
1.54
1.76
1.74
1.60
1.28
0.96

es:

....

,0.68 1.66
>r cool fui

l'.26

i'.io

which make for high silicon make also for low sulphur. Particu-
larly is this true of a high temperature in the smelting zone, and
the term 'hpt iron' has come to be synonymous in the minds of
blast-furnace foremen with iron high in silicon and low in sulphur.

38 THE METALLURGY OF IRON AND STEEL

Manganese is reduced by the following reaction :
(11) MnO2+2 C = Mn+2 CO.

The amount of manganese in the iron is dependent, to a certain
extent, upon the character of the ores charged, but it may be con-
trolled somewhat by the character of slag made, because an acid
slag will carry a large amount of manganese away in the form of
silicate of manganese (MnSiO3) .

With a certain unimportant qualification, the amount of phos-
phorus in the iron is controlled by the character of the ores charged,
and districts or countries having high-phosphorus ores must make
high-phosphorus irons. This is not an insuperable objection, be-
cause the presence of phosphorus, even up to 1.5 per cent., is
desired in certain irons for foundry use, and the basic processes for
making steel can remove this element.

The chemical influence of the blast furnace is a strongly redu-
cing one, and this is produced in order, first, to reduce the iron from
the ore; second, to get rid of the sulphur, and third, to saturate the
iron with carbon. Many attempts have been made to provide a
process wherein the reducing influence was not so strong, and thus
to produce a purer material than pig iron, because it is the intensity
of the reduction which vitiates the iron with carbon and silicon.
The great weakness of all such processes, however, is that they do
not get rid of the sulphur, which is the most objectionable impurity
that iron is liable to contain and which is not satisfactorily re-
moved by any process after once it makes its way into the iron.
Finally, to saturate the iron with carbon renders the blast-furnace
operation very much cheaper, because pure iron melts at a tem-
perature much higher than can readily be obtained in the furnace,
and melted iron is handled much more cheaply than it could be if
allowed to solidify. Even the presence of silicon is an advantage,
as we shall see in Chapter XII.

Drying the Blast. — The water vapor blown into the furnace
(derived from the moisture of the air) is equivalent to from J to 2
gal. of water per 10,000 cu. ft. of blast, or 1J to 8 gal. per minute,
depending on the humidity of the atmosphere". Though this steam
is as hot as the blast, it materially cools the smelting zone of the
furnace by dissociating there:

= 2 H+CO (absorbs 28,900 gram calories);

THE MANUFACTURE OF PIG IRON 39

or 1 Ib. of steam absorbs 7,110,000 calories. The hydrogen and
oxygen reunite in a cooler part of the furnace and return the same
amount of heat, but this does not compensate for that taken away
from the smelting zone, where it is most needed. For this reason
a few American plants, and at least one in England, have adopted
James Gayley's expedient of drying the air by refrigeration before
it is drawn into the blowing engine. This results in greater regu-
larity of furnace working and valuable saving in fuel. In fact, so
great is the economy shown in this respect that there was a tend-
ency at first to receive the results with skepticism. J. E. Johnson,
Jr., has explained this saving, however, in a very ingenious and
skillful manner, by showing that every blast furnace has a certain
'critical temperature/ below which it will not perform any smelt-
ing, and that the theoretical temperature of combustion of the
smelting zone is only a little above this 'critical temperature.'
To increase this small interval between the two, therefore, greatly
increases the 'available heat/ though the change in nominal tem-
perature be small.1

Slag Disposal. — On account of its lower specific gravity, the
slag floats on top of the bath of iron in the hearth and accumu-
lates, frequently until it reaches the bottom level of the tuyeres.
Four or five times every six hours the plug in the cinder-notch is
pierced with a steel rod and the cinder above this level allowed to
run out. It flows down an inclined iron runner for a distance of
15 to 30 ft. and pours into an iron ladle on a standard-gage railroad
track, whence it is drawn away by a locomotive and poured out on
the slag dump. Slag varies in composition according to the will of
the blast-furnace manager, and some typical analyses are given in
Table II. Slags high in lime are sometimes treated with addi-
tional lime to make a good grade of Portland cement, known as
'Puzzolini.' The amount of cinder made will depend on the
amount of silica, alumina, etc., in the ore, the amount of coke ash,
and the amount of flux, which will also depend on the desired slag
analysis. Under favorable circumstances, the slag may weigh
slightly less than half the iron; under other conditions it may weigh
nearly twice the iron.

1 If further explanation of this argument is needed, it may be found in the
following simile: Water boils at 212° F. If the temperature of a boiler is 262°,
there is a certain pressure of steam ; if we increase the temperature only 50°, we
greatly increase the pressure; yet 50° appears small in comparison to 262°.

40 THE METALLURGY OF IRON AND STEEL

Weight of Slag. — The amount of slag may be calculated from
the amount of lime (CaO) in the furnace, which may be calculated
from the percentage of lime in the limestone and other materials
charged into the furnace. Since all the lime charged goes into the
slag, the amount of the latter will be equal to the weight of lime
divided by the percentage of the lime in the slag. Thus, if we use
per ton of iron 1300 Ib. of limestone, containing 50 per cent, of
lime, there will be 650 Ib. of lime charged for every ton of iron
made. If the slag made contains 40 per cent, lime, then the weight
of slag will be ^j5/ = 1625 Ib. per ton of iron made.

Iron Disposal. — Immediately after the last 'flushing/ i.e., re-
moval of cinder, the tap hole or iron notch is opened by several
men "drilling a hole in it with a heavy, pointed steel bar. Out of

FIG. 23.

this notch flows 100 to 150 tons of liquid pig iron, with which is
carried along 30 tons or so of slag. The 'skimmer' is situated
about a dozen feet from the front of the furnace. It is an iron
plate extending down almost to the -bottom of the runner. The
slag is deflected by this plate into a runner of its own, which leads it
off to a slag ladle such as described before. The heavier pig iron
flows under the skimmer and is distributed to six or seven brick-
lined ladles on a standard-gage railroad track. It is then drawn

THE MANUFACTURE OF PIG IRON

away to the steel works, or, if not wanted there, is poured into iron
molds at the pig-casting machine.

Mechanical Pig-Molding Machine. — There are several types of
molding machine, but a common form is illustrated in Figs. 24-5, and
consists of a long continuous series of hollow metallic molds car-

FIG. 24. — UEHLING PIG-CASTING MACHINE.

ried on an endless chain. D is the pig-iron ladle pouring metal into
the spout, from whence it overflows into the molds as they travel
slowly past. The pig iron chills quickly against the metallic
molds, and by the time it reaches the other end of the machine, it
consists of a solid pig of iron which drops into the waiting railroad

FIG. 25. — HEYL AND PATTERSON PIG-CASTING MACHINE.

car as the chain passes over the sheave. The pig iron is now in a
form convenient for transportation or for storing until needed.
The molds travel back toward the spout, underneath the machine
and hollow side down. At the point C they are sprayed with
whitewash, the water of which is quickly dried off by the heat of the
mold, leaving a coating of lime inside to which the melted iron will

THE METALLURGY OF IRON AND STEEL

not stick. This mechanical casting is a great improvement over
the former method of cooling iron in front of the blast furnace,
because of the severity of the work which the former method in-
volved and which, in hot weather, was well-nigh intolerable to
human beings. It also gives pigs which are cleaner, i.e., freer from
adhering sand. This silicious sand is objectionable, especially in
the basic open-hearth furnace.

Sand-Casting. — This method is still used at some furnaces,
because of the capital needed to install machines and their high
cost for repairs. Moreover, foundrymen often prefer the sand-cast

pig because they are able to
tell by the appearance of its
fracture what grade of castings
it will make, which they cannot
well do with iron cast in metal
molds (see pages 336, 337).
In the sand method, the cast
house extends in front of the
furnace and its floor is com-
posed of silica sand, in which
the molds or impressions to
receive the liquid iron are
made. The main runner ex-
tends from the taphole down
the middle of the floor, and
the space on either side of it is
used alternately for alternate
castings. The plan of the ar-
rangement is shown in Fig. 26.
After cooling the iron, the
pigs are broken away from the
sows, which are also broken
into pieces with a sledge, and
then all is carried over and

thrown into a railroad car. In making ' basic iron/ — i.e., iron for
the basic open-hearth steel process, — the molds for the sows and
pigs are permanently made of metal, so that the iron will not
carry acid sand into the basic hearth.

Irregularities in Blast-Furnace Working. — Although the man-
agement and control of the operation is in general as I have de-

FIG. 26. — SAND-CASTING BED.

THE MANUFACTURE OF PIG IRON

scribed it, the blast furnace is by no means a perfect machine, and
great difficulties arise in the working of the furnace and in main-
taining a uniform grade of product. The chief of these difficulties

FIG. 27. — PIG BEDS.

44 THE METALLURGY OF IRON AND STEEL

result from localized chilling of the semi-molten charge. This is
most liable to happen in the upper part of the smelting zone, where
a little lump of pasty material may attach itself to the walls of the
furnace. This has the effect of hindering the descent of that part
of the charge above it and of deflecting the hot gases to other parts
of the furnace. The result of the first action is to disarrange the
order and evenness with which originally horizontal rings of stock
come down into the hearth. The obstruction is also liable to re-
ceive chilled materials from above and to build itself out toward
the center. When the furnace is working badly, these scaffolds may
occur at two or more places at the same time and cantilever out
toward the middle. This will cause a 'hanging' of the charge, and
may become so bad as to cause a complete arch over the smelting
zone, through which it is impossible to drive the blast. Sometimes
the scaffold may be broken down by suddenly cutting off the blast
pressure and allowing the full weight of material in the furnace to
come upon the obstruction; but sometimes it is necessary to cut a
hole in the wall of the furnace and melt it out with a blow-pipe
burning oil or gas, or with some other form of heat.

The i scaffolding7 of a furnace and hanging of the charge is more
liable to happen when large percentages of the earthy Mesabi ores
are used, and in this type of practice localized hanging and slips
are not infrequent. When the slip is extensive in character and a
large amount of material is suddenly precipitated into the hearth,
the upward rush of gases resembles an explosion inside the furnace
and may do damage to the charging apparatus and throw a part of
the stock out of the top of the furnace. Some furnaces are pro-
vided with explosion doors, which fly open under pressure and
relieve the strain; while the practice in other instances is to fasten
everything down as tight as possible and prevent the rapid escape
of the gases.

There is also a large amount of hanging due to the action of the
blast in tending to drive the stock before it up into the stack of the
furnace and thus compress it. This action is more liable to take
place with fine ores

Cooling of the charge, also, results in some cases in the freezing
of material over the mouths of the tuyeres. The solid layer may
sometimes be broken away with a bar, and the blow thus allowed
to continue until more heat can be brought down into the hearth.
Sometimes it is necessary to melt out the frozen material with a

THE MANUFACTURE OF PIG IRON 45

blow-pipe, and in extreme cases it may even be necessary to break
through it with explosives.

Another difficulty sometimes met with is the freezing up of the
metal in the lower part of the hearth, so that it is impossible to
open the tap-hole. Then a new tap-hole must be made by boring
through the front of the furnace at a higher level, from which the
iron is drained, and then the heat gradually worked down until the
whole hearth is melted out and normal conditions reestablished.
The bad work of a furnace is often cumulative in its effects, because
irregularities in the smelting zone have an effect upon the top gases,
which, in turn, derange the work of the stoves and hence impair
the hot blast.

These irregularities in the smelting have a disturbing effect upon
the character of the iron made, and the changes sometimes come
suddenly and without warning. For instance, a sudden precipita-
tion of cold material into the hearth will chill the smelting zone and
cause the silicon in the iron to be low and the sulphur high. The
same effect will be produced by the leakage of several gallons of
water into the hearth through the burning out of a tuyere or the
cooling-ring of one of the tuyeres.1

Dimensions of Blast Furnace. — The size of a modern blast
furnace is limited by the conditions of its work: the hearth may
not be much more than 15 ft. in diameter, else the blast from the
tuyeres will not be distributed evenly to the center; the batter of
the bosh walls cannot be much more nor less than a certain amount,
because they must give support to the charge above them, and yet
allow the solid coke to slip down; the height of bosh is limited, be-
cause its top must be practically the same as the top of the smelting
zone, — that is, no solid material except coke should descend into
the bosh. These conditions therefore limit the diameter of the
top of the bosh to not much more than 22 ft. From the bosh the
stack walls must decrease in diameter upward in order that the
descending charge, which swells in the reactions that take place
from the throat downward, shall not become wedged in the stack;

1 In iny early days at the blast furnace I was once informed by the assist-
ant manager that, on one occasion, he tapped several tons of water from the
tap-hole with the iron. Whether he was himself deceived or whether he was
merely trying to test me, I have never been able to decide; but the fact is
worthy of mention in an elementary treatise to illustrate the character of the
tales to which even the educated men around a plant will treat a novice.

THE METALLURGY OF IRON AND STEEL

as the throat must have a sufficiently large diameter to properly
charge the materials, this limits the height of the stack. Modern
furnaces are therefore usually built about 90 ft. in height, and the
exceeding of that limit has resulted in some cases in a decrease,
rather than an increase of fuel economy.

CALCULATING A BLAST-FURNACE CHARGE

This subject is of prime importance to young metallurgists,
because the ability to calculate a charge is sometimes a cause of
advancement, and the knowledge of the way to do so is not always
obtainable from one's superior.

Assumptions. — Let us assume that we desire to produce a slag
containing 55 per cent, lime, 15 per cent, alumina, and 30 per cent,
silica, these proportions being determined by the experience of the
manager, and that the materials from which the charge is to be
made analyze according to Table III. Assume furthermore that
the coke ash is equal to 10 per cent, of the coke, and that the iron
we are going to make will contain about 1 per cent, silicon.

TABLE III

MATERIAL

Per cent.
CaO

Per cent.
MgO

Per cent.
A1203

Per cent.
SiO8

Per cent.
Fe203

Per cent.
Fe

Ore A
Ore B

5
2

2
12

11
16

Coke ;i sh

Limestone

Silicon in the Iron. — This last assumption necessitates our
allowing a corresponding amount of silica, because the silica
reduced and absorbed by the iron will not be available for slag-
making purposes. One per cent, of silicon is roughly equal to 2
per cent, of silica ; we may therefore make the requisite allowance
by subtracting from the silica in each material an amount equiva-
lent to 2 per cent, of its iron content. Thus we begin to make up
Table IV.

Magnesia. — In considering slags, magnesia is classified under
the head of lime. We cannot do this, however, by a simple addi-
tion of the figures of the percentages, because 1 per cent, of mag-

THE MANUFACTURE OF PIG IRON

jiesia will do the chemical work of 1.4 per cent, of lime, on account
of the difference in molecular weight (CaO = 56 ; MgO = 40) . Thus :

CaO + SiO2 = CaSiO3;
MgO + SiO2 = MgSiOs.

We thus multiply each percentage of magnesia by 1.4 and add
the product thus obtained to the percentage of lime in each ma-
terial, thus obtaining column 2 in Table IV.

TABLE IV

MATERIAL

Per cent.
CaO

Per cent.
A1203

Per cent.
SiOa

Ore 4

Ore B

Coke ash

Limestone

Self fluxing of Materials. — It is evident that in so far as each
of the materials in Table IV contains all the components of the slag,
they will partially flux themselves. For example, the 2 per cent,
of alumina in ore A will theoretically combine with 4 per cent, of
the silica (2 per cent. X-f-J =4 per cent.) and 7 per cent, of the lime
(2 per cent. X-j-j- = 7.3 per cent.) to make a slag of the desired pro-
portions, leaving unfluxed percentages as per the first line of Table
V. In the same manner we may use up all of the lime in ore B by
uniting it with weights of alumina and silica in proportion to the
percentages of these components in the slag. Similar simplifica-
tions in the analyses of coke ash and limestone may then be calcu-
lated, and Table V will be completed.

TABLE V

MATERIAL

Per cent.
CaO

Per cent.
A1203

Per cent.
SiOa

Ore A

Ore B

11 5

Coke ash

13 0

Limestone

Weight of Charge. — Let us assume that we are going to make
one ton of pig iron for every ton of coke used in the charge, and
that the coke will be put in in charges weighing 11,000 Ib. each.
This weight includes about 10 per cent, of moisture, dust, etc., so

THE METALLURGY OF IRON AND STEEL

we calculate with it as if it weighed only 10,000 Ib. Now let us
determine how much ore will be put in each charge: The ores
average 55 per cent, of iron; therefore, g I 0p£0J»t = 18,000 Ib., the
amount of ore that must be in each charge, according to the assump-
tion of this paragraph.

Adjusting the Alumina and Silica. — Next adjust the different
materials so that the weight of alumina shall be -J-§- of the weight of
silica. In the first rough approximation of this we may neglect the
coke ash, because the weight of this ash is so small in relation to
the other materials. Therefore only the two ores need be appor-
tioned, and we quickly find by trying a few mixtures at random
that 60 per cent, of ore A mixed with 40 per cent, of ore B will give
the desired relation i1 60 per cent. X 6 + 40 per cent. X 14 = 920 parts
of silica; 60 per cent. XO + 40 per cent. Xl 1.5 =460 parts alumina;
4}jj =£|. Now draw Table VI, and enter 10,800 Ib. of ore A ( =60
per cent, of 18,000), 7200 Ib. of ore B, 1000 Ib. of coke ash ( = 10
per cent, of 10,000), and the percentages from Table V. All the
weights in this table may then be filled in except those of the lime-
stone and total CaO.

To obtain the total number of pounds of lime :

A12O3: 958 X fl! = 3513 Ib.
SiO2: 2046 X M = 3751 Ib.
Average of 3515 and 3751 is 3632.

TABLE VI

MATERIAL

A13(

Si<

Weight.

Per

Cent.

Lb.

Per
Cent.

Lb.

Per
cent.

Lb.

Ore A

10,800

216

648

Ore B

7200

11 5

828

1008

Coke ash

1 000

13 0

130

390

Limestone

7,940

3416

Total Ib

A
3632

958

2046

1 Try first 50 per cent, of each, and we see that there is too much alumina;
therefore try less than 50 per cent, of the ore having the most alumina, and
correspondingly more of the other, and we have it.

THE MANUFACTURE OF PIG IRON

Adjusting for Lime. — It is now only necessary to determine the
amount of total lime that shall bear the correct relation to the
alumina and silica calculated. This we do by means of the method
shown in the figures above Table VI. We enter this in the square
' A ' of Table VI. The figures at the square ' B ' are then obtained
(3632 — 216=3416), and thence the weight of limestone to be used
— (3416^-43 per cent. =7940).

Checking the Calculations. — We now check up all the calcula-
tions by making up Table VII, in which we go back to the original
percentages found by chemical analysis and given in Table III.
In making up this final table, however, we use our experience in
making slag calculations and estimate slight changes: For ex-
ample, Table VI shows us that the alumina comes a little low in
relation to silica; therefore we increase ore B, say, by 400 Ib. and
decrease ore A correspondingly. But ore A is high in lime; there-
fore we use a little more limestone to offset this reduction.

TABLE VII

MATERIAL

CaO + MgO

Ai20s

Si02

Weight.

Per
cent.

Lb.

Per

cent.

Lb.

Per
cent.

Lb.

Per
cent.

Lb.

Ore A

10400
7600
1000
8200

9
2
20
50

936
152
200
4100

2
12
18
2

208
912
180
164

11
16
50

1144
1216
500
328

60
50
10
2

6240
3800
100
164

Ore B

Coke ash
Limestone

Total weights

5388

3 per
t.

1464

= 14. 9 per
cent.

. .3188 10,304
-206(=2%X10,304)

cen

2982=30

. 3 per cent.

These figures are much closer to those desired than the limit of
accuracy in furnace operation. The chief difference is that we are
making a little more iron with 10,000 Ib. of coke than we intended.
If any change seems necessary it is then well to reduce the weight
of ore A to 10,000, leaving everyth ng else the same. This will
lighten the burden and bring the calculated lime, alumina and
silica even closer to the desired figures.

Phosphorus and Manganese. — No account of the phosphorus
has been taken in the calculation above. This is necessary some-

50 THE METALLURGY OF IRON AND STEEL

times. For example, if ore A happened to be very high in phos-
phorus we could not use so large a proportion of it. It would then
be necessary either to secure another ore low in both phosphorus
and alumina, or else to make a slag with less alumina. The same
line of reasoning applies to manganese.

III

THE PURIFICATION OF PIG IRON

THE large amount of carbon in pig iron makes it both weak and
brittle, so that it is unfit for most engineering purposes. It is used
for castings that are to be subjected only to compression, or trans-
verse or very slight tensile strains, as, for example, supporting
columns, engine bed-plates, railroad car wheels, water mains, etc.,
but the relatively increasing amount of steel used shows the prefer-
ence of engineers for the stronger and more ductile material.
To-day three-fourths of the pig iron made in the United States is
subsequently purified by either the Bessemer, open-hearth, or
puddling process. Each of these will reduce the carbon to any
desired point, while the silicon and manganese are eliminated as a
necessary accompaniment of the reactions, — indeed, we might
almost say as a condition precedent to carbon reduction. Phos-
phorus and sulphur are reduced by the puddling process, and by
a special form of open-hearth process known as 'the basic open-
hearth process.'1 The complete scheme of American iron and
steel manufacture is given in Fig. 29.

Explanation of Fig. 29. — Practically all the iron ore mined is
smelted in about 325 blast furnaces, producing annually 25,000,000
tons of pig iron. About 3 per cent.2 of this pig iron is remelted
and made into malleable cast iron; 20 per cent, is remelted and cast
as gray cast iron; 52 per cent, is purified in 62 Bessemer converters
to Bessemer steel; 20 per cent, is purified in 465 basic open-hearth
furnaces; 2 per cent, is purified in 195 acid open-hearth furnaces,
while the remaining 3 per cent, is purified in 3000 puddling fur-
naces to make wrought iron. The wrought iron may be used as
such for pipe, blacksmith work, small structural shapes, etc., and

1 The basic Bessemer process is not in operation in America.

2 The numbers and percentages given in this figure will change slightly
from time to time, but this will convey to a beginner an idea of the relative
amounts of the different products made.

THE METALLURGY OF IRON AND STEEL

95 per cent, of it is so used; the other 5 per cent, is remelted in
crucibles to make crucible steel. To sum up, about 23 per cent,
of the pig iron made is used without purification/ and 77 per cent,
is purified and converted into another form. In all cases of purifi-
cation the impurities are removed by oxidizing them; and we must

Skeleton of American Iron and Steel Manufacture
1903

50,032,279 Tons of Iron Ore

25,307,192 Tons of Pig Iron

Bessemer
.Converters

Bessemer Steel Basic Open-
(12,275,250 ) Hearth Steel
(9,649,400;

Acid Open-

Hearth Steel

(1,321,613;

Wrought Iron

Used
as Such

Crucible Steel
(118,000)

All Tons are 2,240 Ibs. each

FIG. 29.

again emphasize the rule that unoxidized elements dissolve in the
metal, while those in the oxidized condition pass into the slag, or,,
if there is no slag, form a slag for and of themselves. In consider-
ing the Bessemer, open-hearth, and puddling processes then, we
have to do with oxidizing conditions, \vhereas the opposite was the

1 It is true that the annealing process for malleable cast iron purifies the
outer layers of the castings from carbon, and, if the castings are very thin,
this purification may extend to the center; but this is not primarily a purifi-
cation process and will be treated at length in another section.

THE PURIFICATION OF PIG IRON

case in the blast furnace. The oxidation is effected by means of
the oxygen of the air or that of iron ore, Fe2O3, or its equivalent,
or of both air and oxide of iron.

There is not an exact relation between the amounts of pig iron
used for the different purposes and the amounts of the resulting
materials. In 1906 the following production was made:

PRODUCTION FOR 1906

Cast Iron Used

Made

Malleable cast iron

600,000* tor
5,100,000 '
13,150,000 '
5,150,000 '
500,000 '
800,000 '

750,000* to]
6,000,000
12,275,250
9,649,400
1,321,613
2,000^000*
118,000

Gray cast iron ....

Bessemer steel

Basic open-hearth steel

Acid open-hearth steel
Wrought iron

Crucible steel

* Estimated.

The reason for the discrepancy is found in the scrap iron or
steel mixed with the pig iron in the manufacture of gray-iron cast-
ings and open-hearth steel. Perhaps an average of 25 per cent,
of old scrap will be mixed with 75 per cent, of new pig iron for
making iron castings, and 50 per cent, or so of steel scrap will be
mixed with 50 per cent, or so of pig iron in the open-hearth proc-
ess, while wrought iron is often made by the piling and rerolling
of old wrought-iron scrap.

Bessemer Process. — In the Bessemer process, perhaps 10 tons
of melted pig iron is poured into a hollow pear-shaped converter
lined with silicious material. Through the molten material is then
forced 25,000 cu. ft. of cold air per minute. In about four minutes
the silicon and manganese are all oxidized by the oxygen of the air
and have formed a slag. The carbon then begins to oxidize to
carbon monoxide, CO, and this boils up through the metal and
pours out of the mouth of the vessel in a long brilliant flame. After*
another six minutes the flame shortens or 'drops'; the operator
knows that the carbon has been eliminated to the lowest practi-
cable limit (say 0.04 per cent.) and the operation is stopped. So
great has been the heat evolved by the oxidation of the impurities
that the temperature is now higher than it was at the start,
and we have a white-hot liquid mass of relatively pure metal.
To this is added a carefully calculated amount of carbon to

THE METALLURGY OF IRON AND STEEL

produce the desired degree of strength or hardness, or both;
also about 1.5 per cent, of manganese and 0.2 per cent, of silicon.1
The manganese is added to remove from the bath the oxygen with
which it has become charged during the operation and which would
render the steel unfit for use. The silicon is added to get rid of the

FIG. 30. — SECTION THROUGH BESSEMER CONVERTER WHILE BLOWING.

gases which are contained in the bath. After adding these ma-
terials, or ' recarburizing/ as it is called, the metal is poured into
ingots, which are allowed to solidify and then rolled, while hot,
into the desired size and form. The characteristics of the Besse-
mer process are: (a) Great rapidity of purification (say ten min-
utes per 'heat'); (6) no extraneous fuel is used; and (c) the metal
is not melted in the furnace where the purification takes place.
1 In the case of making rail steel.

THE PURIFICATION OF PIG IRON

Acid Open-hearth Process. — The acid open-hearth furnace is
heated by burning within it gas and air, each of which has been
highly preheated before it enters the combustion chamber. A
section of the furnace is shown in Fig. 32. The metal lies in a
shallow pool on the long hearth, composed of silicious material,
and is heated by radiation from the intense flame produced as
described. The impurities are oxidized by an excess of oxygen in
the furnace gases over that necessary to burn the gas. This action
is so slow, however, that the 3 to 4 per cent, of carbon in the pig

— (ff* •' ,

FIG. 31. — BLOWHOLES OR GAS-BUBBLES IN STEEL.

iron takes a long time for combustion. The operation is therefore
hastened in two ways : (a) iron ore is added to the bath to produce
the reaction

Fe2O3 -f 3C= SCO -f 2Fe,

and (6) the carbon is diluted by adding varying amounts of cold
steel scrap. The steel scrap is added to the furnace charge at
the beginning of the process, and it takes about 6 to 10 hours to
purify a charge, after which we recarburize and cast the metal
into ingots. The characteristics of the open-hearth process are:
(a) A long time occupied in purification; (6) large charges are
treated in the furnace (the modern practice is usually 30 to 70
tons to a furnace) ; (c) at least a part of the charge is melted in the

THE METALLURGY OF IRON AND STEEL

purification furnace; and (d) the furnace is heated with preheated
gas and air.

Basic Open-Hearth Process. — The basic open-hearth opera-
tion is similar to the acid open-hearth process, with the difference

FIG. 32. — DIAGRAM OF REGENERATIVE OPEN-HEARTH FURNACE.

The four chambers below this furnace are filled with checkerwork of brick with hori-
zontal and vertical channels through which the gas and air may pass. The gas enters the
furnace through the inner regenerative chamber on one side and the air enters through
the corresponding outer one. They meet and unite, passing through the furnace and
thence dividing into proportional parts and passing to the chimney through the two regen-
erative chambers at the opposite end. In this way the brickwork in the chambers is heated
up by the waste heat of the furnace. The current of gas, air, and products of combustion
is changed every twenty minutes whereby all four regenerators are always kept hot. The
gas and air enter in a highly preheated condition and thus give a greater temperature of
combustion, while the products of combustion go out to the chimney at a relatively low
heat and thus fuel economy is promoted.

that we add to the bath a sufficient amount of lime to form a very
basic slag. This slag will dissolve all the phosphorus that is oxi-
dized, which an acid slag will not do. We can oxidize the phos-

THE PURIFICATION OF PIG IRON

phorus in any of these processes, but in the acid Bessemer and the
acid open-hearth furnaces the highly silicious slag rejects the
phosphorus, and it is immediately deoxidized again and returns
to the iron. The characteristics of the basic open-hearth process
are the same as those of the acid open-hearth, with the addition
of: (e) Lime is added to produce a basic slag; (/) the hearth is
lined with basic, instead of silicious, material in order that it may
not be eaten away by this slag; and (g) impure iron and scrap may
be used, because phosphorus and, to a limited extent, sulphur
can be removed in the operation.

Puddling Process. — Almost all the wrought iron to-day is
made by the puddling process, invented by Henry Cort about
1780, with certain valuable improvements made by Joseph Hall
fifty years later. In this process the pig iron is melted on the
hearth of a reverberatory furnace lined with oxide of iron. During

FIG. 33. — PUDDLING FURNACE.

the melting there is an elimination of silicon and manganese and
the formation of a slag which automatically adjusts itself to a very
high content of iron oxide by dissolving it from the lining. After
melting, the heat is reduced and a reaction set up between the
iron oxide of the slag and the silicon, manganese, carbon, phos-
phorus and sulphur of the bath, whereby the impurities are oxi-
dized and all removed to a greater or less extent. The slag, be-
cause of its basicity (by iron oxide), will retain all the phosphorus
oxidized, and therefore the greater part of this element is removed.
The oxidation of all the impurities is produced chiefly by the iron

THE METALLURGY OF IRON AND STEEL

oxide in the slag and the lining of the furnace, although it is
probable that excess oxygen in the furnace gases assists, the slag
acting as a carrier of oxygen from it to the impurities.

The purification finally reaches that stage at which the utmost
heat of the furnace is not sufficient to keep the charge molten,
because iron, like almost every other metal, melts at a higher tem-
perature the purer it is. The metal therefore l comes to nature/
as it is called, that is to say, it assumes a pasty state. The iron
is rolled up into several balls, weighing 125 to 180 Ib. apiece,
which are removed from the furnace, dripping with slag, and car-
ried over to an apparatus, where they are squeezed into a much
smaller size and a large amount of slag separated from them.
The squeezed ball is then rolled between grooved rolls to a bar,
whereby the slag is still further reduced, so that the bar contains
at the end usually about 1 or 2 per cent. This puddled bar, or
'muck bar/ is cut into strips and piled 'up, as shown in Fig. 34,

FIG. 34. — METHOD OF PILING MUCK BAR.

into a bundle of bars which are bound together by wire, raised to a
welding heat, and again rolled into a smaller size. This rolled
material is then known as ' merchant bar/ and all wrought iron,
except that which is to be used for manufacture into crucible steel,
is treated in this way before sale. The effect of the further rolling
is to eject more slag, and also to make a cross network of fibers,
instead of a line of fibers all running in the same direction, i.e.,
lengthwise of the bar. The fibers are produced by the action in
rolling of drawing out the slag into strings, long fibers of metal
also being produced, each of which is surrounded by an envelope
of slag.

THE PURIFICATION OF PIG IRON 59

COMPARISON OF PURIFICATION PROCESSES

Acid with Basic Open-Hearth. — Acid open-hearth steel is be-
lieved by engineers to be better than basic, and is usually specified
for in all important parts of structures, although not so rigidly
to-day as a few years ago. This is in spite of the fact that phos-
phorus and sulphur, two very harmful elements, are lower in the
basic steel. The basic process is much less expensive than the
acid, because high phosphorus pig iron and scrap are cheap, and
the lower cost of materials used more than balances the greater
cost of the basic lining and the lime additions and the circumstance
that the acid furnace has a higher output because the heats are
shorter. The reasons for the preference of acid steel are as follows :

(a) A basic slag will dissolve silicon from the metal; we there-
fore recarburize in the basic process by adding the recarburizer to
the steel after it has left the furnace, instead of in the furnace, as
we do in the acid process. Should any basic slag be carried over
with the metal, however, which is liable to happen, there is the
danger that the ingots will be too low in silicon. They are then
impregnated with gas bubbles, or 'blow holes/

(b) Moreover, the recarburizer does not mix with the steel as
well if it is not added in the furnace, and this sometimes produces
irregularities.

(c) A basic slag is usually more highly oxidized than an acid
one ; therefore the metal at the end of the operation is more highly
charged with oxygen. For this reason we add a larger amount of
manganese in the recarburizer, but the remedy is never quite as
good as prevention.

(d) Since we cannot remove the phosphorus from the bath in
the acid process, it is necessary to use only picked iron and scrap,
whereas, in the basic process, good steel can be made from almost
any quality of material. Many engineers believe, however, that a
better grade of steel results from using the picked material.

(e) It occasionally happens in the basic process that, after the
phosphorus has all been oxidized in the slag and the operation is
ended, some of it will get back into the metal again. This is
especially liable to happen when basic slag is carried over into the
ladle before the recarburizer is.all in. If this occurs, and if the bath
is very hot, a reaction may take place between the basic slag and

60 THE METALLURGY OF IRON AND STEEL

the acid lining of the ladle. In this way the slag will be enriched
in silica and phosphorus will be forced out of it.

Basic Open-Hearth with Bessemer. — Basic open-hearth steel
is better than Bessemer steel. The reasons for this are believed
to be:

(a) The open-hearth process being slower, more attention and
care can be given to each detail. This is particularly true of the
ending of the process, for if the Bessemer process is continued only
a second or so too long, the bath is highly charged with oxygen,
to its detriment, and even under normal circumstances there is
more oxygen in the metal at the end of the Bessemer process than
at the end of the basic open-hearth, because there has been so
intimate a mixture between metal and air.

(b) For the same reason the Bessemer metal is believed to con-
tain more nitrogen and hydrogen,1 which are thought to be dele-
terious.

(c) The Heat of the Bessemer process is dependent upon the im-
purities in the pig iron, and especially upon the amount of the
silicon, and can be controlled only to a limited extent by methods
that are not perfect in their operation. Furthermore, the heat is
regulated according to the judgment of the operator and his skill
in estimating the temperature of the flame. Irregularities there-
fore result at times, and these produce an effect on the steel, be-
cause the temperature at which the ingots are cast should be
neither too high nor too low. It is true that the temperature of the
open-hearth steel is also regulated by the judgment of the operator,
but more time is afforded for exercising this judgment and for
controlling the heat.

(d) In the Bessemer process we must get rid of all the carbon
first and then recarburize to the desired point. In the open-hearth
process we may stop the operation at any desired amount of carbon,
and then recarburize only a small amount. Therefore the open-
hearth has the advantage of greater homogeneity when making
high-carbon steel, since a very large amount of recarburizer may
not distribute itself uniformly.

(e) In order to produce the best quality of steel, it must be cast
into ingot molds within a certain limited range of temperature,
which varies according to the amount of carbon, etc., that it con-
tains. Therefore, in casting the very large heats of the open-

1 From moisture in the blast.

THE PURIFICATION OF PIG IRON 61

hearth process, the ingots must be very large, else the first one will
be too hot and the last one too cold for the best results. On the
other hand, if the ingots are large, segregation is liable to be ex-
cessive (see page 180).

For nearly fifteen years the Bessemer process has been fighting
a losing battle to maintain its supremacy against the inroads of
the basic open-hearth, which have been possible because of the
increasing cost of Bessemer pig iron, due to the exhaustion of the
low phosphorus ores. The pig iron for the Bessemer process must
contain so little phosphorus that, after allowing 10 per cent, loss of
metal during the blow, the phosphorus in the steel shall be not
over 0.100 per cent. Ores low enough in phosphorus to make
this grade of metal have, therefore, come to be known as ' Besse-
mer ores/ The requirement of such an ore is that the per-
centage of iron in it must be 1000 times the percentage of phos-
phorus. During the year 1906, the Bessemer process in the United
States yielded very much to the basic open-hearth, and it would
seem as if there was no chance of its ever taking up so impor-
tant a position again unless new iron ores low in phosphorus are
discovered.

On account of its ability to make low carbon steel more readily
than the basic open-hearth, the Bessemer process has a firm hold
on the wire and welded steel-pipe industry, although even here the
open-hearth process has encroached. For rolling very thin for
tinplate, etc., we want a metal relatively high in phosphorus, and
therefore the Bessemer process is largely used here, although in
some cases ferrophosphorus is being added to basic open-hearth
metal to accomplish the same result. The reason phosphorus is
desired is because the plates are rolled very thin by doubling
them up and putting several thicknesses through the rolls at the
same time. Low phosphorus metal welds together too much
under these circumstances.

The chief requisites of railroad rails are lack of brittieness and
ability to withstand wear. The Bessemer process is able to pro-
vide such a material, and it works so well in conjunction with the
rapid, continuous operation of the rail-rolling mill that it has a
decided advantage. It produces a small tonnage of ingots at
frequent intervals (say 15 tons every 7 minutes), while the open-
hearth process provides a large tonnage of ingots, which may come
at irregular intervals and thus alternately delay and overcrowd

THE METALLURGY OF IRON AND STEEL

the rail-mill operations. But notwithstanding these advantages,
an increasing tonnage of basic open-hearth rails is made every year
in the United States.

During 1907 this phase of the industry has attracted wide-
spread interest, owing to reports of an alarming number of rail
breakages and of the action of some railroads in blaming the
Bessemer process therefor. It is true that every year there is a
greater scarcity of Bessemer ores and therefore an increasing
amount of phosphorus in the steel manufactured, so that it is no
secret that many rails have been made within the past year
containing more than the allowable 0.1 per cent, phosphorus.
Phosphorus makes the steel brittle, especially under shock and
in cold weather. It also makes the steel hard and more able to
resist wear; but this hardness is better obtained by means of
carbon, and low-phosphorus, high-carbon steel rails would un-
doubtedly break less often in the track. It is to be remembered
that heavier trains are being run every year, and that this brings
greater strains upon the rails, to meet which they have not been
correspondingly increased in size. At the present time such a
very large amount of capital is tied up in Bessemer-rail mills, and
it would take so long to change them over into open-hearth
mills, that there is no immediate liability of a great replacement.
The acid- and basic-steel production of the principal countries
of the world is shown in Tables IX and X, while the recent
history of open-hearth steel-rail manufacture is shown briefly in
Table VIII.

TABLE VIIL— AMERICAN RAILROAD RAIL MANUFACTURE

Bessemer J
Gross Tons

Open-Hearth 2
Gross Tons

Wrought Iron3
Gross Tons

Total
Gross Tons

1900

2 383 654

1 333

695

2 385 682

1901 . .

2 870 816

2093

1 730

2 874 639

1902

2 935 392

6029

6 512

2 947 933

1903

2 946 756

45054

667

2 992 477

1904

2 137 957

145 883

871

2 284 711

1905

3,188,675

183 264

318

3 372 257

1906

3 700 000*

250 000*

* Estimated.

1 The first Bessemer rails were made commercially in 1867.

2 In 1881, 22,515 gross tons of open-hearth rails were produced. The
first open-hearth rails were made in 1878.

3 The maximum production of iron rails was 808,866 gross tons in 1872.

THE PURIFICATION OF PIG IRON 63

TABLE IX.— STEEL PRODUCTION OF PRINCIPAL COUNTRIES: 1906

United States

Germany

Great Britain

Acid converter

12,275,253

407,688

1,307,149

Basic converter

6,772,804

600,189

Total converter

12,275,253

7,180,492

1,907,338

Acid open-hearth

1 321 613

230,668

3,378,691

Basic open-hearth

9 649 385

3 534 612

1 176 245

Total open-hearth . . .

10,970 998

3,765,280

4,554,936

Crucible and special

118,500

189,313

Total

23,364,751

11,135,085

6,462,274

Proportion steel to pig iron

92.3

89.2

63.7

TABLE X.— MAKE OF ACID AND BASIC STEEL: 1906

ACID

BASIC

Tons

Per cent.

Tons

Per cent.

United States

13 715 366

58 7

9 649 385

41 3

Germany

715 952

6 4

10 419 133

93 6

Great Britain

4 685 840

72 5

1 776 434

27 5

Total

19 117 158

46 7

21 844 952

53 3

Crucible Steel with Others. — Crucible steel is the most expen-
sive of all and costs at least three times as much as the next in
price — acid open-hearth steel. It is also the best quality of steel
manufactured, and for very severe service, such as the points
and edges of cutting tools, the highest grades of springs, armor-
piercing projectiles, etc., it should always be employed. The
reason for its superiority is believed to be because it is manufac-
tured in a vessel which excludes the air and furnace gases, and is
therefore freer from oxygen, hydrogen and nitrogen. Perhaps
the fact that the process is in some ways under a little better con-
trol than any of the others, and receives more care, on account of
being manufactured in small units, assists in raising its grade.
Crucible steels are usually higher in carbon than Bessemer and
open-hearth steels, because the special service to which the cru-
cible steels are adapted is usually one requiring steel that can
be hardened and tempered — for example, cutting tools, springs,
etc., and only the high-carbon steels are capable of this hardening
and tempering.

64 THE METALLURGY OF IRON AND STEEL

Wrought Iron with Low-Carbon Steel. — Wrought iron costs from
10 to 20 per cent, more than the cheapest steel. Its claim to superi-
ority over dead-soft steel consists in its purity and the presence in
it of slag. Just how much advantage the slag is has never been
proven; it gives the metal a fibrous structure which, perhaps, in-
creases its toughness and its resistance to breaking under bending
or under a sudden blow or shock. Some think that the slag also
assists in the welding of the material, but this is doubtful, and it is
probable that the easy weldability of wrought iron is due alone
to its being low in carbon. Some also believe that the slag assists
the metal in resisting corrosion ; hence one reason for the preference
of engineers for wrought-iron pipe for boilers and other purposes.
There are other qualities of wrought iron which may tend to make
it corrode less than steel, chief among which are the absence of
blowholes and possibly the absence of manganese, and the pres-
ence of phosphorus. It is now believed by many that manganese
starts an electrolytic action which hastens corrosion. An ad-
vantage of wrought iron in this connection is its rough surface to
which paint or other protective coatings will adhere more firmly
than to the comparatively smooth surface of steel. Nevertheless
the evidence goes to show that properly made steel corrodes no
more than wrought iron, especially in boilers, pipe, and other
articles which cannot be coated.

The properties of wrought iron are the nearest to those of pure
iron of any commercial material, notwithstanding its slag. This
is because the slag is mechanically mingled with the metal and
does not interfere with its chemical or physical actions. There-
fore wrought iron is greatly preferred for electrical conductivity
purposes and as a metal with high magnetic power, for armatures
of electromagnets, etc.

The advantages I have mentioned, the conservatism of engi-
neers, and the capital previously invested in puddling furnaces
are the chief factors in keeping alive the manufacture of wrought
iron. It was freely predicted that the invention of the Bessemer
and open-hearth processes would bring about the extinction of
the puddling process, but these prophecies have never been ful-
filled, although the importance of wrought iron has waned very
greatly in fifty years. When under strain greater than it can
withstand wrought iron stretches more uniformly over its' entire
length than steel, as shown by the following tests:

THE PURIFICATION OF PIG IRON

STRAIN TESTS ON WROUGHT IRON

Elastic Limit

Ultimate
Strength

Elongation
per cent.

Reduc-
tion of
Area

Per cent.
28.30
51\50

Wrought iron .
Steel

Lb. per sq. in.

31,550
33,150

Lb. per sq. in.
48,810
59,260

In 12 in.
23
39

In 18 ft.
15.22
14.40

Summary. — In order of expense and of quality the different
steels are arranged as follows: (1) Crucible, (2) acid open-hearth,
(3) basic open-hearth, and (4) Bessemer. The amounts of the dif-
ferent kinds made in America to-day and ten years ago are shown in
Table XI. Though I have not made a direct comparison between
certain of the classes, e.g., acid open-hearth with Bessemer, their
relations may be easily learned by collating the other comparisons
given.

TABLE XI

1906..
1896..

Bessemer

12,275,253 52%
4,909,128 78%

Open-Hearth
9,649,385 41%
776,256 12%

1,321,613
522,444

6%
9%

Crucible, etc.
118,500 1%
68,524 1%

Many engineers will be interested in the uses to which the
annual steel and wrought iron production of the United States
is put, which are shown below:

TABLE XII.— USES OF STEEL AND WROUGHT IRON. 1906.

GROSS TONS, 1906

Steel, tons

Wrought
Iron, tons

Railroad rails

3 977 872

Railroad rail splice bars

213 977

10,934

Plates and sheets

4 107 783

74,373

Structural shapes

2 114053

4,719

Merchant bars*

2 510 852

1 481,348

Rods for wire and wire products

1 310 413

1,201

Rods for wire nails

560 000§

Plate for cut nails

37032

17 179

Skelpf

1 137 068

391 517

Hoops, bands and cotton ties

579018

1 332

Blooms and billetsj

205 648

462

All other rolled shapes

648 195

203 477

Castings

700 000 §

Totals

18 101 911

2 186 557

* Merchant bars are small bars to be worked up into other forms.
t Skelp is welded into pipe.

t Blooms and billets are larger pieces to be worked up farther.
$ Estimated.

66 THE METALLURGY OF IRON AND STEEL

DISTINGUISHING BETWEEN THE DIFFERENT PRODUCTS

Low-carbon steel pipe, merchant bars, horseshoe blanks,
etc., sometimes masquerade under the name of wrought-iron ;
high-carbon open-hearth and Bessemer-steel merchant bars,
tool blanks, etc., sometimes masquerade as 'crucible steel/ or
perhaps 'cast steel/ which is the trade name for crucible steel;
other deceptions are not unknown; indeed, even malleable cast
iron is sold oftentimes as 'steel castings.' It is therefore im-
portant for engineers to understand the essential differences
between these materials, although care in the wording of contracts
and specifications should be the important consideration and
should precede watchfulness over the products. The definitions
of iron and steel materials are in such a confused and unsettled
condition that it does not do to rely upon them at all, especially
where a lawsuit may be involved; and contracts in clear, simple
language, free from legal and metallurgical phraseology, are the
best safeguards. But even where it is entirely plain what material
is called for, there is always a temptation to substitute steel for
wrought iron, Bessemer for open-hearth, basic for acid, and
Bessemer or open-hearth for crucible, steel. In case one such
substitution is suspected, there are means by which the material
may be tested, aside from its strength and ductility, which may
or may not be in the contract. The tests are somewhat delicate
and usually require the judgment and experience of an expert—
one who has standard samples of the different grades of material
for comparison, because the details of manufacture vary from
district to district, and still more so with the purposes for which
the products are to be used.

Wrought iron may be distinguished from low-carbon 'steel
by the fact that it contains slag. Usually, there is more than
1 per cent, of slag in iron and less than 0.2 per cent, of slag (in-
cluding metallic oxides) in steel. The slag may be determined
either by chemical or microscopical analysis. Normal wrought
iron is practically free from manganese, while normal Bessemer
and open-hearth steel will contain 0.5 per cent, or more. Wrought
iron generally contains more than 0.1 per cent, phosphorus, while
good steel should never do so.

Crucible steel normally has less than 0.4 per cent, manganese
and more than 0.2 per cent, silicon, while open-hearth and Bessemer

THE PURIFICATION OF PIG IRON 67

steels normally have more than 0.4 per cent, manganese and less
than 0.2 per cent, silicon. In the case of steel castings, however,
this rule for silicon does not apply, as Bessemer and open-hearth
steel castings are sometimes as high as 0.6 per cent, in silicon.
It is possible to make both Bessemer and open-hearth steels low
in manganese, but they cannot be made low in both manganese
and silicon without great danger from blow-holes, while this
difficulty is not met with to the same extent in crucible steel.
When crucible steel, low in carbon, is ordered, there is a much
greater temptation to substitute another steel for it.

Acid open-hearth steel may be distinguished from basic open-
hearth steel by its being normally higher in silicon, and usually
in phosphorus also, but lower in manganese. The same differences
exist between acid and basic Bessemer steel.

Basic open-hearth steel may be distinguished from Bessemer
steel by its lower manganese, silicon, phosphorus and (generally)
sulphur, as well as by the fact that it dissolves much more slowly
in dilute hydrochloric acid.

It is possible to place such physical specifications in a contract
as to practically insure obtaining the grade of material ordered.
For example, such a high degree of ductility may be demanded,
especially the percentage elongation in ten or twenty feet, that
nothing but wrought iron will give it; the strength and ductility
may be put so high as to make it too dangerous to try to supply
anything but crucible steel for the order; or they may be put a
little lower so as practically to preclude Bessemer steel. The
average physical difference between acid and basic open-hearth
steels is not great enough to make this method of assurance so
practicable, but it is possible in the case of basic and acid Bessemer
steel in England, where alone both these kinds of steel are made
in important quantities.

MISCELLANEOUS PURIFICATION PROCESSES

Bell-Krupp Process. — The late Sir I. Lowthian Bell devised a
process in which liquid pig iron is violently stirred up with iron
oxide, producing a slag which carries away more than 90 per cent, of
the silicon and phosphorus in the metal in the course of from 7 to 10
minutes. As soon as carbon begins to burn the process is stopped,
and therefore there is almost no change in this element. The

68 THE METALLURGY OF IRON AND STEEL

operation is conducted on the revolving hearth of a mechanical
puddling furnace, into which the melted iron is poured while the
hearth is rotating at about 11 revolutions per minute. The tem-
perature is lower than that of the open-hearth process, in order
that the elimination of phosphorus may be rapid.

The purified metal is used to some extent in the manufacture of
crucible steel. During recent years, when the low phosphorus
ores of America have become more scarce and the price of Besse-
mer pig iron is consequently increased, the metal has been bought
to a limited extent by foundry men using the acid open-hearth
process or the baby Bessemer process for the production of steel
castings.

Finery Fire. — This furnace is known under various names,
such as 'refinery hearth/ ' running-out fire/ 'finery fire/ etc. It
consists of a shallow, rectangular hearth, surrounded on the sides
by water-cooled, hollow blocks of iron about 3 to 3J ft. long by 2
ft. wide and 24 to 30 in. deep. In and above this hearth is built a
fire of coke upon which is placed 500 or 600 Ib. of pig iron. The
coke is burned by a blast at 2 to 3 Ib. pressure from 2 to 3 tuyeres
on each side, and the pig iron gradually melts and sinks below it.
When this takes place, more coke and pig iron is placed upon the
top, and the operation repeated. A bath of pig iron forms in the
hearth, and upon this the blast from the tuyeres impinges. This
oxidizes the silicon in the metal, and also a large amount of iron,
phosphorus, and sulphur. A slag high in iron oxide and therefore
very basic is formed. As the temperature is low, phosphorus is
eliminated without burning much carbon, and the result is the
production of a purified iron still high in the latter element.

It takes about two hours to perform this purification, and then
the metal is tapped out from the tap-hole in the front. It usually
runs into the long, shallow trough, whence the name 'running-out
fire'; but sometimes the refined metal is not allowed to cool, but is
run directly into the furnace in which the purification is to be com-
pleted. The consumption of coke is about one-eighth of the metal
produced, and the loss from 5 to 20 per cent., depending upon
the purity of the iron treated.

The running-out fire is frequently used in connection with the
charcoal finery known as the 'knobbling fire/ to produce knobbled
charcoal-iron, which is employed especially for boiler tubes and to
-a less extent for boiler-plate, wire, rivets, etc. Running-out fires

THE PURIFICATION OF PIG IRON

for this purpose are often known as 'melting fineries/ because in
them the pig iron is melted before it goes to the knobbling fire.
There are usually two tuyeres in the back, and the melted metal,
after the operation described just above, is run directly into two
charcoal fineries, which are very similar in construction to the

FIG. 35. — MELTING FINERY.

melting fineries, but have only one tuyere, situated in the back,
and take a charge of 250 Ib. apiece. During the transfer the slag
is separated from the metal as well as possible, but some gets into
the charcoal fineries; it is allowed to solidify and then is removed.
Knobbling Fire. — Upon the metal is now charged some damp
charcoal. The cold blast is turned on and the metal constantly
agitated and raised up from the bottom so as to bring it in contact
with the blast. Charcoal is added from time to time and is kept

THE METALLURGY OF IRON AND STEEL

damp to avoid loss, and the slag is removed at intervals, but there
must always be a layer of slag between the metal and charcoal.
As the metal comes to nature, it is pressed together with the pointed
bar, like a crowbar, which is used for the agitation and raising.
At the end of about an hour and a half, the ball is withdrawn and
hammered. The cinder from the knobbling fire is usually charged
into the melting finery.

The great advantages of the knobbled iron as compared to
puddled iron are its softness and relative freedom from slag. Tubes
made of this material may be flanged out very extensively without
showing any cracks, and rivets will flow easily when hammered
cold.

The Lancashire Process. — In the Lancashire process, which
some believe is a descendant in Sweden of the purification in the
two fires last described, the same operations are performed in one
hearth, which is made of iron plates, sometimes cooled with water,
with a tap-hole in the front. The Swedish Lancashire process is

Swedish Walloon
Charcoal Hearth

FIG. 36. From Howe, "The Metallurgy of Steel."

known as the ' Walloon process/ and is used in Sweden for making
bar iron from the very pure pig iron reduced from Dannemora
iron ore. This bar iron is used in Sheffield, England, for conver-
sion into blister-steel, and some of the steelmakers pay a large
price for it, in the belief that it has a certain intangible ' body ' not

THE PURIFICATION OF PIG IRON

contained in wrought iron from any other process, and which
makes a superior quality of steel. It is probable, however, that
this body is wholly imaginary. The process consists of three
stages :

1. The melting down, which is somewhat similar to the opera-
tion in the running-out fire or melting finery.

2. A purification period, during which the metal is nearly
purified and comes to nature.

3. A remelting above the tuyere for further purification.

In America, pig iron heated red-hot in the chamber H-C
(Fig. 37), during the working of a previous charge, is placed be-
tween two layers of charcoal and a little above the level of the

FIG.

37. — AMERICAN LANCASHIRE FURNACE. Tu
"The Metallurgy of Steel."

Tuyere. From Howe

tuyere. It is thus quickly melted, the liquid drops being forced
to trickle down through the blast and thus exposed to strongly
oxidizing conditions.

The second period begins at the end of about 15 minutes, when
the melted metal is all collected in the bottom. It becomes pasty
in contact with the cold hearth, and is raised by a pointed bar and
charcoal allowed to fall under it. Some slag at the same time is
tapped off and some is mixed with the pasty lump, which pro-
duces a reaction between the two that assists in the purification.
Toward the end of this period, which lasts 20 or 25 minutes, car-
bonic oxide comes off very rapidly, and when the metal becomes
so stiff that great pressure is needed to raise it, and the slag has
become thinner and whiter, the third period begins.

The metal is now broken into pieces and raised to its original
position, the action of the first period being substantially repeated.
During this period the workman is careful not to touch the mass

72 THE METALLURGY OF IRON AND STEEL

collecting in the bottom of the hearth lest he mix slag with it. By
the time about two-thirds of the metal is melted, some rich iron
oxide slag is added in order to keep enough in the bottom of the
hearth to protect the metal from being carburized by the charcoal.
When all the metal has melted and dropped down in front of the
tuyere, the pasty ball is pried out of the hearth and hammered.
The third period takes about 25 to 30 minutes.

Walloon Process. — In the Swedish Lancashire or Walloon proc-
ess, the long pigs of metal are fed slowly down into the fire, so that
it is not nece sary to constantly pry them up with bars. The
charges are smaller and the product is more liable to be hetero-
geneous, because the first melted metal is decarburized more than
the last, and this heterogeneousness is not removed by the re-
melting. (See Fig. 36.)

GENERAL REFERENCE BOOKS ON STEEL

30. Henry M. Howe. "The Metallurgy of Steel." Vol. i. 1890.

New York. This is the recognized standard authority on
the metallurgy of the Bessemer and crucible steel processes,
and upon the properties of steel as far as they were known
and understood at the time when this book was written.
It will long remain a classic. There have been many edi-
tions of different dates, but no change in text since 1890.

31. F. W.Harbord. "The Metallurgy of Steel." 1905. London.

With a Section on Mechanical Treatment by J. W. Hall.
Next to No. 30, this is the most complete and thorough
book on steel ever written, and, for English readers, will be
the first source of reference for those desiring recent prac-
tice. The section on mechanical treatment is the best
extant.

32. A. Ledebur. "Handbuch der Eisenhuettenkunde." Fourth

edition. 1902. Leipzig. This is an excellent reference
book for those who read German, and contains a very com-
plete account of the metallurgy of both iron and steel, and
of their properties. There are also classified lists of the
literature upon each of the branches of the subject.

33. Hermann Wedding. " Ausfuehrliches Handbuch der Eisen-

huettenkunde." Braunschweig. 1906. Four volumes.
This is an edited translation of Percy's "Iron and Steel/'

THE PURIFICATION OF PIG IRON 73

brought up to date and greatly enlarged with especial ref-
erence to German practice, which is the second largest in
the world.

34. H. Noble. "Fabrication de 1'Acier." Paris, 1905. As its

name indicates this book deals chiefly with the manufacture
of steel, the section on properties being very small.

35. Leon Gages. "Traite de Metallurgie du Fer." In two vol-

umes. Paris. 1898. The first volume covers the manu-
facture of iron and steel, and the second, foundry, mechan-
ical treatment and properties.

36. Sir I. Lowthian Bell. "Principles of the Manufacture of

Iron and Steel. " London. 1884. This book well accom-
plishes its aim, namely, to elucidate the principles of iron
and steel manufacture, and no man can be either so well
informed or so ignorant as not to understand the metal-
lurgy of these metals better after reading it.

37. John Percy. "Metallurgy. Iron and Steel." London. 1864.

This classical book is now chiefly valuable for historical
reasons, where its usefulness is often unexpectedly advan-
tageous, as, for example, in patent litigations, but at the
time it was written it was a model for wealth of informa-
tion (although badly arranged).

THE MANUFACTURE OF WROUGHT IRON AND
CRUCIBLE STEEL

THE MANUFACTURE or WROUGHT IRON

Pig Iron Used. — The pig iron employed is of the grade known
as 'forge iron7 or 'mill iron/ In the United States we prefer to
have this contain about 1 per cent, of silicon, because the higher
the silicon the larger will be the amount of slag made, while if it is
too low the iron will be oxidized excessively. Manganese is usu-
ally about 0.50 per cent., although it varies anywhere from 0.25
per cent, to over 1 per cent., depending on what the blast furnace
puts in the pig. Phosphorus is preferred to be less than 1 per cent.,
and sulphur not more than 0.10 per cent., because neither of
these elements are entirely eliminated during the process. A
large amount of phosphorus in wrought iron is not, however, as
objectionable as it is in steel, because the slag mechanically mingled
with the wrought iron hinders it from being brittle under shock,
which is the chief damage caused by phosphorus. Pig iron con-
taining 2.50 per cent, and even 3 per cent, of phosphorus, and as
much as 0.35 per cent, of sulphur, is sometimes used. The larger
the amount of impurities the larger the loss of metal in the
process.

Puddling Furnaces. — There are many different varieties of pud-
dling furnace, varying in capacity from 300 to 1500 Ib. and even
more, but the commonest is probably the 500-lb. furnace, built
either single or in pairs, back to back, the latter arrangement hav-
ing the advantage of reducing loss of heat by radiation, which is
always a very large factor. Puddling furnaces are heated by gas
or bituminous coal. The commonest method is a deep bituminous-
coal fire, giving a long flame, and with a large area of grate in rela-
tion to the area of the hearth, in order that a high temperature
may be maintained.

Fettling. — The hearth is lined or 'fettled7 with oxide of iron in

MANUFACTURE OF WROUGHT IRON AND CRUCIBLE STEEL 75

the form of roll scale, or high-grade iron ore, or 'bulldog/ i.e.,
roasted puddle cinder, and this oxide, together with the metal

FIG. 40. — 500-LB. PUDDLING FURNACE.

oxidized during the melting, supplies the base which automatically
maintains a very basic slag and also serves as the principal oxidiz-

FIG. 41. — 1500-LB. PUDDLING FURNACE.
(Two work doors.)

ing agent of the impurities. The fettling is repaired between melts
as often as is necessary, and suffers wear with each operation.

THE METALLURGY OF IRON AND STEEL

Squeezers. — A very common form of squeezer is that shown in
Fig. 42, the distance between the inner and outer circle being

FIG. 42. — ROTARY SQUEEZER.

greater on the entering side than on the outgoing side. As the
inner circle revolves, the corrugations on the surface carry the ball

FIG. 43. — ROE MECHANICAL PUDDLING FURNACE.

around, giving it at the same time a movement of rotation. By
the time the ball exits on the opposite side, it has been squeezed

MANUFACTURE OF WROUGHT IRON AND CRUCIBLE STEEL 77

and kneaded sufficiently to get rid of a large amount of slag. In
European countries the squeezer is rarely used and the ball is
' shingled ' - — reduced under a hammer — to weld its particles to-
gether.

Mechanical Furnaces. — The labor in puddling is very severe
on the men, and many attempts have been made to remedy this
by mechanical furnaces which will work the charge without so
much manual labor. Several forms of mechanical appliances and

FIG. 44. — CHARGING THE PUDDLING FURNACE.

of mecnanical furnaces have been invented, but without any per-
manent success. However, the mechanical furnace shown. in Fig.
43, devised by James P. Roe, of Pottstown, Pa., has given approxi-
mately satisfactory results. It is suspended on trunnions, and
the water-cooled bottom and sides are lined with magnesite brick.
The oil and blast for combustion enter through the two trunnions
and the products of combustion escape through a stack at each end,
which meet above the top of the furnace and discharge into the
atmosphere as shown. The furnace is made to oscillate 65° each
way from the vertical, which keeps the slag and bath uniformly

78 THE METALLURGY OF IRON AND STEEL

mixed and avoids the hand-rabbling of the ordinary puddling proc-
ess. The whole charge, weighing about 4000 lb., is discharged in
one ball by sliding it down the hearth of the furnace toward the
end, and then out into a hydraulic squeezer of special design, in
which it is compressed in three dimensions until it is a slab and
ready for rolling.

Puddling. — The pig iron is usually charged by hand through
the working doors of the furnace, and the puddler's assistant fires
vigorously in order to melt it down as fast as possible, which usu-
ally takes about 30 to 35 minutes. As soon as melted, there fol-
lows a short stage of 7 to 10 minutes, during which iron oxide in the
form of roll scale or very high-grade iron ore is added, in order to
make a very basic slag, the charge being thoroughly mixed and

FIG. 45. — PUDDLING.

cooled, for which purpose the damper is put on and sometimes
even water is thrown on to the bath. The object is to reduce
the temperature to the point where the slag will commence to
oxidize the impurities and especially the phosphorus and sulphur
ahead of the carbon. As soon as the reaction is started, light
flames begin to break through the covering of slag, produced by
burning carbon monoxide from the oxidation of carbon :

1. Fe2O3 + 3C= SCO + 2Fe;

2. CO + O = CO2.

If the slag is not very basic at this time the CO will reduce
"phosphorus and sulphur and cause them to return to the metal.

MANUFACTURE OF WROUGHT IRON AND CRUCIBLE STEEL 79

As the carbon monoxide forms more and more abundantly, the
charge is more violently agitated by its escape, and the 'boil' is in
progress. The formation of gas in its interior causes the charge to
swell greatly, and it thus rises in the furnace and a large amount of
slag pours out of the slag-hole and into a waiting buggy. About
one-half of all the slag produced during the process, and amounting
to about one-eighth to one-quarter of the weight of the metal

FIG. 46. — THE BOIL.

(A removable iron shield down which water flows protects the puddler from the heat.)

charged to the furnace, is removed at this time. The boil con-
tinues from 20 to 25 minutes, and during this time the puddler
stirs or 'rabbles' the charge vigorously with a long iron rabble,
shaped like a hoe. Toward the end of the boil the metal begins to
come to nature, and points of solid metal project through the cover
of slag, while other pasty masses form on the bottom of the fur-
nace. Both of these things must be corrected immediately by the

80 THE METALLURGY OF IRON AND STEEL

FIG. 47. — GETTING A HOLD OF THE BALL.

puddler, lest (1) the iron that is exposed to the furnace gases be-
come too much oxidized, (2) lest the iron sticking to the cold bot-
tom become too much chilled, or (3) lest the charge be not uniform

FIG. 48. — TAKING THE BALL OUT.

MANUFACTURE OF WROUGHT IRON AND CRUCIBLE STEEL 81

in composition. Finally, the whole charge comes to nature and
the 'balling' period begins and occupies about 15 to 20 minutes.
During this period the bath is divided into three or four portions,
which are each rolled up into a ball, consisting of a large number of
particles partially welded together. The balls are rolled up near the
fire-bridge in order, first, to protect them from direct contact with
the flame, and second, to keep them as hot as possible until the
puddler can draw them, so that the slag may be fluid and thus

FIG. 49. — THE BALL ENTERING THE SQUEEZER.

more easily squeezed out of the metal. The balls are then squeezed
in turn, and the furnace hearth repaired for another charge. The
total time between operations is usually from 1 hour and 10 min-
utes to 1 hour and 40 minutes.

Chemistry of the Process. - - The removal of the impurities
during the puddling process are shown in Table XI, which is
quoted because it records probably the first successful attempt
ever made to study in this way the chemistry of an iron or

THE METALLURGY OF IRON AND STEEL

TABLE XL — REMOVALS IN HAND PUDDLING
By Calvert and Johnson, Phil. Mag., 1857.

Time
after
charging

Sample No. 1

Hrs. Min.
0 0

Per cent.

2 275

Per cent.
2 720

Per cent.
0 301

Per cent.
0 645

.. .. .>

0 40

2 726

0 915

« "3

1 00

2 905

0 197

« « ^

1 5

2 444

0 194

« "5

1 20

2 305

0 182

« a Q

1 35

1 647

0 183

a "7

1 40

1 206

0 163

« a g

1 45

0 963

0 163

« « g

1 50

0 772

0 168

Puddled bar

0 296

0 120

0 134

0 139

steel process. A similar study is graphically represented in
Fig. 53.

During the melting-down stage, the silicon and manganese in
the puddling charge are almost entirely eliminated, and these re-
actions are as complete as they will be by the end of the ' clearing '
stage which follows it. Much phosphorus and sulphur are also

FIG. 50. — PUDDLE ROLLS.

removed. The boil period is, of course, the period during which
the carbon escapes together with all the phosphorus and sul-
^phur which was not removed during the first two periods.

MANUFACTURE OF WROUGHT IRON AND CRUCIBLE STEEL 83

Fuel. — The temperature of the puddling process is as high
as can be obtained in furnaces of this type without preheat-
ing the air.1 The result is a very large waste of heat up the

FIG. 51. — ROLLING PUDDLE BAR.

chimney, although some economy in this respect is obtained
by placing boilers, or else furnaces to heat metal for the rolls,
where they will receive the waste heat of the puddling furnaces.
The two greatest items of expense
in the puddling process are the fuel
used and the excessive labor, which,
on account of the strength and en-
durance demanded, receives a high
price. The amount of fuel burned per
ton of iron produced will usually be
about one ton of a soft bituminous
coal, or a little more, although better
figures than this are obtained in some
cases.

Losses. — The loss in the puddling
process usually averages from 4 to 6
per cent, of the weight of the metal
charged, although so much iron oxide
is added or is reduced from the lining by the impurities in the
metal, that in the Roe furnace the wrought iron produced will

1 Indeed, in some cases the air is preheated by the regenerative process,
although this is not the usual practice.

FIG. 52. —WROUGHT IRON
SHOWING STRINGS OF
SLAG MAGNIFIED 50
DIAMETERS.

(Unetched.)

THE METALLURGY OF IRON AND STEEL

actually weigh more than the pig iron charged. The following
table gives a typical example of loss :

TABLE OF LOSSES IN HAND PUDDLING

Percentage of
Loss

Silicon burned 1 . 00

Carbon burned 3 . 50

Sulphur burned .20

Phosphorus burned 50

Manganese burned V. .30

Total 5.50

Iron reduced from oxide1 =1 .00 per cent, gain
Net loss 4.50

1 There is much iron oxidized and carried off in the slag, but there is also
much reduced by impurities. The figure here given represents the excess of
reduction over oxidation; in some cases it runs as high as 6 per cent, or more.

100

10 20 30 40 50 60 70 80 90 100

Time (Minutes)

FIG. 53. — REMOVALS IN HAND PUDDLING.

MANUFACTURE OF WROUGHT IRON AND CRUCIBLE STEEL 85

Slag. — The slag which runs from the tap-hole during the boil
is known as 'boilings/ while that tapped out at the end of the
process is known as 'tappings' or 'tap-cinder/ The characteris-
tics of the boilings are that they contain a larger amount of the phos-
phorus than the tappings, and also that globules of metallic iron
are carried off in the violent agitation of the boil. An analysis of
the two varieties, giving a mean composition from seven heats, is as
follows : 1

Boilings

Tappings

Ferric oxide (Fe-^Os). .

6 94

12 90

Ferrous oxide (FeO) .

62 61

64 62

Silica (SiOa)

19 45

15 47

Phosphoric anhydride (P2O5). .

6 32

3 91

Not determined (MnO, S, CaO, etc.)..

4.68

3.10

100.00

Total iron. .

53 55

59 29

The amount of slag will depend chiefly upon the amount of
silicon in the pig iron. It will average in weight about one-half
the weight of the charge, where the silicon is high, as in English
practice (say 1.70 to 2 per cent.), and about one-quarter to one-
third in American practice, where the silicon is about 1 per cent.

THE CARBURIZATION OF WROUGHT IRON

Wrought iron is converted into steel by the operation of car-
burizing, or the adding of carbon to it. This is to-day accom-
plished in two ways: (1) By the cementation, or steel conversion,
process, in which carbon is allowed to soak into red-hot steel in a
manner like in nature to the absorption of ink by blotting paper;
and (2) by the crucible process, in which wrought iron is melted
in a crucible with carbon, or with iron containing carbon, e.g.,
cast iron.

Cementation Process. — We have already observed in describ-
ing the blast-furnace process that iron at a bright-red heat will
absorb carbon very slowly. The action appears to be a traveling of
solid carbon into the interior of solid iron, forming with it a chemi-

1 Page 297 of Number 40, page 93.

THE METALLURGY OF IRON AND STEEL

cal compound or carbide, FeaC, to which Professor Howe has given
the name of 'cementite.' When the steel is at a proper heat, the
rate of travel is approximately f of an inch per 24 hours.

Steel Converting Furnace. — A section of the type of cementa-
tion-furnace used in Sheffield, England, is shown in Fig. 55. The
superstructure, e, is a mere chimney for the purpose of carrying

off the products of combustion
from the fire at c, and for re-
ducing loss of heat by radiation.
The real furnace is the part un-
derneath this superstructure or
stack, and it has several small
chimneys of its own. The two
converting pots are shown un-
derneath the points a a. In
Sheffield they are built of stone
and are 2J to 4 ft. wide and
deep, and 8 to 15 ft. long. On
the bottom of the pot is first
placed a layer of charcoal in
small pieces freed from dust.
On this is laid a layer of the
wrought-iron bars to be con-
verted, which are 2 to 5 in.
wide, i to J in. thick, and nearly
as long as the pot. A little
space is left (about J in.) be-
tween each pair of bars in order
that they may be completely
surrounded by the charcoal.
On top of the layer of bars is

then placed another layer of charcoal, and then a layer of bars,
and so on until the pots are filled, each one containing from 10
to 30 tons of iron. The top of the pots is then luted air-tight
and the fire lighted.

In about 2 days the temperature has reached a full red-heat of
say 650° to 700° C. (1200° to 1300° F.), and this is maintained
from 7 to 11 days longer, depending upon the grade of steel to be
made. The names of the different grades of steel made in Shef-
field are as follows:

•"•''/A*. - '. ...... ~ /<:,

FIG. 55. — CEMENTATION FURNACE.

MANUFACTURE OF WROUGHT IRON AND CRUCIBLE STEEL 87

No. 1. Spring heat 0.50 per cent, carbon

' 2. Country heat 0.60

' 3. Single-shear heat 0.75

' 4. Double-shear heat 1 . 00

' 5. Steel-through heat 1 . 25

1 6. Melting heat 1 .50

The product is controlled by a series of trial bars which are so
placed that, beginning about 7 days from the full-red heat, they
can be withdrawn from the furnace from time to time, broken,
and examined. The appearance of the fracture denotes the ex-
tent of the cementation.

Discharging the Furnace. — When the cementation has pro-
ceded to the desired point, the fire is withdrawn and the furnace is
allowed to cool for about a week, when the pots are opened and
the bars withdrawn through the door b.

Blister-Steel. — The product of the cementation process is
known as l blister-steel' because its surface is covered with blisters,
due to the formation of gas by a reaction between the carbon of the
cementite and the slag contained in the wrought iron:

C + FeO = CO + Fe.

The blister-steel has gained about 1 per cent, in weight over
the wrought iron and the appearance of the fracture is entirely
different, as the broken surface now shows large bright crystals.

Shear-Steel. — The bars of blister-steel are sometimes forged to
a smaller size, piled up, and then the pile forged down again into a
bar, which makes what is known as* 'single shear-steel/ Single
shear-steel may be again piled and forged into double shear-steel. In
America, however, it is more common to melt the blister-steel in
crucibles, which separates the metal from the slag it contains and
produces the finest quality of cutlery and tool-steel that is made
in America.

Crucible or Cast Steel. — The cementation process, on account
of the length of time and the very large amount of fuel required,
has now been largely superseded by the crucible process. In this
process the wrought iron is cut up into small pieces and melted in
covered crucibles, the desired amount of carbon being placed on top
of the charge before the melting, together with any other alloying
element desired, such as chromium, tungsten, manganese, etc.

Furnaces. — In Sheffield, England, coke-furnaces, or melting-
holes, containing each two crucibles, are almost universally used,

THE METALLURGY OF IRON AND STEEL

while in America gas-furnaces, containing about 6 crucibles each,
are the common type. In the gas-furnace it is necessary that the
gas and air for combustion shall be preheated, in order that we may
(1) obtain fuel economy and (2) reach the desired temperature for
melting quickly.

Regenerative Furnace. — The operation of the regenerative-
furnace is shown in Fig. 57. The gas enters the gas regenerative-
chamber and passes up between the checker-work of bricks, laid

FIG. 56. — SECTION THROUGH SHEFFIELD
MELTING HOUSE.

with many spaces between, and into the melting-hole. The air
enters and passes up through its regenerative-chamber, meeting
the gas above. They there combine and, passing through the
melting-hole, divide into two parts and pass through the regenera-
tive-chambers on the other side. Previous to beginning the opera-
tion, all of the brickwork in the regenerative-chambers has been
heated red-hot by means of a wood fire. The gas and air have
therefore absorbed a good deal of heat from the brickwork
before they meet. As they pass down through the regenerators
on the outgoing side, they will still further increase the heat of
this brickwork, giving up their temperature to the checker-

MANUFACTURE OF WROUGHT IRON AND CRUCIBLE STEEL 89

work. This causes them to go to the stack at a relatively low tem-
perature, and when the current of air and gas is reversed, each
now entering from the other side, they become more highly pre-
heated than before, and now serve to heat the opposite pair of

FIG. 57. — REGENERATIVE GAS CRUCIBLE FURNACE OR "MELTING-HOLE."

regenerators. When all four regenerators have been raised to a
high temperature, the reversal of direction takes place every 20
minutes, and thus a uniformly high heat is obtained with low
temperature of chimney gases and consequent fuel economy.

90 THE METALLURGY OF IRON AND STEEL

Crucibles. — In England, the crucibles are made of fire-clay.
They are usually made by hand at the steel melting plants and
are dried for one or two months, on a shelf next to the chimney of
the melting-furnace. Before being used they are heated to red-
ness in an annealing furnace and are then ready to receive a charge
of 50 Ib. of metal. The clay is deeply cut by the slag, and there-
fore the charge must be reduced to 44 Ib. for the second melt and
38 Ib. for the third, in order that the slag-line may be lower each
time. After the third melt they are thrown away. The advan-
tages of clay crucibles are that the first cost is lower, and that they
do not give up any carbon to the metal, so that the composition
of the final product may be regulated with greater exactness and
a lower carbon steel may be made, if desired.

In America, the crucibles are made of a mixture approximating
50 per cent, graphite and 50 per cent, fire-clay. They are made
and tempered by factories outside of the steel-works and are re-
ceived by the latter ready for use. They last about six heats,
after which the bottom is sawed off and used for the top of a new
crucible. They hold almost 100 Ib. of metal, because they are
stronger than clay and can therefore stand greater strains.

Metal Used. — Although crucible steel is supposed to be made
by the melting of pure wrought iron with charcoal, washed metal,
ferromanganese and other 'physic/ it is not at all uncommon for
the wrought iron to be diluted with varying amounts of cheaper
scrap steel, which unquestionably lowers the quality of the product.
The pieces of wrought iron are put into the crucible first, and on top
of that is placed the charcoal or pig iron, ferromanganese or spiegel-
eisen, and various physics, such as salt, potassium ferrocyanide,
oxide of manganese, etc. The purpose of these physics is not
entirely clear. Probably the salt and oxide of manganese make a
more fluid slag; the ferromanganese puts a little manganese in the
steel; and the ferrocyanide may, perhaps, favor the absorption of
carbon by the steel.

In some cases the material is charged directly into a hot cruci-
ble from the previous melt, but when graphite crucibles are used,
these are sometimes allowed to cool, in order to be examined for
cracks, because the breaking of a crucible in the furnace, allowing
the liquid mass to flow out upon the floor, is very objectionable.
There is a hole in the middle of the floor of the furnace, so that if
such an accident happens the metal may run down into a pit

MANUFACTURE OF WROUGHT IRON AND CRUCIBLE STEEL 91

underneath ; and the floor is also covered with a few inches of coke-
dust to absorb the metal in such a contingency.

Melting. — The crucibles are placed in the furnace and the gas
and air turned on, in the case of a gas-furnace. In the case of a
coke-furnace, the crucibles are placed upon their fire-clay stands
and the coke packed around them upon the fire left from the last
melt. At the end of about an hour, the coke fire has burned down

FIG. 58. — STAGES IN CRUCIBLE STEEL MELTING.

so low that it has to be poked down and more coke added. The
melting takes on the average from two to three hours in both
cases, and the charge is examined by removing the crucible cover,
to make sure that it is entirely molten. The coke fire also requires
further attention at this time, and care must be taken that no coke
falls into the crucible when the cover is off.

Killing. — When the charge is entirely molten, it is kept in
the furnace for one-half to an hour longer in order that it may teem
'dead/ that is, pour quietly without the evolution of gas, and
yield solid ingots. If the 'killing' time is too long, the ingots will
be solid, but the steel will be hard, brittle and weak, probably as
a result of the absorption of too much silicon from the walls of the
crucibles. Graphite crucibles probably yield more silicon than
clay crucibles. Just what takes place during the killing of steel
is not definitely known. Some have suggested that the gas con-
tained in the steel is eliminated from it during this time, but the
alternate suggestion that the principal effect of the killing is to
cause the steel to absorb silicon, becoming sound on this account,
is the more generally accepted one. The amount of silicon in the
final steel will vary greatly, but will average perhaps from 0.10 to
0.50 per cent.

Pulling. — When, in the judgment of the melter, the steel has
been properly killed, the crucibles are removed from the fire by
the puller-out, who straddles the top of the furnace and grasps the

THE METALLURGY OF IRON AND STEEL

FIG. 59. —POURING.

crucible with a pair of tongs, his legs and arms being swathed in
wet cloths to protect him from the heat, and his eyes frequently
being protected by heavy blue glasses. The puller-out then passes
the crucible to the pourer, who pours it as shown in Fig. 59, the slag

first being swabbed off with a
ball of cold slag on the end of
an iron rod. The total time of
operations is 3J to 5^ hours in
England, and
3J to 4 hours
in America.

Ingot Molds.
— Ingot molds
are shown in
Fig. 60, and as
a usual thing
each mold has
a capacity to
take the charge

from one crucible. The metal must be teemed
into this with great care, so that the stream shall
not touch the sides during pouring. In case a
large ingot is to be poured, then several cruci-
bles are poured at once into the same mold, care
being taken that the metal shall be liquid as long
as the pouring continues.

Grading. — The composition of the final metal
is a matter of some uncertainty, especially as re-
gards the carbon and silicon. The former is more
easily adjusted when clay crucibles are used, be-
cause .the amount of carbon dissolved from a
graphite crucible will depend to a large extent
upon the time and temperature of the operation, FIG. eoT^ INGOT
etc. The ingots are therefore always graded after ^RUCI BLE
they have cooled, by breaking off the upper part STEEL. From
of them, which contains the pipe and is therefore MetYiiurg^of
useless, except when remelted, and examining the steel."
fracture with the eye. The skilled steel man can thus estimate the
carbon within 0.10 per cent., and the ingots are put away in the pile
with others of like analysis. At large American works, how-

MANUFACTURE OF WROUGHT IRON AND CRUCIBLE STEEL 93

ever, this grading by eye is always supplemented by chemical
analysis.

Chemistry. — The chemistry of the crucible process is very sim- '
pie, and consists principally in the elimination of the slag in the
wrought iron and the absorption by the metal of carbon, silicon
and manganese. There is also a very slight increase in sulphur,
which perhaps comes from the pyrite in the clay or graphite, or
from sulphurous gases which find their way under the cover of the
crucible. Phosphorus also increases slightly, perhaps from the
slag out of the wrought iron.

Loss. — The loss is due to the elimination of the slag and to
some slight oxidation of metal by oxygen in the gas inside the
crucible. It is counteracted to some extent by the absorption of
carbon, silicon and manganese, and will average slightly more
than 2 per cent, in clay crucibles and somewhat less than 2 per cent,
in graphite crucibles, doubtless due to there being less oxidation
in the presence of the graphite.

Fuel. — In coke fires, the amount of fuel used will be three to
four times the weight of steel produced. In gas-fired furnaces the
amount of fuel used to make producer gas will be equal to or
slightly less in weight than the amount of steel produced. The
high cost in making crucible steel is on account of the cost of cru-
cibles, fuel, labor and raw material.

REFERENCES ON THE MANUFACTURE OF IRON

There is but one American book (No. 47) devoted either to the
manufacture of pig iron or wrought iron. One should refer to
Nos. 2, 32, 33, 35, 36, 37, and the list given below:

40. Thomas Turner. " The Metallurgy of Iron." * London, 1895.

41. A. de Vathaire. "Les Hautes Fourneaux." Paris.

42. H. Bauerman. "A Treatise on the Metallurgy of Iron."*

London, 1868.

43. M. A. Pavlov. "Atlas of Plans for Blast Furnace Construc-

tion." Gekatermoslov (Russia), 1902. Although these
drawings are. lettered in Russian, one gets much valuable
information from them even without being able to read the
language.

* Starred books refer both to pig iron and wrought iron.

94 THE METALLURGY OF IRON AND STEEL

44. Frederick Overman. "The Manufacture of Iron. " * Philadel-

phia, 1850.

45. W. Truran. "The Iron Manufacture of Great Britain/'*

London, 1865. Revised by J. Arthur Phillips and W.
H. Dorman.

46. James P. Roe. "The Development of the Roe Puddling

Process/' Journal, Iron and Steel Institute, No. Ill, 1906,
pages 265-306.

47. Robert Forsythe. "The Blast Furnace and the Manufac-

ture of Pig Iron."

* Starred books refer both to pig iron and wrought iron.

THE BESSEMER PROCESS

Pig Iron Used. — In the large American works, pig iron for
the Bessemer process is preferred to have about Lper cent, of
silicon. This is the chief slag producer and also the chief heat
producer. To keep it at a low figure limits the amount of slag
made, which limits one of the sources of iron loss. Furthermore,
the lower the silicon the shorter will be the time of blow; but it is
usually risky to allow it to fall below 1 per cent., or the blow will
be cold, and it is only by very rapid working and permitting the
least possible delay between operations, so that the converter and
ladles are kept very hot, that we are able to get along with as little
as this. The manganese is below 0.8 per cent. This also fur-
nishes heat;1 but it is now an expensive ingredient of pig iron, and
also has the effect of making very liquid slags, which cause a good
deal of slopping or 'spitting' from the converter (i.e., ejection of
the material by the blast), and also make the steel ingots dirty
and spotted with oxide spots, due to slag carried over with the
steel. Manganese of 1.50 per cent., with silicon of 1.00 to 0.90 per
cent., gives a very ' wet' slag, which follows the metal into the ladle
and boils up through it, oxidizing the manganese in the steel :

(1) FeO+Mn =MnO+Fe.

The phosphorus and sulphur must be below 0.10 and 0.08 per
cent., respectively, in order that the steel may be salable, as
neither of these elements is reduced in the acid Bessemer process.

Mixer. — It takes about two blast furnaces to supply one con-
verter with metal, so that a modern plant of two to four converters
will be operated in conjunction with a large blast-furnace plant.
The product of each of these furnaces, if not too different from the

1 Indeed, formerly, in the Swedish Bessemer practice, the pig iron con-
tained 2 per cent, of manganese, and this element was relied upon as the chief
source of heat, because silicon was necessarily low in the Swedish charcoal
pig iron.

FIG. 65. — A BESSEMER BLOW.

THE BESSEMER PROCESS

desired analysis, will be poured into a huge reservoir, or 'mixer/
capable of holding 150 to 500 tons, which is then used as a source
of supply for the converters.

The mixer serves several very useful purposes: (1) It equalizes
the irregularities of pig iron composition by mixing the product of
several furnaces, and also brings the composition somewhat under
the control of the metallurgist of the Bessemer plant, because he
not only can pick and choose from the different furnac'es, but he

FIG. 66. — MIXER.

has a few large cupolas under his dominion in which he can melt
iron of any desired analysis to pour into the mixer and help regu-
late its contents.

(2) Because of its large size, and the fact that it is continually in
receipt of new fresh metal, the mixer can keep its contents molten
for an indefinite length of time, whereas a ladle containing 15 tons
of pig iron would chill up in a very few hours. Mixers are supplied
with blowpipes which can contribute a small amount of heat to the
charge, but it is not often necessary to use them.

(3) The capacity of the mixer is so large that a delay either at
the blast furnace or at the steel-works will not discommode it
greatly, and thus each operation is independent of the other.

THE METALLURGY OF IRON AND STEEL

(4) Pig iron in the mixer suffers a slight loss in sulphur, be-
cause manganese sulphide forms and, not being very soluble in
iron, slowly passes out of it into the slag.

Construction of the Converter. — The construction and dimen-
sions of the converter are shown in Figs. 67 and 68. It consists of

FIG. 67. — FIFTEEN-TON CONVERTER SHELL. From Howe, "The Metallurgy of

Steel."

a steel shell, riveted together and supported by two trunnions upon
which it can be made to rotate. One of these trunnions is hollow,
and serves as a wind-pipe to connect the blast from the blow-
engine with the wind-box at the bottom of the vessel. On the
other trunnion is fastened a pinion, which engages with a rack
joined to a hydraulic piston and of such a length that its move-
ment can rotate the converter through an angle of at least 270°.
The lining of the bottom is pierced with about 250 half-inch holes,
which connect the wind-box with the inside of the converter and

THE BESSEMER PROCESS

serve for the passage of the blast. The shape of the converter is
such that, when it is lying on its side, the metal will not cover any
of these tuyere-holes. This is necessary, or the blast could never
be turned off without having molten metal run down into the wind-
box. The converter may have either an eccentric or a concentric
shape. The advantage of the eccentric shape is that less heat can
escape from the nose : the advantage of the concentric shape is that
the vessel may contain its charge when turned on either of the
sides.

Lining. — The lining is made of highly refractory acid ma-
terial composed principally of silica. In England, a ganister rock

FIG. 68. — FIFTEEN-TON CONVERTER SECTION. From Howe, "The Metallurgy

of Steel."

is used, or sometimes the lining is rammed up around a pattern
and is composed of silicious material held together by a small
amount of fire-clay. In America, it consists usually of blocks of
ganister or of mica-schist (a silicious rock consisting of pseudo-

100

THE METALLURGY OF IRON AND STEEL

strata, or laminae, formed by tiny plates of mica) laid with a thin
layer of refractory fire-clay between, and in such a manner that the
edges of the laminae will be exposed to the wear to which it is
subjected.1 After a new lining is put in, it is carefully dried, and
every Sunday afternoon, before the converter begins its operation
for the week, a wood fire is kept in it for several hours in order to

FIG. 69." From Howe, "The Metal-
lurgy of Steel."

FIG. 70. — ECCENTRIC CONVERTER.
From Howe, "The Metallurgy of Steel."

heat the lining to a red heat. Between the heats the lining is re-
paired, if necessary, with balls of silicious material and clay. On
Sundays, and with an occasional lay-off for which one extra shell
is provided, more extensive repairs are made, and in this way the
lining is made to last several months, — say 10,000 to 20,000 heats.
The converter slags are always high in silica and corrode the lin-
ing only slightly. If, however, any uncombined oxide of iron
comes in contact with it, it is attacked very rapidly. For this rea-
son the mouths of the tuyeres are rapidly eaten away, and this
part of the converter lasts only about 20 to 25 blows. The bottom

» If we represent the blocks of mica-schist by big books, and lay these
books in a horizontal position with the edges of the leaves exposed, it will
illustrate the method employed.

THE BESSEMER PROCESS

ioi

is therefore fastened to the body with links and keys, so that it
may be readily detached and replaced by a new one. Indeed, in
some works bottoms are changed with an average delay to the
operation of only about 20 minutes for each replacement.

Bottoms. — The lining of the bottom is made by placing the
tuyere-bricks (see Fig. 68) in position and then filling in around
them with refractory material consisting of damp silicious ma-
terial held together with clay and containing usually some coke
breeze, which seems to lessen the chemical activity of the corrosion.
The details of lining vary so greatly that no general rules can be
given. The number of tuyeres is from 18 to 30, the number of
holes in each from 12 to 18, and the size of the holes f in. (Eng-
land) or | to f in. (America).1 The correct lining is of the greatest

FIG. 71. — CORROSION OF THE BOTTOM LINING OF BESSEMER CONVERTER.

importance and is the most influential factor in determining the
life of the bottom, which furthermore depends upon the care in
drying, the temperature of blowing, the pressure of blast, and the
composition of pig iron. A bottom should dry 36 hours or more.
Its life is shortened by (1) hotter blows, (2) longer blows, (3) lower
blast pressure (because the blast holds the metal away from the

1 The acid Bessemer process finds its greatest importance in America, and
next to that in England. Germany is the leader in basic Bessemer practice.

THE METALLURGY OF IRON AND STEEL

mouths of the tuyeres), and (4) more manganese in the pig iron (be-
cause a wet slag is more corrosive) . Between heats, when the vessel
is on its side receiving the recarburizer, pouring into a ladle, or re-
ceiving a new charge, the lining of the bottom can be repaired.
For instance, if one tuyere eats away faster than its fellows, the ex-
cessive corrosion can be prevented by stopping it up with mud,
because, if no air passes through the holes, no oxide of iron is
formed at their mouths. Or a worn tuyere may be replaced by a
new one, etc., etc. These repairs are chiefly made through the
wind-box, the back plate of which is removable.

When a bottom is worn out, it is taken away and a new one
brought on a car and placed under the converter, which is in the

vertical position. Around the
top is piled a ring of thick wet
mud, and, as the bottom is forced
up against the body by hydraulic
pressure, the mud is squeezed

shoulder ^^ / No^e\ into a firm joint. Lack of space

prevents an account of some of
the interesting expedients that
are resorted to to stop an oc-
FIG. 72. — CONVERTER PARTS. casional leak in this joint with-
out delaying the first heat, which

must be avoided if possible, as the first heat on a new bottom is
already too liable to be a cold one.

Operation of the Converter. — Figs. 73 and 74 are sections
through the converter plant at different points. In Fig. 73, A is
the mixer, capable of containing, say, 300 tons of metal; B is the
ladle carriage from the blast furnace, from which the ladle C has
been raised to pour the metal into the mixer. Immediately above
D is the ladle that is to take the metal from the mixer to the con-
verter, for which it is transferred along a level track unti^ it comes
to E (Fig. 74), where it is poured into the vessel, now in the hori-
zontal position, as shown by the dotted line. When the metal
has been poured in, the wind is turned on and the vessel elevated
into the vertical position. The blast now pours through the 18
inches or so of metal in the bottom of the converter in a wide
spray of tiny bubbles until the impurities are oxidized, when it
is turned again into the horizontal position and trie wind cut off.
In anticipation of this a predetermined quantity of spiegeleisen

THE BESSEMER PROCESS

103

has been tapped from the spiegel cupola into the ladle at H (Fig.
73). (For soft steel ferromanganese, not spiegeleisen, is used.)
Spiegeleisen is pig iron very high in manganese. Some analyses

are given in Table XII. It is melted in the spiegel cupola together
with a predetermined amount of high silicon pig iron, and is then
used to recarburize the bath in the vessel, for which purpose the

Tom .

FIG. 74.

ladle is now run into the position E (Fig. 74) and its contents
poured into the bath. The reactions that take place between the

104 THE METALLURGY OF IRON AND STEEL

TABLE XII. — ANALYSIS OF VARIOUS GRADES OF PIG IRON

NAME

Silicon
Per cent.

Sulphur
Per cent.

Phosphorus
Per cent.

Manganese
Per cent.

Carbon
Per cent.

fNo. 1

2.75
2.25
1.75
1.25
0 . 75 to 1 . 75
0.80 to 2. 00
under 1.00

under 1 . 66
0.50 to 1.00
under 1 . 00
under 1 . 00
8.00 to 15.00
50.00
8.00 to 15.00

0.035
0.045
0.055
0.065
0.05 to 0.30
0.03 to 0.08
under 0.10
under 0.050

under 0.03
under 0 . 03
under 0 . 05
under 0.07
under 0.02
under 0. 01

0.30 to 1.50

0.30 to 3. 00
under 0.10
2.00 to 3. 00
under 0 . 05
0.10 to 2.00
0.10 to 1.00
0.10 to 0.50
under 0.15
0.10 to 0.50
under 0 . 08
under 0.15

0.20 to 1.60

0.20 to 1.50
0.30 to 0.50
1 . 00 to 2 . 00
0 . 30 to 0 . 50
1.00 to 2. 00
80.00
40.00
15.00 to 30. 00

3.00 to 4.00

3.50 to 4. 00

5.00 to 7. 00
5.00 to 6. 00
5.00 to 6. 00
1.00 to 2. 00
under 0.40
1.00 to 1.50

Foundry j No. 2
Irons ] No. 3

1 No. 4
Forge

Bessemer — Acid
Bessemer — Basic ....
Open-hearth — Acid. .
Open-hearth — Basic. .
Ferromanganese
Ferromanganese
Spiegeleisen

Ferrosilicon

Silico-Spiegel

15.00 to 20. 00

elements in the recarburizer and the impurities in the bath are as
follows :

(a) Mn + FeO = MnO + Fe;

(6) C + FeO = CO + Fe.

Reaction (b) produces a boil of the bath, which serves to stir it well
and distribute the elements uniformly. Reaction (a) removes a

Converter.

Cupola.
Ladle.
Runner.

FIG. 75. — POURING METAL INTO THE CONVERTER.

large amount of oxygen from the metal and takes some of the
manganese in the recarburizer into the vessel slag. There is also

THE BESSEMER PROCESS 105

a slight loss of silicon from the recarburizer by the following re-
action :

All of these losses are discounted in calculating the composi-
tion and amount of metal tapped from the spiegel cupola, so that
there should be left in the steel the desired percentage of carbon,
silicon, and manganese.

After the 'spiegel reaction' is completed, the steel is poured
from the converter into a ladle held at the point 0 by the jib-crane
/, or by one of the traveling cranes K or L. The ladle is then car-
ried over to a position above the ingot molds into which the steel is
to be teemed. In pouring the steel into the ladle, the slag is held
back in the vessel as much as possible, because this not only fur-
nishes heat for the next operation, but also makes it shorter by as
much as 20 or 25 per cent., because the slag, being an oxidized
substance, assists in oxidizing the impurities in the next charge.

The turning on and off of the blast and the rotation of the con-
verter is all executed by the blower, who stands upon the pulpit at
W and operates the various valves, and also judges by eye the
progress of the purification from the appearance of the flame which
issues from the mouth of the converter.

Every effort is made by him to so arrange the different opera-
tions of the converter, cranes, and ladles and to bring in both
spiegeleisen and iron, that each step shall fit into the others without
delay to the operation in any of the converters. He also has under
his control means for lowering the temperature of the bath, if
necessary : (a) By ordering an amount of cold steel scrap to be
thrown into the mouth of the vessel during the blow1 and (6) by
admitting live steam into the converter with the blast. The de-
composition of the steam very quickly reduces the heat of the
blow. Attempts have been made to dispense with the blower,
who, on account of the long training and experience necessary, is
the most highly paid man in the plant, next to the foreman, by
judging of the progress of the operation with the spectroscope;
but when two or three vessels are blowing at the same time the
reflected light from one interferes with the spectroscopic indica-
tions of the others. Furthermore, the spectroscope gives no indi-
cations of the temperature of the blow, and until the past two or

1 Scrap is so valuable for use in the open-hearth process that (6) is now
used much more than (a).

106

THE METALLURGY OF IRON AND STEEL

three years no reliable pyrometer existed which was suitable for
this use.

Steel Ladles. — The ladles to receive the steel and teem it into
molds are steel shells lined with a cheaper grade of acid refractory
material, because the life of these ladles is limited to about six

FIG. 76. — FIFTEEN-TON STEEL TEEMING LADLE.

heats by the wearing out of the nozzle, and therefore a more ex-
pensive lining would be wasted. The arrangements of nozzles,
stoppers, and handle shown in Fig. 76 are provided in order that a
thin stream of steel may be poured into the molds and may be in-
terrupted when a new mold is being brought into place. The slag
lies on top of the metal, and when this begins to come out of the
nozzle the stopper is let down and the ladle carried over a slag car
and turned upside down to dump out the slag. At this time the
blower observes the lining of the ladle in order to tell whether the

THE BESSEMER PROCESS

107

steel was too hot or too cold. If there is a skull of metal frozen
inside the ladle, the steel was too cold; if there is no frozen metal,
it was too hot ; but if there is a spot of metal here and there on the
bottom, it was just right.

Ingot Molds, Stools and Cars. — The arrangement of the molds
into which the ingots are to be cast is shown in Fig. 78, which also
gives the dimensions of molds commonly used for railroad rails,
wire, and pipe. Molds last about 100 heats, after which they are
so cracked inside that they are with difficulty lifted off the solidi-
fied ingot of steel, and are also very liable to tear it, producing
cracks which are not easily welded up later because they become

FIG. 77. — TEEMING INTO MOLDS.

oxidized on the interior. A continuous series of mold cars are fed
into the steel-mill at one end and drawn out, with the ingot inside,
at the other end. They should be heated so hot that the palm of
the hand will not bear the heat on the outside, and are washed
inside with a thin clay wash which prevents the liquid metal stick-
ing to the cast iron. At the pouring platform they are moved

108

THE METALLURGY OF IRON AND STEEL

forward under the ladle by means of a little finger, which enters a
notch in the side of the car and is itself carried on the rod of a
hydraulic piston.

Stripping. — As soon as they can be run out to the stripping
house/ the ingots are solidified sufficiently on the outside for the
mold to be removed, leaving the ingot standing on the car ready to
be drawn to the rolling-mill. In the majority of cases it is only
necessary to place the jaws of the stripping machine under the
lugs on the mold, raise it up to a sufficient height, and then trans-

FIG. 78. — INGOT MOLDS, STOOL AND CAR.

5000-lb. molds: 7 ft. high, 15f in. square at top, 19i in. at bottom; about 2£ in.
thickness of cast iron.

fer it to a stool on a car in the second track of the stripping house.
Sometimes the ingots do not come out so easily, however, and then
the plunger is rested on top of the ingot, which is held down by
the hydraulic pressure as the mold is drawn upward. Most strip-
ping machines are actuated by hydraulic power, although electric

THE BESSEMER PROCESS

109

ones are common. The empty molds are stored in the yard
until they are sufficiently cool to be drawn back to the steel-mill
for another charge, and during the wait they are washed inside
with clay water, which is quickly dried by the heat of the mold.

Chemistry of the Acid Bessemer Process. — If a stream of air be
made to impinge upon a melted bath of iron, the metal and im-
purities will be immediately oxidized, about in proportion to the

FIG. 79. — STRIPPING.

relative amounts of each present. We may therefore consider
the chemistry of the Bessemer process as a union of the oxygen
from the tuyeres with the first element it meets, irrespective of
relative affinity, and a subsequent attack upon these oxidized
elements by unoxidized ones for their oxygen. As iron is the
predominant element, and as the oxides of iron are readily reduced,
they especially serve as carriers of oxygen between the air and
other impurities:

1. FeO + Mn = Fe + MnO.

2. FeO + C = Fe + CO.

3. 2 FeO + Si = 2 Fe + SiO2.

In the first part of the blow, the oxides of carbon also suffer
reduction :

4. 2CO + Si = SiO2 + 2 C.

5. CO2 + Si = SiO2 + C.

6. CO + Mn = MnO + C.

110 THE METALLURGY OF IRON AND STEEL

Equilibrium is therefore established by the elements which
have the greatest chemical affinity for oxygen getting practically
all of that agent, either directly or by robbing their neighbors.
Some iron oxide survives, however, because it is so predominant
in amount that its neighbors cannot rob it before a part of it has
united with silica, which is either formed by the oxidation of silicon
or else won from the lining :

7. FeO + SiO2 = FeSiO3.

The ferrous silicate forms a slag with manganese silicate, and
this slag will dissolve oxides of iron. Even after the oxide of iron
has been absorbed by the slag, it may be reduced by manganese
and carbon, although not to the same extent. Practically no
oxygen of the air escapes from the bath uncombined, even though
it has but 18 in. of metal to pass through and amounts to more
than 5000 cu. ft. per minute,1 because the blast pressure is made
high in order that the air may be broken up into a fine spray of
bubbles as soon as it strikes the metal and thus offer a large sur-
face of contact for chemical reaction.

Slag Formation. — Manganese oxide unites with silica:

8. MnO + SiO2 = MnSiO8;

and it is perhaps for this reason that manganese, unless very high
(say over 2 per cent.), is removed early in the blow. The silicates
of iron and manganese dissolve in each other and form a slag, and
this slag will dissolve large amounts of iron oxide, manganese oxide,
silica and alumina, the latter coming from the vessel lining. At
all times the bath is so violently agitated that the metal and slag
are intimately mixed, and the reducing effect of manganese, sili-
con and carbon on the iron oxide dissolved in the slag limits it in
amount. The slag itself therefore serves as a carrier of oxygen
to the impurities.

Critical Temperature. — As the temperature rises, chiefly on
account of the oxidation of silicon, the chemical affinity of carbon
for oxygen increases relatively more than that of the other im-
purities, and reaction No. 4 ceases and then reverses:

9. SiO2 + 2C = 2 CO + Si.
What the critical temperature of this reversal is has never been

1 Air is composed of 20.8 parts, by volume, of oxygen and 79.2 parts of
^nitrogen, and the amount of blast per minute is 15,000 to 30,000 cu. ft.

FIG. 80. — SOAKING PIT HEATING FURNACES AT ROLLING MILL WHICH
RECEIVE THE RED HOT, NEWLY STRIPPED INGOT.

FIG. 81. — LINE OF COOLING INGOT MOLDS OUTSIDE BESSEMER MILL.

112

THE METALLURGY OP IRON AND STEEL

determined, but we may estimate it as being somewhere between
1450° and 1550° C. (2642° to 2822° F.), and unless the silicon is
all oxidized before this temperature is reached, we shall have
'residual silicon' in the steel. In English Bessemer practice,
where silicon is often above 2 per cent, of the pig iron, this is not

3.00 %

2.00%

LOO %

Carbon

Metal

Acid Bessemer Blow
American Practice

0 123456789 10

Minutes of Blowing End of Blow

FIG. 82.

rare, because the high silicon takes longer to go and also increases
the temperature:

10. Si + 20 = SiOa (generates 180,000 calories)1.

1 It is immaterial from the heat standpoint whether oxidation takes place
directly or as a result of two reactions:

2 Fe + 2O = 2 FeO (generates 131,400 calories)

2FeO + Si =SiO + 2Fe ( " 48,600 " )

180,000

THE BESSEMER PROCESS

113

There is also a critical temperature for manganese oxidation,
above which reaction No. 6 is reversed and residual manganese is
left in the steel ; but this happens only when the manganese in the
pig iron is very high. No warning is given that the temperature
is approaching these critical points, because no flame — nothing
but sparks — comes from the mouth of the converter until carbon
begins to burn, and therefore no indication is given to the operator
of the degree of heat until too late.

Second Period. — The second period begins when the carbon
commences to burn, which happens in America only after the

50*
40*
30*
20*

1701

Ibs.

/'*

5001

''liW

Ibs.

^MOOlb

V/M

anganese

Oxidi

f»'

.-—

/IK

Olbs

/10

XHbi

JO,

11234 56789 10
Minutes of Blowing End of Blow

FIG. 83.

silicon and manganese have been almost eliminated, as shown by
Tables XIII and XIV and Figs. 82 and 83. During this period the
reactions consist principally of the oxidation of carbon, although
a little silicon passes off at the same time. Further details will be

TABLE XIII.— REMOVAL OF IMPURITIES IN THE BESSEMER CONVERTER,
WITH ACCOMPANYING SLAG-ANALYSES

TIME AFTER
COMMENCE-
MENT OF
BLOW

REMOVAL OF METALLOIDS —
PER CENT.

Authority *

ANALYSIS OF CORRESPONDING SLAGS — PER CENT.

O
n

q
3

§<s

Alkalies

Min. Sec.
Pig iron.
2 0
3 20
6 3
8 8
9 10
Steel.
9 20

Pig iron.
8 0
15 0
17 0
18 0
Steel.

Pig iron.
After slagging
End of boil.
End of blow.

Pig iron.
3 0
4 45
5 45

Pig iron.
2 15
4 30
5 0

Pig iron.
3 0
5 0
5 45

Pig iron.
2 30
5 30
6 30

Pig iron.
4 15
8 35
9 20

Pig iron.
3 0
6 0
9 0
12 0
14 30
16 30

2.98
2.94
2.71
1.72
0.53
0.04

0.45

3.55
3.21
1.25
0.207
0.034
0.370

3.93
2.465
0.949
0.087

4.00
4.30
0.90
0.10

3.94
4.20
1.10
0.05

4.49
3.87
1.30
0.33

4.35
4.10
1.00
0.08

4.22
4.20
1.30
0.55

3.5
3.6
3.4
2.4
0.09
0.075
0.0

0.94
0.63
0.33
0.03
0.03
0.02

0.038

2.39
1.08
0.11
0.06
0.04
0.06

1.96
0.443
0.112
0.028

1.02
0.03
0.03
0.03

1.14
0.04
0.03
0.01

1.08
0.03
0.03
0.02

0.88
0.10
0.05
0.04

1.06
0.43
0.12
0.07

1.70
0.80
0.28
0.05
0.01
0.0
0.0

1.43
0.09
0.04
0.03
0.01
0.01

1.15

0.49
0.15
0.13
0.13
0.10
1.17

3.46
1.645
0.429
0.113

1.83
0.22
0.12
0.09

0.64
0.12
0.12
0.06

0.83
0.11
0.09
0.07

1.15
0.15
0.15
0.08

5.12
3.26
0.85
0.43

0.100
0.104
0.106
0.106
0.107
0.108

0.109

0.09

6.'09
0.08

6.09'

0.040
0.040
0.045
0.045

0.06
0.06
0.06
0.06
0.06
0.06

0.059

Howe.
40 American 39

42.40
50.26
62.54
63.56

5.63
5.13
4.06
3.01

40.29 4.36
34.24 0.96
21.26 1.93
21.39 2.63

6.54
7.90
8.79
8.88

1.22
0.91

0.88
0.90

0.36
0.34
0.34
0.36

0.008
0.008
0.010
0.014

0.009
0.009
0.014!
O.OOSi

62.20

.276

17.44

2.90

13.72

0.87

0.29

0.010

0.011

A12O3
&P205

0.018
trace

*§

.S.2
Wfc

W3)
c
W

oM .

1*
.Is

•Bit

J50

•5*
^

J*
•&»

*j»

*|!

.2*

62.65
73.24
75.63
61.30
64.15

40.95
46.78
51.75
46.75

50.20
55.26
47.20
40.50

46.50

48.76
59.82
48.48

51.00
53.44
57.80
55.76

47.16
53.26
62.34
44.52

46.72
45.87
39.07
37.63

7.98
4.51
5.19
4.24
5.71
A1208

8.70
4.65
2.98
2.80

3.86
2.86
2.70
2.24

12.90
0.78
0.98
0.72

2.56
1.84
1.94
1.58

5.83
2.28
3.90
2.14

4.36
3.08
2.49
2.94

1.93
0.00
0.00
13.47
13.95

0.60
6.78
5.50
16.86

1.80
14.20
18.52
31.19

0.90
34.72
21.08
35.82

0.90
20.24
17.04
18.48

0.77
13.50
9.54
30.60

0.70
4.20
6.24
9.45

'.'.'.'.

10.52
8.72
7.70
9.12
2.39

15.78
11.83
10.92
10.82
12.81

2.18
37.00
37.90
32.23

5.44
26.31
31.01
25.43

0.58
13.95
15.48
12.29

3.40
23.90
22.80
22.23

2.14
29.76
23.70
21.39

11.16
46.38
52.26
48.92

0.65
1.11
0.96
0.75
0.75

30.35
2.98
1.76
1.19

27.22
0.62
0.38
0.32

32.07
2.60
3.25
2.35

31.80
0.44
0.46
0.36

21.79
0.42
0.60
0.38

19.10
1.26
0.70
1.00

0.52
0.64
0.38
0.29
0.24

16.31
1.53
0.45
0.52

10.88
0.22
0.14
0.11

6.75
0.24
0.30
0.21

9.03
trace

.....

0.01
0.03
0.02
0.01

0.34
0.04
trace

0.32
trace

,...

:::::

: : : : :

::::

:::::

22.18
0.23
0.28
0.21

18.37
0.54
0.29
0.46

.!...

'.'.'.'.'.

1.50
1.50
1.63
1.43
1.42
1.20
0.08

0.05
0.05
0.05
0.05
0.05
0.05
0.05

P2Cv

32.6
42.6
36.0
35.6
33.0
15.6

7.26
2.57
5.91
6.17
7.89
13.43

0 60

••••••

0 15

'.'.'.'.'.

:::::

•••••

:::::

1.60
2.61
5.66
15.06

::'.:'.

* See No. 52, page 125.

THE BESSEMER PROCESS

115

TABLE XIV. — REMOVAL OF IMPURITIES IN THE BESSEMER

PROCESS

TIME

AFTER

COM-
MENCE-
MENT
OF BLOW

REMOVAL OF METALLOIDS — PER CENT.

References and Remarks

Min. Sec.
Pig iron .
5 0
10 0
15 0
20 0
25 0

Pig iron .
5 0
10 0
15 0
20 0
25 0

Pig iron.
4 30
13 0
16 0
Steel... .

Pig iron .
6 0
12 0
18 0
Steel....

Pig iron.
6 0
9 0
13 0
Steel....

Pig iron .
5 0
10 0
15 0
20 0
25 0

Pig iron .
5 0
10 0
15 0
18 0

Pig iron .
4 30
9 15
11 15
13 0

3.52
3.6
3.3
2.5
1.0
trace

3.5
3.6
3.3
2.5
1.0
trace

3.52
2.78
0.43
0.05
0.23

3.57
3.95
1.64
0.19
0.37

3.270
2.170
1.550
0.097
0.519

3.5
3.6
3.3
3.25
2.0
trace

3.50
3.55
2.35
0.07
trace

2.97
2.480
0.811
0.049
0.0

3.00
2.0
1.25
0.75
0.65
0.35

3.0
1.75
1.25
0.9
0.7
0.5

1.85
1.21
0.93
0.28
0.27

2.26
0.95
0.47

trace
a

1.952
0.795
0.635
0.020
0.033

2.25

1.0
0.5
0.2
0.1
trace

1.50
0.50
0.09
trace

1.25
0.60
0.20
0.10
trace

German method. "Leav-
ing Silicon in the
Bath," Stahl und
Eisen, vol. iii, pp.
262-264.

German method. Same
reference.

German method. Carl
Rott. 31

English method. Carl
Rott. 31

English method. Snelus.
See Encyclopaedia
Britannica, Ameri-
can ed., vol. xiii, p.
334.

.Stahl und Eisen, vol.
iii, pp. 262-264.

["Basic Process," Stahl
\ und Eisen, vol. iii,
pp. 262-264.

"Basic Process," by
Muller, at Horde,
Encyc. Brit., p. 346.

0.75
0.25
trace

1.93
1.69
1.00
0 37

0.62

0.04
trace

u
0.54

0.086
trace

0.309

1.0
0.35
0.2
trace

0.048
0.051
0.064
0.067
0.053

0.014
trace

u
u
u

0.71
0.56
0.27
0.12
trace

0.61
0.247
0.0
0.0
0.123

1.57
1.60
1.43
1.22
0.08

1.22
1.250
1.320
0.786
0.021

0.16
0.14
0.13
0.12
0.10

0.15
0.206
0.262
0.262
0.206

0.53
0.009
0.0
0.0
0.0

116

THE METALLURGY OF IRON AND STEEL

teamed from the tables and figures. It is interesting to note that
the phosphorus and sulphur in the metal are not eliminated, be-
cause the acid slag will not dissolve them even if they become
oxidized. For this reason the percentage of these impurities in-
creases slightly during the blow, because their actual weight
remains the same, while the weight of the bath decreases.

TABLE XV. — ANALYSES OF BOTTOM-BLOWN CONVERTER-GASES

TIME AFTER

PEJ

t CENT.

STARTING BLOW

CO2

Reference *

2 min

10 71

0.92

88.37

3 95

8.59

0.88

86.58

fi "

4 52

8 20

2 00

85 28

Sir Lothian

10 "
12 "

19.59
29 30

3.58
2 30

2.00
2 16

74.83
66.24

Bell. 37

14 "

31 11

1 34

2.00

65.55

18 "
3 to 5 min

End of

blow.
9 127

4.762

86.111

9 to 10 " ...
21 to 23 " ...
26 to 27 " ...

2 to 3 min

17.555
19.322
14.311

5.998
4.856
1.853

6 608

1.699
0.967
0.550

7 256

0.908
1.120
1.699

73.840
73.735
81.587

86.137

Y38
)

8tolO " ...
12 to 15 " ...
17tol9 " ...

15.579
25.580
25.606

5.613
4.144
2.995

1.296
0.980
1.318

1.112
1.040
1.120

76.400
68.256
68.961

jaa '

* See No. 52.

Gases. — In connection with Table XV it is interesting to note
what a large proportion of the carbon is oxidized to the monoxide,
and this is the more important because this formation generates
only 29,160 calories, while the higher oxidation generates more
than three times as much:

C + 2 O = COa (generates 97,200 calories).

Thus a large amount of heat is wasted, and is generated only when
the flame passes out of the mouth of the converter and unites with
more oxygen. When the heat of a charge is too low, it is custom-
ary at some plants to tip the converter forward or backward, so
that a few tuyere holes will be above the level of the bath and will
blow free air into the interior of the converter. This results in a
portion of the carbon monoxide being oxidized to carbon dioxide

THE BESSEMER PROCESS

117

inside the converter, and is a limited means of making the blow
hotter. Oxidation of additional iron produces the same effect.

Slag. — In Table XIII the lime and magnesia in the slag come
from a small amount of blast-furnace slag which finds its way to
the converter through the mixer, in spite of efforts made to hold it
back at all points when pouring. Because the blast-furnace slag
is basic, it has the effect in the converter of making the slags wet
and sloppy, and therefore increasing the loss. Although the iron
as shot is only shown in one case, this is not because it is absent
in the other cases, but merely because it was not determined. At
all times the slag carries a great many pellets of iron, which should
be added to the combined iron, since they represent a loss in the
process. It is interesting to note how closely the amount of iron
in the slag, after adding the spiegeleisen, approximates 15 per
cent.,1 and there seems to be a chemical balance which fixes this
amount as a condition of equilibrium. When the silicon in the
pig iron is higher, and therefore the amount of slag made is
'larger, there is a slightly lower percentage of iron oxide in it.
The practice of 'side blowing for heating/ described above, has
the effect of increasing the amount of iron oxide in the slag by in-
creasing the oxidizing influences in the interior of the converter.
The rise in manganous oxide in the slag during recarburizing
is, of course, due to the formation of MnO by the action of the
manganese on the oxygen of the bath, while the iron oxide in
the slag is reduced at the same time by the action of manganese
and carbon.

The weight of slag at different periods of the Bessemer process
has been calculated by H. H. Campbell 2 from its analyses, with
the following average results:

TABLE OF SLAG WEIGHTS IN BESSEMER PROCESS

Percentage of blow
finished

Pounds of slag

Percentage
of charge

Silicon flame

1035

4.5

Brightening

1146

5.1

Carbon flame

1255

5.5

Full carbon flame

1385

6.1

1 Not including pellets, which average 6 to 8 per cent, of the slag.

2 See page 158 of No. 2. Old edition, 1904, on page 8 herein.

118 THE METALLURGY OF IRON AND STEEL

The final amount of slag made will probably average about 7J to
8 per cent, of the weight of the metal produced, or, roughly, 7 per
cent, of the weight of the pig iron charged.

Flame. — A flame is the result of burning gas, and as practically
the only gas in this process is carbon monoxide, there is no flame
except during the period when the carbon is burning. In the first
part of the blow a large number of small sparks issue from the
mouth of the converter, consisting mainly of pellets of iron and
slag ejected by the blast. At the end of the first two or three
minutes, a small tongue of reddish-yellow flame begins to pour
from the mouth, showing that the carbon is beginning to be oxi-
dized. This soon increases in size and brilliancy until a white-hot
flame, 30 ft. in height, pours from the vessel with a loud roaring
sound caused by the boiling of the bath and the passage of the
blast and gas through it. This boil lasts until the end of the
operation, and the bath is at all times violently agitated and inti-
mately mixed with the slag, which greatly facilitates the reaction
between the two. The process at this period presents a spectacle
which is almost unmatched as a pyrotechnic display. Soon the
flame begins to flicker or 'feather' at the edges, as a warning that
the carbon is becoming low, and finally it shortens or drops,
whereupon the converter is immediately turned down and the
blast stopped. The carbon at this time will be about 0.03 to
0.10 per cent., depending on how ' young' or how 'full' was the
blowing.

At all times there is a varying amount of slag and metal thrown
out of the mouth, so that the converter may be likened to a foun-
tain of sparks, the great bulk of which consists of slag. Through-
out the blow there is also a constant stream of fume issuing with
the flame. It is a brownish-red smoke, which rises to a good
height and consists principally of oxide of iron and manganese.
Dr. Charles F. Chandler, Professor of Chemistry at Columbia Uni-
versity, while observing this smoke at the Homestead steel-works,
suggested to me that it might be due to the formation in the bath
of iron carbonyl, a volatile compound of iron, carbon and oxygen
(Fe(CO)5), and perhaps of manganese carbonyl.

Loss. — The difference in weight between the pig iron charged
into the converter and the steel ingots made will be 8 per cent, in
good practice, although running above that (say to 10 per cent.)
in some mills. This is distributed as follows:

FIG. 84. — BESSEMER FLAMES DURING A BLOW.

120 THE METALLURGY OF IRON AND STEEL

Carbon burned 3.5 per cent.

Silicon burned 1.0

Manganese burned 0.5

7 per cent, slag @ 15 per cent. Fe 1.0

7 per cent, slag @ 7 per cent, iron pellets 0.5

Volatilized and ejected 1.5

8.0 "

Recarburizing. — It might be thought that a more economical
method could be found than that of burning up all of the carbon
in the pig iron and then adding the desired amount, and in fact
such a method is employed in Sweden, where the carbon is burned
down to the desired point, as estimated by the appearance of the
sparks issuing from the converter, the vessel being then turned
down and the charge held until a hammer-test confirms this esti-
mate. Such a complicated procedure, however, requires a hotter
bath and very slow working, and it is much cheaper to burn the
carbon until the drop of the flame and then add the requisite
amount. In making rail-steel to contain about 0.50 per cent,
carbon, we will add to 15 tons of blown metal about 3000 pounds
of melted spiegeleisen, containing roughly 6 per cent, of carbon,
12 per cent, of manganese, and 1.50 per cent, of silicon. The
mixture charged into the spiegel cupola for melting must be higher
in manganese than this, as there is a loss into the cupola slag by
oxidation.

In making dead-soft steel for wire, material to be welded, etc.,
we add ferromanganese as high in manganese as possible. This
material will contain about 7 per cent, of carbon, 80 per cent, of
manganese, and 13 per cent, of iron. It is only necessary to add
about 500 pounds to a 15-ton bath, and therefore it will dissolve
without being melted, although it is customary to heat it to a red
heat in order to lessen the chilling of the metal. This dead-soft
material will then have the requisite amount of manganese, but
will be low in carbon and frequently less than 0.01 per cent, in
silicon. It is not improbable that pure manganese metal, if it
were readily obtainable, would be used in many cases instead of
ferromanganese, in order that the carbon might be still lower.
Ferromanganese and spiegeleisen are made in the blast furnaces
by smelting very high manganese ores in a manner somewhat simi-
lar to the smelting of pig iron.

Calorific Equation of the Acid Bessemer Operation. — The pre-
dominant part that silicon plays in furnishing heat for the acid

THE BESSEMER PROCESS 121

Bessemer process is shown by a calorific calculation. In making
this, let us assume that we have a charge of pig iron weighing
30,000 lb., and that we burn: silicon = 1.00 per cent. ; manganese =
0.40 per cent.; carbon = 3. 50 per cent.; iron = 2.00 per cent. And
let us assume further that the average temperature during the
operation will be 1500° C., and that the atmosphere is at 0° C.
Then:

Si + 2O = SiOa produces .................. 1,900,000

28.4 2X16

300 lb.!+ 338 lb. =638 lb.

338 lb. of oxygen =1470 lb. air.

Specific heat air =0.268 cals. per lb.

1470 lb.X!500° C.X0.268 = ............... 591,000

- - - 1,309,000
Mn + O =MnO produces ................ 198,000

55 16
120lb.2+35lb. =155lb.

35 lb. oxygen =152 lb. air.
152 Ib.X 1500° C.XO. 268= ................. 61,000

- 137,000

C + O. =CO produces ............... 2,552,000

12 16

1050 lb. + 1400 lb. =2450 lb.
1400 lb. oxygen =6087 lb. air.

6087 lb. X15000 C. X0.268 = ................ 2,447,000

- - - - 105,000
Fe + O =FeO produces ................ 704,000

56 16

6001b. + 1711b. =771 lb.

171 lb. oxygen =743 lb. air.

743 lb. X 1500° C. X0.268 = ..... . ........... 299,000

- • - 405,000

Total net heat from chemical reactions = 1,956,000

1 30,000 lb. X 1 per cent.= 300 lb.

2 30,000 lb. X 0.40 per cent. = 120 lb.

Now let us suppose that the specific heat of the metal is 0.20
calories, per pound, per degree Centigrade; then how many degrees
will it be raised by the heat produced in the chemical reactions
of the blow?

30,000 lb.X0.20 cals. = 6,000 cals. per 1° C.
1,956,000 cals. H-6,000 cals. = 326° C. Answer.

This simple calculation neglects the heat lost by radiation through
the vessel lining, and the heat necessary to raise the silicon, man-
ganese and carbon of the bath from their temperature at the begin-
ning of the blow, to 1500° C., and also leaves out of account the

122

THE METALLURGY OF IRON AND STEEL

heat produced by the combination of FeO and MnO with Si02 to
form the slag. All these figures are relatively less important,
but those who desire to calculate with greater delicacy should
consult J. W. Richards' very thorough little book, No. 53, Part
II, pages 307-354.

Basic Bessemer Process. — The basic vessel is almost the same
as the acid, except that the lining is made of calcined dolomite held

3.50
3.25
3.00

10 11 12 13 Umin.

.2.50
2.25
2.00
1.75
1.50

1.00

CU-T™

Mn.
C

Fll,

0 1 2 3 4 5 6 7 8 9 10 11 12 13 Umin.
FIG. 85. — REMOVAL OF IMPURITIES IN BASIC BESSEMER PROCESS.

together by about 10 per cent, of tar and rammed in around a pat-
tern while still warm. In ramming up the bottom wooden pins
.are rammed in with the lining, and when withdrawn they leave
J-in. tuyere-holes, through which the blast enters. The object
of the lining is to resist the chemical action of the slag, and it is not
desired to have it enter in any way into the chemical reaction.
Before beginning the blowing, lime is added to the bath, equivalent
in weight to 14 to 20 per cent, of the iron, in order that it may form
a basic slag, take up all the silica formed, and prevent this uniting
with the lining.

THE BESSEMER PROCESS 123

The basic blow is similar to the acid one up to the point where
the carbon is eliminated and the flame drops, before which prac-
tically no phosphorus is oxidized, as its chemical affinity for oxygen
is less than that of carbon. The blow is then continued for a few
minutes to form what is known as an 'after-blow/ during which
the phosphorus is oxidized and absorbed by the basic slag, prob-
ably in the form of calcium phosphate,. There is also an elimina-
tion of sulphur, at the same time, although this is never a very
satisfactory action. The indication given by the flame during
the after-blow is not a good guide, and the operation is usually
controlled by continuing the blow for the given number of minutes,
or the given number of revolutions of the blowing engine, after the
drop of the flame which experience has proved will produce the
desired dephosphorization. Before the recarburizer is added, a
sample of metal is ladled out of the bath, quickly cooled, and
broken, so that the appearance of its fracture may be used as a
final estimation of the elimination of phosphorus.

Recarburization of the basic heat cannot take place in the
converter, because the carbon monoxide formed at this time is
liable to reduce phosphorus from the slag and cause rephosphori-
zation of the metal. At the end of the blow the slag is therefore
poured off the bath as completely as possible, and then the rest is
held back in the vessel when the metal is poured into the ladle.
The amount of basic slag made will be about 25 per cent, of the
weight of iron charged and will contain about 9 per cent, of iron.
The loss of metal will be much higher than in the acid process,
averaging perhaps 13 to 17 per cent., a part of which is due to the
fact that the pig iron used contains a larger amount of impurities
to be oxidized:

Carbon , 3.7 per cent.

Silicon 0.5 '

Manganese 1.5

Phosphorus 2.5

25 per cent, of slag @ 9 per cent. Fe 2.3

Pellets 1.0

Fume and ejected 1.5

13.0 "

The demand for a given composition of pig iron for the basic
process is even more rigid than for the acid process. The silicon
is kept as low as possible in order to decrease the necessary lime
addition, and also because the combustion of phosphorus is here

124 THE METALLURGY OF IRON AND STEEL

relied upon to furnish the greater part of the heat. The phos-
phorus is not less than 1.8 per cent., and the manganese is high
in order to aid in the production of heat and the elimination of
sulphur. The calorific equation for fourteen tons of metal is shown
in Table XVI.

TABLE XVI. — CALORIFIC EQUATION OF THE BASIC BESSEMER

PROCESS
Charge: 30,000 Ib.

Analysis: 0.50 per cent, silicon burned; Temperature of atmosphere =0° C.;
1.60 per cent manganese Specific heat of air =0.268 pound-
burned; calories per 1° C.;

3.50 per cent, carbon Specific heat of metal =0.20 pound-

burned ; calories per 1° C. ;

2.50 per cent, phosphorus Average temperature of blow =

burned; 1500° C.

3.00 per cent, iron burned;

Principle of calculation same as in Acid Bessemer Operation, p. 121.

•p ^ Surplus

REACTIONS oXies P°und-

calones calories

2 P + 5 O =P2O6 produces ................ 4,420;000

62 80

750 Ib. 968 Ib.

4209 Ib. air X15000 C. X0.268 = .............. 1, 692,000

Si + 2O =SiO2 produces ................ 950,000

28.4 32
150 Ib. 169 Ib.

735 Ib. air X 1500° C.X0.268 = .............. 295,000

" 655,000

Mn + O =MnO produces ................ 794,000

55 16
480 Ib. 140 Ib.

608 Ib. air X 1500° C.X0.268 = ........ . ..... . 244,000

C. + O =CO produces ................. 2,552,000

12 16

1050 Ib. 1400 Ib.
6078 Ib. air X15000 C. X0.268 = .............. 2,447,000

Fe + O =FeO produces ................ 1,056,000

56 16
900 Ib. 257 Ib.

1118 Ib. airX!500° C. X 0.268 = ............. 449,000 607,000

Total net heat from chemical reactions = 4,645,000

Thus, much more heat is generated in the basic process, but
more is required because a large amount is absorbed by melting
the additions of lime to form the basic slag, and also there is
greater radiation because the blows are longer and the operation
is slower.

THE BESSEMER PROCESS 125

The basic process is much more expensive to operate than the
acid, on account of the longer time in blowing, delay at the end of
the operation to test for dephosphorization, and greater cost for
repairs, because (1) the basic lining costs more than the acid, and
(2) it lasts but a fraction as long. To balance this expense we
have the much decreased cost for the high phosphorus pig iron,
and also the slight return from the sale of the slag, which is high
enough in phosphorus to be used as a fertilizer. The process is no
longer operated in America, and it seems improbable that it ever
will be again. The high phosphorus ores of the Minette district
produce a pig iron which is, however, especially adapted to the
basic process, and the skill of the Germans in producing a high-
grade structural steel by this method, which is the predominant one
in Germany, excites the admiration of the other iron and steel
countries.

REFERENCES ON THE BESSEMER PROCESS

See Nos. 2, 30, 31, 32, 36, and the following list:

50. Richard Akerman. "The Bessemer Process as Conducted

in Sweden." Trans. American Institute of Mining En-
gineers. Vol. xxii, 1893, pages 277 et seq.

51. Henry M. Howe. "Notes on the Bessemer Process." Jour-

nal, Iron and Steel Institute, No. 11, 1890, pages 100
et seq.

52. Bradley Stoughton. "The Development of the Bessemer

Process for Small Charges." Trans. American Institute of
Mining Engineers. Vol. xxxiii, 1903, pages 846 et seq.

53. Joseph W. Richards. "Metallurgical Calculations." Part I,

Introduction, Chemical and Thermal Principles, Problems
in Combustion. 1906. Part II, Iron and Steel. 1907.
These problems not only teach how to calculate many very
important things in connection with furnaces and their
efficiency, but give a good insight into the principles of the
processes themselves.

54. Friedrich C. G. Miiller. Untersuchungen uber den deutschen

Bessemer process. Zeitschrift des Vereines deutscher In-
genieure. Vol. xxii, 1878, pages 384-404 and 454-470.

126 THE METALLURGY OF IRON AND STEEL

This is one of the most comprehensive studies of the metal-
lurgy of the Bessemer process in any language.
55. Hermann Wedding. "The Basic Bessemer, or Thomas,
Process." Translated into English by William B. Phillips
and Ernst Prochaska. New York, 1891. This is the fullest
account of the basic Bessemer process.

THE OPEN-HEARTH OR SIEMENS-MARTIN PROCESS
OPEN-HEARTH PLANT

THE arrangement of the open-hearth plant is not of such vital
importance as that of the Bessemer, because the open-hearth proc-
ess is so much slower that it is easier to arrange the different
cycles so that one will not delay another and lessen the output of
steei. The different open-hearth cycles are as follows:

1. Getting the stock to, and in, the furnace.

2. Supplying the furnace with fuel and air and preheating both
of these.

3. Working the charge, repairing the furnace, etc.

4. Recarburizing.

5. Disposing of the steel and slag.

6. Repairing and preparing ladles, ingot molds, etc.

A plan and elevation of a typical open-hearth plant is shown
in Figs. 90 and 91. It is not to be presumed that all plants are
laid out in the same manner, but that shown is a modern type
which is much favored in America. The furnaces are arranged in a
long row, with often as many as ten in one house and with the level
of the hearth several feet abo>ve the general ground-level of the
plant.

Melting Platform. — On the same level as the hearth, and in
front of the furnace, is the melting platform, or working platform,
upon which are placed one or more charging machines, depending
upon the number of furnaces to be served, running upon tracks ex-
tending the entire length of the working platform. The space
above this is usually spanned by one or more electric traveling
cranes which assist in repairs to the charging machines, in handling
materials on the melting platform, and in pouring molten pig iron
into the furnace where such practice prevails. The melting plat-
form has a small extension around the back of the furnace to afford

127

THE OPEN-HEARTH PROCESS

129

access to the tap-hole and the ladle into which the steel is poured,
and for putting the recarburizer into this ladle when necessary.
Upon the working platform are the valve handles for regulating
the admission of gas and air to the furnace and for reversing the
current of these at the proper time.

Gas Producers. — The gas producers are situated outside of the
furnace house and in a long line parallel to it. This arrangement
has the great disadvantage of placing the men on the working plat-

FIG. 92. — OPEN-HEARTH MELTING PLATFORM AND CHARGING MACHINE.

form between the smoke of the gas producers and the heat of the
furnace, but there seems to be no good way of avoiding it. As the
regenerators are usually placed underneath the working platform,
this situation of the producers gives the least possible distance
which the gas has to be carried, and therefore the least possible loss
by deposition of tarry components.

Stock. — The stock yards are oftentimes placed between the
furnace house and the gas producers, but this nearness is not
necessary and stock is frequently stored by the end of the house,

130

THE METALLURGY OF IRON AND STEEL

or even at some distance. For its transfer to the furnace, the
stock is loaded into steel boxes similar to those in Fig. 93. Three
or four boxes are supported on a little car, which is transferred in a
train to the melting-floor, passing over a pair of scales on the way,
where the weight is taken. Between the track of the charging
machine and the line of furnaces runs the track upon which these
cars are transferred, and if a constant supply of boxes is brought
to the machine, it can empty them upon ihe hearth of the furnace
at a rate of about 50 boxes (equivalent to about 125 tons) per hour.

FIG. 93.

Charging Machine. — A view of a Wellman charging machine
is shown in Fig. 94. Its essential feature is a long charging-bar
with a foot on the end which can be dropped into a socket on the
charging-box. By this means the charging-box is raised off the
car, thrust into the open door of the furnace, and turned upside
down to empty its contents of pig iron, steel scrap, limestone, iron
ore, or other material, upon the hearth. The operator is seated in
a little cage, which moves backward and forward with the charging-
bar, and has within his reach five levers: (1) To move the charg-

THE OPEN-HEARTH PROCESS

131

ing-bar inward and outward; (2) to move the charging machine
backward and forward on its track in order to serve any of the
furnaces, or to place itself opposite any of the doors of a single
furnace, or to push the train of charging-boxes along by means of
the charging-bar; (3) to lock the foot of the charging-bar in the
socket of the charging-box ; (4) to raise the charging-bar up and
down ; and (5) to turn the box over.

FIG. 94.

Casting-Pit. — The casting-pit extends all the way behind, the
furnaces; it is on the general ground-level of the plant and there-
fore several feet below the melting platform. This pit is spanned
by one or more electric traveling cranes of large capacity, which
are used to hold the ladles while the steel is running into them from
the tap-hole of the furnace and while it is being teemed into the
ingot molds. They are also used to serve the pit for several pur-
poses, such as carrying away the slag, transferring empty ladles
to and from the point where they are lined and dried, etc. Through
the casting-pit extend 2 to 4 railroad tracks. One of these, and
sometimes two, are used for the passage of the cars carrying the
ingot molds, and they therefore run along the side of the teeming
platforms upon which the ladleman stands to pour the ingots;
another is used for the railroad cars into which the slag, dirt, and
other waste material is dumped; and still another, if present, is kept
clear for transfer purposes.

132 THE METALLURGY OF IRON AND STEEL

OPEN-HEARTH FURNACE

The form and dimensions of a modern 50-ton rolling open-
hearth furnace is shown in Figs. 95 to 97. It consists of a long
shallow hearth suitably enclosed in fire-brick and bound together
with steel, and can be rolled forward in order to pour material
out of the tap-hole. A stationary furnace is shown in Fig. 100.
There are more stationary than tilting furnaces, but the general
principles of the steel-making operation are the same in both.

Regenerators. — With the furnace are connected two pair of
regenerators which preheat the gas and air for combustion. The
internal volume of each of these chambers is equal to -f to -fv
of that of the working-chamber itself. The larger the chamber
the greater will be the amount of heat intercepted in them, and
therefore the lower the temperature of the gases that go to the
stack. The amount of space actually occupied by the bricks, or
checkerwork, is the important consideration, however, and this
should be from 5000 to 10,000 cu. ft., total, for all four regenera-
tors in a 50-ton furnace, the capacity of the two gas regenerators
usually being less than that of the air regenerators, because the
volume of gas used is less than that of the air, and also because
the gas does not require to be preheated so much, since it is already
somewhat warm from the gas producer. During the operation of
the furnace more or less slag, dirt, and dust are carried over with
the outgoing gases. To intercept this the slag-pockets or dirt-
pockets A A (Fig. 97) are provided ; but in spite of them the space
between the bricks of the checkerwork becomes partially choked,
and for this reason, as well as because the deposit of dust makes the
surface of the bricks rough, the total area between the bricks must
be much larger than the area of the ports, so that the velocity of
the gas will not be lessened. The furnace must be laid off for re-
pairs when the passages between the bricks are choked by dirt,
but, on the other hand, the space is limited, because the bricks
must be laid in such a way that the maximum amount of surface
shall be exposed and the gases forced to the greatest possible con-
tact with them. The modern construction makes the regenera-
tors as tall as possible in order that incoming gas and air may be
forced into the furnace by the draught, and also because this chim-
ney effect causes the incoming gas and air to naturally seek the

Longitudinal Section and Elevation

Main Gas Flue

*xSf

i i ! i

H!"— <4i

J[ v iL

jrt-^^M

! ** — i

1ist»ckjr

l'^-'^1

' « ' I ,T I '

,--f====-^. ^jcat — A J/x-M -

ijiiSl" "7 1 "x?T?\

I —

-so'o- >]

"T"I

Casting Pit

•

FIGS. 95 AND 96.

134

THE METALLURGY OF IRON AND STEEL

hottest places and the outgoing gas to naturally seek the coolest
places, in this way equalizing the temperature in the different
parts of the regenerators.

The space underneath the checkerwork should be so large that
the incoming gas and air will distribute itself nearly uniformly
through the different passages, and the temperature of the fire-

n n r-i n r-i n

FIGS. 97 AND 98.

bricks at this lower part will be, say, 400° C. (752° F.), although
varying, of course, at different furnaces and at different times.
When the regenerator is receiving the waste gas from the furnace,
the temperature of these bricks will be that of the gases that go to
the chimney, say 400° to 600° C., and when the air or gas is passing
through the regenerator on its way to the furnace, these bricks

THE OPEN-HEARTH PROCESS

135

will be somewhat cooler, depending upon the length of time that
the regenerator has been in this phase of the operation. The tem-
perature of the bricks at the top of the regenerator will be about
1000° C. (1832° F.), and therefore the air and gas entering the
furnace will be the same.

Ports. — The ports are so arranged that the flame shall be
deflected away from the roof and yet not impinge upon the bath,
or impinge only very slightly, because the bath would thereby be
oxidized excessively. The gas should be spread out all over the

FIG. 99. — REMOVABLE BUGGIES TO CATCH SLAG IN SLAG POCKETS.

width of the hearth beneath the air, and the two should be brought
together just before they enter the laboratory, or work-chamber.
The air especially must be kept away from direct contact with the
bath, and for this reason the gas-ports are placed below the air-
ports, which arrangement has the further advantage of promoting
a better mixing of the two, since the gas is lighter and therefore
rises. In America the favorite arrangement is two gas-ports,
above which is situated a long slit, extending almost the entire
width of the furnace, which serves as an air-port. This is not uni-
versal practice, however, for in some cases there are two, or even
three, air-ports. In one case the two gas-ports are built wide and
low, so that they will deliver thin streams and get a better mix-

136 THE METALLURGY OF IRON AND STEEL

ing of the two materials for combustion. The area of the air-port
in a 50-ton furnace should be about 18 sq. ft., and the combined
area of all the gas-ports on one end should be from 8 to 10 sq. ft.,
depending upon the quality of gas used.

The roof must be protected from the direct impact of the flame,
because even the most refractory silica bricks would be melted by
the intense heat. The heat in the regenerators and uptakes gives
the gas and air a velocity which causes them to enter the furnace
with some force, and the construction of the ports directs the stream
in the desired manner. The mouths of the ports are gradually
melted away by the intense heat of combustion of the outgoing
gases, until they finally cease to serve this purpose and it is im-
possible to get the proper mixing and the proper kind of a flame in
the laboratory of the furnace. The ports must then be repaired,
else the temperature cannot 'be maintained.

Draught and Chimney. — The draught must be sufficient to catch
the flame about in the middle of the laboratory and drag it out
through the ports on the opposite side from which it entered with-
out allowing it either to drop down and touch the bath (as this is to
be heated almost altogether by radiation) or to impinge upon the
roof. This draught also has to do the work of overcoming the fric-
tion of the outgoing regenerators and flues. Its force will depend
upon the height of the chimney and the temperature of the products
• of combustion after they have left the regenerators, which should
be about 400° C. (752° F.), though even better economy (i.e., a
lower temperature) than this is obtained in many cases. All the
heat carried away by these flue gases is of course wasted, but the
very great cost of the refractory bricks in the regenerative cham-
bers makes it unwise to reduce the temperature of the flue gases
too much by enlarging the checkerwork. The calculation of the
amount of draught produced by any given height of chimney with
any given temperature of flue gases, according to the method of
Professor J. W. Richards, is given in Table XVII.

Roof. — The roof is made very thin and of the most refractory
bricks that can be obtained, i.e., almost pure silica, with enough
lime to hold it together in a compact mass. The walls are also
thin, and the radiation from the furnace chamber is great, but
this has to be endured, as thicker walls and roof produce endless
trouble by expansion and contraction. The roof is arched and
" suspended from beams independent of the side walls.

THE OPEN-HEARTH PROCESS 137

TABLE XVII. — CALCULATION OF DRAUGHT OF CHIMNEYS

The draught in a chimney is produced by the gases within it being lighter
than the air outside, and this lightness is due to their being hotter. In this
calculation we assume the following data:

Air weighs 1.29 oz. per cu. ft. at 0° C. and 760 mm. pressure;
Chimney gases weigh 1.03 times air at 0° C. and 760 mm. pressure;
Data Water weighs 772 times air;
assumed Mercury weighs 13.6 times water;

The friction of a chimney absorbs the equivalent of 0.1 in. of water
on a water-gauge showing draught, for every 100 ft. height
of chimney.

As the gases rise in the chimney they cool; therefore their average tempera-
ture will be about 50° C. cooler than their temperature at the foot of the
chimney. Call this average temperature T° C. Then:
9700 /^

1.29 oz. X 1.03 X — — = the weight of a cu. ft. of the chimney gases

T° C. + 273° C. at the temperature T° C.

Subtract this weight from 1.29 oz. and we get the difference in weight of air
and the furnace gases at the temperature T° C. Then:

Height of chimney X difference in weight = ounces draught pressure per square

foot of chimney area.

Ounces^draught X 12 in. = equivalent height of a water-gauge.

From the height of a water-gauge so found, we must subtract the amount
lost by friction, or 0.1 in. Xheight of chimney -7- 100.

For further modifications of this type of calculation see No. 53, Part I, pp.
164, 166, 194, 200, and 201.

Life of the Furnace. — The ' life' of an open-hearth furnace
means the number of heats that it can make continuously without
stopping for any more extensive repairs than can be made in the
usual week's-end shut-down. No figure can be given for this
except in the most general way. The life of the furnace will be
ended usually in one of three ways: (1) The falling in of the roof,
(2) the eating way of the ports, so that the flame can no longer be
maintained properly, or (3) the giving out of the regenerators,
which may occur either through the choking of the checkerwork,
or through a crevice formed by the contraction and expansion of
the bricks, so that there is a serious leak between the gas-chamber
and the air-chamber, and premature combustion takes place. If
a basic furnace makes 350 heats, it is considered good work, and we
may perhaps tentatively consider this figure as the "three-score
years and ten" of a furnace making steel for structural work and
similar purposes. Three hundred and fifty heats would mean

138

THE METALLURGY OF IRON AND STEEL

about 18 to 24 weeks' work in America. An acid furnace will last
about 1,000 heats.

Construction of Hearth and Bottom. — The hearth is made with
a thickness of 18 to 24 in. inside the furnace shell, in the form of a
shallow dish whose sides reach up to the level of the charging doors,
and so constructed that the depth of the metal will be from 12 to
24 in. — the former figure in the case of a very small furnace, say 5
to 15 tons, and the latter in the case of one of 50-ton capacity. If
the bath is too shallow the oxidation will be excessive and the
wear of the lining by oxide of iron, with consequent production

It is now considered best to have
a 12 * roof through-out

= 1 Course chrome brick'
Silica brick

^These checker brick
rest on tile 18 x 12 xt

Silica brick from hen i

clay brick below line

|4 __ Loam, fettling etc.

2 Courses magnesite brick on edge
1 Course chrome

Granite or concrete

These checkers are not in the
uptake but in a chamber just
beyond and opening into in.

FIG. 100. — MATERIALS OF CONSTRUCTION AND LINING FOR BASIC
OPEN-HEARTH FURNACE.

of slag, will be great;1 if the bath is too deep, the melting and
oxidation will be slow. In the case of an acid bottom, that portion
of the lining next to the shell will be made up of refractory clay
brick, and the upper portion will be formed by shoveling in silica
sand, spreading it out in a thin layer about ^ in. thick over the
entire hearth, and then allowing it to sinter at the full heat of the
furnace for about ten minutes, so as to set it firmly in place.
Upon this layer will be set another layer in like manner, until
the whole hearth is constructed, and then it will be 'washed' with

1 In an open-hearth furnace the lining suffers the greatest wear at the
side of the bath, where it is thin, because the oxidation of the metal is the
greatest there.

THE OPEN-HEARTH PROCESS 139

a melted bath of old slag to fill up all crevices and give a glazed
surface.

In the case of the basic hearth, the bottom is made of calcined
dolomite,1 held together with 10 per cent, or less of anhydrous tar.
In this case the layer of brick next to the lining is very thin and
the dolomite and tar are set in in layers by the heat of the furnace
as in the case of the acid lining. The tar burns to a strong coke,
which holds the mass together in a firm hard form. In some cases
no tar is used, and the calcined dolomite is fritted slightly to hold it
together; in other cases 15 per cent, of old slag is used as a bond.
Pure magnesite gives a more permanent lining than dolomite, and
is now much used since its longer life more than compensates for
its greater expense. Even when magnesite is used for the bottom,
the topmost layer, or working bottom, and the repairs, or 'fettling/
put in during the intervals of the furnace life, are made of dol-
omite, because this sets more quickly. As the sides and roof even
of the basic furnace are made of silica bricks, it is customary,
although not absolutely necessary, to put a layer of neutral
material between these bricks and the basic hearth, and also to
protect this joint from excessive heat. The neutral material
commonly used is chromite bricks, which are made of ground
chromite (FeO, C^Oa), held together with tar and then burned
to form a firm, hard mass.

Repairing Bottoms. — Between the heats, bottoms are re-
paired by filling up holes with acid or basic material, as the case
may be, and by more extensive attention at the end of the week.
In this way the bottom may be made to last almost indefinitely,
unless a part of the charge works its way down a crevice and
forces up whole sections of the bottom lining, which not infre-
quently happens; while sometimes the charge even works its way
out through the bottom of the furnace. A sticky or viscous slag
is also liable to bring the bottom up by sticking to it. In the
tilting furnace, the bottom may be repaired along the point where
the worst corrosion usually takes place — Le., at the edge of slag
line — even during the operation, by tipping the furnace until
this place is uncovered.

Tap-Hole. — In the stationary furnace, the tap-hole is made
according to the section in Fig. 101. During the operation this is

1 Dolomite is a magnesia limestone, and after calcining consists of a mixture
of lime and magnesite (CaO, MgO), with a little silica and other impurities.

140

THE METALLURGY OF IRON AND STEEL

closed by material rammed into it from the door of the furnace.
It sets quickly into a solid mass, which is pierced with a pointed
bar when it is desired to allow the metal to run out. After tap-
ping, the hole must be entirely freed from metal, and then it is
made up and filled anew, ready for the next operation. This
always delays work to some extent, and occasionally the tap-hole
causes trouble by the charge working through it prematurely, or,
on the other hand, by its becoming so hard that a hole is pierced
in it only after a long delay and much difficulty, during which the
oxidation continues in the charge beyond the desired point. In
the tilting furnace there is no tap-hole, strictly speaking, but the

FIG. 101.

opening into the metal-spout is closed by loose material, which is
scraped away before the furnace is tipped to pour the charge.

Tilting Compared with Stationary Furnaces. — Tilting furnaces
are more expensive to install than stationary ones, and require
repairs to the machinery and power to operate them. They also
require special arrangements of ports and uptakes, to be described
later, and means for cooling the junction between the movable
and stationary parts of the furnace. Their advantages are that
they do away with troubles and delays from the tap-hole, the
slag can be poured off at any time, and the charge may be tapped
at a moment's notice, which is especially advantageous when
making steels within very narrow limits of composition. Further-
more, the back-wall of the bottom may be repaired more easily
between heats when the furnace itself is still white hot. In the
stationary furnace the angle of this back-wall will be about 60°
or 70° from the horizontal, and loose material thrown in from

THE OPEN-HEARTH PROCESS

141

the front doors will not rest on this angle; but the tilting furnace
can be tipped until the back-wall is more nearly horizontal, and
then loose material will remain upon it until sintered into place.

The tilting type is of more advantage in the basic process than in
the acid, because it enables the slag to be poured off at will, which is
occasionally advantageous in basic practice.

142

THE METALLURGY OF IRON AND STEEL

Tilting Furnaces. — There are two types of tilting furnaces,
known respectively as the Campbell and the Wellman. In the
Campbell type the hearth of the furnace is arranged so that the
center of tilting is coincident with the center of the ports, and
therefore the furnace can be oscillated without cutting off the
supply of gas and air. In order to facilitate this, there is a little
clearance between the uptake and the furnace proper, and these
parts are surrounded by water-cooled castings. In the Wellman

FIG. 103. — TILTING FURNACES.

type the gas and air supply must be cut off when the furnace is
tilted. In tipping the Wellman furnace the ports move with the
hearth, and they are therefore seated in a water-tank, which makes
an air-tight connection with the regenerators when the furnace
is in a horizontal position, but breaks it when it is tipped forward.
The Wellman type is not as expensive to build as the Camp-
bell, and probably requires less repairs. The Campbell type has
the advantage that the bottom can be repaired along the slag line
without interrupting the operation, and that the bottom can be

THE OPEN-HEARTH PROCESS

143

sintered into place by the heat of the flame when the hearth is in
any position. This is more important in the acid furnace than in
the basic, where the mixture of dolomite and tar can be set by

the heat contained in the furnace walls themselves. In the Well-
man type, when the furnace is tipped for pouring, cold air can
enter it through the port-holes, and may oxidize the manganese in
the final product after recarburizing. Finally, a great advantage

144 THE METALLURGY OF IRON AND STEEL

of the Campbell type is the fact that a great deal of ore can be
used during the operation, and although the boiling of the charge
is violent on this account, metal does not flow out of the furnace
doors, because the hearth can be tipped in the opposite direction.
The slag which runs off during this period is allowed to pass
through a hole in the bottom of the port-opening, at the joint be-
tween the fixed and the rotating portion, where it is continually
exposed to the flame and therefore not liable to chill up.

Temperature Calculation. — There is almost no limit to the
temperature that can be obtained in the open-hearth furnace by
frequent reversing of the valves, except the danger of melting the
roof. The temperature is controlled by reversing and by throt-
tling the amount of gas and air admitted to the furnace. It is the
almost universal practice to reverse the valves every twenty min-
utes, in order to maintain a uniform temperature of the regenera-
tors. Thp melter endeavors to keep the charge always in a very
liquid state, and at the same time to have a slight excess of air, in
order that the atmosphere of the furnace may be slightly oxidizing
to burn the impurities in the metal. Every excess of oxygen,
however, causes a loss of heat, because each volume of oxygen in
the air is accompanied by four volumes of nitrogen, which carries
heat out of the furnace, but does not assist in any way in the re-
actions.

The actual temperature of the furnace will depend upon the
length to which the decarburization is carried, because as the metal
gets lower and lower in carbon, it requires a higher heat to keep it
fluid. It will average about 1600° to 1700° C. (2912° to 3092° F.).
An operation which is interrupted when the carbon is still 0.50
per cent, will take much less fuel than one in which dead-soft steel
is manufactured. The amount of coal burned in the gas pro-
ducers, per ton of steel produced, will therefore vary greatly, but
will average perhaps 400 to 1000 Ib. Figures for oil and natural
gas fuel will be given on pages 167 and 169.

Size of Open-Hearth Furnace. — The so-called ' standard' open-
hearth furnace has a capacity of 50 tons. The bath in such a fur-
nace will have a length of about 30 to 35 ft., a width of about 12 to
15 ft., and a maximum depth of about 24 in. In America, how-
ever, there is now (1907) a tendency to increase the size to a
capacity of 60 or even 75 tons. This innovation is due largely to
Mr. T. S. Blair, Jr., of the Lacka wanna Steel Company; some of the

THE OPEN-HEARTH PROCESS 145

furnaces under his control are 43 ft. between the ports. The re-
sult is a much better opportunity for complete combustion within
the laboratory of the furnace, and therefore less deferred combus-f
tion as the gases escape through the ports and downtakes. This
lengthens the life of the ports and promotes fuel economy.

The smallest practicable size of an open-hearth furnace is about
15 tons, and this is very expensive to operate. There are furnaces,
however, as small as 5 tons capacity, but this is not good practice,
even under the special circumstances which alone justify the use
of the 15-ton furnace. The maximum practical size will not be
far above 75 tons, and the real governing factor is the ability of
the mechanical apparatus to handle the raw material and the
product, besides which it is difficult to cast so much metal out
of one ladle without having the casting temperature of the first
metal too hot, or else that of the last metal too cold.

BASIC OPEN-HEARTH PRACTICE

Formerly, the open-hearth practice was divided into two types,
known respectively as the ' pig-and-ore process7 and the ' pig-and-
scrap process.7 In the pig-and-ore process the charge consisted
entirely of pig iron, and the oxidation was hastened by the addi-
tion of as much ore as the charge would stand without boiling
over; in the pig-and-scrap process the charge consisted of pig iron
with large amounts of steel scrap, the proportion of the latter being
so large in some cases that the operation became a mere remelting
process, there being only enough pig iron for its silicon, manganese
and carbon to protect the metal from excessive oxidation. At the
present time the pig-and-ore process does not exist in its original
form, but normal charges consist of pig iron and steel scrap, in pro-
portions determined by financial considerations, and the opera-
tion is hastened by judicious additions of ore. The amount of
steel scrap will vary all the way from 0 to 90 per cent., and will
average not far from 50 per cent, in America (1906). By decreas-
ing the proportion of impurities in the raw material in the charge
the scrap enables the process to be completed in less time.

Charge. — A difference of opinion exists as to whether it is
desirable to charge the steel scrap first upon the hearth and then
cover it with the pig iron, in order that the impurities in the latter
may lessen the oxidation of iron during the melting period, or

146 THE METALLURGY OF IRON AND STEEL

whether it is better to put the pig iron on the hearth and cover it
with the steel, upon the theory that the hearth is in this way cor-
roded less by oxide of iron. The former practice is in vogue at
several American works, and the latter at some American and
many English works. It is probable that the scrap, unless small
in size, does not corrode the hearth much, and the practice of put-
ting the pig iron on top is preferable. In acid open-hearth practice
it is more common to charge all, or most, of the scrap on top; in
basic practice we may charge, first limestone, then scrap, and
lastly pig, or else, limestone, most of the scrap, most of the pig,
and then a layer each of scrap and pig.

Composition of Pig Iron Used. — Basic pig should be free from
adhering sand, and therefore only machine-cast metal will ordi-
narily be used, or else that cast in metal molds. Its silicon should
be below 1 per cent., and its manganese above 1 per cent., though
the price of manganese ore makes this ingredient an expensive
luxury. It is desired because it makes the slag more fluid and
aids in removing sulphur. The phosphorus will be almost any
figure up to 2 per cent., but a regular supply of pig with more
phosphorus than that would tempt the manager toward the basic
Bessemer process. Silicon and phosphorus increase the amount of
slag, lengthen the operation, and require lime to flux them.

Fluxes. — The function of the basic lining is to remain inert
and serve simply as a container for the bath. In order that the
slag may be at all times rich in lime (35 to 45 per cent. CaO ordi-
narily), from 5 to 30 per cent, of lime is added with the pig and
scrap, the exact amount depending upon the purity of; {the metal
charged. Especially if the charge is high in sulphur, the slag must
be kept as rich in lime as possible without making it too infusible
and therefore viscous, the limit being about 55 per cent. CaO in
this respect. The lime is usually added in the form of calcined
limestone, and, as already noted, the higher the silicon and phos-
phorus in the metal the greater must be the amount of lime used.
It is also customary to charge ore with the pig and scrap, in order
to increase the amount of oxidation that takes place during the
melting. This ore has very little effect on the basic lining,
although it would rapidly corrode a silicious one. Further addi-
tions are made during the operation if necessary. The average
amount of ore added to a 50-ton charge will be between J of a ton
and 2J tons.

THE OPEN-HEARTH PROCESS

147

Chemistry. — It takes about 3 or 4 hours for the charge to
melt, and during that time the silicon is almost entirely eliminated,

10-

Hours

FIGS. 105 AND 106. — CHEMICAL CHANGES IN BATH AND SLAG IN A BASIC
OPEN-HEARTH FURNACE.

while the proportion of carbon and manganese is somewhat reduced,
as shown by Figs. 105 and 106, which also show the chemical

148 THE METALLURGY OF IRON AND STEEL

changes that take place during the entire operation. These reac-
tions must be controlled by the melter, because it is necessary
that the carbon shall be eliminated last, and therefore it is occa-
sionally necessary to 'pig up7 the charge, i.e., add pig iron in order
to increase the amount of carbon. On the other hand, in case
the phosphorus is being eliminated very fast, the oxidation of the
carbon may be hastened by 'oreing down/ as it is called, i.e.,
adding ore to produce the reaction.

Fe,O3 + 3C = 2Fe + SCO.

If the carbon is eliminated before the phosphorus, a great deal
of iron will be oxidized, because the phosphorus does not protect
it as much as the carbon.

When the charge is melted, and at intervals thereafter, the
melter takes a spoon-ladle full of metal and pours it into an iron
mold. As soon as this is set hard it is cooled in water, and from
the appearance of its fracture the melter can estimate very closely
the amount of carbon and phosphorus it contains. In many
plants it' is customary to tap heats on these estimates alone and
astonishing accuracy can be obtained by these methods as to the
amount of carbon contained in the bath of steel. In all large
plants, however, the laboratory analysis is made for phosphorus
and usually for carbon as well, and these determinations may be
made in 20 minutes or less.

Functions of the Slag. — The functions of the slag are: (1) To
absorb and retain the impurities in the metal, particularly silicon,
manganese and phosphorus, and as much sulphur as possible; (2)
to lie upon the top of the bath as a blanket and protect it from
excessive oxidation by the furnace gases; and (3) to oxidize the
impurities of the bath by means of its dissolved iron oxide, which
serves as a carrier of oxygen from the furnace gases to the impuri-
ties in the metal.

For the retention of phosphorus and sulphur, the slag must be
rich in bases. For the oxidizing it is necessary that the s7ag shall
be fluid, so that it will mix easily with the bath, and therefore the
content of lime must not be too high. During the boil, when the
carbon is passing off, there is an intimate mixture of metal and
slag, and there is even such a violent agitation of the bath that the
metal itself is frequently uncovered in places and exposed to direct
oxidation by the furnace gases.

THE OPEN-HEARTH PROCESS 149

The removal of sulphur is a variable quantity and cannot be
altogether depended upon, but there is usually a large loss of this
element during the operation, in spite of slight additions of sulphur
from the coal through the gas. This sulphur reduction is greatly
assisted if the slag is thinned out (i.e., its melting-point reduced)
by the addition of calcium fluoride (CaF2) just before the end of
the operation. The higher the slag is in lime, provided it remains
at the same time fluid, the more complete in general will be the
sulphur elimination, but over 55 per cent, of lime usually makes
the slag viscous, unless the calcium be added in the form of fluoride,
or of chloride (CaCl2) . The latter agent was added in many cases
by the recommendations of E. H. Saniter, but the practice was
never general in the United States, and the use of calcium fluoride
(known as fluorspar) answers most purposes, unless the sulphur
in the charge is excessive.

It is of very great importance that the final slag in the open-
hearth operation shall be a nonoxidizing one, lest the metal itself be
full of oxide and very 'wild/ i.e., give off gas abundantly during
solidification. Therefore no ore should be added within two or
three hours of the finish of the operation. For this reason there is
some difficulty in making steel of under 0.20 per cent, carbon by
the open-hearth process, and the difficulty increases almost in
geometric ratio as the carbon is reduced lower and lower, be-
cause iron retains with tremendous affinity the last traces of
carbon, and it requires a most powerful influence to separate them.
Open-hearth managers usually refuse to take orders for steel
under 0.10 per cent, carbon, except where consumers own their
own plant.

Slag. — Basic slags will contain :

10 to 20 per cent. SiO2
45 to 55 per cent. CaO
10 to 25 per cent. FeO
5 to 15 per cent. P2O6.

Weight of Slag. — The slag made will depend upon the silicon
and phosphorus and dirt in the metal, and will average from 10 to
30 per cent, of the weight of metal. Since the slag takes practi-
cally all the lime charged into the furnace, its weight can be calcu-
lated with sufficient accuracy by the same formula given for calcu-
lating the blast-furnace slag: Divide the total lime (CaO) used

150 THE METALLURGY OF IRON AND STEEL

with the charge (plus 30 per cent, of itself to allow for wear of the
lining) by the percentage of lime in the slag.

Loss. — The weight of steel produced will be a variable propor-
tion of the weight of metal charged, depending upon the amount of
pig iron in the charge, the amount of ore used, the extent to which
the carbon was eliminated before tapping, etc., but it will average
perhaps 93 to 96 per cent., the difference being made up as follows,
in a typical example :

ANALYSIS OF CHARGE
(50 per cent, pig; 50 per cent, scrap)

Pig Iron Scrap Average Loss
per cent, per cent, per cent, per cent.

Carbon 3.75 0.25 2.00 2.00

Silicon 0.70 0.06 0.38- 0.38

Manganese 1.10 0.40 0.75 0.75

Phosphorus 1.30 0.04 0.67 0.67

Sulphur 0.37 0.03 0.20 0.20

20 per cent, slag @ 12£ per

cent. Fe 2.50

6.50
Iron reduced from ore 1 . 50 per cent, gain

Net loss 5 . 00

Recarburizing. — Steel must not be recarburized in the pres-
ence of a basic slag, lest the carbon, silicon, and manganese of the
recarburizer reduce phosphorus from the slag and cause it to pass
back into the metal :

2(3 CaO. P2O&) + 50= 6CaO. P206 + SCO + 2P
4(3 CaO. P2O5) + 5 Si = 2(6 CaO. P2O6) + 5SiO2 + 4 P.

Therefore in basic practice the recarburizer is added to the stream
of metal while it is pouring from the furnace into the ladle, and
special arrangements are made for allowing the slag which floats
on top of the metal to overflow at the top of the ladle and thus be
gotten rid of in a large part. In careful practice ' rephosphoriza-
tion' need not exceed 0.01 to 0.02 per cent, of the steel, although a
much larger increase may take place through accident.

The recarburizer usually consists of ferromanganese, together
with anthracite coal, charcoal or coke which is broken into small
pieces and loaded into paper bags. About 45 per cent, of the
broken coke is burned and the other 55 per cent, will be dissolved

THE OPEN-HEARTH PROCESS

151

by the steel. It is practically impossible to melt spiegeleisen in
the cupola, because a cupola must be run continuously in order to
do satisfactory work, while the open-hearth process is too slow and
too irregular to use a continuous supply of molten recarburizer.

In making soft steel the carbon is reduced to about 0.10 or 0.15
per cent, in the bath, and then the necessary amount of recarburizer
is added. In making high carbon steel, on the other hand, two

.90

.80

Diagram Showing Variations in Composition
of a Normal Acid
Open Hearth Heat

FIG. 108.

THE OPEN-HEARTH PROCESS 153

methods are possible: We may reduce the bath to 0.10 or 0.15 per
cent, and then add sufficient recarburizer to bring the carbon up
to the desired point, or we may bring the bath down to only slightly
below the desired point and recarburize it to the desired percentage.
The former practice usually frees the steel more completely from
gas and therefore makes it less wild, besides reducing the danger
of phosphorus being left in the steel.

ACID OPEN-HEARTH PRACTICE

Acid open-hearth practice is in many respects similar to basic,
but the operations are shorter because: (1) a much larger propor-
tion of steel scrap is used; (2) phosphorus is not removed; and (3),
no fluxes are added, except in rare instances, when a little silica is
charged at the beginning to prevent iron oxide cutting the lining.

Chemistry. — The chemistry of the acid process is much sim-
pler, because neither phosphorus nor sulphur is removed; there-
fore it is necessary to start with pig iron and scrap low in both of
these elements. The progress of the operation is shown in Fig.
108. The manganese in the pig iron for acid work is usually not
so high as it is for basic work: (1) Because this element is some-
times costly; (2) because it increases the amount of slag made;

(3) because it forms a base which requires silica for fluxing it; and

(4) because it increases the waste, since all the manganese burned
represents a loss in weight of metal purchased. Therefore the
manganese, as well as the silicon in the bath, is usually reduced
to only a trace by the time the charge is melted. H. H. Campbell
has shown that the open-hearth slag at the end of the acid opera-
tion automatically adjusts itself to contain about 46 per cent, of
bases (FeO + MnO), the remainder being principally silica, and
that this ratio remains almost the same even when very varying
amounts of iron oxide are added.

Loss. — The loss in the acid process will not be as large as in
the basic, because the pig iron and scrap charged are not so im-
pure, and because the amount of slag made and the amount of iron
oxidized and retained by the slag are not so large. The loss will
vary on an average from 3 to 5 per cent., so that the final metal
will weigh 95 to 97 per cent, of the weight of the charge. The
analysis of this difference is as shown on page 154.

154

THE METALLURGY OF IRON AND STEEL

ANALYSIS OF CHARGE
(50 per cent, pig; 50 per cent, scrap)

Carbon

Pig Iron
per cent.

3 75

Scrap
per cent.

0.25

Average Loss
per cent, per cent.

2.00 2 00

Silicon

.... 0 90

0.10

0.50 0 50

Manganese

0 . 60

0.40

0 . 50 0 . 50

Phosphorus

0 . 035

0.045

0 . 04 0 . 00

Sulphur

0.028

0.032

0.03 0.00

8 per cent, slag (c
Fe. - . .

& 25 per cent.

. 2.00

5.00
Iron reduced from ore = 1 . 20 gain

Net loss 3.80

Recarburiz ng. — Recarburization may take place in the fur-
nace in the acid process, if desired, because there is practically no
phosphorus in the slag to be reduced and absorbed by the metal.
This method has the advantage of the recarburizer being more thor-
oughly mixed with the bath, but it has the disadvantage of more

FIG. 109.

manganese being burned out of the metal after the recarburizer is
added and before the charge is all run into the ladle. The recar-
burizer consists of ferromanganese, broken in pieces not larger

THE OPEN-HEARTH PROCESS 155

than a silver dollar. It is sometimes heated red-hot, but not
always so, because the heat of the metal will readily melt it. As
in the basic practice, recarburization of high carbon steel may be
effected either by burning the bath down to a low point and then
bringing it up again with carbon and ferromanganese, or by redu-
cing it only slightly below the desired amount and then adding
ferromanganese only.

In acid practice, if the carbon is burned to a low point and then
brought back, it is customary to accomplish this by dissolving the
requisite amount of pig iron in the bath before adding the ferro-
manganese. This could not be done in the basic furnace, because
the impurities in the added pig iron would be liable to cause a
rephosphorization of the metal from the slag.

Weight of Slag. — The weight of slag made in the acid process
may be determined, according to the method of H. H. Campbell,1
from the tota amount of manganese in the furnace, which will
include that put in with the charge (including any that may have
been in the ore) and that added with the recarburizing. First we
subtract from this total the amount of manganese in the metal
tapped; the weight of the remainder is then divided by the per-
centage of the manganese in the slag, which gives the weight of the
slag. The amount of acid slag will depend primarily upon the
amount of silicon in the charge and will vary on the average from
6 to 18 per cent, of the weight of the metal charged, about three-
fourths of which is formed during the melting period.2

Coal Burned. — The amount of fuel used per ton of steel made
in the acid open-hearth furnace will be perhaps 50 to 100 Ib. less
than in the basic furnace, but will depend again upon many vary-
ing conditions, so that figures should be used only with caution.
(See pages 144, 167, and 169).

SPECIAL OPEN-HEARTH PROCESSES

It might seem as if the use of molten pig iron direct from the
blast furnace, or through the medium of a mixer, might be as
advantageous in open-hearth practice as in Bessemer; but this is
not so, because the metal in the open-hearth is subjected to oxida-
tion at the same time that it is being melted, so that charging

1 See page 275 of No. 2. Old edition, on page 8 herein.

2 In the case of making structural steel (say 0.18 per cent, carbon).

156 THE METALLURGY OF IRON AND STEEL

molten material does not shorten the operation much. The time
and labor cost for charging is less, but this advantage is partly
neutralized by the scorification and wear of the hearth produced
by the inpouring stream of melted metal. About one-half the
open-hearth steel of America is made from molten pig. Where
it is used the limestone is charged first, on top of that the steel
scrap and any other cold metal, and finally the molten metal is
poured in. In this way the hearth is protected as much as possi-
ble from being cut.

The old pig-and-ore process is abandoned very largely, because
of the length of time required to burn off the impurities unless
they are diluted by steel. It is difficult for a novice to understand
why the reactions are not more rapid in the open-hearth furnace,
when the entire purification of pig iron in the puddling-furnace is
accomplished in an hour and a half, including melting; but the
difference is due to the very shallow bath in the puddling opera-
tion and its extensive contact with the fettling. If we should
attempt to purify under such strong oxidizing conditions in the
open-hearth furnace, the molten metal would boil violently, be-
cause of the high temperature, and for the same reason would also
become so charged with oxygen as to be worthless. Even at the
low temperature of the puddling-furnace, the boiling is so violent
as to increase the height of the bath, and this action would be pro-
portionately increased at the temperature of the open-hearth fur-
nace, which, at the end of an operation producing dead-soft steel,
will be about 1650° to 1700° C. (3002° to 3092° F.) . The increase in
volume of the metal when carbon monoxide gas is escaping from
it may be likened to that of champagne when the drawing of the
cork allows a rapid escape of gas. There is another reason why the
boil causes more of an increase in the volume of the bath in the
open-hearth furnace than in the puddling-furnace: in the latter,
the carbon monoxide has only molten metal and slag to bubble
through, but in the open-hearth process, where cold ore is added to
the charge, it produces a certain amount of chilling of the metal
and slag adjacent to it, and the gas having to bubble through this
somewhat pasty material causes a greater increase in its bulk.
By the time the puddling charge becomes pasty, the carbon is
largely gone and therefore there is not a violent action.

Various attempts have been made by different metallurgists
to adapt the open-hearth process to the use of all pig iron rapidly

THE OPEN-HEARTH PROCESS 157

oxidized by iron ore or other agencies, and this has led to the
Talbot and the Monell processes, each of which is carried on in a
single furnace, molten pig iron being acted upon by a highly
oxidized liquid slag, formed prior to the addition of the pig iron
in the Talbot process and coincident with it in the Monell process.
It has also led to the Duplex process, whereby a large proportion
of the oxidation is effected in an acid Bessemer converter, the
operation being completed in a basic open-hearth.

Talbot Process. — The Talbot process has a basic lining and
contains a charge as high as 200 tons in some cases, as, for ex-
ample, at the Jones & Laughlin plant, in Pittsburg, Pa. As
the bath of metal is over 3 ft. deep, however, which is about twice
that of an ordinary bath, the furnace is only as large in other
dimensions as a 100-ton furnace. The tilting furnace is used in
order that any desired quantity of metal or slag may be poured
out at will. The operation is continuous and the furnace is drained
of metal only once a week. After the charge has been worked
down to the desired percentage of carbon, the great part of the
slag is poured off, and then about one-third (52 tons) of the steel
is poured into the ladle, recarburized, and teemed into the ingot
molds in the usual way. To the charge of metal left in the bath
is now added iron ore and limestone to produce a basic and highly
oxidized slag, and through this slag is then poured an amount of
pig iron equal to the steel removed. The reaction between the
impurities in the pig iron and the iron oxide in the slag is very
vigorous, but does not cause a frothing or foaming, because all the
materials are in the liquid form and the gas bubbles through them
without great difficulty.

The oxidation is so rapid that the silicon and manganese are
said to be oxidized almost immediately, and then the phosphorus
and carbon are worked off in the usual way, using more ore and
limestone, if necessary. The temperature is low at first, in order
that the phosphorus may be more readily burned. At the end of
about four to six hours, the bath has again become purified, and
50 tons of metal are removed, the whole operation being then re-
peated. The yield of steel is 106 to 108 per cent, of the weight of
the pig iron charged, because of the large amount of iron reduced
from the ore by the impurities in the pig iron.

3 C + FeaO3 = 2 Fe + 3 CO (absorbs 108,120 calories).

158 THE METALLURGY OF IRON AND STEEL

The advantages of the process are: We obtain three or four
heats of 50 tons each in 24 hours without the use of steel scrap ; the
yield is large (though this advantage is somewhat neutralized by
the cost of the iron ore used); and the temperature of the final
metal can be more easily controlled. The disadvantages of the
process are : The very large cost of furnaces 1 and the slightly
higher cost for repairs.

Monell Process. — Mr. A. Monell, when metallurgist of the
Carnegie Steel Company, developed a pig-and-ore process in which
highly heated oxidizing and slag-making materials were made to
react with the impurities in molten pig iron without the necessity
of having a reservoir of metal into which to pour it. Upon a
basic hearth he heats limestone and a relatively large amount
of ore until they begin to melt, and then pours molten pig iron,
equivalent to the capacity of the furnace, upon it. The tem-
perature of the bath is necessarily low, since pig iron direct from
a blast furnace or from a mixer will not be more than 200° or
300° C. above its melting-point, and therefore the phosphorus
will be oxidized very rapidly. The slag foams up and pours
out of a slag-notch that is provided for the purpose, and in an hour
the bath is practically free from phosphorus, silicon, and man-
ganese, and most of the slag is removed. The operation is then
continued in the usual way to eliminate the carbon, and the
metal is tapped when this has been reduced to the desired point.
The American rights to the process are owned by the Carnegie
Steel Company and they are operating it at many of their fur-
naces. No details are known generally, but it is to be presumed
that the results are favorable. The apparent disadvantages of
the process are excessive cutting of the hearth and a heavy loss
of iron in the rich slag which flows off at the beginning of the opera-
tion. The Monell process has been used successfully in England,
with pig iron containing up to 2 per cent, of phosphorus.

Duplex Process. — At Witkowitz, Austria, at Ensley, Alabama,
at Monterey, Mexico, at Sidney, Nova Scotia, and at Pueblo,
Colorado, the combined Bessemer and basic open-hearth process
is in operation, an acid converter being used to oxidize the silicon,
manganese, and most of the carbon, while the phosphorus and the

1 It is stated that the first 200-ton furnace at the Jones & Laughlin works
cost $1,000,000 to build; but with the present experience they can be installed
for about one-fourth of that sum.

THE OPEN-HEARTH PROCESS 159

remainder of the carbon is eliminated in a basic open-hearth fur-
nace. In the different localities there are different ways of carry-
ing out this combination, but these divide themselves into two
general methods: In one method, the metal is blown in the con-
verter until it is purified to the point where it is practically equiva-
lent to so much high-phosphorus, molten steel scrap, which is then
mixed with either liquid or solid pig iron in the open-hearth furnace
and worked as any ordinary pig-and-scrap heat after melting. In
another, and more common, method, the pig iron is blown in the
converter until it contains about 1 per cent, or so of carbon, and
this product, with little or no additional pig iron, is then dephos-
phorized and completely decarbonized in the open-hearth furnace.
It is stated that the total time of the purification is less than if the
ordinary basic open-hearth process were used (this is 4 to 7 hours
in Alabama, while the open-hearth part of the process is about 3
hours at Witkowitz), and that the total loss isjonjy 10 per cent.

At Alabama, the metal from two 20-ton converters is poured
into a mixer furnace of 250 tons capacity, which supplies four 100-
ton basic open-hearth furnaces writh metal. The blown metal con-
tains about 1 per cent, of carbon, about one-half the heats being
blown 'full/ while the remainder are blown to about 1.75 to 2.50
per cent, carbon. In this connection it is interesting to note that,
if the metal is blown to about 1.75 per cent, or so of carbon, there is
less loss of iron as shots in the slag than if the carbon is higher.1

Processes in Two Open-Hearth Furnaces. — At a low tempera-
ture, phosphorus is very easily oxidized and absorbed by a basic
slag, even in the presence of carbon, but when the heat is high the
oxidation of phosphorus is hindered by the carbon, for the reason
already explained, — that the affinity of carbon for oxygen in-
creases more rapidly with the temperature than the affinity of the
other elements in the bath. This explains the rapid elimination
of phosphorus in the puddling process, where the slag is strongly
basic with oxide of iron and the temperature is low. We could
obtain the same conditions in the beginning of the open-hearth
process, but the operation would be extremely slow at this low
heat, and the carbon would pass away slowly. These difficulties
have been met by the Campbell No. 2 and Bertrand-Thiel proc-
esses, the former of which was developed at Steelton, Pa., and the
latter at Kladno, Bohemia.

1 For this information I am indebted to Mr. Hugh P. Tiemann.

160 THE METALLURGY OF IRON AND STEEL

Campbell No. 1 Process. — The pig-and-ore process using
molten metal has long been in operation in the Campbell tilting
furnace, and the frothing of the bath is taken care of by tipping the
furnace backward so that no slag or metal will pour out of the door,
though a large amount of the former flows from the slag-hole be-
tween the ends of the furnace and the ports. The operation is
continued in this way for two or three hours, since, as already
noted, the furnace can be tipped without cutting off the supply of
gas and air, and the yield of steel is 104 to 106 per cent, of the pig
iron charged.

Campbell No. 2 Process. — At the same plant there is also a
combination process in which the charge, consisting of all pig iron,
or of pig iron1 and scrap, is placed in a basic open-hearth furnace,
and the purification carried on at a high temperature until almost
all the silicon and phosphorus and part of the sulphur and carbon
are eliminated. The bath is then tapped from the basic furnace
and poured into an acid-lined furnace, care being taken that none
of the basic slag goes with it. At this period the metal is low in
phosphorus and sulphur, and contains about the same amount of
carbon that a cold charge would have contained as soon as melted.
The conditions are therefore the same as if low-phosphorus low-
sulphur material had been charged into an acid furnace and
melted there, and the process is now continued at a higher tem-
perature, in the usual way to make acid open-hearth steel. The
disadvantage of this process is that the transferring of molten
metal from one furnace to another is not an easy matter, nor,
in fact, is it possible with the arrangements in many plants.

OPEN-HEARTH FUELS

The commonest fuel used in the open-hearth furnace is pro-
ducer gas, because more heating power and more gas can be ob-
tained for a dollar in this variety than in any other, except in
those favored localities where natural gas is found.

Producer Gas. — If air be blown through red-hot carbon the
following reaction takes place:

C + 20 = C02;

but if the bed of fuel is deep, the carbon dioxide enters into a
further reaction, as follows :

CO2 + C = 2 CO.
1 The pig iron charged may be either in the liquid or solid state.

THE OPEN-HEARTH PROCESS

161

In other words, if air be made to blow through a deep bed of red-
hot carbon, there will be produced carbon monoxide gas which has
combustible value:

CO + O = CO2 (generates 68,040 calories).

Producer gas for open-hearth furnaces is usually made from
bituminous coal, because the hydrocarbons contained in this coal
enter into the gas and thus give it illuminating power, which
makes it much more efficient in
the furnace, because the heating
takes place by radiation chiefly.
Such a gas will contain 3 to 5
per cent, of hydrocarbons, 20 to
25 per cent, of CO, 55 to 60 per
cent, of nitrogen, and 2 to 8 per
cent, of carbon dioxide.

The two latter components
produce no heat and are there-
fore worse than useless, because
they carry heat away from the
furnace up the chimney stack.
The nitrogen comes from the air,
of course, but the CC>2 is theoreti-
cally absent. Its presence is due
to irregularities in the gas pro-
ducer operation, such as vertical
channels forming in the bed of
fuel, up which the CO2 gas passes
without being brought into con-
tact with carbon; or the rapid
passage of the gas does not per-
mit time for the reactions to be
completed; or irregularities in the fuel bed, whereby the fuel will
be red-hot much higher on one side than on the other.

The air is usually blown through the fuel by means of a steam-
jet, which results in a certain amount of steam passing into the
producer with the air; but this is rather an advantage than other-
wise, as the steam is decomposed by the red-hot carbon and en-
riches the producer gas :

H2O + C = 2H + CO;
(2H + O = H2O: generates 58,060 calories).

FIG. 110.— TAYLOR REVOLVING
BOTTOM GAS PRODUCER.

162 THE METALLURGY OF IRON AND STEEL

Gas Producers. — The gas producers are the furnaces in which
the fuel is contained while the air is passed through it. The main
objects to be accomplished are: (1) To pass the air uniformly
through the bed; (2) to remove ashes and charge fresh fuel with-
out interrupting the production of gas; and (3) to preserve the
deep bed of incandescent carbon, having level upper and lower
surfaces. There are three horizontal zones in the gas producer:
The first is the ash zone, which is deep in order that the air may be
slightly preheated in passing through it and that any unburned
carbon which gets into it may have a strong liability of being
burned. Next above that is the C(>2 zone, in which the oxygen
and carbons are first combined; and above that the CO zone, in
which the CO2 is reduced by more carbon. The top of this zone
should be at a dull-red heat.

There are many different forms of producer, which are exten-
sively used for open-hearth work, but these may be divided into
two general types : In- the first or water-sealed type, the bottom
of the producer dips into a pool of water and thus the tools may
be introduced for the removal of the ashes at will. In this type
there are sometimes steel arms extending into the bed of fuel,
either from the top or from a central shaft, by the rotation of which
the bed is polled,, lumps and channels are broken up, etc. The
second type has; a mechanical grate, by which the ashes can be
scraped down into the chamber underneath without interrupting
the producer operation for the purpose.

Grate Area. — The total grate area of all the producers supply-
ing gas to a furnace should be about 3.5 sq. ft. per ton of furnace
capacity; some producer plants run even higher than this, and up
to 6.25 sq. ft. Another method of figuring the grate area is that
1 sq. ft. should be supplied for every 7.5 to 12.5 Ib. of coal burned
per hour, although with expert gas makers and good coal the com-
bustion may be much greater than this, and higher values (to 22
Ib.) are claimed by the makers of the gas producers.

Volume and Calorific Power. — The volume of producer gas
obtained from a ton of coal will be about 150,000 to 170,000 cu. ft.,
having a calorific power of 33 to 36 Calories per cu. ft., or 130 to
145 B.t.u. per cu. ft.1 These figures will, of course, depend
upon the quality of the coal gasified, but the calorific power is no

1 For calculation of this relation see page 171.

FIG. 111. — MORGAN WATER-SEALED GAS PRODUCER.

164

THE METALLURGY OF IRON AND STEEL

more important than the amount of heat that it will radiate, which
depends upon the luminosity of the flame.

Luminosity. — The luminosity of flames depends upon the
amount of hydrocarbons, and especially of heavy hydrocarbons,

FIG. 112. — HUGHES MECHANICALLY POKED GAS PRODUCER.

burned to produce them. It is therefore necessary, if the pro-
ducer gas is made from bituminous coal low in hydrocarbons or
from coke or anthracite, to increase its illuminating power by

THE OPEN-HEARTH PROCESS 165

spraying oil into it. The luminosity is produced by the deposition
of a myriad of tiny particles of carbon, which are heated to incan-
descence and then radiate energy in the form of light. It is prob-
able that this action is produced by the relatively light hydrocar-
bons, such as methane (CH4), breaking up first into ethylene
(C2H4) , and then into acetylene (C2H2) , which deposits the carbon
particles or soot. It is for this reason that the pure acetylene
flame has such intense luminosity.

FIG 113. — BUTTERFLY REVERSING VALVE.
(See also Fig. 32, page 56.)

Gas Mains. — The gas mains leading from the producer plant
to the open-hearth furnace should be lined with brick and be at
least large enough for a man to pass through. Beyond this, a
good rule is 1 sq. ft. of area of cross-section of gas main for every
8 sq. ft. of total combined area of gas-producer grates. The gas
loses heat by radiation in the mains and deposits the tarry constit-
uents, i.e., the nearly solid hydrocarbons, in both ways losing
heating power.

166

THE METALLURGY OF IRON AND STEEL

Valves. — The reversing valves of the open-hearth furnace are a
cause of large loss in producer gas, and sometimes the leak amounts
to 10 or 20 per cent. Ordinary leaks may be prevented by having
water-sealed valves, but sometimes the pressure of gas reaches the
point where it overcomes the water pressure and escapes, causing
a heavy loss. Moreover, water-sealed valves are open to serious
objections: (1) The water may freeze, causing an endless amount
of annoyance and trouble; (2) some water on the inside of the
valve is vaporized and the vapor carried into the regenerator or
into the furnace, where it absorbs heat; (3) the hoods are liable to
warp with the heat; and (4) valves of this type are so heavy that
elaborate mechanism is necessary to reverse them. The common
butterfly valves burn out rapidly and warp badly. If water-

FIG. 114.

sealed, the hoods warp and leak. Lining the hoods with brick
makes additional weight for shifting and adds very largely to the
repairs. The mushroom type requires eight valves to a furnace
and an elaborate arrangement for reversing them properly. They
burn out badly unless water-cooled, when difficulty is met with
from freezing.

Natural Gas. — In those districts, like Pittsburg, where natural
gas occurs, it is a great boon to the open-hearth steel industry,
because of its high calorific power and the cheapness with which it
may be obtained, and about 80 wrought-iron plants and 90 steel
plants in America use it. It is drawn from the earth, and has a
calorific power of 970 to 1010 B.t.u. per cu. ft. (equivalent to 225
to 250 Calories per cu. ft. or 8600 to 9000 Calories per cu. meter).1

1 The calculation of this relation will be found on page 171.

THE OPEN-HEARTH PROCESS

167

In the Pittsburg district the amount of natural gas used per ton
of steel made ranges from 4000 to 11,000 cu. ft., averaging about

FIG. 115. — MUSHROOM REVERSING VALVE.

~ -Furnace Flue ' Chimney Flue

FIG. 116. —WATER-SEALED REVERSING VALVE.

5500. This includes the gas used in the furnace and the small
amount necessary for heating the ladles. The natural gas is usu-

168

THE METALLURGY OF IRON AND STEEL

ally introduced through two pipes at each end of the furnace,
which are located close to the bottom of the gas-ports and deliver
the fuel about 3J feet back from the hearth. The natural gas is
therefore applied without any loss in gas mains, reversing valves,
regenerators, etc., and is cheaper in labor on the furnace floor
itself. The Pittsburg natural gas averages about 70 to 90 per
cent, of methane (CH4) or marsh gas, the remainder being hydro-
gen with a fraction of a per cent, of carbon dioxide and a very few
per cent, of nitrogen.

On account of its high calorific power, it is not necessary to pre-
heat this gas, and this is the more fortunate, because the amount of

Detail of Washer

Rubber Hoee /

Connects-Here

•plG. 117. — OIL BLOW-PIPE FOR OPEN-HEARTH FURNACE.

hydrocarbon is so great that if the gas be passed through regenera-
tive chambers, it decomposes and deposits soot on the checker-
work. All the open-hearth furnaces in the Pittsburg district are
so arranged that they can be put on producer gas in case it should
become necessary, because the supply of natural gas has been de-
creasing for many years. In the early part of 1908, some furnace
plants of a very important company in this district began the use
of the manufactured gas, because of the increased cost of natural
gas.

Oil. — On the continent of Europe, and in parts of the United
States distant from the natural gas and bituminous coal regions,
many open-hearth furnaces are heated by petroleum. The

THE OPEN-HEARTH PROCESS

169

United States has many deposits of fuel oil, besides which it is
sometimes possible to obtain a refuse from the oil refineries which
is excellent for this purpose. There are therefore many different
grades employed, but they will usually average from 7.8 to 8.3 Ib.
per gallon, with a calorific power of 14,000 to 17,000 B.t.u. per
pound. H. H. Campbell l states that a rough comparison may be
made by assuming that 50 gallons of oil will give the same amount
of heat as about 1000 pounds of soft coal, and he has had a valu-
able amount of experience with this kind of fuel. It would seem,
however, as if this value for oil was somewhat high for safety in

Foreplate- Level

Charging Floor

FIG. 118. — FURNACE ARRANGED TO USE OIL BLOW-PIPE.

making calculations, and that a more conservative estimate would
be to say that from 35 to 60 gal. of oil would be required per ton of
steel treated in the open-hearth furnace. Eight furnaces using oil
averaged 38 to 42 gal. per ton of steel made.

The crude petroleum is vaporized by atomizing it with a jet of
steam or compressed air, and it is not common practice in the
United States to pass this vapor through a regenerative chamber.
The usual method of application is by a blow-pipe introduced
through the brickwork at the end" of the furnace, as shown in
Fig. 118, and a special form of blow-pipe is now on the market for
this purpose. It is necessary to pump the oil to the blow-pipe or

1 See page 247 of No. 2. Old edition, on page 8 herein.

170 THE METALLURGY OF IRON AND STEEL

else to store it in an overhead tank, from which gravity will carry
it, but the labor in connection with this is much less than the labor
on gas producers. In the rare cases when the oil vapor is pre-
heated, it must be introduced into the hot part of the regenerative
chamber, because if it cools it condenses. Moreover, when intro-
duced into a cold furnace or with a cold charge in the furnace,
combustion will be retarded.

Whether oil is used or not will depend principally upon freight,
because it may be transported much more cheaply than any other
form of fuel. It gives a longer flame than either natural or pro-
ducer gas, and one very great advantage of using it is a saving of
the roof of the furnace; the oil flame may be directed so accurately
by means of the blow-pipe that it does not impinge directly on the
roof, and the brickwork therefore lasts very much longer. On the
other hand, it spreads out horizontally and causes a greater wear
on the front and back walls of the furnace. It gives a more uni-
form heat, a more oxidizing flame, and no danger of losses or dif-
ficulties, in case there is a leak in the walls between the gas and
air regenerators, which is not an infrequent occurrence with gas.

Water-Gas. — If steam be made to pass through a bed of red-
hot carbon, the product is a gas containing slightly less than 50
per cent, each of hydrogen and carbon monoxide and having a
high calorific power:

C + H2O = CO + H2 (absorbs 28,900 calories).

The result of this reaction is a reduction of the temperature of
the fuel bed, which is rapidly reduced until another reaction begins
to take place:

C + 2H2O = CQs + 4H (absorbs 18,920 calories).

Therefore some means must be employed to raise the tempera-
ture at intervals, and this is ordinarily accomplished by inter-
rupting the passage of steam and passing air through the fuel bed,
which raises the temperature and at the same time forms producer
gas which is used for other purposes. Usually, it is necessary to
blow air through for 12 to 15 minutes, and then steam for 4 or
5 minutes. Consequently, the manufacture of water-gas is
intermittent and the method is not very satisfactory for open-
hearth work, although the gas is used to some extent on account
of its high calorific power. This operation produces roughly
;35,000 cu. ft. of water-gas and 80,000 cu. ft. of producer gas per

THE OPEN-HEARTH PROCESS

171

ton of coal. Water-gas has about 2600 Cals. per cu. meter, or,
say, 290 to 300 B.t.u. per cubic foot.1

Dellwik-Fleischer System. — The Dellwik-Fleischer system is
a modification of the ordinary water-gas system, in that the amount
of air blown through for heating up the fuel is so very large that
carbon dioxide is produced instead of carbon monoxide. This

Air

FIG. 119. — WATER-GAS PRODUCER.

gas is therefore altogether wasted, but the formation of carbon
dioxide generates so much more heat that the blowing up does
not take so long, and usually lasts from 1£ to 2 minutes, after
which water-gas is made for 8 to 12 minutes. In this way about

1 One cubic meter = 35.3 cubic feet.
One Calorie = 3 .968 B. t. u. (See page 466.)
One Calorie per cubic meter = 8 . 9 B. t. u. per cubic foot.

172 THE METALLURGY OF IRON AND STEEL

80 per cent, of the calorific value of the fuel is converted into
water-gas, and several steel-works in Europe have adopted the
method.

Mond Gas. — In the Mond process a mixture of water-gas and
producer gas is made continuously. For every ton of fuel burned
there is forced into the producer about 3 tons of air and 2.5 tons
of steam, the latter being produced by absorbing the sensible heat
of the gas in the boiler. It gives about 150,000 cu. ft. of gas per
ton of fuel burned, containing about 25 per cent, of hydrogen,
12£ per cent, of CO, 45 to 50 per cent, of nitrogen, and 12£ per cent,
of CO,. This gives a slightly higher calorific power than ordinary
producer gas, and is used in some steel-works.

REFERENCES ON THE OPEN-HEARTH PROCESS

61. M. A. Pavlov. "Album of Drawings Relating to the Manu-

facture of Open-Hearth Steel." St. Petersburg, 1908.
(See also page 93.)

62. W. M. Carr. "Open-Hearth Steel Castings." 1907. This

contains a concise, simple and readily intelligible discussion
of the acid and basic open-hearth processes and practice.

VII
DEFECTS IN INGOTS AND OTHER CASTINGS

BESIDES the dangers already mentioned in connection with
improper methods of manufacture, excessive amounts of impuri-
ties, etc., iron or steel may suffer from damage caused or developed
during casting. The commonest defects which may appear at this
time are: (1) Blow-holes, or gas bubbles enclosed in the body of
the metal; (2) a pipe, or shrinkage cavity; (3) ingotism, or the
formation of large-sized crystals; (4) segregation, or the concentra-
tion of impurities in localities; (5) checking or cracking of the cast-
ing because of strains produced when the metal is hot and tender.
Avoiding these defects cannot make bad steel good, but their pres-
ence may make good steel bad, and therefore to guard against
them is an important part of the processes.

Blow-holes. — Blow-holes are especially liable to occur in steel,
particularly in low-carbon steel. When the metal is in a molten
state, it readily dissolves certain gases, such as hydrogen, nitrogen,
oxygen. Upon solidification these gases come out of their state
of solution, but may become entangled in the steel and cause a gas
bubble or cavity varying all the way in size from microscopic
proportions up to an inch or more in length. The formation of
these blow-holes is precisely similar to the formation of air bubbles
in ice: water dissolves a good deal of air while in the liquid state
and, as we all know, it is well-nigh impossible to freeze the water
without obtaining a great many air bubbles in the ice, due to the
separation of this air during freezing. The danger in the case of
steel is not so great as in the case of ice, and it is by no means im-
possible to obtain steel absolutely free from this defect. Appar-
ently, the reason for this difference is that the gas separates from
steel a short time before solidification is complete, and thus the
bubbles have some opportunity to escape before they are enclosed
in the solid. Also, steel passes through a pasty stage during solid-
ification, as we shall learn later, and therefore gives a better
opportunity for the gas bubbles to pass away.

173

174 THE METALLURGY OF IRON AND STEEL

Another cause of blow-holes in steel is undoubtedly the presence
of oxide of iron in the metal. This oxide of iron reacts with the
carbon added in the recarburizer and forms carbon monoxide gas
(CO), which may be produced during the entire solidification pe-
riod and thus cause many blow-holes.

Prevention of Blow-holes. — That oxide of iron is one of the
chief causes of blow-holes is shown by many things; for instance,

(1) steel known to be highly oxidized is very liable to blow-holes;

(2) cast iron, which from its chemical composition can never be
much oxidized,1 is never subject to blow-holes; and (3) the addi-
tion to steel of deoxidizers prevents the formation of blow-holes.

Chief among the deoxidizing elements which are added for this
purpose are manganese, silicon, and aluminum. These elements
seem to act partly by deoxidizing the iron and carbon, in both
ways preventing the formation of carbon monoxide, and partly by
increasing the solvent power of the solid metal for gases, so that a
less amount separates. The amount of these deoxidizing sub-
stances necessary to be added will depend largely upon the extent
to which we desire to prevent the formation of blow-holes. In the
case of steel castings it is often necessary that blow-holes be en-
tirely prevented, but in the case of ingots which are to be subse-
quently forged or rolled it is not necessary that blow-holes should
be absent altogether, because they will be closed up under the
pressure of the mechanical wrork and their sides welded together.
Indeed, their presence is sometimes desired, because when they
separate from the steel they occupy space, thereby counteracting
to a certain extent the shrinkage of the metal during solidification
and tending to reduce the volume of the shrinkage cavity or pipe.
For this reason a small number in some harmless locality is often
intentionally allowed to form in steel ingots. Mr. Brinell found in
his steel- works that if the percentage of manganese plus 5.2 times
the percentage of silicon is equal to 2.05 or more, the steel will
be entirely free from blow-holes. In this case, however, the pipe
will be large. If this sum is equal to 1.66, the steel will contain a
harmless number of minute blow-holes, but the pipe will be small.

JThat cast iron is sometimes partially oxidized is claimed by several
eminent authorities, and the evidence presented makes us hesitate to deny that
a certain variety of wild cast iron owes its peculiar behavior to the presence
of some partially oxidized constituent (perhaps the oxysulphide of iron, as
suggested by J. E. Johnson, Jr.).

DEFECTS IN INGOTS AND OTHER CASTINGS

175

This figure is therefore about the correct amount for ingots under
conditions similar to those of Mr. Brineirs experiments, and not
far different in any event. Mr. Brinell also found that the addi-
tion to the steel of 0.0184 per cent, of aluminum will give ap-
proximately the same result as that given by the amount of man-
ganese and silicon last mentioned, 1.66.

Location of Blow-holes. — The number and size of blow-holes is
no more important, however, than the position they occupy in in-
gots in relation to the external surface. Even in castings, blow-
holes, if present, should be deep-seated, as they are then less liable
to be exposed by machine work performed on the surface. In the
case of ingots the deep-seating is of still greater importance, be-
cause then the blow-holes may be closed up before they have an
opportunity to break through to the surface and thus become
oxidized on their interior. The normal gases in blow-holes are
reducing in effect, and thus the interior surfaces of the holes are
bright and silvery in appearance and readily weld together ; but if
they become oxidized they will never adhere firmly. For instance,
a blow-hole near the surface, as in Figs. 120 and 122, is liable

FIG. 120. — SKIN
BLOW-HOLES.

FIG. 121. — DEEP-SEATED
BLOW-HOLES.

FIG. 122.

to break through to the exterior when the ingot is put under pres-
sure. This not only causes a crack in the steel but allows the air
to oxidize the interior of the hole and thus prevent the crack being
welded up by the rolling. It is not at all uncommon to see a
number of these openings form during rolling when the blow-
holes are near the surface.

As the percentage of manganese plus 5.2 times the percentage
of silicon decreases from 1.66, the blow-holes become correspond-
ingly deeper-seated. Finally when this sum becomes as low as
0.28, the blow-holes are harmlessly located in the interior. It is

176 THE METALLURGY OF IRON AND STEEL

usually impracticable, however, to get the manganese and silicon
as low as this in steel, because manganese is needed to counteract
the bad effect of sulphur and oxygen.

The location of the blow-holes is also very largely dependent
upon the fluidity of the metal when first cast into the molds. The
more fluid it is, other things being equal, the nearer will the blow-
holes be to the surface of the solid ingot. On the other hand, if
the casting temperature is too low there will be a dangerously large
number of blow-holes in the steel (see Fig. 122), because it solidifies
so quickly that very little opportunity is afforded for any part of
the dissolved gases to escape. The fluidity of the steel depends
partly upon its temperature and partly upon the amount of im-
purities in it. For instance, pig iron is fluid at a temperature at
which steel is solid; high-carbon steel is fluid at a lower temper-
ature than low-carbon steel. Therefore every different kind of
steel has a different correct casting temperature; but we have al-
ready learned (p. 107) how to determine this by means of the skull
left in the casting-ladle, and it is evident that this test applies
equally well to all grades of steel. It is to be observed that low-
carbon steel suffers greatly from blow-holes, because the more the
carbon the less oxidized will be the steel.

Pipes. — When steel is poured into a mold, it forms almost im-
mediately a thin skin of frozen metal against the cold surface of
the sand or iron. The radiation of heat thereafter necessarily
takes place through these surfaces, and therefore a casting will
usually complete its solidification by the formation of thicker and
thicker layers of solid metal around all the sides. The top, how-
ever, will usually remain molten longer than the rest because the
hottest metal is usually at this point, having been the last to leave
the ladle, and also because the heat is not conducted away by the
air as fast as by the walls of the mold. This is especially true where
the casting is poured into an iron mold — for example, in the case
of ingots (see Fig. 123). But it is evident that at some period a
stage will be reached when all the outside of the ingot, or casting,
will be covered by a skin of solid metal while the interior will still
be liquid. The liquid interior will continue to freeze and will, at
the same time, contract. The result will be the shrinking of the
molten mass away from the solid walls and consequently the
formation of a cavity, known as a ' pipe/ in the interior. This pipe
will be filled with the gases evolved by the steel during solidification.

DEFECTS IN INGOTS AND OTHER CASTINGS

177

Professor Howe has shown that the volume of the pipe is too large
to be accounted for altogether by the shrinkage of the steel during
solidification, and has shown that the rate of contraction of the
inner walls of the ingot being greater than the rate of contraction
of the outer walls, a virtual expansion of the outer walls is caused
and a consequent enlargement of the pipe.1

The portion of the steel containing the pipe is of course defective
and should be discarded at some time subsequent to casting. In
the casting of ingots the upper part, which contains the pipe, is

^/^%

= |

M^i

Outside, J,
Solidifies

Outside
gets Cool

Pipe
Begins

Pipe

Increases

FIG. 123. — SOLIDIFICATION OF AN INGOT.

cut off during the rolling or forging and goes back to the furnace to
be remelted as scrap. In the casting of steel castings there is a
large adjunct to the castings situated above it, and so regulated in
size and otherwise that it freezes after the casting itself, and thus
always contains a supply of molten metal which runs down and fills
any cavity that forms in the casting. This adjunct is cut off when
the casting has cooled. In other words, the 'riser' or 'feeder/ as
this extra part is called, serves the same purpose for a steel casting
as the upper part of an ingot does for the ingot.

Cast-iron castings do not form a pipe under ordinary circum-
stances, because cast iron expands during solidification on account
of the separation of graphite, as we shall learn later. Under cer-
tain circumstances, however, there may be enough difference in
expansion between the inside and outside of cast-iron castings to
produce a porous spot which, while not exactly a pipe, is due to
similar causes. We shall discuss this matter further in Chapter
XII.

1 No. 71, page 183.

178 THE METALLURGY OF IRON AND STEEL

• Lessening the Volume of the Pipe. — If the steel were poured
into the mold so extremely slowly that it would solidify in layers
from the bottom upward there would be no pipe. Therefore one
method of lessening the volume of the pipe is by slow casting.
We have already noted another method, namely, permitting a small
number of blow-holes to form, which causes a certain amount of ex-
pansion of the steel during solidification and thus diminishes its
shrinkage. Another way is to use wide ingots, because this reduces
the difference in contraction between the inner and outer layers of
the ingot, which, as I have already stated, caused a virtual expan-
sion of the outer walls and thus enlarged the cavity. Casting in
sand molds has the same effect, because radiation is not so rapid
through sand as through metal. Still another method is to pre-
vent the steel forming a solid skin over the top by constantly
stirring and breaking it up with an iron rod. This method is
often resorted to with the risers of steel castings, with very
beneficial results.

Bottom Casting. — In the case of steel castings, and less fre-
quently in the case of ingots, the metal is poured from the ladle
into a runner which delivers it at the bottom of the casting (see
Fig. 193, p. 241). With steel castings this is often necessary in
order to prevent dirt getting into the casting. It also has a similar
effect on ingots, because it prevents slag getting into the molds
and also prevents metal from spattering up on the side of the mold
and forming what is known as a 'cold-shut/ that is, a part of the
metal which is not melted in with the rest. In both cases, how-
ever, this bottom casting has the effect of increasing the volume
of the pipe and also of making the pipe extend deeper, because at
the end of casting the hottest metal is at the bottom instead of at
the top.

Casting with the Large End Up. — Risers on castings are al-
most always made with the top end larger than the bottom, in
order that the pipe may be less in volume and shorter in depth. At
steel- works, however, the ingot molds are always tapered slightly
with the large end at the bottom, in order that the mold may be
easily drawn off the top (see p. 109) . This results in the large end
of the ingot being down, and consequently in the pipe being larger
in volume and very much greater in depth.1 Because of this
advantage Professor Howe has proposed certain mechanical ar-

1 No. 72, page 184.

DEFECTS IN INGOTS AND OTHER CASTINGS

179

rangements by which the ingot may be cast with the large end
upward.1

Liquid Compression of Ingots. — If the pipe is caused by the
difference in expansion between the inside and the outside of an
ingot, it is evident that putting sufficient pressure upon the outside
when the walls are solid but the interior is still liquid will prevent
the formation of a pipe. Numerous processes have been devised
for effecting this liquid 'compression/ some of which are in opera-
tion at steel-works and produce ingots that are entirely free from
pipes. In Whitworth's system the ingot is raised and compressed
lengthwise against a solid ram situated above it, during and shortly
after solidification.2 In Harmet's method the ingot is forced up-
ward during solidification into its tapered mold.2 This causes a
large radial pressure on its sides. In Lilienberg's method the
ingots are stripped and then run on their cars between a solid and
movable wall. The movable wall is then pressed against one side
of the ingots.3

Ingotism. — When iron and steel freeze they crystallize, and
these crystals grow with great rapidity, so that if the passage
through the solidification period is slow they will attain a very

FIG. 124. — INGOTISM.

large size. This formation of large crystals is known as 'ingotism/
It is especially liable to occur if the metal is cast at too high a
temperature or is allowed to cool through the solidification period
1 See No. 71, page 183. 2 See page 373 of No. 1, page 8. * See No. 73, page 184.

180 THE METALLURGY OF IRON AND STEEL

at too slow a rate.1 In the case of steel, ingotism may be detected
by breaking the casting, when the large size of the crystal faces or
facets may be observed.

Damage Due to Ingotism. — Large crystals always produce
weakness and loss of ductility, for the large crystals do not adhere to
one another as firmly as when there is a more intimate association ;
consequently steel that shows ingotism will be tender. In the case
of steel castings, they will not give as good a result in the testing
machine; in the case of ingots they will be liable to tear during roll-
ing or under the hammer.

Remedy for Ingotism. — Ingots in which large crystals have
formed during solidification may be brought to a high degree -of
strength and ductility by forging or rolling, because the mechanical
work crushes the crystals and reduces them to a smaller size. The
work must be done very carefully at first, however, or cracks will
be formed that are not afterward welded up. Ingotism in steel
castings is not so easy to cure; indeed, it is maintained by some
authorities that its bad effects are never entirely obliterated. I
am inclined to agree with this opinion, although annealing the steel
at a proper temperature (see Chapter XIV) will produce a very
beneficial effect.

Segregation. — When either iron or steel is molten, the various
impurities are dissolved in it, and some of them, especially carbon,
phosphorus, and sulphur, make the metal more fusible, that is, they
lower its melting-point. But the impurities are not as soluble in
the solid metal, and therefore tend to separate on solidification; so
it can readily be conceived how each layer that freezes, beginning
at the outside, rejects some of its impurities to be dissolved by
the still liquid mass in the interior. When the next layer freezes
that too will tend to reject a part of its impurities into the contigu-
ous molten layer, and thus the concentration will proceed so that as
a general thing the portion of the metal richest in impurities, es-
pecially in carbon, phosphorus, and sulphur, will be that which
freezes last. With ingots, this will evidently be at a point just be-
low the bottom of the pipe, and it is found to be so in the great
majority of cases; but the location of the richest segregate is very

*In the case of cast iron, large crystals formed during solidification pro-
duce what is known as an ' open grain ' ; we shall consider this more particularly
in Chapter XII. The name 'ingotism' is not usually applied to this open
grain in cast iron.

DEFECTS IN INGOTS AND OTHER CASTINGS

181

liable to vary, and rules can only be used for general guidance.
For example, in Fig. 125, l the most impure metal is found at a
point higher than the bottom of the pipe; and other unexpected

exceptions occur. In the case of iron
and steel castings, the most impure point
will generally be near the top of the
thickest section of metal. The riser is
calculated to be the last portion to
freeze and the richest segregate should
be located in it.

In iron castings which contained on
an average less than 1 per cent, phos-

— B

Top in Ingot

o ® o

.25 .20 .24 -26 .20

O O O O O

.24 .25 .22 .24 .24

O O O O O

.23 .23 .22 .23 .24

O O O • ®

FIG. 125. — LINES OF
EQUAL CARBON-
PERCENTAGE IN
A STEEL INGOT.

FIG. 126. — CARBON-PENCENTAGE AT
DIFFERENT PARTS OF A STEEL
INGOT.

phorus and 0.1 per cent, sulphur, I found on one occasion a

segregated portion containing 1.856 per cent, phosphorus and

1 From page 205 of No. 71, page 183.

182

THE METALLURGY OF IRON AND STEEL

0.144 per cent, sulphur; and on another occasion I found one
containing 2.43 per cent, phosphorus and 0.236 per cent, sulphur.
An extreme case of segregation in steel is shown in the following
analysis : l

Carbon
Per cent.

Silicon
Per cent.

Manga-
nese
Per cent.

Phosphorus
Per cent.

Sulphur
Per cent.

Average

0 24

0 336

0 97

0.089

0.07

Segregate

1 27

0 41

1.08

0.753

0.418

Treatment of Segregated Steel. — Segregation cannot be pre-
vented, although, of course, it seldom takes place to the degree
shown in the extreme cases that I have cited above. Nevertheless,
there are always certain portions of the ingot or casting which are
richer in impurities than others. An attempt is made to get this
richer portion into the upper part of an ingot, or into the riser of a
casting, and then it is cut off when the ingot is rolled, or when the
riser of the casting is removed. It is therefore advantageous to
cause the segregate to go as high up in the ingot or casting as pos-
sible. Whatever tends to raise the whole pipe higher up in the cast-
ing would, in general, tend to raise the segregate also; but wide
ingots, or ingots cast in walls with low conducting power, though
they tend to decrease the volume of the pipe, would not necessarily
raise the segregate to a higher point. Furthermore, wide ingots
will probably have a much greater degree of segregation than nar-
row ingots, other things being equal, because the wider the ingot
the greater will be the number of layers of solidification, and con-
sequently the greater concentration of impurities in the center.

Lessening Segregation. — Benjamin Talbot 2 has shown that
quieting the steel by adding aluminum to it will lessen the segre-
gation. J. E. Stead 3 argues that this result is due to the branches
of crystals (commonly called 'fir-tree crystals'), which grow per-
pendicularly to the cooling surfaces when steel solidifies and me-
chanically entangle some of the impure metal, thus preventing it
from traveling inward. Professor Howe calls this type of freezing
the ' land-locking type/ When the steel is violently agitated by
the escape of gas its rapid movement washes off the fir-tree crys-

1 From page 373 of No. 1 , page 8. 2 Pages 204 to 223 of No. 74.

3 Pages 224 to 228 of No. 74.

DEFECTS IN INGOTS AND OTHER CASTINGS 183

tals and prevents them from growing out into the liquid mass and
entangling the impure metal. The quietness produced by alu-
minum, however, makes this growth possible.

Another important means of lessening the segregation is by
making ingots narrow,1 that is, by reducing the area of the hori-
zontal cross-section; but this is often difficult of accomplishment.
For example, if we cast fifty tons of open-hearth steel out of one
ladle, it will take a very long time to cast all of this into small in-
gots, and therefore the first ingots cast will be too hot or else the
last ingots will be too cold. There is a difference of opinion as to
whether or not rapid cooling decreases or increases the degree of
segregation, and it seems probable that it acts in both directions,
sometimes prevailing one way and sometimes in the opposite. On
first thought it would seem that slow cooling must necessarily in-
crease segregation, because it would allow more time for the im-
purities to separate from one layer of metal and dissolve in the next.
On the other hand, slow cooling also favors the growth of the fir-
tree crystals, and therefore opposes segregation. It does not seem
possible, at the present time, to tell under what conditions we
should have the one influence prevailing or the other.

It seems to be pretty well established that the greater the per-
centage of impurities present the greater will be the extent of the
segregation. Therefore high-carbon steel should be cast with due
care and narrow ingots used wherever possible. Generally when
the phosphorus and sulphur are low (say, not more than 0.05 per
cent, each), much segregation is not liable to occur, especially in
low-carbon steels.

REFERENCES ON DEFECTS IN INGOTS

70. C. A. Caspersson. Reviewed by Richard Akerman. "The

Influence of the Temperature of the Bessemer Charge on
the Properties of the Steel Ingots." Stahl und Eisen, 1883,
pages 71-76.

71. Henry M. Howe. "Piping and Segregation in Steel Ingots."

Transactions, American Institute of Mining Engineers,
1907. Advance proof.

1 Some metallurgists disagree with this and believe that the large ingots
do not segregate so much. Nevertheless, I am inclined to think that the
greater weight of evidence is against them in this contention.

184 THE METALLURGY OF IRON AND STEEL

72. Henry M. Howe and Bradley Stoughton. "The Influence

of the Conditions of Casting on Piping and Segregation, as
Shown by Means of Wax Ingots/' Transactions, American
Institute of Mining Engineers, 1907. Advance proof.

73. N. Lilienberg. "Piping in Steel Ingots. " Transactions,

American Institute of Mining Engineers, vol. xxxvii, 1906,
pages 238-247.

74. Benjamin Talbot. "Segregation in Steel Ingots." Journal,

Iron and Steel Institute, No. 11, 1905, pages 204-247.

75. N. Lilienberg. "The Compression of Semi-liquid Steel

Ingots." Journal of the Franklin Institute, February, 1908.

76. E. von Maltitz. "Blowholes in Steel Ingots." Transactions,

American Institute of Mining Engineers, vol. xxxviii, 1907,
pages 412-447.

VIII
THE MECHANICAL TREATMENT OF STEEL

METALS may be shaped either by pouring them whilst molten
into a mold, as described in the following chapter, or by mechanical
pressure. The choice of the casting or the mechanical method of
shaping will depend on the size and form of the finished product
and the purpose for which it is intended. Some shapes must be
produced by casting, because they are either too intricate or too
large to be shaped by pressure; others must be produced by pres-
sure, because the service in which they are to be used demands
the higher strength and ductility which mechanical work pro-
duces. Between these two classes, however, is a large number of
forms, each of which is a study by itself. Financial considera-
tions will govern in some cases, and the importance of quality
in others. The advantage of quality is usually with the pressed
material.

Effect of Work. — Mechanical work will multiply the strength
of steel from two to five times. In order to accomplish as much as
this, however, it is necessary to reduce the material to very small
sizes, in order that the beneficial effect of the kneading action may
extend throughout the mass, and to finish the work cold, in order
that the metal may have no opportunity to recrystallize. The
ductility also will be increased at first by working, but again de-
creases if the metal is worked cold. The increase in strength and
ductility is due (1) to the closing up of blow-holes, both large and
small, which are almost all welded together under pressure at high
temperatures, unless they are near enough to the surface to be-
come oxidized inside, and (2) to increasing the cohesion and ad-
hesion of the crystals. The first effect is contributive both to
strength and ductility, while the second is chiefly contributive to
strength, and, if excessive, will greatly diminish ductility. Both
increase the specific gravity and hardness of the metal, and are
more effective in these respects, as well as in increasing strength,
if hot work is followed by cold work.

185

186

THE METALLURGY OF IRON AND STEEL

Crystallization of Steel. — Metals are crystalline substances,
the individual components arranging themselves in regular forms
unless opposed by the rigidity of the mass in which they form.
Indeed, the metallic crystals grow with astonishing rapidity when
the metal crystallizes from the molten state (i.e., solidifies), or
even when it is in a mobile condition (i.e., at temperatures near
or above a red heat). Once crystals have formed they cannot be
reduced in size except by annealing (see Chapter XIV) , or by break-
ing them up with mechanical crushing. These facts are important,
because large crystals do not adhere to each other firmly, and
thus they cause a weak and brittle mass. Iron and steel follow
the same laws as other metals in these respects.

Effect of Strain. — When a metal is strained, the crystals first
stretch, and the amount of this stretching is directly proportional
to the strain; when the metal is relieved, the crystals and the mass

STRAIGHT SLIP BANDS

IN WROUGHT IRON
MAGNIFIED 60 DIAMETERS.

Unetched.
(William Campbell.)

CURVED SLIP BANDS

IN WROUGHT IRON
MAGNIFIED 60 DIAMETERS.

Unetched.
(William Campbell.)

as a whole return to the original dimensions. If the strain is
greater than the ' elastic limit/ however, the crystals yield, and
the particles composing them slip along the cleavage planes, so
that a permanent deformation or extension occurs in the direction
of the strain. This 'elongation' is accompanied by a 'reduction
in cross-sectional area/ and gives warning that the material is
suffering from excessive strain. The extent to which these two

THE MECHANICAL TREATMENT OF STEEL 187

forms of distortion precede rupture is usually taken as the meas-
ure of the 'ductility' of the metal.

Rationale of the Effect of Work. — Mechanical pressure upon a
metal crushes the crystals, mixes them intimately together, and
breaks up the cleavage planes along which they would yield. If
the work is finished above a red heat when the mass is still mobile,
the crystals reform to a certain extent, decreasing the strength.
The elastic limit of a structural steel rolled in this way will be a
little more than one-half its ultimate strength. If the work con-
tinues while the metal is cold, there is no opportunity for the re-
formation of the crystals, and the strength, hardness, and brittle-
ness are much increased.

Methods of Applying Pressure. — Aside from the differences of
hot and cold working, the mechanical pressure may be exerted in
one of three ways: (1) Instantaneously, by a blow, in which method
the pressure is relieved before the metal has fully yielded to it;
(2) more slowly, by rolling or wire-drawing, in which the pressure
is relieved almost as soon as the metal has yielded to it; and (3)
slowest, by presses, in which the pressure remains for a second or
so after the metal has yielded.

THE FORGING OF METALS

The instantaneous application of pressure is man's first method
of shaping metals and is accomplished by a blow from a falling
weight, frequently aided by some other force. Examples of the
first practice are found at the present time in the helve-hammer
used at many small forges and steel-works. After the introduc-
tion of steam, however, this was used to raise the weight, and very
soon steam-power was employed not only to raise the weight but
also to force it downward for the blow, whose momentum was thus
greatly increased. Hammers of this type, which is now the most
prevalent, have been built capable of delivering a blow estimated
as equivalent to 150 tons weight. Such large sizes are not now
-approved of, however, because of the inordinate expense for foun-
dations, which must be deep and powerful in order to take up the
force of the blow, while the constant jarring disturbs the founda-
tions and alinement of machinery, even at distant parts of the
plant. For very heavy forging work, such as armor-plate, etc.,

188

THE METALLURGY OF IRON AND STEEL

the hydraulic press is therefore preferred, and hammers are not
often built in sizes above 30 or 50 tons.

Effect of Hammering. — A blow creates in a metal practically
nothing but compressive strains, which act chiefly in the vertical
direction, and, by transmission, in the two horizontal directions.
Because the pressure is relieved almost as soon as felt, the amount

FIG. 127. — STEAM HAMMER.

of yield to it is not great in proportion to its force, and therefore it
takes more pressure to accomplish a result than would be the case
if the application was slower. This makes hammering a slow
process of reduction, but results in a better and more uniform
working of the crystals, which is one of the chief reasons for the
superiority of hammered over rolled material. On the other hand,

THE MECHANICAL TREATMENT OF STEEL 189

the effect of forging extends only 1 or 2 inches beyond the upper
and lower surfaces. Another and, perhaps, even more potent
reason is the exact control of the operation which can be exercised
by the expert forger, and more especially his control over the tem-
perature at which the work is finished, and over the varying force
of pressure applied at different stages and temperatures.

Finishing Temperatures for Forging. — Forging seldom con-
tinues after the red heat is lost, but the exact temperature will
depend upon the article and the properties which it is desired to
have. The colder it is finished the closer to the exact size required
it can be made, because it has less shrinkage to undergo; but it
will also be harder, less ductile and stronger. The relation be-
tween the finishing temperature of mechanical work and the critical
points of steel will be discussed in Chapter XIV.

Drop-Forging. — There is a large variety of articles, such as
parts of machinery, hammer-heads and similar tools, which are
formed by the process known as 'drop-forging.' In this operation
a piece of metal of the desired size is forged by repeated blows
between a lower die, upon which it rests, and an upper die attached
to the head of the hammer. These two dies are made in the de-
sired form of the finished article, and the metal is squeezed into
them until it has assumed the proper shape. Sometimes several
pairs of dies are necessary to complete the finished shape (see Fig.
129) . Drop-forgings are directly comparable with steel castings, to
which they are superior in quality on account of the beneficial
effect of the working. To be economically made, they must be
ordered in large quantities, so that it will pay to make the costly
dies of hardened steel — often an alloy. Even then, castings are
usually cheaper, though sometimes forgings are still preferred on
account of their quality. There are cases, however, in which
drop-forgings may be made more cheaply, either because the
shape is one that lends itself to rapid production in this way, or
because it is one liable to cheeking, or requiring a large riser, if cast.

Forging Bars. — Crucible-steel ingots are often forged out into
bars for the market, because the material will bring a price high
enough to pay for the superior method of working it. The ingot,
after the top third has been broken off to remove the pipe and
segregate, is heated to a bright-red heat, out of contact with the
flame and fuel, and then tilted down under a hammer of about 10 to
15 tons size until it is about one-half as large on the sides and four

190

THE METALLURGY OF IRON AND STEEL

times as long. The piece is held in a handle which fits over one end
of it, and usually a second heating is necessary before both ends are
down to the correct size. One end is then reheated and drawn

FIG. 128. — SOME AUTOMOBILE DROP-FORCINGS.

down to a bar of the desired size, under the same hammer, or under
one of less weight, and the long bar is then used as a handle while
the other end undergoes heating and reduction. The finished size
is produced by light taps of the hammer just before the blue heat

FIG. 129. — SOME STAGES IN THE MAKING OF A DROP-FORGING.

191

192 THE METALLURGY OF IRON AND STEEL

appears, and often a piece of cold steel is laid beside it on the anvil
to more correctly arrest the downward blow. The finished bar
will be so straight and true as to lead one to believe that it was
produced by drawing through a die, or rolling in grooved rolls.
Sometimes it is finished in a square shape, and sometimes as an
octagon, by turning it upon the corners and drawing them down.

Forging Razors. — Flat bars for razor stock, made of cemented
steel melted in crucibles without additional carbon, are produced
in a manner similar to that outlined above, and then forged down
by hand to the rough size of a razor. They are then stamped with
the appropriate name and mark, drilled with a hole, heated to the
correct temperature and hardened in water, after which the temper
is drawn to the light or medium straw color (see page 387). The
exact shape is then produced by grinding, care being taken not to
heat the razor during this operation, lest it be tempered thereby,
and the blade polished and fitted with a handle.

Forging Cannon. — Large cannon tubes are made from open-
hearth steel ingots weighing perhaps 65 tons or so, more than one-
half of which is discarded or 'scrapped' during the process. In
France and Germany, cannon-tube ingots have been made of
crucible steel by pouring many crucibles into one mold, but the
expense and the liability to heterogeneity because of the many
small units is believed to outweigh the advantages due to the
quality of steel, which is superior on account of its process of
manufacture. The heating of the ingots must be done with great
care, lest a crack or hollow be formed by too rapid expansion or
by the expansion of the outside away from the interior. More-
over, as reducing a flame as possible must be
maintained, lest the carbon be oxidized in the
outer layers of steel during the many hours
required to attain the bright-red heat neces-
sary. Ingots of the form shown in Fig. 130
are usually employed, and those cast in sand
molds are preferred because they are not liable
to contain surface cracks produced by tearing
FIG. 130. the steel when the iron mold is withdrawn.

Only one end of the ingot projects into the
furnace, and when the desired temperature is reached, the handle
or 'porter bar7 is fitted over the cool end, and the crane which
supports this transfers the whole over to the hammer.

THE MECHANICAL TREATMENT OF STEEL 193

By the time the heavy blows from a hammer of perhaps 75
tons force have effected a certain reduction, the ingot must go
back to the furnace for another heating of about an hour or so, and
this reheating is necessary at intervals. When the top end of the
ingot has been drawn down to a convenient size, this is used as a
handle to the other end, which is to make the completed cannon
tube, and not more than the lower two-thirds of the original ingot
is allowed to be present in the tube at the finish of the forging
operation. The blows are delivered on all sides of the ingot in
order that the center of the tube shall be the same as the center
of the original ingot, for this center portion is to be drilled out
in the subsequent operations, and, as we have already seen, the
center of the ingot is of looser texture and contains more of the
segregate.

After the inner tube is forged, an outer tube is produced in a
similar manner, but of larger size, so that it may be bored out to
fit over the carefully turned inner tube. After the boring the
outer tube is too small to pass over the inner one, and it is therefore
heated to a temperature of about 280° C. (550° F.) in a tall vertical
furnace, which expands it so that it may be passed over the inner
tube and ' shrunk' upon it, greatly increasing its compactness and
reinforcing it against the tremendous strains it is subjected to in
service. Cannons are now frequently forged under presses instead
of under hammers.

THE REDUCTION OF METALS IN ROLLS

If two rolls rotating as shown in section in Fig. 131 be made to
grip a piece of metal, A, they will drag it between them and force
it out on the other side reduced in thickness. The metal between
the points 0 0 and N N is being compressed vertically, while its
outer layers are suffering tension. In the case of a deep section,
the unequal strain is liable to tear the steel (see Fig. 165). At the
points N N the metal is being forced back upon itself. The
mechanical pressure is therefore not as uniform as in hammering,
and acts for a longer period of time. Reduction can only take
place vertically, as in forging, there being always a certain amount
•of expansion sidewise, and a large amount of extension in length.
The metal at the points N N being forced backward, and that at
the points 0 0 being forced forward, the ends of the rolled section

194

THE METALLURGY OF IRON AND STEEL

assume a shape somewhat like that shown in Figs. 132 and 133,
when the outside of the piece is very hot. The reduction in
area at each 'pass' will vary between 5 and 50 per cent, of the
original, and the work is very rapid. For example, a railroad rail

FIG. 131.

may be produced, from an ingot having a section 18 in. square, in
22 passes, varying in amount of vertical squeeze from 8 to 52 per
cent., only about five minutes being required for the whole opera-
tion, the piece traveling through some of the passes at a rate of ten
miles per hour, and not being reheated after the ingot comes to the

FIG. 132. — SHAPE OF ENDS OF
ROLLED METAL WHEN THE
INSIDE IS THE HOTTER.

FIG. 133. — SHAPE OF ENDS OF
ROLLED METAL WHEN THE
OUTSIDE IS THE HOTTER.

first pair of rolls. Some American rolling-mills produce about a
mile of single" rail per hour for 24 hours a day and 25 days per month.
The temperature at which the rolled material is finished is gaged
with much less accuracy than in forging operations, and is always

THE MECHANICAL TREATMENT OF STEEL 195

too high for the best quality of the steel, because economy of power
urges the manufacturer to work the metal hot.

Pull-Over Mill. — In a single pair of rolls, such as shown in
Fig. 131, the metal, after passing between them once, must be
handed or pulled over the top of the mill, to be fed in for a second
pass. This type of train is known as a 'pull-over' or 'pass-over'
mill. It can be used only for shapes small in size and that can be
handled readily, and the action is slower than in a continuous
operation, such as in a 'three-high mill.' The pull-over mill is
simple and cheap to construct and operate, and is used especially
for the rolling of plates and shapes from crucible steel, whose high
price renders it less important to seek rapid output. It is also
used very largely for the rolling of steel to be used for tinplate.
The upper roll is adjustable, so that any thickness may be pro-
duced.

Multiple-Ply Plate, etc. — Three-ply plate for plowshares,
and five-ply plates or bars for burglar-proof safes and jail bars,
are often made in this type of mill. We first roll independently
thin plates of high-carbon, crucible, chrome steel (or an equiva-
lent alloy steel capable of becoming very hard upon quenching in
water from a red Heat), and thicker plates of wrought iron. For a
plowshare, a plate of wrought iron will then be sandwiched be-
tween two plates of chrome steel, tied into a bundle with wire, and
raised to a welding heat. The wire burns off in the furnace, but
the bundle is grasped with a pair of tongs and fed into a pair of
plain rolls, where it is welded into a plate of three-ply steel which
is reduced to a thickness of a little over a quarter of an inch. This
plate is trimmed to about the desired size and shape and then
hardened and used for a plow, the hard outer layers resisting the
wear of service, and the ductile core resisting the shocks which
would shatter the brittle outside. For safes and jail bars we have
an inner layer of wrought iron, then two layers of chrome steel,
and then two layers of wrought iron, welded together and then
hardened. A burglar can neither drill through this, on account of
the hardened chrome steel, nor break it with a sledge, on account
of the ductile iron, which will not be hardened by the quenching.

Three-High Mills. — When a piece is passed over a two-high
mill, it is often rested upon the top of the upper roll, whose travel
assists somewhat in the transfer. While watching this operation
at the Cambria Iron Company's mill in 1857, John Fritz con-

196

THE METALLURGY OF IRON AND STEEL

ceived the idea of the three-high mill, which is shown in section
in Fig. 134. It will be seen that passing the piece over the top of
the middle roll in this mill will result in its receiving work, and
thus the output of the mill will be increased. At the present time
the great bulk of the tonnage of steel and wrought iron produced,
consisting of structural shapes, railroad rails, plates, wire rods,

FIG. 134.

billets and bars, is finished in this type of mill. The output is
large, because the rolls can be run very fast indeed (rod mills run-
ning 600 to 1200 revolutions per minute and sometimes passing the
rod through at the rate of half a mile a minute in American prac-
tice 1), and two or more pieces maybe passing through at the same
time. The disadvantage of the three-high mill is the power neces-

1 The reason for this rapid rolling is not only large product, but that the thin
rods may not radiate their heat during the operation and thus be finished too
cold. This rapid work actually raises the heat of the metal faster than it can
be radiated, and rods are hotter at the end than at the beginning of the rolling.

THE MECHANICAL TREATMENT OF STEEL 197

sary to raise large weights up to pass over the middle roll. Most
Bessemer ingots are cast two tons or more in weight, and most
open-hearth ingots from three tons to ten or more tons.

Reversing Mills. — Therefore ingots are often ' cogged' in two-
high reversing mills to avoid this consumption of power. More-
over, the two-high mills, which have an adjustable upper roll, have
the advantage of being able to work an ingot gently at first, in
case it shows a tendency to be ' tender/ that is, to crack in spots

FIG. 135. — THREE-HIGH MILL.

when the pressure is applied. Three-high mills may, however, also
have an adjustable middle roll, and plate mills are frequently made
in this way. The disadvantages of the two-high reversing mill are
its slowness and the severe strain on the engines, which are often
reversed while running full speed.

Universal Mill. — During the rolling of metal there is a cer-
tain amount of expansion sidewise, which gives the piece a cross-
section somewhat bulging on the sides, and makes the edges
uneven, unless the rolls have collars which form a groove through
which the metal passes. In 1855 R. M. Daelen, at Hoerde, Ger-
many, devised a mill in which even edges could be produced at any
width by having an auxiliary pair of vertical rolls, between which
the piece passes immediately after it emerges from the horizontal
rolls. These vertical rolls are adjustable to any width up to the
capacity of the mill, and give only enough pressure to keep the
edges even without producing any reduction. They are usually
made to rotate with a surface velocity greater than that of the hori-

200 THE METALLURGY OF IRON AND STEEL

zontal rolls, so as to prevent the serious buckling that would take
place if the conditions were reversed. As this tension is not good
for the edges of the metal and wears out the vertical rolls, some
mills have independent control of drive for each pair of rolls, and
others have friction-clutches connected with the vertical rolls,
which allow them to run faster if pushed by the metal, but ordi-
narily run them at a slower speed. Universal mills are made
two-high or three-high, and with vertical rolls on one or on both
sides of the horizontal rolls. With the two-high mills (which are,
of course, reversing), there is one set of vertical rolls. Slabbing-
mills are usually made this way. With three-high mills, there is
a pair of vertical rolls on both sides of the horizontal rolls. Plate-
mills are sometimes of this kind. In England, the Universal mill
is not in favor, as rolling-mill managers believe that the faster work
of simple mills more than makes up for the necessity for changing
grooved rolls at intervals when a new width is to be produced.

PARTS OF ROLLING MILLS

Rolls. — Rolls may be plain cylinders, by which plates and
rectagonal shapes are produced, or they may be cylinders with
'collars' at intervals, as shown in Fig. 138, in which large rectagonals
with even edges may be produced ; and the collars may be on both
rolls, giving an 'open pass/ or may be on only one roll and extend
into grooves on the other roll, as shown in Fig. 139, giving a ' closed
pass.' With open passes, the collars cannot be made to quite
touch, hence the name; and the pressure may squeeze some metal
between them, forming a 'fin' along the side of the piece. This
results from 'overfilling the pass.' The closed pass makes the
upper roll weaker, and there is also a liability of the metal becom-
ing wedged tightly between the collars and thus drawn all the way
around the roll, with the result that something will be broken.
Wedge-shaped grooves may be cut in the rolls, producing the
'diamond' pass, in which small squares are made (see Fig. 139); or
oval grooves make nearly round bars which are finished round in
the last pass with almost no draft. Other forms of passes are
shown in Figs. 139 and 165. In case rolls are weakened by deep
cutting, as shown in Fig. 139, they may be strengthened by stiffen-
ers, D, while long rolls for producing wide plates are sometimes

THE MECHANICAL TREATMENT OF STEEL

201

stiffened by an idle roll running on top, lest the springing of the
roll make the plate thicker in the middle than at the edges.

Cast-iron versus Steel Rolls. — Cast-iron rolls are chilled upon
the outside so as to produce a surface layer of white iron (see Fig.
264), which, after turning in a lathe, makes a very smooth surface
for rolling and is especially advantageous for finishing-mills. They
are not so good for the mills which do preparatory work, however,
because they are not so strong, and because in preparatory work we

FIG. 138. — CC = COLLARS. W = WOBBLERS.

want a rough surface to assist in gripping the metal and drawing
it through. Furthermore, they cost more to turn to the desired
shape, and they cannot be turned down many times (see page 203) ,
lest we get below the ' chill/ The greater cheapness of cast iron
over steel, however, counteracts these factors of higher cost.
Where the rolls must be very strong and yet not too large in
diameter, and for sharp corners, which would crumble if made of
cast iron, steel rolls are often used. The steel employed should
be high in carbon, — say 0.50 to 0.75 per cent. ; but any case-
hardening of steel is useless here, because the heating of the rolls

202

THE METALLURGY OF IRON AND STEEL

by the material passing through will soon draw their temper.
This heating cannot be prevented altogether, though it is custom-
ary to have a stream of water flowing over the rolls. Sometimes
nickel-steel rolls are used for strength. An analysis of roll metal
of a very large American company is: 0.40 to 0-50 per cent, car-
bon, 0.65 per cent, manganese, 3.25 per cent, nickel, and 0.15 to
0.20 per cent. siLcon.

Diameter of Rolls. — With smaller rolls, the amount of power
consumed is less because the area of metal under vertical pressure

A AND C = OPEN PASSES.
B = CLOSED PASS.

C = DIAMOND PASS.

FIG. 139.

is less. There is a limit below which the diameter cannot go,
however, either because the rolls will not be strong enough to give
the desired pressure, or they will not grip the bar. In order to be
gripped the upper and lower edge of a piece must touch the rolls
at a point not more than 30° from the center line of the two rolls
(see Fig. 140). Every effort is made to use smaller rolls, because
the size of all the mills is regulated by them. The surfaces of all
but the finishing-mill are usually 'ragged7 (i.e., made rough), to
make the rolls give a better grip. Those to receive the ingots
are ragged the most, with deep indentations somewhat like the
of cog-wheels, whence the name of ' cogging rolls' for this

THE MECHANICAL TREATMENT OF STEEL

203

mill.1 The next trains, known as the 'roughing rolls/ are also
deeply marked, but even then the piece must come within the 30°
line, or time is lost in trying to make them bite the piece.

Speed of Rolls. — The more work the rolls do, the slower must
they revolve, because the piece entering the train gives a shock to
the mechanism that is depend-
ent upon the power exerted
and the momentum of the mov-
ing parts. Thus the larger the
pieces treated the colder they
are, and the larger the rolls the
slower must be the speed. In
America, speeds are at the high
limit. Reversing slab-mills may
do the work at 20 or 30 r.p.m. ;
three-high blooming rolls may
run over 50 r.p.m.; mills for
finishing rails, 100 r.p.m.; and
rod mills from 550 to 1200 r.p.m.

Making of Rolls. — Cast-iron
or steel rolls are cast in ap-
proximately the desired shape
and then turned accurately in
a lathe, being fitted exactly to a
templet when completed. After

rolling some thousand tons of material, they become worn and
produce too large a size of finished shapes. They may then be
used for a larger size of the same kind of article by putting them
back in the lathe and turning to another templet. For example,
a roll for a 20-inch I-beam, with a certain thickness and width of
flange, may be converted to one for a 20-inch I-beam with thicker
web and longer flange.

The Mill. — The different parts of a rolling-mill may be seen in
Fig. 142. The wobblers are made of the same cross-section as the
spindle, some examples being shown ir. Fig. 143. The coupling-

1 In America, the train that produces blooms (i.e., pieces of steel usually
about 6 to 8J in. square) from ingots, is sometimes, but not always, known as
'bloom rolls/ or 'blooming rolls/ instead of cogging rolls; and the train that
produces 'slabs' (i.e., thick, wide, rectangular pieces that are to be rolled
into plates) from ingots is known as the ' slabbing-mill.'

FIG. 140.

FIG. 141. — INDENTATIONS ON A COGGING MILL.

FIG. 142.

B, Coupling boxes; C, collars; EEE, roller table engine; F, fingers, or horns on manipulator;
H, housings; HC, housing cap; M, manipulator or 'Go-devil'; R, roll; S, spindle; T,
roll table; TR, table rollers.
204

THE MECHANICAL TREATMENT OF STEEL

205

boxes fit over the spindle and wobblers, so that neither can turn
without the other. In some mills the coupling-boxes are made
of cast iron in order that, if any shock comes upon the driving
mechanism, the boxes shall give way and relieve the strain. In
other mills the boxes are made of cast steel, as it is thought that the
constant delays due to broken couplings are more costly than
breakages in other parts of the mill. The spindles are at least

FIG. 143.

B, coupling boxes; C, collars; E, roll engine; G, guides; H. housings; S, spindles; TR, table
roller; HC, housing cap; W, wobblers; TM, table motor.

twice as long as the coupling-boxes, in order that they may carry
both of them at once when the train is uncoupled. Both boxes
slip back upon the spindle.

Pinions are now usually made of steel for the sake of strength.
Housings are made of either steel or of cast iron, depending on the
strains to which they are subjected and the opinion of the manager.
In America, they are usually made so that the top can be removed
and the whole train of rolls removed at once, together with the
chocks, and several mills have spare sets of rolls all made up ready

206

THE METALLURGY OF IRON AND STEEL

and carried in a sling, so that a new set may be dropped into place
with a crane with the least possible delay to the mill. Delays in
rolling-mills are very costly, because of the idle labor and capital,
and because other parts of the plant may be delayed thereby.

FIG. 144. — A PINION.
W = wobbler.

The screw-down mechanism which adjusts the distance* be-
tween the rolls is operated usually by hydraulic pressure, though

electric motors are coming into
vogue. It is connected with a
telltale gage which advises the
roller exactly as to the distance
separating the rolls.

Guards are of steel and serve
to peel the piece off the roll and
prevent it encircling the roll
(called t collaring') in case it be-
comes wedged between the col-
lars. They must be upon the
lower roll, as shown in Fig. 145,
or upon the upper roll, and coun-
terbalanced to hold them in posi-
tion, when they are called ' hang-
ing guards.' Guides are on the opposite side of the train, and assist
in conducting the piece straight into the groove.

Roll Tables. — Heavy pieces are handled at the rolls by sup-
porting them upon a series of rollers, situated in front of and behind
the roll train, and known as the 'tables/ At two-high mills the
tables are stationary ; at three-high mills the front and back tables

FIG. 145.

THE MECHANICAL TREATMENT OF STEEL

207

are sometimes raised and lowered together by hydraulic or elec-
tric mechanism, and sometimes they are pivoted near the middle,

FIG. 146.

A, hanging guards; B, coupling boxes,; H, housings; P, pinions; PH, pinion housing; R,

rolls; S, spindles.

so that the end next the rolls can be tilted upward in order to
bring the piece between the guides which direct it into the groove.

FIG. 147.
G, guides; HC, housing cap; PH, pinion housing; RR, rolls.

The rollers to handle large pieces are 'live/ that is, they are made
to revolve by electric motors and thus move the piece back and

FIG. 148.
D. Screw-down mechanism; EEE, table engine; TM, table motor; TR, table roller.

FIG. 149. — TWO-HIGH, REVERSING UNIVERSAL MILL.

VVV, Vertical rolls; RR, Horizontal rolls.
208

OQ H

210 THE METALLURGY OF IRON AND STEEL

forth. 'Dead' rollers are used where pieces are to be moved by
hand.

Transfer Tables. — Roller tables are sometimes made so that
they may be moved bodily from one roll train to another, carrying
the piece of metal with them, and so connected electrically that the
rollers can be caused to revolve when the table is in any location.

Manipulators. — If two or more posts, supported on a carriage
which can be moved laterally, project between the rollers of a
table, their sidewise motion will transfer the piece from one pass
to another. If the table is of the lifting type, the posts, or 'horns,'
or ' fingers' can be brought to such a position that the lowering of
the table will bring the edge of the piece upon the horns and thus

FIG. 151.

tip it on to the other side. This form of manipulator is much used
at three-high blooming rolls, and is very efficient and rapid in its
work. The same type is used at reversing blooming rolls, but the
pie<?e is more usually tipped over by the roller with the tool shown
in Fig. 151.

Roll Engines. — The service on rolling-mill engines is very
severe, because the full load comes upon it when the piece enters the
rolls, and then leaves it as suddenly again. To equalize these
sudden variations of power, all but the reversing engines are built
with very large and heavy fly-wheels and run at a high rate of
speed (from 30 'to 250 r.p.m.), with governors of a quick-acting
type. The Allen engine with the Porter governor serves these
purposes, and the Porter- Allen type is much used. The ordinary
slide-valve is used on the smaller engines. Corliss valves are com-
moner in America for engines doing heavy work (1000 to 3500 H.P.) ,
while piston-valves are favored in England. The fly-wheel is
placed upon the crank-shaft, to which the roll train is directly

212

THE METALLURGY OF IRON AND STEEL

coupled. The fly-wheels very exceptionally weigh as much as 75
and 100 tons or more.

Piston-valves are used almost always for reversing engines
which are compounded, so they may never come to rest at a dead

FIG. 153. — UNIVERSAL MILL.

E, Engine for horizontal rolls; EE, engine for vertical rolls; EEE, roller table engine;
TT, roll tables; TR, TR, table rollers.

point. There is, of course, no fly-wheel, and the engine is directly
coupled from the crank-shaft to the roll train in the large American
mills, but is geared down so that the engine can develop a higher

FIG. 154. — MOTOR-DRIVEN ROLLING MILL.

speed than is desired for the rolls, thus requiring less power.
Reversing slabbing-mill engines have capacities up to 25,000 H.P.
each.

THE MECHANICAL TREATMENT OF STEEL

213

Electric- Motor Drive. — During the year 1906, very important
installations were made of electric-motor-driven roll trains. The ad-
vantages of electricity over steam are a lower operative cost, greater

security of operation, fewer breakdowns, and a more flexible rela-
tion between the prime mover and the load, the result of electric
motors receiving a sudden shock more elastically. On the other
hand, the advantage of steam is that, although it receives the
load less elastically, it adjusts itself quicker and better to the ex-

214

THE METALLURGY OF IRON AND STEEL

treme variations in load that always occur in rolling-mills. This
is especially true of reversing mills.

As already noted, the smaller the mill the less will be the load,
and therefore the variation in load. Consequently, in England,

Sweden and Germany there are many motor-driven roll trains of
the smaller size, and a few up to several hundred horse-power, in-
cluding one reversing motor of 1200 H.P. In America, there are
some small roll trains operated by electricity, but up to 1906 there

THE MECHANICAL TREATMENT OF STEEL 215

was only one with as high a capacity as 1500 H.P. During 1906,
at some of the most important works in the country, motor-driven
rail mills were begun, including two motors of 1500 H.P. each for a
three-high train, and also including a rail mill operated throughout
by electricity, from the bloom rolls to the finishing train, with six
motors, varying in capacity from 2250 H.P. to 6000 H.P. each.
The most startling innovation, however, is a reversing Universal
plate-mill operated by two motors on the same shaft, which, when
running at full speed (150 r.p.m.), will develop approximately
10,000 H.P. It is supposed that this mill, when built, will be
capable of reversing from full speed forward to full speed backward
in the space of three seconds.

Almost all American rolling-mills of most modern equipment
now use electric power for driving the table rollers, the screw-
down mechanism, the shears, and in fact for all purposes except
driving the rolls.

ROLLING-MILL PRACTICE

Troubles in Rolling. — There are more difficulties met with in
rolling-mill practice than we can discuss here, but it may be said
that the seriousness of a difficulty is estimated almost altogether
in proportion to the delay it causes in the operation of the mill,
rather than in the loss of a small amount of material or of a part
of the mill itself. For example, the. breaking of a table engine,
roller, or even a roll, is regretted more because of the time necessary
to put in a new one than because of the loss of the part. This is
one reason why electric motors to operate the tables have, in many
cases, replaced small steam engines. The same conditions have also
resulted in different parts of the mill being made interchangeable.
In many mills it is customary to have spare table
engines or motors, etc., always ready, and the least
accident to one of these machines would result in its
being immediately replaced by a whole new one.

The most important common troubles in rolling-
mill operations, probably, are: (1) Bending and
breaking of the rolls, due to their being placed under too severe a
strain, either because the draft is too heavy or because the piece is
cooled too much ; (2) the fins caused by metal being squeezed out be-
tween the collars of the rolls, as shown in Fig. 157 ; these fins, besides
spoiling the material, are liable to break the rolls; (3) collaring.

FIG. 159. — UNIVERSAL MILL.

FIG. 160. — THREE-HIGH PLATE MILL.

216

THE MECHANICAL TREATMENT OF STEEL

217

Rolling Plates. — In the rolling of plates an ingot, usually of
open-hearth steel and weighing 2 to 10 tons, is first cogged down
in the slabbing-mill, producing a long, flat piece of metal. The
slabbing-mills are frequently of the two-high, reversing, Universal

FIG. 161. — THREE-HIGH PLATE MILL, TWO STANDS OF ROLLS.

B, Coupling boxes; D, screw-down mechanism; E, roll engine; H, housings; P, pinions;
PH, pinion housing; RR, rolls; S, spindles.

type. The front end of the piece is cut off in a huge hydraulic or
electric shear to remove the pipe and then it is cut up into slabs
of the desired size or into slabs of a size such that each one will make
one plate. The slabs are then transferred to the heating furnace,
heated to about 1300° C., and rolled in a three-high or, more rarely,

FIG. 162. — OTHER VIEW OF FIG. 161, SHOWING METHOD OF RAISING
THE ENDS OF THE TABLES NEXT TO THE ROLLS.

a two-high reversing plate-mill, in some cases there being a pair of
vertical rolls to keep the edges straight. During the rolling a
shovelful of salt is occasionally thrown upon the surface of the
plate, which carries in between the rolls some of the water which
is always trickling over them to keep them cool. Sand may be

FIG. 163. — PLATE STRAIGHTENING ROLLS.

218

FIG. 164. — PLATE SHEARS.

THE MECHANICAL TREATMENT OF STEEL

219

used for the same purpose, and in England heather is sometimes
used. As soon as this water is pressed against the hot plate it is
converted into steam, causing a rapid series of explosions which
blow the scale off the upper surface of the plate and give it a
smoother finish. As the process continues, the operator tests the
thickness of the plate with a gage, and when it is of the desired
thickness, it is passed up to the straightening rolls and then to a

-flH

FIG. 165.

cooling table, being marked with a distinguishing mark on the way
to indicate the heat of steel from which it was manufactured.
When cooled it is sheared to the desired size and shape. The weight
of finished plate will probably be not more than 80 per cent, of
the weight of the steel sent to the rolling-mill in the form of ingots.
Rolling Rails. — An ingot of about three tons in weight is sent
to the rolling-mill, where it is kept in the heating furnace for 50
minutes or more until the interior is entirely solid and it is of a uni-
form temperature throughout. It is then rolled into blooms,
either in a three-high mill, such as shown in Fig. 142, or in a two-

220

THE METALLURGY OF IRON AND STEEL

high reversing mill. In the three-high mill, an ingot 18J in. square
at the middle (tapering about J to j inch to a foot in order that
the mold may be more easily removed) will be reduced to a bloom
of about 8 in. square in nine passes, the amount of reduction in each
pass being about 12 to 18 per cent, of the original area. The top
end is then cut off to remove the pipe, the bottom end to remove
the irregularity due to the rolling, and the piece cut in two to make
two blooms. The blooms are then generally reheated in a heating
furnace and passed through the series of changes shown in Fig. 165,
until they have assumed the proper size and form, the greatest

FIG. 166.

amount of draft being usually not more than 22 per cent., except
upon the middle portion of the web. In some cases the blooms
are not reheated, but go directly from the bloom rolls to the first
roughing train. This makes the metal crack more in rolling, how-
ever, and these cracks will ultimately show as a mark on the fin-
ished product, which causes the rails to be classified by the in-
spector in the second or third class. Railroads will accept only
5 or 10 per cent, of their order in second-class rails, while third-
class rails are not acceptable and must go into the tracks of the steel
company itself or else be reheated and rolled into smaller sizes,
whereby the marks will often be eliminated. If the blooms are
reheated before going to the roughing train, many of the cracks

THE MECHANICAL TREATMENT OF STEEL

221

formed during blooming will be seemingly closed up, or in any
event will not show. Furthermore, if this reheating is to take
place, the ingots need not be heated so hot in the first instance,
and therefore will not be so tender and so liable to crack.

Making Lap-Welded Tubing. — The wrought iron or steel low
in carbon is first rolled out into skelp about 20 to 25 ft. long, and
of a width a little more than three times the intended diameter of
the tube. The skelp is then rolled up into a rough form of a pipe,

222

THE METALLURGY OF IRON AND STEEL

as shown in Fig. 168, by passing it sidewise through rolls, which
bend it roughly to the shape of a pipe, with edges overlapping.
The same is done in the case of small 2-to-8-in. tubes, by draw-
ing them through a die. It is
then passed at a welding heat
through a pair of rolls, with
the seam that is to be welded

FIG. 168.

FIG. 169. — MANDRIL.

upward. Between the rolls is a mandril, on the end of a long
rod and of the size of the inner diameter of the tube. The rolls
press the two parts of the weld together over the mandril, and the
pipe, after another rolling to give true size and after straightening
and testing, is ready for service.

Making Seamless Tubes. — Seamless or weldless tubes are
made either by distorting a steel plate between dies, as shown in

FIG. 170. — PIPE-WELDING ROLLS.

Fig. 172, or else by piercing a hole through the center of a hot steel
billet and then rolling it successively between rolls over a mandril.
The hole is sometimes first of small size and then expanded by
pressing larger and larger expanders through it. The pierced

FIG. 171. — PIPE-WELDING ROLLS WITH MANDRIL IN POSITION.
1

FIG. 172.

223

224 THE METALLURGY OF IRON AND STEEL

billet is then rolled over mandrils constantly decreasing in size
until the inner and outer diameters are brought to the desired size.
Butt-Welded Tubes. — Butt-welded tubes are made by heating
the skelp to a welding temperature and then drawing it out of the
furnace through a bell, as shown in Fig. 173, which
curls it up and welds the edges together, with-
out lapping. Butt-welded tubes are not so strong
as lap- welded tubes, and are not usually used for
boilers or high pressures, or where they will be
expanded much by heat during service. They
are mostly made in the small sizes.

In the United States, over a million and a
half tons of pipe are made each year. It is used
principally for the transmission of oil, water,
steam and gas, and for conduits for electric wires. About 30
per cent, of this is made of wrought, and 70 per cent, of soft
weld-steel. The steel is usually, but not always, made by the
Bessemer process on account of the difficulty of making very low-
carbon material in the open-hearth.

WIRE DRAWING

Wire is a product formed by being drawn cold through a die.
The commonest shapes are 'rounds/ and the next, hollow tubes,
but a great variety of forms may be produced at will.

Effect of Drawing. — The effect of the drawing is to produce a
very exact size of material and to increase the strength, hardness,
and brittleness of the metal. In the drawing of steel, the crystals
of the metal are actually pulled apart and flow by each other, the
outer layers of the metal being dragged back over the central core,
there being at the same time a pressure exerted in all directions
toward the center, which results in a certain amount of backward
flowing even there. Because the crystals are so broken up during
the operation, and because the metal is never heated above its
critical temperature during annealing, the grain of the steel is very
fine and the crystals are intimately mixed, which is probably the
cause of the great strength of wire.

Annealing. — With each draft the wire becomes harder and
more difficult to draw. As it is pulled through the die by a
force equal to 40 to 80 per cent, of its tensile strength, it is neces-

THE MECHANICAL TREATMENT OF STEEL

225

sary to soften it at intervals by annealing, iest it break. The an-
nealing is accomplished by enclosing the wire in some receptacle

FIG. 174. — WIRE ROD ROLLING TRAIN.
These mills will roll steel down to from J to ^ inch rods, which are then drawn into wire.

that protects it from oxidation and then heating to a low-red heat.
In the case of steel, it is required after every 8 to 3 passes, de-

FIG. 175. — WIRE ROD FRAME.

pending upon the amount of carbon in the metal and the amount
of draft.

226

THE METALLURGY OF IRON AND STEEL

Dies. — Wire dies are usually made of high-carbon steel (say
about 2 per cent.), through which a tapered hole is made, as shown
in Fig. 176. The object of using this material is that, as it becomes
worn in service, it can be reformed and used for larger sizes, which
could not be done with white cast iron.

Bench. — A ' bench ' on which wire is drawn consists of a
reel which holds the coil of undrawn wire, a die support, and a
second reel which draws the wire through the die and coils it up,

and which is driven by bevel gears.
The die rests against the support, and
the wire, having a tapered point, is
thrust through the hole and grasped
by a pair of tongs, which pulls it out
until it can be attached to the reel.
This is then set in motion and draws
the wire through. The die-holder is
heaped up with lubricant of some kind,
in order that the metal may pass more
easily through the hole. The speed at
which wire is drawn will vary from
75 to 750 feet per minute, depending
upon the size and hardness of the
material drawn and the amount of

reduction during each draft. In many cases there is more than
one die, and the wire passes successively through two, three or
more, being constantly reduced in each one. Between each pair
of dies is a reel, around which the wire passes two or three times,
since the strength of the wire emerging from the last die would
not be sufficient to draw it through all of the holes.

Draft. — The. heavier the draft the greater is the hardness
produced in the wire and the greater the wear of the dies. The
average amount of draft will probably be from 20 to 25 per cent.
Drawing Tubes. — Hollow wire or small tubes are drawn
sometimes over a mandril. This mandril may be a wire of about
the size of the inner diameter of the finished tube. After several
drafts, the tube is wedged so tightly on the mandril that it cannot
be separated. It is then given an unbalanced squeeze between a
pair of rolls, so that the tube is reduced in thickness, whereby its
diameter is increased and the mandril may easily be withdrawn.

FIG. 176. — SECTION OF
WIRE DIE.

THE MECHANICAL TREATMENT OF STEEL
PRESSING

227

Steel may be pressed either hot or cold, the latter method
being used chiefly for thin and soft steel, and the former for very
large work, such as armor-plate, cannon, etc., for which hydraulic
presses have now largely replaced the heaviest steam-hammers.

FIG. 177. — FOURTEEN-THOUSAND-TON ARMOR-PLATE PRESS.

Effect of Pressing. — The effect of pressing upon the metal is
almost exactly the same as that of hammering, except that its
action extends a little deeper into the material, thus giving a
somewhat superior texture to this part of the body. Tests cut

228 THE METALLURGY OF IRON AND STEEL

from the center of large pieces forged under the press are very
much superior to those cut from the same place in pieces forged
under the hammer.

Hot-Pressing. — Presses vary in size usually from 600 to
14,000 tons. They may be either of the continuous or of the
intermittent type. In the latter, the amount of pressure exerted
increases step by step as the work progresses. The amount of
work that can be done by the press in large-sized pieces is greater

FIG. 178. — DROP-FORGING PRESS FOR PLATES.

than that done by hammers for the same amount of power used.
This results in a double saving of fuel, since more work can be
accomplished with one heating. By means of the 10,000-ton
press at the Homestead Steel Works, a 50-ton armor-plate has
been reduced 2 in. in thickness and moved forward 6 in. for each
squeeze, while a 3000-ton press at the Firths Works in England
has reduced a 30-ton ingot from 49 to 28 in. diameter in 30 minutes,
and from 51 to 26 in. diameter in 65 minutes. Small pieces can
be turned out a little faster under the hammer.

Cold-Pressing. — Thin plate for steel railroad-car construction
is often formed by pressing it cold between dies under hydraulic
presses of from 30- to 800-tons capacity. In this way bolsters,
braces, and many other parts are formed with great economy.
Sometimes two or three presses are required with different dies to
complete the shaping, and occasionally it is necessary to press some
of the work hot, because the distortion is so great that the steel

THE MECHANICAL TREATMENT OF STEEL 229

would otherwise be torn. .Cold-pressing is also known as ' flanging/
It has one great difference from hot-pressing, in that there is no re-
duction in the sizes of pieces treated.

COMPARISON OF MECHANICAL METHODS

Hot-Rolling with Cold-Rolling. — Cold-rolling gives a harder
material and more accurate finish as to size than any form of
hot- working. Furthermore, it produces a finer grain in the metal.
If the cold-rolling is followed by annealing at a temperature below
the critical range of the steel (see pages 388 to 389), the material
retains its fine grain, and is stronger and more ductile than metal
that has been worked hot. Before cold-rolling, the metal is
pickeled in dilute sulphuric acid to remove the scale, and is there-
fore produced with a bright surface which is suitable, without
machining, for use as shafting, for nickel-plating, etc. The
annealing is usually effected inside closed vessels, in a reducing
atmosphere of illuminating gas or some similar medium, which
prevents the formation of scale. Cold-rolled steel is used for
articles that are to be drawn or stamped to shape — watch and
clock springs, hacksaw blades, corset steels, etc.

Hammering versus Rolling and Pressing. — Of all the mechanical
methods, rolling gives by far the largest output per day, per unit
of power, and usually per unit of fuel for heating. It is therefore
the cheapest method, especially for labor. It does not work the
metal as well as either hammering or pressing, both of which pro-
duce a much better crystalline structure, beside affording a better
control of the temperature at which the operation is ended.
Pressing works the metal at greater depths than hammering, and
is therefore especially advantageous for producing large pieces,
and the more so because small presses are very costly to install as
compared to steam-hammers. Where a shape is intricate, rolling
is more liable to tear the metal than hammering or pressing be-
cause, at the point where the roll is deeply cut, its surface velocity
is much less than where the diameter is greater, and thus it tends
to drag the metal through at different speeds.

HEATING FURNACES

Heating furnaces are usually of the reverberatory type, burn-
ing soft coal or gas. The flame produced must be as reducing as

THE MECHANICAL TREATMENT OF STEEL 231

possible in order to produce a small amount of scale. Much better
control is obtained if the ash-pit is enclosed and forced draft is
used to burn the fuel. In this way about half a ton of fuel will be
required to heat a ton of steel from the atmospheric temperature
to that necessary for mechanical work, with a loss of about 4 to 5
per cent, of the metal as scale. The gases must necessarily leave
the furnace at a high temperature, and therefore it is not uncom-
mon to have boilers situated over the heating furnace, in which
steam is raised by means of the waste heat. With this economy
the amount of fuel chargeable against heating the steel will be
from 350 to 450 pounds per short ton of steel heated.

Regenerative Furnace. — If the heating furnaces are fired with
producer gas and the regenerative method is employed, we get a
far better control of the temperature and of the reducing influence
of the furnace gases. By this means a short ton of steel may be
heated with from 150 to 200 pounds of fuel and with a loss of
metal of from 1 to 5 per cent, by oxidation.

Continuous Furnaces. — Billets and other small pieces may be
heated in furnaces whose action is continuous. Such a one as
this is shown in section in Fig. 180. Along the hearth stretch
two lines of pipe, which are kept cool by a stream of water inside.
Upon the pipes is laid a long series of billets, which are gradually
moved forward toward the end at which the gas and air enter. In
this way the flame is always met by colder material and finally
leaves the furnace at a relatively low temperature. As the gases
pass out, they go through a series of pipes, B, B, around which
circulates the air that is afterward led to the fire and used for com-
bustion. As soon as the billet nearest the fire end is heated to the
desired temperature, a new one is pushed in at the bottom, caus-
ing the hot billet to be shoved onto the inclined plane, whence it
rolls out of the furnace to the point A, whence it is transferred to the
rolling-mill. In this type of furnace a short ton of steel may be
heated with from 120 to 145 pounds of fuel, with a loss in weight
of less than 1 per cent, by oxidation.

Soaking-Pits. — Ingots with molten interiors must be put in
some form of furnace in which they will stand upright until they
have solidified throughout and are ready to roll, in order that the
pipe may form in the upper portion. The type of furnace used
for this is known as a soaking-pit and is shown in Fig. 181.
The original intention of soaking-pits was to have the heat in the

THE MECHANICAL TREATMENT OF STEEL

233

ingot itself bring the interior of the furnace and the mass of metal
to the desired temperature ; but this is not found practicable in the
United States, and soaking-pits are usually heated by regenerated
gas and air. The ingots must be kept in these soaking-pits long
enough to be entirely solid in the interior, and for this purpose at

•rt-^+lv^v/:;.:^.^;;:^

FIG. 181. — BRICKWORK OF REGENERATIVE GAS SOAKING-PIT.

least 55 minutes are required for 3-ton ingots when stripped and
charged as soon as possible after teeming, and more for ones of
larger size.

Furnace Bottoms. — Heating-furnace bottoms must be of some
material not readily corroded by oxide of iron scale, and basic bot-
toms are very commonly employed with success. Heating-fur-
naces are also frequently supplied with a tap-hole, from which the
slag, composed chiefly of oxide of iron, can be tapped at intervals.
Soaking-pit bottoms are frequently covered with a layer of coke
breeze to absorb the slag and prevent corrosion of the furnace
bottom. This is shoveled out when an opportunity is afforded,
and new breeze substituted, or is knocked out through a hole in the
bottom, for which see Fig. 181.

Heating Practice. — In heating steel for rolling, the lower the
temperature the better will be the quality of the product. On

234 THE METALLURGY OF IRON AND STEEL

the other hand, if the metal is to undergo many passes before it
receives another heat, it must be correspondingly hot, in order
that the finishing temperature may be high enough to avoid ex-
cessive power for reduction. There is no doubt that rolling tem-
peratures at the present time are higher than they should be, for
the metal when finished should be only just above the critical

FIG. 182. — CHARGING A SLAB INTO A HEATING FURNACE.

temperature of the steel. Until within recent years no suitable
pyrometers for measuring the temperature have been available,
and the temperature for drawing the material is judged by eye, so
that no figures can be given. There can be no doubt that more
careful attention to this point will result in less waste in rolling
(on account of the production of cracked or second-class material
because of too hot steel at the starting) and in the production of
a higher quality of steel.

It must be remembered, however, that the steel must be hot
enough to cause it to weld together wherever it has become cracked.
'This is especially to be observed in low-carbon steel whose welding,
as well as its melting-point, is higher than that of high-carbon steel.

THE MECHANICAL TREATMENT OF STEEL 235

The casting temperature and the absence of ingotism which the
author has discussed elsewhere is probably more important than
any other factor in preventing cracking during rolling, as properly
made steel can stand without injury a high temperature which
would be very harmful otherwise. High-carbon steel is very
delicate to roll especially when the silicon is also high.

REFERENCES ON MECHANICAL TREATMENT

See especially No. 31 and the current technical literature,
especially Nos. 8, 9 and 12.

IX
IRON AND STEEL FOUNDING

FOUNDING is a mechanical art, and consists in pouring melted
metal into a mold of any desired size and form, which the metal
assumes and retains when cold. The mold is made of some kind
of sand, with rare exceptions to be mentioned hereafter. The art
is a very complex one, added to which it is now passing through an
important transition period in which science is very rapidly taking

Sweeping up a mold.
FIG. 185. — VIEW IN AN IRON FOUNDRY.

the place of rule of thumb. It is impossible to treat the subject
adequately in a single chapter, but several books are now avail-
able, to which foundrymen, metallurgists and chemists are referred,
and which are also recommended to all engineers, to whom a
knowledge of the art is of prime importance.

236 •':

IRON AND STEEL FOUNDING

237

THE MAKING OF MOLDS

There are various kinds of sand molds made for foundry work,
but the three principal kinds are loam molds, dry-sand molds, and
green-sand molds.

Loam Molding. — In molding with loam, sand is usually built
up into the required shape by hand, aided by machines. Fig. 185

FIG. 186. — MACHINE FOR FORMING THE TEETH OF A BEVEL-GEAR.

shows the molding of a gear in which the parts are built up of
brick and sand and then ' swept' into the proper shape by means
of the wooden sweeps. Large wheels and gears are often swept
up in this way, the teeth being formed subsequently by means of a
small pattern that is moved around as the molder progresses, or

238

THE METALLURGY OF IRON AND STEEL

by means of a machine, as shown in Fig. 186. In the case of a gear,
the arms are usually formed by placing within the swept-up mold
forms of sand known as l cores/ as shown in Fig. 187. Loam
molding is common in iron foundries, but almost never used for
steel castings.

Pattern-Molding. — To only a limited class of work is loam
molding applicable, and the commonest manner of making a mold
is to press or ram sand around a pattern, which is subsequently

FIG. 187. — PLACING CORES IN A MOLD.

removed, leaving the desired cavity. Usually the pattern is en-
closed by a l flask ' much larger than itself, between which and
the pattern the damp sand is rammed. The pattern (sometimes) is
split into halves, one half being in the lower part, or 'drag/ of the
flask, and the other half being in the upper part, or 'cope.'1 The
cope is now taken off and turned upside down, after which a lifting-
screw is inserted into each half of the pattern in turn, by means of

1 The old English word ' cope/ meaning a covering for the head, which
has now largely been replaced by the name ' cap.'

240

THE METALLURGY OF IRON AND STEEL

which it is drawn from the sand ; and when a ' gate ' is cut through
the cope, the flask is again fastened together, and a receptacle is
formed of the shape of the pattern into which the metal may be
poured.

The art does not consist of these simple operations alone, how-
ever, for in drawing the pattern from the sand, even though the
lifting-screw be lightly tapped with a hammer in four horizontal
directions to loosen the pattern, the slightest tremble of the mold-
er's hand, or of the crane used for lifting, may cause the sand to be

FIG. 192. — PATTERN IN SAND.

broken in places, and the chief skill of the molder as well as a large
share of his time is employed in repairing the damage thus pro-
duced. Furthermore, the mold may be 'washed/ that is, painted
inside; the proper cores must be put in place; parts of the sand
liable to drop off must be nailed in place with thin, large-headed
wire nails thrust in with the thumb ; before the pattern is taken from
the sand the cope must be l vented/ that is, made porous, by
jamming a wire into it many times and pulling it out again, so that
the air and gases will escape when the metal is poured in; and so on.

IRON AND STEEL FOUNDING

241

Furthermore, it may readily be imagined that the parts of the
pattern shown in Fig. 193 might be of such a shape, with flanges
on the bottom, or something of that kind, that they could not be
drawn without breaking the sand. In the case of such a design
the pattern and flask must be split into three or more parts,1 or
else a core must be put in to make an offset. It will be evident to

Cheefc

Cheek

Drag
FIG. 193. — SECTION OF FLASK AND PATTERN.

every engineer that he will have to pay more for making a casting
so designed.

Ramming. — In pattern molding, it is essential that the pres-
sure of the sand around the pattern shall be nearly uniform in all
places; because (1) when the metal is poured into the mold, it
drives out the air already there by forcing it through the inter-

1 The bottom and top parts being still known as the 'drag' and 'cope,'
respectively, while the intermediate parts are known as 'cheeks.'.

Finishing Trowel

Yankee

Finishing Trowel

Lifter

Bench Lifter

Finishing Trowel

Square Trowel

FIG. 194.

Slick and Spoon

IRON AND STEEL FOUNDING 243

stices between the particles of sand, and if the sand is too hard in
any place, the pressure of air collected there is liable to form a de-
pression, or 'scab/ in the casting; and because (2), if the sand is
too loose in any place, the pressure of the metal upon it is liable to
'swell' it outward and thus cause an enlargement of the casting at
that point. To obtain uniformity it is necessary that the sand be
packed around the pattern, and not the pattern pushed into the
sand. This packing is accomplished by the hands for the sand
immediately adjacent to the pattern, and by rammers for the lay-
ers further distant. In the case of bench molding hand rammers
are used, and for making larger molds on the floor long iron ram-
mers are employed. The molder's skill is shown in applying the
proper amount of power in ramming each different kind or part of
pattern.

Dry-Sand Molds. — After ramming up the mold, drawing the
pattern and applying the 'wash/ the mold may be used green
(when it is called a 'green-sand mold'), or it may be put in the
ovens and dried (when it is called a ' dry-sand mold ') . The drying
has the effect of driving off the moisture and leaving a firm, hard
mass, semi-baked into a sort of brick. Sand for these molds should
have slightly more clay than for green-sand molds; otherwise,
instead of baking into a hard mass, they would be liable to crum-
ble with the heat. The temperature of drying-ovens should be
about 350° to 400° F. (170° to 200° C.), and they are heated by
coke, coal, gas, or oil. If the temperature is too high, the mold
will be burned, that is, it will crumble under the fingers after drying;
if not hot enough, the mold will not be baked hard. The length
of time in the oven will depend upon the size of the mold, and will
vary from an hour or so to a day or so. During the drying the
molds are liable to shrink somewhat, due to the action of the clay
in binding together.

Green-Sand Molds. — Green sand requires less clay than dry
sand, because it has a certain coherence due to its dampness.
Many natural sands are found suitable for both the green-sand
mold and the dry-sand mold, or they can be made up as desired
by mixing a good clay with a sand rich in silica. Green-sand
molds must not be rammed as hard as dry sand, so the moisture
may more readily evaporate.

Washes. — For iron castings the common wash is graphite
dust, which is made up into a paint with water and applied with a

IRON AND STEEL FOUNDING 245

brush or dauber to the inside of a dry-sand mold before it goes to
the oven. In the case of a green-sand mold, pulverized coal is
dusted onto the surface through a piece of cloth, and then spread
uniformly with the l slicker/ In the case of a dry-sand steel casting,
the wash is composed of pulverized silica rock, running from 98 to
99 per cent, silica, which is made up into a thick paint with water,
thickened with molasses, and applied to the inside of the mold
with a brush or dauber before the mold is dried. Green-sand
molds for steel castings cannot ordinarily be washed. Graphite
washes cannot be used for steel molds, because the hot metal
attacks the graphite and becomes rough upon its surface.

The functions of washes are : (1) To make a very smooth face
on the sand, so that the surface of the casting shall be smooth
(this they accomplish by the very fine size of their particles) ; (2)
to give a surface that shall resist the melting and chemical action
of hot metal, and so be more easily cleaned of sand.

Skin-Dried Molds. — The inside surface of green-sand molds is
occasionally dried by painting or spraying it with some inflammable
liquid, such as gasolene, and then applying a match. This is more
common in steel-foundry than in iron-foundry practice, and pro-
duces a little better surface to the casting. Herbert B. Atha
patented a mold wash for green-sand steel castings which would
enable them to be skin dried.1 The formula for this wash was
devised by me with the aid of Parker C. Mcllhiney. It consists
of ordinary gasolene in which is dissolved as much rosin as it will
take up without warming. The rosin increases the specific gravity
of the gasolene so that it forms a paint with the silica wash, which
would otherwise sink to the bottom and not adhere to the brush.
After the wash is applied, it is touched with a match, the burning
resulting in giving a dry skin to the mold and leaving it coated with
the silica. The rosin does not completely burn off, but binds the
sand together and gives a tough skin, so that the sand is not so
liable to drop when the cope is turned over to place it upon the
drag.

Dry- versus Green-Sand Molds: For Iron Castings. — (1) Dry-
sand molds are often cheaper to make and require less molding
skill, because the sand does not have to be tempered so carefully,
that is, brought to the proper condition of dampness, since the

»U. S. Letters Patent No. 686,189.

246

THE METALLURGY OF IRON AND STEEL

moisture is eventually to be driven off by the drying. In green-
sand molds, if the sand is too wet, it is liable to ' cut' (be eroded by
the stream of metal) and get dirt into the casting, and also to be
impervious to the gases. (2) The sand requires less care in ram-
ming, because, whether too hard or too soft, the expansion and
contraction in drying will adjust its firmness and porosity. (3)
Furthermore, the dry sand is stronger, which is an advantage,
especially in large castings or when the sand is liable to be under

FIG. 196. — CORES FOR FORMING THE INSIDE OF A GAS-ENGINE
CYLINDER CASTING.

pressure from the metal, or to have the metal fall upon it from a
height. (4) Dry-sand castings are also more liable to be sound,
because there is less gas in the pores of the sand.

The disadvantages of dry-sand castings are: (1) The mold is
liable to shrink during drying and therefore be less true to the pat-
tern; (2) the castings are more liable to 'check' (that is, crack in
cooling), because the mold is firmer and so does not give way so
easily to the crushing action when the casting contracts; (3) molds
exposed to the direct action of the flame during drying, or heated

IRON AND STEEL FOUNDING

247

too hot, are liable to be burnt and therefore rendered useless, caus-
ing a loss; (4) in handling, the molds are liable to be damaged;
and, furthermore, it is more costly to repair a dry mold than a
damp one, because the adjacent sand must first be damped, the

FIG. 197. — CORE.

damage repaired, and then a flame applied to dry the wound; (5)
it takes longer to complete an order. The actual cost of heat is
not very great, and usually is less of an item than the extra labor
of handling for drying.

For Steel Castings. — Dry-sand steel castings have a surface
much superior to those made in green sand.1 They are also
stronger and more liable to be sound. Soundness is much more

FIG. 198.

difficult to obtain in steel castings than in iron castings. Green-
sand molds, however, have the great advantage of allowing their

1 The reason that dry-sand steel castings have a better surface is because
they can be washed. In practice, however, the drying and washing is often
improperly performed; therefore green-sand castings often have the better
surface.

248

THE METALLURGY OF IRON AND STEEL

sand to crush more easily when between two parts of the casting
that are being drawn together by the shrinking of the metal.
This is doubly advantageous in steel work, because steel shrinks
twice as much as cast iron and is therefore more liable to checking.

FIG. 199.

Special means may be employed for making the sand easily col-
lapsible after the metal is poured, such as mixing with it an in-
flammable substance like flour, chopped hay, hay-rope, sawdust,
etc., which burns away after the hot metal is poured in.

FIG. 200.

Cores. — The function of cores has already been referred to:
they are set inside the mold proper to assist in forming the metal.
The commonest use for them is to extend through a casting in some
place to make a hole, as, for instance, the inside of a cylinder, the

IRON AND STEEL FOUNDING

249

bore of a pulley, etc. In this position they are subjected to
great crushing strain when the metal shrinks, and therefore the
bond which keeps the sand together, consisting of linseed-oil, or
flour paste, a patented core compound, etc., must be of such a
nature that when subjected to the heat of the liquid metal it will
burn away and allow the sand to disintegrate, which both prevents
it bursting the shrinking casting and permits of its being more
easily cleaned out. Cores are often built up around an iron pipe

FIG. 201. — OVEN FOR DRYING SMALL CORES.

riddled with holes, so that the gases formed may readily escape
through this 'vent.' In the case of large cores, the pipe is fre-
quently wound with hay-rope, or some similar material that will
burn away and make the sand more collapsible. Some cores have
coke breeze or cinder in the center to make them light as well as
porous.

Cores are supported sometimes by being set in the drag, some-
times by being hung in the cope (see Fig. 202) ; but it is more com-

250

THE METALLURGY OF IRON AND STEEL

mon to have a hollow adjunct to the mold, known as a ' core-print/
into which an extension of the core fits (see Fig. 190). Some-

FIG. 202. — CORES HUNG FROM THE COPE.

FIG. 203. — CHAPLETS.

times both ends of the core are so supported, and sometimes only
one end is thus supported, while the other end rests upon a metal

FIG. 204. — POSITION OF MACHINE WHEN PRESSING THE DRAG.

FIG. 205. — POSITION OF MACHINE WHEN PRESSING THE COPE.

FIG. 206. — CUTTING SPRUE WITH TUBULAR SPRUE CUTTER,

PIG. 207. — RAPPING THE PATTERN BEFORE SEPARATING THE MOLD.

IRON AND STEEL FOUNDING

253

chaplet that is absorbed in the casting when the metal is poured.
Cores are often dried, lest their gases make the casting unsound
or cause it to blow, that is, boil with the rapid escape of gas through
the metal.

Chill Molds. — It is often desired to chill certain parts of a
mold, or cool them more rapidly than the remainder, in order either
to make a thick part of a casting solidify as soon, or sooner, than
the thinner portions, or else to produce white cast iron at that
point. The former may be desirable in the case of either an iron

FIG. 208. — STRIPPING PLATE
MOLDING MACHINE.

FIG. 209. — STRIPPING PLATE
MOLDING MACHINE.

or steel casting, because the shrinkage cavity occurs in the last
portion to freeze, and therefore hastening local freezing may be
necessary to bring the pipe into the riser or feeder. The latter
applies only in iron-casting work in which it is desired to make
the outside of a casting very hard. The chilling is usually accom-
plished by embedding pieces of metal in the sand, against the face
of which the casting is poured. This metal is oiled, blackened or
' washed/ so that it does not stick to the casting.

Permanent Molds. — A great deal of expense in foundry work
is due to the fact that a sand mold must be made anew for every
casting, and the subject of permanent molds has occupied the
attention of foundrymen for a great many years without the prob-

FIG. 210.

FIG. 211.

IRON AND STEEL FOUNDING 255

lem being solved. When a casting is knocked out of the mold,
the sand is usually knocked out also and its form destroyed. In
the rare case of a smooth cylinder, or something of that kind, the
casting may be withdrawn without damage except to the face of
the sand, and this can sometimes be repaired and swept up anew
without reforming the entire mold. Again, molds for railroad car
wheels, which have a metal ' chill' all the way round the tread and
flange, in order that the cast iron may be white at that point to
withstand the grinding action on the rails, have a certain amount

FIG. 212. — ROCK-OVER MOLDING MACHINE.

of permanency. Finally, molds carved out of carbon which has
been preheated to a very high temperature are said to withstand
the action of the melted metal and to last for a large number of
castings.

Gated Patterns. — Where the castings are very small, a large
number of them will be made into one pattern, fastened onto a
common 'gate' through which they are poured, which produces
very great economy in molding.

Molding Machines. — At the present time various types of

Upper head swung back to receive flask. Drag patterns on upper head.
Cope patterns on lower head. Yoke swung by power.

Sand frame on flask.
FIGS. 213 AND 214. — VIBRATOR MOLDING MACHINE.

Mold removed, flask frame raised, showing method of drawing cope patterns.

Mold lowered away, drawing drag patterns. Drag patterns may be

returned with absolute accuracy.
FIGS. 215 AND 216. — VIBRATOR MOLDING MACHINE.

258 THE METALLURGY OF IRON AND STEEL

molding machines are being extensively introduced into foundries,

in order to save some of the labor or skill required in molding, or
both. The simplest form of machine is the 'squeezer/ which
may be described by reference to Fig. 204. The correct amount of
sand is poured into the flask and by means of a long lever the
'presser board' is forced down on top of this sand, squeezing
it around the pattern and producing a half mold. In taking
the flask off the pattern by hand, damage may be done the sand,
and the molder's skill is still required to repair it.

: : ^l^i^HR

FIG. 217. —VIBRATOR.

The stripping-plate machine obviates part of this difficulty,
however. In this type the half pattern may be pushed up or
down through a close-fitting hole in a plate known as the ' strip-
ping-plate' (see Fig. 211). After the sand has been rammed
around the pattern, a lever draws the pattern down through
the stripping-plate. As this drawing is mathematically exact,
no damage results to the sand and no repairs to the mold are
necessary, so that unskilled laborers may be employed for the work.

A still further extension in the line of machines is the ' vibra-

IRON AND STEEL FOUNDING

259

tor/ whereby the pattern is vibrated an extremely small amount,
some 5000 to 30,000 times per minute, during the drawing of the
pattern. No stripping-plate is necessary to separate it from the
sand, since the vibrator frees it perfectly and without damage.

In all these molding machines the operation may be conducted
either by means of levers, or by a mechanism operated by hydrau-
lic or pneumatic power, and several hundred patterns may be
made per day by one man who is very little, if any, more skilled
than a common laborer; and machines are made in which castings

FIG. 218. —CORE MACHINE AND CORES.

weighing from a couple of ounces to several hundred pounds may
be molded with great economy.

Core Machines. — There are also on the market several machines
for making cores, an example of which is shown in Fig. 218.

Multiple Molds. — During the past year or two a new type of
molding, known as ' multiple molding/ has come into use, in which
several flasks are placed in a pile and poured through a common
gate of sprue, as shown in Fig. 219. This type of molding saves
mold costs, flasks, sand, floor space, and the weight of metal
wasted in sprues, and many difficulties have been overcome, so

260

THE METALLURGY OF IRON AND STEEL

that the castings are now made accurately to size and good in all
respects.

Shrinkage. — A bar of cast iron 12 in. long will contract about
0.125 in. during solidification and cooling (i.e., it will be about llf
in. long when cold. See page 346 for further details), while a bar

FIGS. 219 AND 220. — MULTIPLE MOLD AND CASTING.

of steel will contract about twice as much. In both cases the
contraction in sectional dimensions will not be as great as in length.

DESIGN OF PATTERNS

The foregoing description will show what a great financial
advantage it is to a purchaser if he designs castings that can be
easily molded, and if he can order a large number of castings of
exactly the same design. It is certain that a hundred castings
of one design can be made with very much greater cheapness than
the same number all of different designs, and of this economy the
purchaser obtains his full share, because the foundry is glad to
encourage such a customer and to make concessions in order to do
his work. I cannot recommend too strongly to engineers the
practice of making the castings in all similar machines inter-
changeable, both for the sake of economy and of avoiding some

IRON AND STEEL FOUNDING 261

delay and expense in replacement after a breakdown. The correct
design of castings is furthermore one of the most important
branches of engineering work, since the number of castings used
is almost one-half of the total number of pieces used in engineering
work, while their weight is equal to about one-sixth of the weight of
all the iron and steel employed. The following, general hints are
therefore offered to assist in this design ; but each casting is a study
in itself, in order that the various desiderata referred to may be
obtained.

To Avoid Checks. — The commonest error in engineering de-
signs of castings is to make the corners too sharp, which makes
them very liable to check, because of the crystalline character of
iron and steel. This is the more important, because the greatest
leverage comes at the corners, which therefore should be made as
strong as possible. Metals are crys-
talline substances and the crystals grow
during solidification. As solidification
usually extends from the surface in-
ward, the crystals grow in a direction
perpendicular to the cooling surfaces.
As shown in Fig. 221, this results in a
line extending inward from all corners,
marking the junction of many crystals.
As the junction lines of crystals are FIG. 221.

not as strong as the crystals themselves,

this makes a line of weakness on corners, which is the more marked
the sharper the corner is. In case a casting is to be machined,
it is much better to put a large fillet in all the corners, even if the
rounded metal must be cut away later, as greater strength is
obtained in this way.

The checking of castings comes from the strain produced by
the shrinkage of the metal tending to crush the sand. This is
the more intense the greater the distance is between the two
crushing parts, because they must approach each other by an
amount exactly proportional to the length of metal between them.
It is therefore wise, wherever possible, to avoid long lengths of
metal connecting two parts which project into the sand.

Unequal cooling strains will also cause a check. This may be
illustrated by a pulley with thick arms and a thin face. The face
will solidify first and therefore yield very little to the subsequent

262 THE METALLURGY OF IRON AND STEEL

shrinkage of the arms. Moreover, the face, being cooler, will be
stronger and the tendency will be for the arms to tear themselves
in two. To avoid this it is common practice to chill the arms
either by setting metallic pieces in the mold, or by means of a
' water gate/ A water gate is a loose column of coke molded into
the sand of the cope down which water may be poured.

To Avoid Shrinkage Cavities. — The formation of a ' pipe' or
shrinkage cavity has already been explained. Such a defect in a
casting would be intolerable, and is commonly avoided by having
a reservoir of metal situated above the casting proper and large
enough to keep it supplied with molten metal until it has com-
pletely solidified. This reservoir is known as the ' riser/ or
' header' (sometimes merely 'head'), or 'feeder.' The riser is
included as a part of the mold when it is made, but is cut off
the finished casting and used over again as scrap. Sometimes
castings are so designed by engineers that a heavy section of metal
must be molded underneath a thinner section. As the thinner
section will solidify first, it cannot l feed ' this lower, heavy section ;
and therefore a special form of riser is required, or else the heavier
section must be artificially cooled.

CUPOLA MELTING OF IRON FOR CASTINGS

Iron for castings is melted either in the cupola or the air fur-
nace, although 'direct castings,' i.e., castings made from the metal
just as it comes out of the blast furnace, are used in many cases,
and especially for cast-iron ingot molds at steel-works. There
are men with sufficient expertness to be able to judge by eye the
character and the analysis of the liquid iron as it flows from the
furnace, and this is necessary where direct castings are to be made,
because the metal may vary greatly and without warning from
one cast to another. In this work, however, metal mixers are
sometimes used, similar to those at steel-works.

The Cupola Furnace. — The design and principle of operation
of the cupola furnace bears some similarity to that of the blast
furnace, the chief difference being that the coke of the cupola fur-
nace is desired only for its melting influence, and that the only
chemical reactions are minor ones and unintentional. The cupola
affords the cheapest method of melting metals, because there is
direct contact between the metal and fuel and therefore the maxi-

Layer of
Iron

Layer of
Coke

Layer of
Iron

Layer of
Coke

Layer of
Iron

iim

Iron

FIG. 222. — IRON CUPOLA.

264 THE METALLURGY OF IRON AND STEEL

mum absorption of heat. The most usual amounts of fuel burned
will be from one-fourth to one-twelfth of the weight of the iron
melted, the former figure prevailing where exceedingly hot metal
is desired — as, for example, for very small castings for malle-
able cast-iron work — and the latter figure where the melting is
continued for several hours and the metal is not made very hot,
but is to be poured into large castings.

Cupola Zones. — The cupola should be so operated that cer-
tain well-defined zones of action are maintained, in order that
rapid, hot and economical melting may be obtained and that loss
by oxidation may be small. If proper conditions prevail all of
these desiderata may be obtained together, while, if otherwise,
wasteful methods may be accompanied by slow, irregular melting
and 'dull iron/ i.e., iron not sufficiently hot. The management
of the cupola is too often neglected and left in the hands of men
who understand nothing of its proper operation, and whose only
skill consists in knowing how to perform certain manual opera-
tions. If the iron is coming too cold (which may actually be due
to too much coke having been used in the charges), their stock
remedy is to put in more coke. The result is slow melting and
probably still colder iron. At once an earnest complaint goes to
the office of the bad quality of the coke. To get faster melting,
however, the blast pressure is increased ; now the iron comes faster
and hotter, and the office is informed that "we have worked off
that bad lot of coke." If the manager becomes doubtful and
orders less coke used, every wrong thing that happens in the foun-
dry thereafter is blamed to that order, until the manager decides
that coke is cheaper than dissatisfaction and tells the cupola man
to follow his own discretion.

All this is wrong. The cupola deserves the oversight of a man
who is capable of understanding its operation and who will give
real thought to it, and not be satisfied with blind rules of thumb.

The zones of action which, I have said, are so important in this
connection are, beginning at the bottom of the cupola and going
upward: (1) The crucible zone, or hearth; (2) the tuyere zone;
(3) the melting zone; and (4) the stack. The cupola is filled with
alternate layers of coke and iron, as shown in Fig. 222, x and the
different zones are produced by the action of the blast and the

1 Except, of course, during the intervals of starting up arid blowing out.

FIGS. 223 TO 227. — FOUNDRY LADLES.

266 THE METALLURGY OF IRON AND STEEL

heat upon these different layers. The thickness of these layers
should be the same in large as in small cupolas.

Crucible Zone. — The crucible extends from the bottom of the
cupola to the level of the tuyeres. The sole object of this part is to
form a place in which the iron and slag may collect after they have
melted and trickled down to the bottom. If the tap-hole is kept
open all the time and the metal allowed to flow out of the cupola
and collect in an outside ladle as fast as it melts, the crucible zone
will be very shallow, and the tuyeres will be situated not more
than two to five inches above the bottom. If, on the other hand,
the crucible is used as a reservoir for a large amount of metal, the
tuyeres are placed correspondingly high. Hotter metal may be
obtained by collecting the iron in an outside ladle.

Tuyere Zone. — The tuyere zone is the place in which the blast
comes in contact with, and burns, the red-hot coke. It is the zone
of combustion, and all the heat of the operation should be pro-
duced in this place. It is of course situated near the tuyeres and
wherever the blast may come in contact with coke. As there is
always a column of coke extending from the melting zone to the
very bottom of the cupola, combustion will begin immediately
above the reservoir of melted metal. The upper limit of the com-
bustion zone will depend upon the pressure of blast, because the
greater the blast pressure, everything else being the same, the
higher will it extend its zone of combustion. The blast pressure
should be such, however, that the top of the tuyere zone, or zone
of combustion, should never be more than 15 to 24 in. above the
uppermost tuyeres.1

Melting Zone. — The melting zone is the space in which all
the melting of iron takes place; it is situated immediately above
the tuyere zone. During the melting the iron is supported on a
column of coke which extends to the bottom of the cupola, and
which is the only solid material below the melting zone. When
each layer or charge of iron enters the melting zone it should be
about 15 to 24 in. above the uppermost tuyeres. As fast as it
melts it trickles down over the column of coke to the bottom. It
takes about five to ten minutes for each layer of iron to melt, how-
ever, and during this time the column of coke is burning and sink-
ing. Therefore, the last of the iron will melt at a point about

1 There are sometimes two rows of tuyeres in cupolas; see page 268.

IRON AND STEEL FOUNDING

267

7 in. lower than the first. Consequently, the melting zone over-
laps the upper limit of the zone of combustion. If the layers of
iron and coke are properly proportioned to the pressure of blast,

Tuyere
Tap Hole

FIG. 228.

FIG. 229.

each charge of iron will enter the top of the melting zone just
before the next previous charge is completely melted at the bottom,
and thus a continuous stream of iron will collect in the crucible
or run from the tap-hole. Also, the coke burned from the column

268

THE METALLURGY OF IRON AND STEEL

will be exactly replenished each time by the layer of coke coming
down, and the position of the melting zone, which is the important
consideration, will be maintained within constant limits.

The actual position of the melting zone may always be learned
when the cupola is emptied, because the iron oxide formed there
will corrode the acid lining, which will therefore be cut away some-
what at this point. Corrections may then be made, if necessary, in
the next charge of the cupola.

Stack. — The stack extends above the melting zone to the level
of the charging door. The function of this part of the furnace is

FIG. 230. — POSITIVE PRESSURE IMPELLOR BLOWER.

to contain material that will absorb heat and thus prepare itself
for the actions at lower levels, and that will also keep the heat
down in the melting zone as well as possible.

Tuyeres. — The blast enters the cupola through the tuyeres, of
which there is usually one or two rows. The position of the upper
row of tuyeres determines the position of the melting zone in the
cupola. Two rows of tuyeres give faster melting in the cupola
than one row, but cause greater oxidation and the consumption
of more fuel on the bed, because of the melting zone being higher
in the cupola.

Blast. — The blast pressure will depend somewhat upon the
size of the cupola, but the present prevailing opinion is in favor

IRON AND STEEL FOUNDING

269

of pressures not exceeding a pound, even for the very largest
cupolas, and diminishing to half a pound or so for the smaller
sizes. Fan-blowers are not approved of, because if they are
opposed by pressure in the cupola stack, they revolve with-
out blowing any wind. The common type of blower used in
America is of the two-impellor type, an example of which is shown
in Fig. 231. It takes about 60 cu. ft. of air to burn a pound of coke,
from which may be calculated the size of blower necessary for each

FIG. 231.

cupola, allowing about 50 to 100 per cent, excess for leaks and
incomplete combustion.

Makers of cupolas and blowers give all the necessary data in
their catalogues, but advocate too high blast pressures and vol-
umes, for obvious reasons. If the blast volume is too large, or the
pressure is too great, the position of the melting zone will be too
high. This means that the bed of coke must be larger to reach to
the upper level of the melting zone, which is wasteful. It also

270 THE METALLURGY OF IRON AND STEEL

means 'that the melted iron will have a greater height to drop
through. It therefore oxidizes more, corrodes the cupola lining
more, and consequently causes more waste of iron and more slag.
The volume of blast is the most important consideration, but this
is difficult to measure, so the pressure is the thing that is calculated
upon. It must be remembered, however, that this is only a
makeshift arrangement at the best.

Cupola Charge. — In the cupola is first placed shavings and
wood, on top of which is placed the bed of coke, which should be
large enough to reach 15 to 24 in. above the uppermost tuyeres
after the kindling is burned off. On top of this is placed a layer
of pig iron about 6 in. thick, then another layer of coke about 7 in.
thick, another layer of iron, and so on. The actual weight
of the coke for the bed and of coke and iron for each charge will
therefore depend directly upon the diameter of the cupola inside
the brick lining, which varies from about 32 to 120 in., or even
more in some cases. The weight of the coke in each layer will be
about one-sixth to one-twelfth of the weight of iron in each
layer. The tuyeres and front of the cupola around the tap-hole,
known as the 'breast/ are left open for an hour or so after the
kindling is lighted, in order that the draft may draw air in at that
point for combustion. When the kindling is thus burned off
and the bottom coke well lighted; the breast is closed and the
wind turned on. It is very necessary that the bed should be
well lighted and level.

Cupola Melting. — The heat now generated by the combustion
of coke begins to melt the iron, and in less than 15 minutes after the
wind is put on the metal should begin to run from the open tap-
hole. If it takes longer, then the coke bed was too high and waste-
ful. In another 8 to 10 minutes the first layer of iron should
be all melted. Now the second layer of iron lies upon the column
of coke, whose top should again be 15 to 24 in. above the upper-
most tuyeres. If the layers of coke are too thick there will be a
delay in the iron entering the melting zone and the extra coke
burned will not have been used to the best account. If the layers
of iron are too thick, the last of the layer will melt too near the
tuyeres, which will oxidize it excessively and make it cold. This
can be observed during the run by noting if the iron nms first hot
and then cold. It is very important to watch the flame that comes
off the top of the stack in the cupola. When the blast volume is

IRON AND STEEL FOUNDING 271

too large this flame will be ' cutting/ — i.e., oxidizing in character.
Too great oxidation may also be observed if sparks of burning iron
are projected from the slag-hole. If the layers of iron and coke
are both too thick, there may be a correct relation between the
weights of the two, but both of the irregularities mentioned above
will be observed. If the layers of iron and coke are both too thin,
we will have two charges of iron in the melting zone at the same
time, and this may be learned by watching the iron from the tap-
hole, because it will run at some times faster than at other times.
This does not produce such bad results, however, as having the
layers too thick. Of course, if very hot iron is required it will
be necessary to have thicker layers of coke, and slower melting
must be expected.

Chemical Changes. — As the iron drops down over the coke, it
absorbs sulphur, the exact proportion depending chiefly upon the
relative amount of coke and iron used and the per cent, of sulphur
in the coke. It will vary from 0.02 to 0.035 per cent, of the iron;
that is to say, if the pig iron charged contained 0.08 per cent, sul-
phur, there will be from 0.1 to 0.115 per cent, in the castings. The
sulphur in the first iron will be higher than in that of the middle of
the run, because of the extra amount of coke burned before the
iron begins to come from the tap-hole. The last iron will also be
somewhat higher in sulphur, because there is a larger loss of metal
during the last of the run, when the oxidizing conditions are more
intense, and therefore a concentration of sulphur. The best
practice is to cut the blast off progressively as there is less stock
in the cupola.

In many foundries it is customary to charge limestone, in the
form of oyster shells, marble chippings, or crude limestone, and
sometimes with it a little fluorspar (CaF2), into the cupola. The
amount of limestone varies greatly, but will average perhaps J to 1J
per cent, of the weight of the metal. This limestone fluxes the dirt
on the metal and the ash of the coke and carries off some sulphur in
the slag. Fluorspar makes a somewhat more liquid slag than
limestone alone and the more liquid slag is believed to absorb a
little more sulphur, and also to make the cupola 'drop' more
easily, i.e., dump its contents when the campaign is ended, and
the bottom is allowed to fall. It also cuts the lining more.

As the metal melts and falls from the melting zone down in
front of the tuyeres, it suffers oxidation, which carries iron oxide

272 THE METALLURGY OF IRON AND STEEL

into the slag and also burns up silicon. The melted metal there-
fore contains from 0.25 to 0.4 per cent, less silicon than the original
pig.1 In other words, if the mixture charged contains 2.25 per
cent, of silicon, the castings will contain 1.85 to 2 per cent, silicon.
Cupola Gases. — The gases coming out of the top of the cupola
charge consist principally of nitrogen from the air, while the
remainder is carbon dioxide — COa — and carbon monoxide — CO —
with sometimes a little free oxygen. The latter is evidence of a
'cutting flame' and shows too great oxidation in the melting
zone. Such a flame may be recognized without the aid of chemical
analysis after a little practice by means of the eye. It is ' sharper '
looking than a richer flame and burns close to the top of the stock.
One can identify it exactly by holding an iron rod in it for a
while; after the iron becomes red hot it will oxidize much more
rapidly in a cutting flame than in a reducing flame. A reducing
flame will usually not burn until it becomes mixed with the air
sucked in at the charging door. All the carbon monoxide that
goes out of the charge represents incomplete combustion and a
waste of heat. It seems to be impossible to prevent this here,
however, just as in the blast furnace, whose operations cupola
melting resembles in some general respects. Several analyses
of cupola gases are given in Table XVIII.

1 With good practice it should be no more than 0.30 per cent. less.

IRON AND STEEL FOUNDING

273

TABLE XVIIL— ANALYSIS OF CUPOLA GASES
COLLECTED ABOUT 3 OR 4 FEET BELOW THE CHARGING DOOR

Time Elapsed Since
Blast was Put on

ANALYSIS BY VOLUME

Oxygen
O

Carbon
Dioxide
CO2

Carbonic
Oxide
CO

Nitrogen (by
Difference)

Ratio
CO2 is to CO
as 1 is to:

10 minutes

0.0

13.8

9.9

76.3

0.717

1 hour 13 "

0.0

9.5

16.9

73.6

1.780

2 hours 17 "

0.4

9.2

16.6

73.8

1.804

3 " 13

0.0

6.7

21.7

71.6

3.239

4 " 15

0.1

7.8

22.3

69.8

2.859

1.8

7.6

15.5

75.1

2.04

1 hour 42

2.8

7.5

13.3

76.4

1.77

2 hours 50 "

1.9

7.1

15.8

75.2

2.225

3 " 40

0.1

7.2

19.0

73.7

2.64

38 "

0.0

10.2

14.6

75.2

1.431

3 "

2.9

5.4

14.3

77.4

2.65

0.2

13.1

7.7

79.0

0.588

3 " 10

0.3

10.3

11.7

77.7

1.136

0.0

7.1

15.4

77.5

2.17 •

1 hour 40

0.0

8.3

13.5

78.2

1.626

2 hours 40

0.0

8.2

12.1

79.7

1.475

3 " 40

0.0

6.0

15.0

79.0

2.50

0.0

13.0

12.6

74.4

0.97

1 hour 40

0.0

13.0

11.2

75.8

0.862

2 hours 40

0.0

8.2

20.0

71.8

2.44

3 " 50

0.4

6.0

22.1

71.5

3.683

4 " 43

1.2

5.1

21.2

72.5

4.157

1 hour

0.0

9.8

15.4

74.8

1.571

1 " 53 "

0.0

9.1

16.8

74.1

1.846

2 hours 45

0.0

8.8

16.8

74.4

1.91

3 " 45

0.0

7.5

19.7

72.8

2.627

4 " 45

0.0

7.5

18.7

73.8

2.500

ANOTHER CUPOLA

0.1

16.7

7.3

75.9

.437

Ihour 30

0.1

13.1

10.7

76.1

.817

2 hours 38

0.4

11.8

11.0

76.8

.932

3 " 20

0.0

12.8

7.7

79.5

.602

Burdening the Cupola. — It should be the duty of the foundry
metallurgist or chemist to learn from his records, or other ap-
proximations, the amount and analysis of all the metal in the

274

THE METALLURGY OF IRON AND STEEL

yard. The following table will, for example, show a convenient
form of this record.

TABLE XIX

KIND

Weight
Tons

Price

High Sulphur, Southern
High Silicon, Bessemer .
XNo. 1
No. 3 Foundry

500
60
100
150

0.70
2.50
3.00
1.75

0.100
0.025
0.030
0.070

1.50
0.07
0.80
0 30

0.30
0.60
1.25
0 60

$18.00
25.00
24.00
22 50

Ferrosilicon A

10.00

0.040

0 50

0.10

35 00

Ferrosilicon B

50 00

0 003

0 04

105 00

Machinery Scrap

100

1.70?

0.100?

1.00?

0.60?

19 00

Miscellaneous Scrap
Cast-iron Borings . . .

300
100

1.50?
1 50?

0.20?
0 20?

1.40?
1 40?

0.60?
0 60?

15.00
11 00

Steel Scrap

100

0 10

0 07

0 10

0 60

13 00

The price should always be in evidence. It should not be the
price at which the material was purchased but the market price
at the time the iron is to be used. For instance, if a large amount
of high grade pig iron had been contracted for a year previously
and if meanwhile the price of pig iron had been rising, the pur-
chase price of that pig iron would not represent its present value.
From the current numbers of such trade periodicals as the Iron
Age and the Iron Trade Review, one can always obtain the pre-
vailing prices for the different grades of iron.

Suppose now with these irons it is desired to burden a 72-in.
cupola with a mixture for making heavy hydraulic pumps for
which a satisfactory analysis might be 1.60 per cent, silicon,
0.70 per cent, phosphorus, less than 0.10 per cent, sulphur and
about 0.50 per cent, manganese. The first step is to calculate
the cupola charges, and the chemist knows by experience with
this particular cupola that it will lose 0.25 per cent, silicon and
0.10 per cent, manganese and it will gain 0.03 per cent, sulphur.
The average analysis of the mixture put into the cupola must then
be 1.85 per cent, silicon, 0.70 per cent, phosphorus, less than
0.07 per cent, sulphur and about 0.60 per cent, manganese. The
chemist also knows by calculation that about 5200 Ibs. of iron
will give a layer of the proper thickness in a 72-in. cupola. His
problem now is to make such a mixture of the available pig irons
that their collective weight will be 5200 Ibs. and their average
analysis as given above. Moreover if he is a good metallurgist

IRON AND STEEL FOUNDING 275

he must aim at using as large an amount as possible of the cheapest
materials.

He first considers the steel scrap, but knows he cannot use
very much of this because too much coke would be required for
getting iron of the requisite fluidity, but he estimates that 5 per
cent, (say 200 Ibs.) will not increase harmfully the fuel necessary.
This figure therefore comes at the top of his list (see Table XX).

Next he considers the use of machinery scrap, because he
knows that his miscellaneous scrap and borings are too uncertain
in analysis to be used in a mixture which must give pretty strong
and non-porous castings. One thousand pounds of machinery
scrap would be about 20 per cent, of his mixture, and he knows
from experience that this is a fairly satisfactory proportion for
scrap, so that figure goes down second in Table XX. The low
price of the High Sulphur Southern Pig tempts him, but he
realizes that he must offset the use of this material by some high
silicon low sulphur iron. And in casting about for such a one
he naturally considers first the X No. 1. He cannot use much
of this either because of its high manganese and it seems reason-
able to mix an equal amount of these two; the only question is,
How much of this mixture will the cupola stand ? To get an idea
of this, he first calculates their average analysis and finds it to
be 1.85 per cent, silicon, 0.065 per cent, sulphur, 1.15 per cent,
phosphorus, and 0.78 per cent, manganese. Evidently the phos-
phorus is the only element in this mixture that gives him difficulty.
Indeed, if that were not high he could make almost his whole
charge up of these two irons and the scrap. The phosphorus in
this mixture is 0.45 per cent, higher than that of his desired
mixture. Therefore he knows that he must use a good deal of
No. 3 Foundry iron to bring this element down. The phosphorus
in the No. 3 Foundry iron is about as much below the desired
phosphorus as that in the mixture of the High Sulphur Southern
and the X No. 1 is above it He must not forget, however,
that he has already used 1000 Ibs. of machinery scrap containing
probably 1 per cent, of phosphorus. Therefore he must use a
correspondingly larger amount of No. 3 iron to offset this also.
As a first estimate he therefore considers using 800 Ibs. of High
Sulphur Southern, 800 Ibs. of X No. 1, and 2400 Ibs. of No. 3
Foundry — that is, once and a half as much No. 3 as the mixture
of the other two. But a little reflection tells him that this mixture

276 THE METALLURGY OF IRON AND STEEL

is going to be too low in silicon, because the mixture of High
Sulphur Southern and X No. 1 only gave us 1.85 per cent, silicon,
while the No. 3 foundry and the machinery scrap are both below
that. There are then three ways open to him. He may use a
little Ferrosilicon A or he may pound up a little Ferrosilicon B
and dissolve it in the ladle of iron or he may use a little High
Silicon Bessemer iron. Either of these methods would do, but the
writer would prefer to use the High Silicon Bessemer because this
will have the effect of cutting the sulphur and phosphorus down
and the expense is practically the same. (It requires such a small
amount of ferrosllicon to give the desired silicon in the mixture
that the expense of using it is very small, in spite of its price.)
Consequently we put down the weights shown in the second column
of Table XX, and we now figure out the weight of silicon, sulphur,
phosphorus, and manganese in the mixture by the methods
indicated there, and the average percentage of each element.
The latter figures show us that the silicon is too low, and a simple
calculation shows us that we need 5 Ibs. more in the total weight
of silicon. We can get this by increasing the amount of either
High Silicon Bessemer or of X No. 1, and correspondingly de-
creasing the No. 3 Foundry. The High Silicon Bessemer has
0.75 per cent, more silicon than the No. 3, so it would take (5 Ibs.
-5-0.75 per cent.=) about 650 Ibs. change to make up the difference
in this way. The X No. 1 has 1.25 per cent, more silicon than the
No. 3, so it would take (5 Ibs. -7-1.25=) 400 Ibs. change to make
up the difference in this way. We naturally would prefer to use
the latter, being cheaper, and if we think we can stand all that
extra manganese in our castings we probably will do so; if not,
we will have to use altogether 1000 Ibs. of High Silicon Bessemer
and only 1400 Ibs. of No. 3 Foundry. We then make up a new
table similar to Table XX and figure out the average analysis as
before. It should now come about right.

IRON AND STEEL FOUNDING

277

<M O "tf O O rJH
1-4 CO <N O (M <M

i— 1 i-H

O CO

^ o
co

o C

(M 0 0 T* 0 CO

o o 01 co co d .

C5 CO

CO °

WEIGH

d TH d d I-H d

00 t^

CO O

co d

d
1

(M O CO O O O

d i> 10 •*' »o o'

i-H <N CO i-H

i— i *>

C5 ^

jj*

O O O iO O O
CO CO CO CQ CO CO

d d d i-I d d

If*

e^.

o o o o o t~>-

i— I O 10 00 CO O

d 1-1 i-J d d d

3
1

o o o o »o

l>. O O CO !>• C^l

d d d d d d

o d o o ^0 o

1-1 t^ l> O t^. »O

d I-H d co 1-1 <M

•

Mi!|i

»O

I !

1 i

09 3 'S fl

i r i J n

5 J « ., * «

13**''".*

M J^ W M ^ W

Total weights
Average percentage 2

278 THE METALLURGY OF IRON AND STEEL

LOSSf __ The loss in melting will average about 2 to 4 per' cent.
It is made up of the silicon burned and the iron oxidized and car-
ried away in the slag. There are other sources of loss in the
foundry, such as a second loss of metal remelted — the sprues,
risers, etc., which go back to the cupola in the form of scrap;
metal spilled during pouring (which may amount to as much as
5 or 6 per cent, more), etc. In some foundries it is customary to
pass the used floor sand through a magnetic concentrator, in
order to recover the pellets of iron spilled
during pouring, and important economy
is sometimes obtained in this way. The
total loss, that is, the difference in weight
between pig iron bought and castings
made, will probably be about 7 to 8 per
cent, of the weight of the iron bought.

Scrap Used. — Scrap pig iron is often
mixed with new pig iron for the manu-
facture of castings, both for the sake of
economy and because the scrap iron has a
somewhat closer grain or texture, which

FIG. 232. — MAGNETIC ,, ^ f ,-, . , i™

CONCENTRATOR. increases the strength of the mixture. I he

amount of scrap used will depend upon the

materials to be manufactured. Cast-iron pipe is usually made
without scrap, this industry amounting to between 500,000 and
800,000 tons per year in the United States alone. Stove foundries,
on the other hand, use a very large amount of scrap as a rule, and
jobbing foundries, in general, would probably use an average of
30 to 40 per cent, of outside scrap, besides the gates, sprues, bad
castings, etc., made in their own foundries. The total production
of gray-iron castings in the United States will represent about
75 per cent, pig iron and 25 per cent, bought scrap.1

Cupola Run. — The campaign of an ordinary foundry cupola
is only three or four hours long. As a general thing, the kindling
is started about noon and allowed to burn with a natural draft
until shortly after one o'clock, when the breast is closed and the
blast put on. Metal is then received until four or five o'clock in
the afternoon, when the last charge is melted. The supports are
then pulled out from underneath the door closing the bottom of

1 For these estimates, I am indebted to Henry M. Lane, editor of Castings,
in a private communication of Nov. 30, 1906.

IRON AND STEEL FOUNDING 279

the cupola, and the sand bottom, slag, coke, etc., left in the cupola
is allowed to drop and is quenched with water. In order to allow
plenty of room for the 'drop' to fall, the cupola is usually ele-
vated above the foundry floor.

COMPAKATIVE CUPOLA PRACTICE

After writing the foregoing discussion, some very valuable
figures on "Comparative Cupola Practice'7 were presented by
W. S. McQuillan to the Philadelphia Convention of the Ameri-
can Foundrymen's Association. The figures there given were
of very great interest and value (see the Foundry, July, 1907,
pp. 370 to 373), and confirm in a striking way the rules I have
laid down above. I have copied a part of this table in my ac-
companying Table XXI and added several lines to it. I have
also had all the calculations in the table checked up by two in-
dependent observers.

Fuel. — The first lesson we learn from the table is that a
mixture of coal and coke and inferior coke give slow melting
and a poor fuel ratio. Indeed, the work of the cupola using these
grades of fuel is so far inferior to the others that I have separated
them in Table XXI and omitted them from all my calculations.
After a very careful study of the figures, I am strengthened in the
opinion which I have long held and expressed that a mixture of
coal and coke has nothing to recommend it except a deceptive
first cost.

Tuyere Ratio. — The next most striking evidence produced
by the figures is the relation between the tuyere area and the
speed of melting: if we average up the iron melted per minute in
the cupolas whose area is less than 6.56 times the tuyere area,
we obtain a figure of 22.56 Ibs.; if we get the corresponding figure
for the cupolas with lesser proportionnte tuyere area, we obtain
18.57 Ibs. Indeed so striking is the relation that there is only
one exception, namely, cupola No; 8, and we need not look far
for a reason for the slow melting in this cupola. It is evidently
due to the short height of stack which causes the iron to reach the
melting zone before it has been sufficiently preheated. A large
proportionate tuyere area evidently means that the wind will pass
through the tuyeres with less resistance and a lower velocity.
That is to say, we will get more wind and it will not be driven so

280 THE METALLURGY OF IRON AND STEEL

TABLE XXI.— COMPARATIVE CUPOLA PRACTICE

Diameter of cupola, inches. . . .
Height of tuyeres from sand
bottom, inches

27
12

42
11

44
12

54
7.5

Height of charging door above
tuyeres, inches

106

120

139

109

103.5

Height of charging door above
tuyeres divided by diameter.
Number of tuyeres

4.0
6

3.4
6

3.3
6

2.5

1.9
6

Size of tuyeres, inches, vertical
X horizontal. . .

4x6

5x5

4x6

4^x13

Area of tuyeres, square inches .
Cupola area is how many times
tuyere area

144
3.97

170
5.661

144
9.62

336
4 53

348
6 58

Diameter of blast pipe, inches .
Blast pressure, ounces, 20 min-
utes after start

8
9

16
16

14
8

15
<U

17
13

Class of work made •<

Job

Plate

Light

Boiler

Stove

Lined up how often? Months

Job
6

& Plate
6

Plate
6

Weight of fuel on bed, pounds.
Weight of iron on bed, pounds.
Weight of fuel in charges sub-
sequent to the bed, pounds. .
Weight of iron in charges sub-
sequent to the bed, pounds. .
Total weight of fuel, pounds. . .
Total weight of iron, pounds. . .
Ratio of fuel on bed, above
tuyeres, to iron on bed, 1
is to

350
700

400
2,000
8,000

3 2

650
1,300

125

1,300
1,850
14,000

2 8

650
1,400

100

1,000
2,150
16,000

Q K

1,300
3,000

175

2,000
3,075
27,000

3 2

1,400
3,000

300

3,000
4,800
33,000

3 7

Ratio of fuel to iron, later
charges

4 4

10 4

10 0

11 4

10 0

One pound fuel melts how
many pounds iron

4 0

7 5

7 6

8 8

•_^
6 9

Kind of fuel used -j

Coke

Fuel weighed or measured
Height of fuel bed above
tuyeres, inches. . .

W
20

W
21 1

M
1^

W
24

Thickness of fuel, charges after
the bed, inches

8 4

7 0

Q 8

6 1

6 9

Thickness of iron charges after
the bed, inches

3 3

6 5

3 5

6 3

Time before iron comes after
blast is on, minutes
Time to melt each iron charge
after the bed, minutes ....

7
3 25

15
8 5

5 7

15
9

10
10

Total iron melted per minute,
pounds

123

1 Pjr;

22 *>

QOO

Total iron melted per minute,
per square foot cupola area .

30.98

23.20

18.19

21.30

18.87

1Two rows of tuyeres.

IRON AND STEEL FOUNDING 281

TABLE XXL— COMPARATIVE CUPOLA PRACTICE.— Continued

Diameter of cupola, inches ....
Height of tuyeres from sand
bottom, inches

54
14

56
25

58
16

60
2

60
12

72
24

Height of charging door above
tuyeres inches

113

141

112

114

142

Height of charging door above
tuyeres divided by diameter .
Number of tuyeres

2.1
6

2.5
12

1.6
6

1.9

1.9
6

2.U

Size of tuyeres, inches, vertical
X horizontal

10x7

6x12

7x12

4Jx7i

7x10

Area of tuyeres, square inches .
Cupola area is how many times
tuyere area

420
5.45

864
2 85

546

4.831

270
10 47

420
6 73

706
5 76

Diameter of blast pipe, inches .
Blast pressure, ounces, 20 min-
utes after start

18
14

24
16

20
16

18
8

24
13£

Boiler
&

Rolls

Pipe

Ft'g&

Plate

Sanitary
&

Elec-

Lined up how often? Months .
Weight of fuel on bed, pounds .
Weight of iron on bed, pounds .
Weight of fuel in charges sub-
sequent to the bed, pounds . .
Weight of iron in charges sub-
sequent to the bed, pounds . .
Total weight of fuel, pounds . . .
Total weight of iron, pounds. . .
Ratio of fuel on bed, above
tuyeres, to iron on bed, 1
is to

Radiator
10
2,000
6,000

450

4,000
8,500
64,800

4.3

8 to 10
3,000
9,000

450

4,500
12,000
100,000

5 0

Job
7 •
2,000
6,000

400

4,000
4,400
30,000

5 0

6
1,700
6,000

200

2,000
6,600
49,500

3 8

Plate
12
1,800
5,000

350

4,000
6,600
50,000

4 7

trical

6
2,500
7,000

600

7,600
12,700
127,500

5 4

Ratio of fuel to iron, later
charges

8.8

10 0

4 1

10 0

11 4

11 7

One pound fuel melts how
many pounds iron

7.5

8 3

6 8

7 5

7 6

10 0

Kind of fuel used •]

Coke

Solvay

Fuel weighed or measured ....
Height of fuel bed above
tuyeres, inches

M
32

14 1

M
30

W
22

Coke
W

Thickness of fuel, charges after
the bed, inches

10.4

9.6

8.0

3.8

6.5

7 8

Thickness of iron charges after
the bed, inches

8.4

8.8

7.3

3.4

6.8

Time before iron conies after
blast is on, minutes
Time to melt each iron charge
after the bed, minutes
Total iron melted per minute,
pounds

20
11
360

5
12
370

15
12
333

1
5.5
367

15
11
360

15
13.5
567

Total iron melted per minute,
per square foot cupola area.

22.64

21.60

18.14

18.80

18.40

20.06

JTwo rows of tuyeres.

(For averages, see next page.)

282 THE METALLURGY OF IRON AND STEEL

TABLE XXL— COMPARATIVE CUPOLA PRACTICE.— Continued

Averages

Diameter of cupola inches.. .

Height of tuyeres from sand
bottom, inches

9.5

12. 61

Height of charging door above
tuyeres inches

133

132

84 5

103

116

Height of charging door above
tuyeres divided by diameter .
Number of tuyeres

4.1
6

3.1
6

1.9"

2.1
6

2.52

Size of tuyeres, inches, vertical
X horizontal

4£x6

4x10

2Jxll

4x6

Area of tuyeres square inches

162

240

198

144

Cupola area is how many times
tuyere area .

4 96

5 77

7 68

12 56

6 56 l

Diameter of blast pipe inches

Blast pressure, ounces, 20 min-
utes after start

12 l

Class of work made . •<

Gas

Job

Pumps

Medium

Lined up how often''* Months

Engine
6

& Job
3

Light
6

Weight of fuel on bed, pounds .
\Veight of iron on bed pounds

450
1 000

1,000
4500

1,000
1 000

1,300
4000

Weight of fuel in charges sub-
sequent to the bed pounds

110

250

110

200

Weight of iron in charges sub-
sequent to the bed pounds

1 000

2000

1 500

2000

Total weight of fuel pounds

1 250

1 935

2 040

4 100

Total weight of iron pounds

8 000

12 000

14800

32 000

Ratio of fuel on bed, above
tuyeres, to iron on bed, 1
is to

3 5

8 6

1 6

5 0

4 2

Ratio of fuel to iron, later
charges

8 7

8 0

13 6

10 0

9 5

One pound fuel melts how
many pounds iron

6 4

6 2

7 2

7 8

7.31

Kind of fuel used •<

Fuel weighed or measured ....
Height of fuel bed above
tuyeres, inches

Light
Coke
M

Coal&
Coke
M

Coal&
Coke
W

Coke&
Coal
M

22 5

22 51

Thickness of fuel, charges after
the bed, inches

7 3

9 5

3 9

5 8

7 O1

Thickness of iron charges after
the bed, inches

4 7

5 3

6 41

Time before iron comes after
blast is on, minutes
Time to melt each iron charge
after the bed, minutes
Total iron melted per minute,
pounds

7.7
57

15
9.5
164

7
13.8
145

11. 31
9.46i

Total iron melted per minute,
per square foot cupola area .

10 2

15 53

11 54

21. II2

Including Al, A2, A3, A4.

2 Excluding Al, A2, A3, A4.

IRON AND STEEL FOUNDING 283

much to the center of the cupola. If the publication of these
figures did no more good than to point out the advantage of the
large tuyere area, they would have already contributed a very
great deal to the foundry industry. It may be said to be established
for the types of practice here exhibited that the collective area of
the tuyeres should not be less than one-sixth, and preferably not
less than one-quarter, of the horizontal area of section of the
cupola.

Height of Stack. — There is also an important relation between
the speed of melting and the height of the charging door above
the tuyeres divided by the diameter of the cupola. The average
speed of melting of the cupolas, where this ratio is greater than
2.5, is 24.12 Ibs. per minute, while the average speed of melting
of those whose ratio is less than 2.5 is only 19.15 Ibs. per minute.
There are only two exceptions to this rule: (1) Cupola No. 6 is
a fast melter, but this is doubtless due to the large proportionate
tuyere area. (2) Cupola No. 3 is a slow melter, but this is doubt-
less due to the very small proportionate tuyere area. With these
exceptions the rule is universal and a comparison of different
cupolas one with another only strengthens it;. for example, 4 with
8, the proportionate tuyere area being nearly the same; also
2 with 6, etc. A comparison of 3 with 9 is an apparent exception
which is perhaps explained by the low tuyeres in No. 9.

Blast Pressure. — The average speed of melting of the cupolas
with more than 12 oz. blast pressure is only 20.75 Ibs. per minute,
while that with less than 12 oz. is 21.53. This relation is not
so striking as to establish a rule, especially as a single cupola
(No. 1) throws such a large influence into the low-blast column.
Nevertheless, the evidence is sufficiently striking to discredit
the theory that higher blast pressure necessarily gives faster
melting. Indeed, if cupolas 3, 9, and 10 had a larger tuyere area,
we should expect an average result very favorable to low blast
pressure.

Height of Fuel Bed. — The original height of fuel bed is no
criterion with which to figure as it in many cases is raised or
lowered during the first few minutes of melting, and thereafter
occupies some other position. It is the thickness of the fuel and
iron charges after the bed which is the important consideration
and, as already observed, these should be regulated to about
6 to 8 in. for fuel and 6 in. for iron irrespective of the diameter of

284 THE METALLURGY OF IRON AND STEEL

the cupola. In case hot iron is desired the layer of fuel should be
at the upper limit of thickness and vice versa.

Speed of Melting. — The speed of melting is very important,
because everything else being equal the faster the melting per
square foot of cupola area the greater will be the efficiency of
operation of the cupola, and therefore of economy under the
given conditions.

OTHER MELTING FURNACES

Air-Furnace. — The air-furnace is a reverberatory furnace on
the hearth of which pig iron is melted by radiation from the flame
of a soft-coal fire, or, more rarely, from a gas flame. The furnace
is charged by means of a large side door, or by removing the roof
in sections bound together with iron, or by taking out the end, and
in some cases, to effect an economy in labor, mechanical devices
are employed for charging. Several designs of air-furnaces are
shown in Figs. 266 to 269 inclusive.

Cupola versus Air-Furnace Melting. — The air-furnace is not
as economical of fuel as the cupola, and the amount of coal used
will average about one-fifth to one-third of the weight of the iron
melted. The air-furnace has, however, some very important ad-
vantages which enable it to produce a higher quality of castings:
(1) The metal does not absorb so much sulphur, since it is not
in contact with the fuel and can only take up sulphur from the
furnace gases; air-furnace iron will therefore increase only about
0.001 to 0.008 per cent, in sulphur; (2) the control of its com-
position in silicon and carbon is much better, as by retaining
it for a longer time in the furnace after melting we may burn
out any desired amount of these elements. Therefore, when
making iron which must be reduced in silicon or carbon, as,
for example, metal for malleable cast-iron castings, or when
making iron that must be very close to a certain analysis, as,
for example, cast iron that is to be chilled on the surface (for
instance, railroad car wheels, rolls for reducing the size of metal
bodies, etc.), the air-furnace is very commonly used; (3) air-*
furnace iron is stronger than cupola iron on account of lower total
carbon and sulphur and better control generally.

Operation of Air-Furnace. — Forced draft is usually used for
the fire, and sometimes additional air is introduced just above the

IRON AND STEEL FOUNDING 285

fire-bridge, in order to increase the oxidizing effect of the flame.
There are also doors in the side walls for the same purpose. It
takes from about 3J to 9 hours to melt down a charge of 5 to 35 tons,
respectively, and the metal is tapped as soon thereafter as it is of
the proper composition and sufficiently hot. The composition is
determined by taking a test ladleful from the bath, casting it into
a mold, and examining a freshly broken fracture, to determine
either the amount of combined and graphitic carbon, or the depth
of chill, or both. American furnaces usually vary in size from 10
to 45 tons capacity each. The loss will be about 2 to 5 per cent,
of the weight of metal used.

Lining. — The lining of the bottom of the air-furnace is
made of silica sand of about the same composition as the acid
open-hearth furnace, i.e., containing 95 per cent, or more of
silica, with just enough lime to frit the mass together at the heat
of the furnace. A layer about 1 to 2 in. thick is spread all over
the hearth and then set on as described in the chapter dealing
with the open-hearth process. About five layers are put on
in this way, and the bottom lasts on an average of 6 to 12 heats,
although some foundries regularly make up a bottom after the
third heat; and in other cases the bottoms last as many as 30
heats. Longer life will be obtained if the material is charged
carefully so as not to break the sand, and if a strongly reducing
flame is maintained during the melting period, when there is
not a bath of liquid iron to protect the bottom from the corrosive
iron oxide. Mixing broken fire-bricks of good quality and re-
fractoriness with the sand seems to give more durable bottoms.

Scrap Used in Air-Furnace. — Cast-iron scrap is very liable
to be high in phosphorus and sulphur and to vary greatly and un-
suspectedly in all impurities. Also it is difficult to sample for
chemical analysis and therefore presents some uncertainty.
As the chief objects of foundrymen in undergoing the additional
expense of air-furnace melting are to obtain low sulphur or a close
approximation to a desired analysis or both, very little iron scrap
is used. Steel or wrought iron scrap is sometimes used, however,
because it reduces the total carbon and the impurities are always
pretty low. I estimate the average amount as between 0 and
10 per cent.

Open-Hearth Furnaces. — Open-hearth furnaces of small size,
but in other respects exactly like the open-hearth steel furnaces,

286 THE METALLURGY OF IRON AND STEEL

are used for melting iron for malleable castings in a few important
foundries in the United States. The great drawback of this fur-
nace is that it must be operated continuously, day and night,
which means more floor space on which to set molds ready for
pouring, because molding cannot well be done during the night, as
the artificial light casts shadows that make the work of finishing
up molds more difficult and confusing.

The advantage of the regenerative open-hearth furnace over
the air-furnace is better control of the operation, and especially
of the temperature, and greater fuel economy. Figures are not
given out by the companies using the process, but we cannot be
far wrong in estimating that the average time of melting in the
open-hearth furnace will probably be somewhat less than four
hours, and the amount of fuel not more than 300 to 350 pounds
per ton of iron, or twice as much if melting only on the day turn.
The lining will certainly last much longer on account of the better
control of the character of flame. Moreover, oil can be used for
melting in a regenerative furnace, because it is introduced into a
very hot atmosphere, which is not practicable in the air-furnace,
since the fuel condenses in the cooler atmosphere of this furnace,
especially when the furnace is cold or when there is a cold charge
on the hearth. Open-hearth linings are best made of dolomite,
which enables a basic slag to be produced in the furnace and a
reduction in phosphorus and sulphur. Cheaper pig iron and iron
scrap can then be used. This basic lining is not attacked by the
iron oxide and slag produced in melting and will last for hundreds
of heats if not allowed to cool off too frequently. Basic linings
cannot be employed in air-furnaces because dolomite contracts
and expands so much on cooling and heating that the bottoms
are soon cracked to pieces, for air-furnaces are not operated con-
tinuously.

MELTING STEEL FOR CASTINGS

Steel castings are made in (1) acid open-hearth furnaces, (2)
basic open-hearth furnaces, (3) small Bessemer converters of
special design, to produce steel at a higher temperature, and (4)
crucibles. The castings are made from the metal just as it comes
from the steel furnace, and the processes are mentioned above in
the order of relative importance.

Open-Hearth Furnaces. — The making of steel for castings is

IRON AND STEEL FOUNDING 287

practically the same as making steel for ingots, except that foundry
furnaces are of smaller size, varying on the average from 15 to 30
tons. In some cases furnaces smaller than this are used, but it is
generally believed that no circumstances warrant this, since the
expense of running the small furnaces is large in proportion.
The chief difference in practice is that the temperature for steel-
casting work is hotter than when ingots are made. Therefore
furnace repairs are higher and the life shorter. In ordinary open-
hearth foundry work the purification is continued to the point
where the steel contains about 0.18 to 0.28 per cent, carbon after
recarburizing, while the silicon will be usually 0.20 to 0.30 per
cent.

Acid versus Basic Open-Hearth Steel. — In steel castings it is
necessary to have somewhat lower phosphorus and sulphur than in
ingots, because the metal is not to receive the beneficial effect of
mechanical work, and therefore must be purer in order to have a
good degree of strength and ductility. Consequently, if we use
an acid steel-making process, we must start with very low phos-
phorus and low sulphur pig iron, which is costly and becomes more
so each year in America. For this reason the Bessemer and the
acid open-hearth steel-making processes are more expensive for
casting work than the basic open-hearth. The result is a present
rapid increase in the use of basic open-hearth steel in America as
well as in Germany, with a probability that in a few years it will
be the predominant process for this purpose. This is in spite of
the fact that basic steel has very serious disadvantages, chief
among which are the amount of oxygen contained in it at the end
of the process and the difficulty of getting the desired amount of
silicon in it with the recarburizer, both of which conditions in-
crease the liability to blow-holes, which are especially objection-
able in castings, as there is no opportunity of their being welded up
and as castings may have to be discarded on account of them after
a good deal of expensive machine-work has been done. On this
account the acid open-hearth process has long held the predomi-
nant position in the steel-casting industry. In fact, this is the
only place where the acid open-hearth process now finds important
employment.

Bessemer versus Open-Hearth. — The open-hearth furnace
gives a large amount of steel at long intervals, which is very incon-
venient for foundry work, because the molds necessary to take all

288 THE METALLURGY OF IRON AND STEEL

the metal must be stored upon the foundry floor until the heat is
ready to pour, and then those who are to do the teeming must
interrupt their other work for half an hour or more for this pur-
pose. Even where the foundry is large enough to have many
furnaces, there is no surety that they will come out at regular and
short intervals, because the operation in one may be delayed.
Another disadvantage of the open-hearth process is that, in order
to be economical, it must be operated continuously day and night,
which also is inconvenient for foundry work. Further, it is not
possible to get the metal as hot as desired without great damage
to the furnace, which is subjected to a higher temperature than
the metal. Lastly, since hot metal is desirable for all castings
except those of very large size, it is usually necessary to tap all the
metal from the furnace at once, and recarburize it at once, which
prevents castings of special analysis being made, unless ordered
in very large quantities. Nowadays, a great many nickel-steel
castings are made in open-hearth furnaces, but this requires nice
calculations, so that the castings molded shall be just equal to the
capacity of the heat, and usually results in a certain amount of
scrapping of high-class material.

All these objections are avoided in making castings in Besse-
mer converters, but they too have their great disadvantages, chief
among which are the slight inferiority of the Bessemer steel and its
greater cost. The latter is true only as compared with basic open-
hearth steel (not acid), and is due principally to the amount of
waste in the side-blown converters used for this purpose, and the
greater cost of the pig iron used, which must be low in phosphorus
and sulphur. Small converter plants are so very cheap — costing
less than $5000 for apparatus — that a great many iron-foundries
in the United States are putting them in as an adjunct to their
cupola process, in order that they may be able to make steel cast-
ings at will. The amount of capital tied up is so small, and there
being no important expense in starting and stopping the converter
as often as desired, they can do this economically.

Tropenas Converter. — The largest number of converters for
steel-casting work are of the Tropenas type, in which the wind en-
ters the vessel from seven tuyeres on the side, and the converter
is tipped in such a manner that the streams of air are deflected
onto the top of the bath. The impurities are oxidized as in the
regular Bessemer process, except that the action is not quite

IRON AND STEEL FOUNDING

289

so rapid, and the carbon is burned to carbon dioxide instead of
carbon monoxide, which generates a much larger amount of heat:

C + O = CO (generates 29,160 calories).
C + 2O = CO2 ( " 97,200 " ).

In order to assist in the superoxidation, there is a second row
of tuyeres above the first, connected with a separate wind-box.
The wind is not turned onto these upper tuyeres until the carbon
begins to burn. The pig iron used for these converters runs fre-
quently above 2 per cent, of silicon, and this, together with the
formation of carbon dioxide, results in very hot and fluid steel,
which can be poured into castings of almost any small size. The
blows usually last about 15 to 20 min-
utes, and the loss is from 17 to 20 per
cent, of the weight of pig iron charged
into the cupola in which it is melted.
The vessel can be started up and stopped
with very little expense, and this ad-
vantage over the open-hearth, together
with the small amount of capital nec-
essary to build the converter plant and
the other conditions already mentioned,
has caused about 20 of these conver-
ters to be installed in America, and
several in France, England and other
countries.

The great disadvantages of the Tro-
penas converter are the waste and the
cost for making repairs. Slight patch-
ing can be done through the mouth,
but there is a hole in the front of the
converter shell, closed by a movable steel
plate, through which the operator can
dig his way into the interior to perform
the necessary repairs. As the lining in

the neighborhood of the tuyeres is usually worn out in less than
twenty blows, this costly method of lining is a serious drawback.

Long-Tuyere Converter. — Next to the Tropenas, the greater
number of converters at work in America are of the Long-Tuyere,
or Stoughton, type, devised by the writer. The bottom part of

FIG. 233. — SECTIONS OF
LONG-TUYERE CON-
VERTER.

290 THE METALLURGY OF IRON AND STEEL

this vessel is attached to the trunnion-ring by a method similar
to that used for regular Bessemer bottoms, and may be removed
with great ease, thus cheapening and facilitating repairs. More-
over, the chief repairs are in the bottom part at the mouths of the
tuyeres, and therefore the lining of the upper part does not have
to be relined completely for several months, although slight patch-
ing is necessary every twenty-five to thirty blows, when the bottom
is changed. This converter is arranged to have but one row of
tuyeres, discharging the blast immediately at the surface of the
metal, and the lining on the tuyere side is thicker in order that the
tuyeres maybe increased in length, which decreases the loss of metal
during the process. The excessive loss in the side-blown converters
is due chiefly to the spitting, which is very large, especially when
the tuyeres have become worn away to a short length and the
streams of air are badly directed and set up interfering currents.

Another cause of the Excessive loss is the large amount of slag
formed, because more iron is oxidized, and this corrodes the lining
very rapidly. Part of the iron is oxidized at the mouths of the
tuyeres, and another part is oxidized when the violent agitation
of the bath, which occurs in all of these converters, especially dur-
ing the boil, throws the metal up into the stack, where it meets
free oxygen. For this reason the upper row of tuyeres in the
Tropenas converter results in an increase in the loss and in the cor-
rosion of the vessel lining. After many experiments with the
Tropenas vessel, I learned that the maximum amount of carbon
can be oxidized to CC>2 without the use of the upper row of tuyeres,
and therefore economy could be obtained by omitting them. In
fact, at some Tropenas plants the valve of the upper row of tuyeres
is often not opened at all during the heat, and no diminution of
the temperature of the resulting steel is observed. The loss in the
Long-Tuyere modification is 14 to 16 per cent., including cupola
loss.

Sizes Used. — Most of the small converters have a capacity of
about 2 tons, because this is economical and well proportioned to
the capacity of an ordinary foundry. There are some 4- and 5-ton
converters, however, and some of less than 1 ton capacity. Sizes
less than 2 tons are very costly to operate in proportion to their
output. The blast pressure usually employed is 3 to 5 lb., with an
average of about 3f lb.

IRON AND STEEL FOUNDING

291

TABLE XXIL— ANALYSES OF SIDE-BLOWN CONVERTER-GASES

ANALYSES — PER CENT.

CALCULATIONS FROM
ANALYSES — PER CENT.

TIME AFTER BEGINNING

M 01

OF BLOW

CO2

and
H

Total
0

N»

O
Enter-
ing

3 fl'rt

PQ * a

4 min., flame starts. . .

0.0

8.2

1.1

90.7

7.1

88.7

23.6

16.5

10 " boiling

0.3

24 3

0.4

75.0

18.3

73.0

19.4

1.1

12 " shortening2....

0.4

8.8

0.2

90.6

6.8

88.6

23.5

16.7

17 " after first drop.

10.7

13.0

0.2

76.1

15.8

74.1

19.7

3.9

21 " end of blow...

1 By difference, H being estimated as 2 per cent.

2 I.e., just before the first drop. There are two drops to the flame in this
operation, the second marking the end.

92.

REFERENCES ON FOUNDRY PRACTICE

90. George R. Bale. "Modern Iron Foundry Practice." Part I.
Foundry Equipment, Materials Used, and Processes Fol-
lowed. 1902. Part II. Machine Molding and Molding
Machines, Physical Tests of Cast Iron, Methods of Cleaning
Castings, Foundry Accounting, etc., etc. 1905. London.
This is the most scientific book on iron foundry practice
published.

The Foundry. Published monthly. Vol. xxxi, 1907. Cleve-
land, Ohio.

Thomas D. West. "American Foundry Practice." Treating
of Loam, Dry-Sand and Green-Sand Molding, and con-
taining a Practical Treatise upon the Management of
Cupolas and the Melting of Iron. Published in New York.
This book is tremendously popular among foundry men and
editions appear at frequent intervals.

William J. Keep. "Cast Iron." A Record of Original Re-
search. New York and London, 1902.

Thomas Turner. " Lectures on Iron-Founding." London, 1904.

"Penton's Foundry List." Cleveland, Ohio, 1906. A Direc-
tory of the Foundries of the United States and Canada.

For an account of small converters for steel casting work see
Reference No. 52 and the files of Nos. 91 and 8.

94
95

THE SOLUTION THEORY OF IRON AND STEEL

SCIENTIFICALLY considered, all of the members of the iron and
steel series are alloys of iron and carbon. Therefore a study of
the general theory of alloys leads to important information upon
iron and steel. According to the authoritative definition, "a
metallic alloy is a substance possessing the general physical
properties of a metal, but consisting of two or more metals, or of
metals with non-metallic bodies, in intimate mixture, solution, or
combination with one another, forming, when melted, a homo-
geneous fluid." It will be seen from this that the essential feature
of an alloy is that, when melted, it shall form a homogeneous fluid.
In plain language, this means that, when melted, the different
components are dissolved in one another. Melted alloys, there-
fore, come under the general head of solutions. In fact, the
great bulk of our alloys, and especially of iron and steel, are pro-
duced by first dissolving the melted components and then allowing
them to freeze. The laws governing this freezing, or solidification,
have only been known a few years, and if this new knowledge has
made great revolutions in physical chemistry, it has led to no less
important discoveries in regard to the nature of iron and steel.

Solid Solutions. — Suppose first we have two metals that are
soluble in each other when liquid, and also when solid. In other
words, the metals of the alloy will be just as completely dissolved
in each other after solidification as before. They will then form a
"solid solution," and a solid solution bears practically the same
relation to a liquid solution as a solid pure metal does to the same
metal when liquid. For example, gold and silver dissolve in each
other when liquid, and also when solid, in any proportion. Con-
sequently any solution of these metals will cool to the freezing-
point and then solidify without there being any important change
(from the metallurgical or practical standpoint) in the relations of
the two metals after the freezing. The reason that these solid

292

THE SOLUTION THEORY OF IRON AND STEEL 293

solutions form in any proportion is that the two metals crystallize
alike. It is, perhaps, a new thought to the reader, but it is never-
theless true, that a metal forms a crystal whenever it solidifies.
Furthermore, each metal has a particular, general shape which its
crystals assume, and there is almost no force powerful enough to
prevent them from taking that same shape in preference to any
other.

Tiny as the crystals sometimes are — often requiring the highest
powers of the microscope to reveal them — their crystalline forces
are very powerful. If, therefore, two metals do not form like crys-
tals, they cannot solidify in solution, i.e., in the same crystal, but
crystallization (i.e., freezing) must be accompanied by precipita-
tion.

By European metallurgists solid solutions are often called
"mixed crystals," or "isomorphous mixtures/7 but this termi-
nology is objected to because the relation of two substances when
dissolved is far more intimate than any mixture possibly could be.
In a mixture, the microscope will always be powerful enough to
distinguish the different components, but a solution always appears
like a simple uniform body. Furthermore, the properties of a mix-
ture are intermediate between the properties of its components,
but a solution — either liquid or solid — has some properties
which are different from any of the properties of either of its com-
ponents. Again, the components of a solution are held together
by chemical forces, while the components of a mixture are either
not held together at all, or only because of close mechanical asso-
ciation. In brief, a solution has some of the peculiarities of a
chemical compound, and differs from such a compound chiefly
because the latter must always be composed of definite amounts
of each component, and in some multiple of their atomic weights,
while a solution may contain widely varying amounts of each
component.

Freezing of Solid Solutions. — When metals form solid solu-
tions, the solutions freeze more or less like pure metals. In Fig. 234
I have shown graphically the freezing of all the alloys of gold and
silver, extending from no silver (i.e., 100 per cent, gold) to no gold
(100 per cent, silver). This figure has been drawn from results
obtained by experiment. The proportion of silver is shown by
the abscissae, or the horizontal distance away from the axis OC,
and the temperature is shown by the ordinates, or the vertical

294

THE METALLURGY OF IRON AND STEEL

distance away from the axis 00. In other words, any point
between these two axes represents by one ordinate the composi-
tion of the alloy in silver, and by the other ordinate the tempera-
ture at which the alloy is at the moment. For example, the point
marked a is a certain distance from the axis 0 C, which shows that
the alloy in question contains 50 per cent, of silver; it is also a cer-
tain distance above the axis 00, which shows that the alloy is
now at a temperature of 1100° C. (1012° F.).

Since there is nothing in the alloys but silver and gold, the
amount of gold in each will be the complement of the amount of

1200C

noo

2192 F

900

0 10* Silver
100 H Gold

FIG. 234. — FREEZING CURVE OF THE GOLD-SILVER ALLOYS.

silver. That is to say, the percentage of silver plus the percentage
of gold must always equal 100 per cent. ; if there is 25 per cent, of
silver there must be 75 per cent, of gold; if there is 50 per cent, of
silver there must be 50 per cent, of gold, etc. The horizontal dis-
tances therefore show the percentage of gold as well as of silver,
and this is given in the second line of figures under the table.

Suppose we have a solution of 50 per cent, silver and 50 per
cent, gold at a temperature of 1100° C. It is represented at the
point a, and is now liquid. When it cools to about 1035° C. it begins
to freeze. It does not all freeze at the same time, but first some
solid crystals freeze out and we have a mixture of solid crystals
with liquid solution. The more the mass cools the more solid
crystals there will be, each crystal being a solid solution of gold
and silver. It is not until we reach a temperature of 985° C., how-
ever, that the last liquid freezes, and then we have a solid solution
of gold and silver, the two solid metals now bearing practically
the same relation to each other chemically that they did when
liquid.

It will be noticed that when this alloy cooled from the point a

THE SOLUTION THEORY OF IRON AND STEEL 295

to 1035°, it met the line A E B. This is the line that represents the
beginning of freezing for all the gold-silver alloys. It has been
drawn after many experiments have been made to show where it
lies in the diagram. It will furthermore be noticed that when this
alloy cooled to 985°, it met the line ADB. This is the line that
represents the completion of freezing for all the gold-silver alloys,
and its position has been learned from many experiments.

Take now the alloy containing 60 per cent, silver and 40 per
cent, gold, at a temperature of 1150° C. — the point b. As this
cools to a temperature of 1025° C., it meets the line AE B and
freezing begins in the same way as before. Solid crystals form
more and more as the alloy cools further, until we meet the line
ADB at a temperature of 980° C. At this point the last liquid
freezes.

THE FREEZING OP ALLOYS OF LEAD AND TIN

1. Suppose now, on the other hand, we have two metals which
are soluble in each other when melted, but not when solid. Evi-
dently there cannot be the same results as those described in the
case of the gold-silver alloys. Such a series is found in the alloys
of lead and tin, whose freezing is shown graphically in Fig. 235.
Here again I have shown the percentage of lead and tin by the
horizontal ordinates, and the temperature by the vertical ordi-
nates. And, again, the diagram has been drawn from results ob-
tained by experiment. The following facts are not based upon
reasoning or logic, but are to be accepted because it is known that
the different actions described actually took place upon trial.

2. Consider first a solution containing 83 per cent, of lead and
at a temperature of 300° C. This will be represented by the point
a in Fig. 235 (page 299), and will be a solution of lead in tin, still
liquid, although about 26° below the melting-point of lead itself.
If cooled, the solution will remain liquid until it reaches a tem-
perature of about 275°, which brings it, we see, to the line A B.
This is the lowest temperature to which it will go and retain that
much lead in solution. If cooled any more, the lead will begin to
precipitate and of course as much as is precipitated will immedi-
ately solidify, being already well below its melting-point. The
lead comes out as crystals of solid lead, which remain mixed with
the mass of the molten solution, but form no longer a part of it

296 THE METALLURGY OF IRON AND STEEL

chemically.1 The solution, i.e., the part still liquid, becomes en-
riched in tin in proportion as the lead precipitates. Leaving this
alloy here for the moment, let us next consider one with less lead.

3. Now take an alloy with 67 per cent, of lead and 33 per cent,
of tin at a temperature of 250° C. This will be represented by the
point b in Fig. 235. We have again liquid lead dissolved in liquid
tin, although the alloy is 76 per cent, below the melting-point of
pure lead. Suppose this alloy cools until it reaches a tempera-
ture of about 240° C., where it meets the line A B. It will not cool
below that temperature and retain all the lead in solution. If we
cool it a few degrees, lead separates out and becomes mechanically
mixed with the liquid solution.

4. Consider next an alloy of 55 per cent, lead dissolved in 45
per cent, of tin at a temperature of 250° C. As this cools to a
temperature slightly below 225°, it again meets the line A B and
crystals of lead separate out and solidify. And so on: any solu-
tion, when it cools to a temperature where it meets the line A B,
reaches its limit of solubility ; it cannot cool m6re and retain all of
the lead in solution ; but if it does cool, then some of the lead must
be precipitated.2

5. To sum up the preceding paragraphs, we may then say that
the less lead we have in solution,3 the lower the temperature can
go without any of it being precipitated. Or, in other words, the
less the lead in solution the lower the temperature will go before
any freezing begins. This knowledge is the result of experiment,
and it has been shown that the line A B in Fig. 235 represents the
equilibrium between the amounts of lead in the different solutions
and the temperature to which each will cool before any lead is
precipitated, or, in other words, the amount of lead that saturates
the solution at each temperature.

1 It is a fact that when the lead is precipitated from the solution, it car-
ries with it a few traces of dissolved tin, but for the present we neglect this
slight impurity for the sake of simplicity, and consider that pure lead sep-
arates.

2 Any lead that is precipitated must of course freeze, because the tem-
perature is already below the freezing-point of lead. Conversely, any lead
that freezes must be precipitated from the alloy, because we know as a matter
of experiment that frozen lead will not retain tin in solution (omitting, of
course, the few traces of tin retained by the lead and which, for the sake of
simplicity, we omit in the discussion).

•Within limits to be afterward denned.

THE SOLUTION THEORY OF IRON AND STEEL 297

6. With this knowledge let us consider again the first solution
— containing 83 per cent, of lead. We have stated that when
this solution was cooled below 275° C., some lead was precipitated.
Have we any evidence now as to how much lead would be precipi-
tated with each unit drop in temperature? Evidently we have,
for we know how much lead is normally in the saturated solution
at each temperature ; this evidence is given to us by the line A B.
For example, assume that the solution containing 83 per cent, of
lead is cooled to 240° C.; how much lead will be left in solution,
and therefore how much will have been precipitated? From
paragraph 3 we know that 67 per cent, of lead remained in solution
down to a temperature of 240° C. and that there was an equilib-
rium between the amount of lead and this temperature, or, in
other words, the solution is saturated with lead at this point.
Therefore it would be reasonable to suppose that the solution
which started with 83 per cent, of lead would retain exactly 67
per cent, of lead by the time the temperature of 240° is reached.
That this reasoning is correct is proved by experimental evidence.

7. Now let us cool the same alloy to a temperature of 225°;
how much lead will this retain in solution and how much would be
in a precipitated form? We have already seen (paragraph 4)
that an alloy containing 55 per cent, of lead will retain all of that
lead in solution until it falls to a temperature of 225°. Therefore
it is reasonable to suppose that the cooling alloy we are considering
will retain just 55 per cent, of lead and no more by the time its
temperature has fallen to 225°, and this is in fact the case. In
short, each solution will always retain enough lead to saturate itself
after it once becomes saturated.

8. We have learned above that every point on the line A B
represents the maximum amount of lead that can be retained in
one of these solutions at that temperature. Therefore as soon as
any solution meets the line A B, it follows along this line as it cools
and lead precipitates in amounts proportional to the temperature,
so that the solutions will always be saturated with lead.

9. It is therefore evident that if we start with any of these liquid
alloys containing an amount of lead represented on the diagram
to the left of the point B, lead will precipitate from it during cool-
ing and the amount of lead in solution will continually decrease,
so that the composition of the alloy at the different temperatures
will be represented by a point traveling down the line A B. There-

298 THE METALLURGY OF IRON AND STEEL

fore, in every such case we shall reach finally a solution with 31
per cent, of lead and 69 per cent, of tin (the proportions repre-
sented by the abscissa of the point B) when the temperature has
fallen to 180° C. (the ordinates of the point B). Mixed with the
solution at this temperature will be the amount of lead which has
separated during cooling and which will depend upon the percentage
in the original solution. (If this is not clear on the first reading, a
little thought will make it so, especially if the reader follows in
Fig. 235 each action I have described, step by step.)

10. We have considered up to now only the solutions con-
taining large amounts of lead and from which lead is precipitated
on cooling. Let us now consider a solution containing 90 per
cent, of tin at a temperature of 225° C. (point d, Fig. 235). It is
still liquid, although below the normal melting temperature of
both lead and tin. This will cool until a temperature of about
210° is reached and it meets the line C B. If it cools any more,
tin will be precipitated and will of course freeze, because it is
already below its normal freezing-point (231° C).

11. Consider next an alloy containing 85 per cent, of tin at a
temperature of 225° (point e in Fig. 235). This will cool to about
200° C., where it meets the line C B, but if it cools any further, tin
will be precipitated and will freeze.

12. In other words, the line C B represents the conditions of
equilibrium between the temperature and the amount of tin that
will be retained in solution, just as the line A B represented the
equilibrium between the temperature and the amount of lead that
would be retained in solution. That is to say, every point on the
line C B represents the amount of tin that will saturate a solution
at that temperature.

13. Returning then again to the alloy containing 90 per cent,
of tin, how much tin will have been precipitated by the time the
alloy cools to, let us say, 200° C.? Obviously 5 per cent, of tin
would have been precipitated, leaving a solution containing 85
per cent, of tin, because we have already seen that 85 per cent, of
tin will saturate a solution at 200° C.

14. Whatever solution we may have had to start with, pro-
vided there was always more than 69 per cent, of tin, as soon as
that solution met the line C Bit would travel down this line, precip-
itating tin progressively in proportionate amounts such that the
tin left in the solution at the varying temperatures would corre-

THE SOLUTION THEORY OF IRON AND STEEL

299

spond to the ordinates* of the line C B. In other words, this solu-
tion follows the line C B. Therefore, in all such alloys we shall
finally arrive at a solution having 31 per cent, of lead and 69 per
cent, of tin (the abscissa of the point B) when the temperature
has fallen to 180° C., and with this solution will be mixed precipi-
tated tin in amount depending upon the amount of tin in the
original solution. This solution is known as the 'eutectic' solu-
tion.

15. We may now sum up paragraphs 9 and 14 by saying that
whatever solution of lead and tin we have to start with, we will

C. 350

S> 200

150

100

Tin 0 10

Lead 100 90

8C2 F.

572

392°

212°

32°

90 100'Per-cent

10 0 Per cent

FIG. 235. — THE FREEZING OF ALLOYS OF LEAD AND TIN.

always have one containing 31 per cent, of lead and 69 per cent,
of tin by the time the temperature has fallen to 180°, and with
this solution will be mixed some precipitated tin or precipitated
lead, as the case may be (unless, of course, we started with exactly
31 per cent, of lead and 69 per cent, of tin).

Now let us consider the further cooling of these alloys. Evi-
dently, no change will take place in the precipitated lead or tin, as
either of them would be already in the solid form. What will
happen, however, to the solution containing 31 per cent, of lead
and 69 per cent, of tin, which we have called the 'eutectic' solu-
tion? It is evident that when this solution cools below 180°, it
must cross the point B, and therefore both the lines A B and B C

300 THE METALLURGY OF IRON AND STEEL

at once. Obviously, it cannot cross either of these lines without
precipitating lead or tin. In point of fact, on crossing both of the
lines at the same time, it precipitates at once all the remaining
lead and all the remaining tin, and therefore completes the decom-
position of the solution and the solidification.

The lead and tin separate in tiny solid crystals or flakelets,
which arrange themselves in a parallel banded structure similar
to that shown in Fig. 240, page 313. This structure is known as
the 'eutectic' structure, after the alloy which always results when
any one of the solutions of the series is cooled. The term ' eutectic
alloy' means etymologically 'well-melting alloy/ because it remains
melted longer than any other alloy of the same metals, and every
solution the components of which are soluble in each other in the
liquid state and insoluble in the solid state will form a eutectic
solution in the way I have described. This applies not only to
solutions of metals in each other, but to solutions of metals in
liquids which are afterward frozen, and even of salts in water, etc.

Freezing-Point Curves. — The lines A B and CB are often
spoken of as the "upper freezing-point curves" of the alloys, be-
cause any alloy which cools to this line will then commence, to
freeze. Furthermore, freezing once having commenced when
either of these lines is reached, will continue progressively as
the excess metal separates out. The line D E is often called the
'lower freezing-point curve/ because this line represents the
temperature of freezing of the eutectic of the series, and we have
already shown that every alloy in the series automatically forms a
eutectic by 'selective' precipitation; therefore every alloy in the
series will not be entirely solid until it reaches the temperature at
which the eutectic solidifies,1 which is always the same.

Cooling Curves. — There are certain thermal changes which
accompany the chemical changes I have outlined in the preceding
paragraphs. These thermal changes are of importance, because
it is by means of them that we are usually able to obtain the first
evidence of the precipitation of excess metal, the formation and
solidification of a eutectic, etc. Consider the alloy containing
83 per cent, of lead and 17 per cent, of tin, at 300° C., and let us
observe by means of a thermometer or pyrometer the rate of

1 Some metallurgists prefer not to draw the line D E, but to represent the
freezing of the eutectic merely by the point B.

THE SOLUTION THEORY OF IRON AND STEEL

301

cooling. At first the thermometer will fall pretty fast, but when
we reach 275°, where the line A B is met, the rate of fall is suddenly
retarded. It thus becomes evident to us that some event coun-
teracts the fall in temperature. What this event is we learn from
microscopic evidence, and, as has already been explained, it is the
precipitation of lead. This explanation might have been ex-
pected, because the precipitation of lead at a temperature below
its normal freezing-point would of course be accompanied by
freezing and, during the freezing, the metal would liberate its
latent heat of fusion and thus oppose the cooling of the mass as a

300C

100

b Eutectic

freezing

572 F

212°

10 Minutes

20 Minutes

FIG. 236. — FREEZING CURVE OF AN ALLOY CONTAINING 83 PER CENT.
OF LEAD AND 17 PER CENT. OF TIN.

whole. The rate of fall of temperature is, moreover, retarded all
the way down to 180°, because lead is being continuously precipi-
tated as the solution travels down the line A B. When we reach
180°, the fall of temperature is not only retarded but actually
ceases; in some cases the temperature may rise slightly. This
change is due to the large amount of latent heat of fusion liberated
by the freezing of the eutectic. This arrest continues until the
eutectic is entirely solid, after which the rate of fall becomes rapid
again and proceeds without important change until the atmos-
pheric temperature is reached, because now we have merely the
cooling of a solid alloy.

These changes are represented diagrammatically in Fig. 236,
in which the abscissae show the time, in minutes, from the begin-

302 THE METALLURGY OF IRON AND STEEL

ning of the cooling, and the ordinates show temperatures. The
change in direction at the point a shows the retardation due to
the precipitation of the lead, while the long horizontal part at b
shows that for several minutes the temperature was not falling
at all, because the eutectic was freezing.

Next, experimenting upon the alloy with 67 per cent, of lead,
we find that it cools rapidly until it meets the line A B, when the
rate of fall is retarded continuously until we reach 180°, where
again an arrest (or perhaps an actual rise) of the temperature is
observed, after the completion of which the fall in temperature
becomes rapid again and proceeds normally.

Similar thermal changes are observed in the alloy containing
55 per cent, of lead, but at temperatures of 225° and 180°. Now,
by plotting the upper changes in the several solutions (i.e., at
275° C., at 240° C., and at 225° C.), the position of the line A B is
determined, and by plotting the lower changes (i.e., at 180° C. in
each case) the line D E is determined.

In studying the solutions of the series rich in tin, similar
thermal changes are observed. Consider the alloy containing
90 per cent, of tin and 10 per cent, of lead; this will cool rapidly
until a temperature of 210° C. is reached, when a retardation will
occur and will persist until the temperature 180° is reached. By
this time the excess tin will all have been precipitated and the
residual solution 1 will have travelled down the line C B to the
point B. Thereupon an arrest (for an actual reversal) in the rate
of cooling will occur until the eutectic has solidified, after which
the cooling will proceed at a rapid rate. In the case of the alloy
containing 85 per cent, of tin, the first retardation will begin at
200°, and then an arrest at 180°. By plotting the points at 210°
and 200°, we obtain the position of the line C B 2, and by plotting
the two points at 180° we obtain further points to determine the
line D E.

Properties of Eutectics. — As already said, not only the lead-

1 These residual solutions are technically known as ' mother liquors ' or
'mother metals,' because it is out of them that the solid metal is being
'born/

2 It should be understood that in actual experimental determination of lines
corresponding to A B, C B and D E, in any series of alloys, not a few, but a
large number of different solutions are studied, and a great many retardation
points are found before the lines are drawn.

THE SOLUTION THEORY OF IRON AND STEEL

303

tin alloys, but every alloy of which the components are soluble
when liquid, and insoluble (or nearly so) when solid, will form
eutectics, and it happens that this is the case with the great
majority of our metallic alloys. For this reason the properties
of the eutectic are very important in all metallurgical work. It
will be seen that a eutectic is not a chemical compound in atomic
proportions. For example, in the case of the lead-tin alloys, the
point B comes at 31 per cent, of lead and 69 per cent, of tin merely

Lead 100 90

90 100 Per-
ID 0 Cent

FIG. 237. — FREEZING-POINT AND STRENGTH CURVE OF THE LEAD-TIN

ALLOYS.

because the line A B crosses the line C B at this ordinate, and this
has no relation to atomic ratios. If the melting-point of tin hap-
pened to be higher, or if it did not precipitate from lead so fast
upon cooling, the line C B would cut the line A B at a different
point, and therefore the composition of the eutectic would be
different.

The structure also of the eutectic is very important. The
majority of eutectics have a structure similar to the banded form

304 THE METALLURGY OF IRON AND STEEL

shown in Fig. 240. The tiny crystals in this structure are inter-
mingled very intimately, and this close association has a bene-
ficial effect on the strength of the mixture. In Fig. 237 we may
see how the curve showing tensile strength rises to a maximum
at the point corresponding to the eutectic of the series of lead-tin
alloys.

THE FREEZING OF IRON AND STEEL

1. When the iron and steel alloys are liquid, they are com-
posed of liquid carbon dissolved in liquid iron, and their freezing-
point curves are shown in Fig. 238. It will be noticed that the
lines in this figure bear a close similarity to those in Fig. 235. This
similarity is real, and the laws governing the freezing of this series
of alloys are very similar to those governing the freezing of the
lead-tin alloys. There is a eutectic of this series when the line
A B crosses the line C B, and the components of this eutectic are
95.7 per cent, iron and 4.3 per cent, carbon. In the study of the
iron-carbon alloys, however, we have to take into account the
solid solution which forms. That is to say, we must remember
that iron never separates from the liquid state without carrying
2.2 per cent, of carbon with it in solid solution.1 It will be remem-
bered that when lead separated from the liquid solution, it carried
with it a small amount of tin as an impurity; and that when tin
separated, it carried with it a small amount of lead as an impurity;
but we disregard these traces of impurity for the sake of simplicity
in outlining the laws of solutions. In the case of iron-carbon
alloys, however, the 2.2 per cent, of carbon carried out with the
iron in solid solution, substantially as an impurity, is too impor-
tant in its effect upon the material to permit us to neglect it.

2. The line X Y therefore divides the diagram in Fig. 238 into
two parts. Everything to the left of this line freezes as a solid
solution and the laws are similar to the freezing laws of the gold-
silver alloys. Everything to the right of the line X Y freezes
selectively, according to the same laws as those given in the case
of the lead-tin alloys. It is because of this difference in the freez-
ing of the alloys that the line X Y is arbitrarily considered as the

solid solution may consist of carbon dissolved in the iron, or of
iron carbide dissolved in the iron. We do not know definitely which, but it
is not necessary to discuss this question just yet.

THE SOLUTION THEORY OF IRON AND STEEL 305

dividing line between steel and cast iron. That is to say, all the
alloys with less than 2.2 per cent, carbon are defined as steel, and
alj with more than 2.2 per cent, carbon are defined as cast iron.

3. Freezing of Steel. — All the steels freeze as solid solutions.
// Let us consider a solution of 99.5 per cent, iron and 0.5 per cent.

carbon at 1650° C. This will be represented by the point b in Fig.
238. This cools until it meets the line A B, and now it commences
to solidify. For a few degrees of temperature it is part liquid and
part solid, but by the time it has fallen to a temperature where it
meets the line A a, it has become entirely solid, and it is now a
homogeneous solution of 0.5 per cent, of carbon in iron. Consider
next a solution containing 99 per cent, of iron and 1 per cent, of
carbon, at the point c in Fig. 238. When this cools to the tempera-
ture where it crosses the line A B it commences to solidify, and it
is in a partly liquid and partly solid condition until it drops to a
temperature where it crosses the line A a, upon which solidification
is complete, and it now becomes a homogeneous solution of 1 per
cent, carbon in iron. The same actions take place with an alloy
containing 98.5 per cent, iron and 1.5 per cent, carbon (the point d
in Fig. 238), and also in the case of 98 per cent, iron and 2 per cent,
carbon. In all these alloys we finally arrive at a solid solution of
carbon in iron.1 To this solid solution the name of 'austenite'
is given, and this name applies no matter how much or how little
carbon is in solid solution. In other words, all the steels are in
the condition of austenite as soon as their solidification is complete.

4. Freezing of Cast Iron. — The freezing of cast iron is shown
by the diagram to the right of the line X Y, and if we should con-
sider this part as a separate diagram, then it would be similar to
Fig. 235, the freezing of the lead-tin alloys. There is one differ-
ence to be borne in mind, however: along the line A B in Fig. 235
we had a selective precipitation of lead; along the line A B in Fig.
238 we have a selective precipitation, not of pure iron, but of iron
containing 2.2 per cent, of carbon. In other words, this entity,
consisting of a solid solution of iron with 2.2 per cent, of carbon,
behaves as if it were an elemental substance. Suppose we have,

1 The formation of the solid solutioh from the liquid solution is discussed
in detail in Professor Howe's 'Iron, Steel and Other Alloys/ but it requires
too much space to be discussed here. For our purpose it is sufficient to know
that when freezing is completed, we have a homogeneous solid solution of all
of the carbon in all of the iron.

306 THE METALLURGY OF IRON AND STEEL

for example, a liquid solution of 2.5 per cent, of carbon in iron at
a temperature of 1400° C. This will be represented by the point e
in Fig. 238. The liquid solution will cool until it reaches a tem-
perature of about 1320° C. At this point there will begin to pre-
cipitate the entity of which I have spoken, namely, iron contain-
ing 2.2 per cent, carbon in solution. This precipitation will cause
the liquid solution to be impoverished in iron, and it will therefore
move to the right in the diagram as the temperature falls ; or, in
other words, it will travel down the line A B. By the time the
temperature of 1135° C. is reached, a large amount of the entity
containing 2.2 per cent, of carbon will have precipitated, and the
small amount of liquid solution left will be at the point B, that is,
the eutectic point, where there is 4.3 per cent, carbon. With
further cooling the eutectic will cross the point B, and therefore
will complete its precipitation and its freezing. It breaks up into
crystals, part of which consist of tiny flakelets of the entity before
mentioned — iron containing 2.2 per cent, carbon — and the
other part of crystals of graphite — i.e., carbon.

5. A similar result will be obtained in the case of a 3 per cent,
liquid solution of carbon in iron. This will cool until it reaches
1280° C., where the same entity will precipitate, decreasing the
residual solution in iron so that it travels down the line A B and
finally reaches the point B, after which this eutectic solution at
the point B will precipitate as before.

6. We may sum up paragraphs 4 and 5 by saying that any
solution of iron containing more than 2.2 per cent, and less than
4.3 per cent, of carbon will consist, after freezing, of a eutectic
together with a certain amount of previously precipitated entity
(consisting of iron with 2.2 per cent, of carbon in solid solution).

7. What will occur in case the solution contains more than 4.3
per cent, of carbon?1 Take, for example, a liquid solution con-
taining 4.7 per cent, of carbon at a temperature of 1200° C. This
will cool until a temperature of about 1170° is reached and the line
C B is met. As cooling proceeds to lower temperatures, carbon (i.e.,
graphite) precipitates out and the liquid solution remaining moves
to the left in the diagram. That is to say, it travels down the
line C B. When the temperature 1135° is reached, so much graphite
is precipitated that the residual solution is now of the eutectic pro-

1 It is very seldom that solutions contain more than 4.3 per cent, of carbon,
because this much carbon does not readily dissolve in iron.

THE SOLUTION THEORY OF IRON AND STEEL

307

portions. With further cooling, this eutectic breaks up as before,
consisting thereafter of crystals of graphite and of the entity com-
posed of iron with 2.2 per cent, of carbon in solid solution.

8. Summary. — To sum up, then, all the solutions of iron and
carbon containing less than 2.2 per cent, of carbon will consist,

1600°C
A

1400°
1300°
1000°
800°

600°

i
1C

2912°F
2552°
2192°
1832°
1472°
1112°

SS»»

d
"fcfcfc

^B^

Ellt€

ctic fre

ezes

)* 1% Carbon 2%
K) % Iron 99£lron

3% 4* 5*Carbon
95*Iron

FIG. 238. — THE FREEZING OF ALLOYS OF IRON AND CARBON.

after solidification, of a solid solution of iron and carbon having
the same chemical composition as the original liquid solution, and
being a homogeneous solution of one in the other. There can be
no eutectic formed if there is not more than 2.2 per cent, carbon.
All the solutions with more than 2.2 per cent, of carbon will con-
sist, after solidification, of a eutectic together with a certain
amount of previously precipitated graphite or of the previously

308 THE METALLURGY OF IRON AND STEEL

precipitated entity mentioned (iron with 2.2 per cent, of carbon in
solid solution).

9. Effect of Silicon and Sulphur. — Silicon and sulphur have
an important effect upon the changes that occur in the solidifica-
tion of the iron-carbon alloys. Silicon has the effect of pushing
the point B, Fig. 238, to the left. In other words, it causes the
eutectic to have less carbon than 4.3 per cent. It thus lessens
the amount of total carbon that cast iron usually contains. Silicon
seems to have the effect of pushing the point a also to the left and
thus decreasing the amount of carbon that solidifies in the solid
solution. In fact, we may have iron containing 2 or 3 per cent, of
silicon in which the iron will solidify with only a small amount of
carbon dissolved in it. Sulphur has the opposite effect and tends
to push the point B to the right and to increase the total carbon.

10. Rate of Cooling. — The changes represented in Fig. 238
take place very slowly. This is of great importance, because it
renders it necessary, in order that these changes may occur, that
the cooling of the solutions shall be very slow. In other words,
if we cool rapidly we will not get a precipitation of graphite, even
when the eutectic freezes, but will obtain a solid solution contain-
ing all the carbon. This solid solution, of course, is not in a normal
condition, and theoretically should never be formed to this ex-
tent, but is none the less present in many cases, because the cooling
was not slow enough. As already noted, however, the presence
of silicon hastens the precipitation of graphite, so that silicon in
this sense takes the place of slow cooling, and in our high-silicon
pig irons we often have a precipitation of almost all the carbon
as graphite. In other words, we have only a very small amount
of solid solution formed, which probably does not contain as much
as 2.2 per cent, carbon.

THE SOLID SOLUTION OF IRON AND CARBON

We have seen that every alloy of iron and carbon contains,
after solidification, a varying amount of the solid solution of iron
and carbon. For example, if we started with 1 per cent, of carbon,
then, immediately after solidification, we would have a solid solution
of 1 per cent, carbon in iron ; if we started with 1 .5 per cent, of carbon,
we would have a solid solution containing 1.5 per cent.; if we
started with 2.2 per cent, of carbon, then we would have a solid

THE SOLUTION THEORY OF IRON AND STEEL 309

solution of 2.2 per cent, carbon. Even if we start with more, than
2.2 per cent, carbon, then the alloy, after solidification, will con-
sist partly of a solid solution containing 2.2 per cent, of carbon
and partly of graphite.

Now, what becomes of these solid solutions which make up a
part or a whole of the cast iron and steel alloys when they freeze?
Does the carbon remain in solid solution down to the atmos-
pheric temperature, or does it precipitate, or does it undergo
some other change? We find by experiment that the solid solu-
tions do not survive, but precipitate at a lower temperature, and
the laws governing the decomposition of these solid solutions are
similar to the laws governing the decomposition of liquid solu-
tions — for example, of the lead-tin solutions. In short, we have
again a series of curves showing the selective precipitation of the
constituents of these solid solutions, and the only difference be-
tween the nature of these curves and the lead-tin curves is that
these represent changes taking place in the solid state, while the
lead-tin curves represent changes taking place in the liquid state.

Nature of the Solid Solution. — In the footnote on page 304 I
called attention to the fact that the solid solution of iron and
carbon might be a solution of pure carbon in iron, or a solution of a
carbide of iron in iron, for example, of FeaC in iron. Several
authorities hold this view,1 while others maintain that the solution
is of elemental carbon in iron. The question is of more academic
than practical interest. The important thing is that, when the
solution decomposes, it is a carbide of iron which precipitates.
Those who believe that the solid solution is composed of elemental
carbon and iron explain the precipitation of the carbide by main-
taining that when the carbon separates from solution, it immedi-
ately unites with iron and forms a carbide, usually FesC. With
this explanation I shall hereafter, for simplicity's sake, discuss the
solid solutions as if they were FeaC in iron. 2

1 Indeed one authority believes that there are probably several different
carbides (Fe3C, FeaC, FeC) dissolved in the iron.

2 It may be that when the solid solution is very hot, it consists of elemental
carbon dissolved in iron, but that as it falls to near the point where it begins
to decompose, it consists of iron carbide dissolved in iron. This is the case
with the solutions of table salt in water which, at a high temperature, con-
sists of sodium and chlorine dissolved in water; near the freezing-point, how-
ever, the sodium and chlorine combine and the solutions consist of sodium
chloride dissolved hi water.

310

THE METALLURGY OF IRON AND STEEL

The Decomposition of the Solid Solutions. — The curves of de-
composition of the solid solutions are shown in Fig. 239. The line
G 0 S is the line upon which there is selective precipitation of pure
To this pure iron the name of ' ferrite ' has been given by

iron.

Professor Howe, and this name meets with universal acceptance.

1600°C
1400°
1200°

1000°
G

800°

600°

C
10

2912F
2552°
2192°
1832°

1472°
K
1112°

a
/

\JL

1* Carbon 2%
0£Irou 99 4 Iron

3% M 5* Carbon
95 £ Iron

FIG. 239. — DECOMPOSITION CURVES OF THE SOLID SOLUTIONS OF
IRON AND CARBON. (ALSO KNOWN AS THE CRITICAL POINTS.)

Consider first a solid solution containing 0.40 per cent, of carbon
at a temperature of 800° C. This will be at the point h in Fig. 239.
It will cool until it reaches a temperature of about 780° C. upon
which ferrite will begin to precipitate. As the temperature con-
tinues to fall more and more ferrite precipitates, which impover-
ishes the solid solution in iron and causes it to travel down the

THE SOLUTION THEORY OF IRON AND STEEL 311

line 0 S. By the time the temperature has reached 690°, the solid
solution has reached the point S, corresponding to 0.90 per cent,
carbon.

Consider next an alloy containing 1.60 per cent, carbon at
1000° C. : this will be at the point k. It will cool until it reaches a
temperature of about 970°, at which carbide of iron (FeaC) will
begin to precipitate. This precipitation continues as the tem-
perature falls, constantly decreasing the amount of carbon in
the solid solution, which therefore travels down the line ES
until, at 690°, it reaches the point S, where there is 0.90 per cent,
carbon.

A similar precipitation will take place with all of the solid
solutions of iron and carbon: if they contain less than 0.90 per
cent, carbon, they will begin to precipitate out ferrite when they
fall to the line GO S. If they contain more than 0.90 per cent, car-
bon, they will begin to precipitate carbide of iron when they meet
the line E S. In either case the residual solid solution will travel
down the line G 0 S, or else E S, until it reaches the point S when
the temperature has fallen to 690° C. ; we will then have some solid
solution left containing 0.90 per cent, carbon, and mixed with
this some previously precipitated ferrite or cementite, as the case
may be.

Eutectoid. — The alloy containing 0.90 per cent, carbon is
known as the 'eutectoid alloy/ a name invented by Professor
Howe to indicate that the formation of this alloy, which results by
selective precipitation of the solid solution, is similar to the forma-
tion of the well-known eutectics of liquid solutions. When this
eutectoid solid solution cools below 690° C., it is completely de-
composed into its constituents, ferrite and cementite. These
constituents precipitate in tiny flakelets, which arrange themselves
in the banded structure already familiar to us as the structure of
eutectics. A magnified view of the structure is shown in Fig. 241,
while Fig. 242 shows the magnified structure of a piece of 'steel com-
posed of a eutectoid together with previously precipitated ferrite ;
and Fig. 243 shows the structure of a steel consisting of the eutec-
toid with previously precipitated cementite.

It will be evident that there will be some of the eutectoid in
every piece of iron or steel, for even the cast irons contain, after
solidification, a certain amount of solid solution which precipi-
tated either while the liquid alloy was traveling down the line A B,

312 THE METALLURGY OF IRON AND STEEL

or during the freezing of the liquid eutectic (containing 4.3 per cent,
of carbon), or both. The formation and characteristic of this
eutectoid are therefore of very great importance. Its presence in
iron and steel was known long before the theories that I have out-
lined in this chapter had begun to be understood, and the name
of 'pearlite' was given to it because, under certain circumstances,
it has the appearance of mother-of-pearl. The name pearlite is
only used to designate the eutectoid after its complete decomposi-
tion and the separation of the ferrite and cementite into the banded
structure shown in Fig. 241.

THE COMPLETE ROBERTS-AUSTEN, ROOZEBOOM DIAGRAM

We may now take the diagrams of Figs. 238 and 239 and com-
bine them into one diagram which shall represent all the known
changes in the heating and cooling of iron and steel.1

All the lines drawn in Figs. 238 and 239 show the temperatures
of the changes during cooling.

Explanation of Fig. 246. — In Fig. 246 we have an assembly of
the curves showing the changes in the liquid and in the solid state.
It is to be observed that the dotted line E F is somewhat doubtful.
It was drawn in by Prof. H. W. Bakhuis-Roozeboom in the
belief that the solid solution of iron containing 2.2 per cent, of
carbon began to decompose at 1000° C. during cooling. There is
no direct evidence in favor of this opinion, however, although
certain theoretical considerations are in favor of it ; but we are not
warranted in assuming them as proven. Indeed, the changes that
take place in the alloys and at the temperatures between a C and
S K are yet very much in doubt, and several different theories are
held. If we leave out the line E F, then the line S E would be

1 For it is to be understood that the changes that I have spoken of as
taking place during the cooling are exactly reversed during heating. It is to
be observed, however, that the changes that take place during heating occur
at a somewhat higher temperature than the reverse changes during cooling.
This difference is probably due to a slight lag, so that the changes occur a
little below the normal temperature during cooling and a little above the
normal temperature during heating. This lag is especially noticeable in the
changes that take place in the solid solution, and there is a difference of 20° or
more Centigrade between the lines GOS, PS K, etc., if they are observed
during heating.

FIG. 240. — EUTECTIC OF COPPER

AND SILVER.
(William Campbell.)

FIG. 241. — PEARLITE EUTECTOID
OF IRON AND CARBON.

(F. Osmond.)

FIG. 242. — PEARLITE AND FERRITE.
250 diameters.

FIG. 243. — PEARLITE AND CEMEN-

TITE.
250 diameters.

FIG. 244. — PEARLITE. FIG. 245. — PEARLITE.

250 diameters. 1000 diameters.

Picric Acid. — Large crystallization due to long overheating. Polished in Relief.

314

THE METALLURGY OF IRON AND STEEL

extended to a, as I have shown in the diagram, and the solid
solution containing 2.2 per cent, of carbon will commence to pre-
cipitate its cementite immediately after solidification is completed
(1135° C.), instead of 135° lower.

Other Lines in Fig. 246. — Certain retardations in the cooling
curves would indicate that there was a line extending all the way

2912F
2552°

2192°
D

1832°

1472°
K
1112°

10UUC
A

1400°

•

1200°

iooo

800°
M

600°

(
1(

,**-*•"

) 1* Carbon 2* 3% M 5sS Carbon
NtitlOtt 99 # Iron 95^ Iron

FIG. 246.— THE FREEZING AND SOLID DECOMPOSITION CURVES
OF THE IRON-CARBON ALLOYS.

across this diagram at about 775° C., and another at 600° C. The
significance of these lines is as yet a matter of speculation merely,
and I believe it to be well to disregard them in this brief treatise
until we are able to offer some explanation.

Roberts-Austen. — The diagram of Fig. 246 is often known as the
Roberts- Austen diagram, after Sir William Roberts- Austen, be-

THE SOLUTION THEORY OF IRON AND STEEL 315

cause the cooling curves that located the lines were first determined
in his laboratory.1

1 See " Proceedings of the Institution of Mechanical Engineers " (England),
1897, Fourth Report of the Alloys Research Committee, Plate 2. See also
' Le Fe et 1'Acier au Point de Vue de la Doctrine des Phases," H. W. Bakhuis-
Roozeboom, in Zeitschrift filr physikalische Chemie,vo\. xxxiv, p. 437: French
translation, " Contribution a TEtude des Alliages," pp. 327-386, Paris, 1901.

REFERENCES
For further data on this subject see especially Nos. 1, 8, 9; 10.

n ^t

XI
THE CONSTITUTION OF STEEL

THE properties of steel depend upon the chemical composition
of its constituents as well as upon their size and relation to one
another. Enough has been said to show that steel is not a simple
homogeneous union of iron with varying proportions of carbon,
silicon, manganese, etc.; but is built up of individual crystals
somewhat in the same way as crystalline rocks are formed —
granite, for example. But while the crystals of granite are gen-
erally visible to the naked eye, and its structure may therefore
be determined by a more or less cursory examination, the structure
of steel is visible only by means of the microscope and after
careful polishing, sometimes followed by chemical treatment to
differentiate between the various grains. Nevertheless, the struc-
ture of steel is of great importance, and in some cases, perhaps, is
even more so than the chemical composition.

THE MICRO-CONSTITUENTS OF STEEL

In this chapter I shall speak only of slowly cooled steel except
where I have indicated the contrary. We have already learned
that slowly cooled steel must necessarily contain ferrite and
cementite, resulting from the decomposition of the solid solution
of iron and carbon. There are also other constituents which are
found under the microscope, or separated by chemical analysis,
or in both ways. These latter constituents are compounds of iron
with various other impurities, such as iron sulphide, iron phosphide
and iron silicide; or of two impurities with each other, such as
manganese sulphide.

Ferrite. — Ferrite is theoretically pure iron, and especially
iron free from carbon. It is weak as compared with several of the
other constituents, having a tensile strength of about 45,000 to
50,000 Ibs. per square inch; it is also very soft and ductile, re-

316 *

THE CONSTITUTION OF STEEL 317

sembling copper in these properties, and has, furthermore, a high
degree of malleability. It has a very high electrical conductivity
as compared with the other constituents of iron and steel, and
about one-seventh the conductivity of copper and silver, the best
conductors known. (Copper and silver are of nearly the same con-
ductivity.) Its magnetic force is the highest of any known
substance, its magnetic permeability being high, and its hysteresis
low. It crystallizes in the isometric system.
. Ferrite is an important constituent of all steels and the pre-
dominant one in all the low-carbon steels. The industrial product
approaching nearest to pure ferrite is wrought iron, if we disregard
the slag, which, being mechanically mixed with the mass, does
not appreciably alter its chemical and physical behavior. It is
for this reason that wrought iron is so useful where a soft and
ductile material is necessary, as in boilers, for instance; or where
high electrical conductivity is demanded, as in telegraph wire;
or a high degree of magnetism, as in the cores of electromagnets.
The wrought iron made in Sweden, and known as 'Norway
iron/ is greatly preferred for this latter purpose, on account
of its purity.

Under the microscope ferrite may be distinguished from
cementite by its softness. If steel containing these two con-
stituents be polished on damp, rough parchment, or on chamois
skin stretched over a soft background (as wood), the ferrite will
wear away below the carbide and appear in intaglio. The same
effect will be obtained by Osmond's ' polish attack.'1 Ferrite
is also distinguished from carbide of iron by the fact that, after
being subjected to the brief action of certain reagents, such as
2 per cent, nitric acid, or ordinary commercial tincture of iodine,
the ferrite is seen in darker grains and the carbide in bright thin
plates, unattacked by the reagent. When the two are intimately
associated in minute grains, as in pearlite, the carbide appears
bright and the ferrite dark, because eaten away below the surface
by the reagent.

Allotropic Modifications. — There is one peculiarity of pure
iron, or ferrite, which, on account of its importance, deserves
special attention, namely, its ability to assume different allotropic
modifications at different temperatures. The nature of allo-
tropism has already been explained (see page 486). To the allo-

1 See page 453.

318 THE METALLURGY OF IRON AND STEEL

tropic modification of iron the names of 'alpha/ 'beta/ and
' gamma ' have been assigned. The alpha modification is the
one existing at atmospheric temperatures, and it is familiar to
all who make or use iron. If this be heated, however, it under-
goes a sudden change at about 760° C. (1390° F.). This change
is evidenced by an absorption of heat and the circumstance
that the iron loses almost entirely its power to attract a magnet;
that is to say, it becomes about as non-magnetic as lead, copper,
etc. The change in magnetism is accompanied by a change in
electrical conductivity, in specific heat and in other properties.
In short, the iron has changed in many of its properties without
undergoing any alteration in chemical composition. This new
allotropic modification of iron is known as beta iron.

If, now, beta iron be further heated to a temperature of about
890° C. (1634° F.), it again changes several of its properties and
becomes what is known as gamma iron. Gamma iron differs
from beta iron, especially in electrical conductivity and in crystal-
line form. Ferrite crystallizes always in the cubic system, and
Osmond 1 and Stead 2 have studied the variations of form assumed
by it and by its alloys with carbon. Osmond especially has
studied the crystallography of the gamma, beta, and alpha
modifications of the pure metal. Gamma iron does not crystallize
isomorphously with either beta or alpha iron, which crystallize
identically in cubes; but it assumes all the combinations of the
cube and octahedron, and, in the latter form would be isomorphous
with carbon in the diamond form. Therefore, in the isomorphous
mixtures (i.e., solid solutions) of iron and carbon one would
expect to find some carbon in the diamond form, which has indeed
been accomplished by Osmond. This is used as an argument in
favor of the belief that the solid solution is with carbon, not
with cementite ; for Osmond has shown 3 that cementite does not
assume any crystalline form which would be isomorphous with
ferrite. Beta and alpha iron do not crystallize isomorphously
with either carbon or cementite, which accords with the observed
tendency of ferrite to begin to separate from the solid solution
at the same temperature at which it changes from gamma to beta
iron.

If ferrite is in the gamma form, say at 1000° C. (1832°
F.), then it will undergo upon cooling the changes which, as I

1 See No. 110, page 332. 2 See No. 181, page 457. » See No. 181.

THE CONSTITUTION OF STEEL

319

have outlined, it undergoes on heating, but reversed. That is to

say, at about 890° C. (1634° F.) it will change from gamma

to beta iron; at 760° (1390° F.) it will change from beta to

alpha iron, in each case receiving again the properties which it

had before heating. In other words, the change 'from one allo-

tropic form to another is a reversible change,

taking place in one direction on cooling and

in the opposite direction on heating. The

change on cooling takes place at a slightly

lower temperature than on heating. This

is not because it is other than a true reverse

action, but because the change in either

direction is necessarily slow and lags a little

behind the temperature. The amount of

this lag will vary directly with the speed of

heating and cooling.

We never get pure gamma iron at any
temperature unless we start with iron free
from carbon, because gamma ferrite never
separates from the solid solution. If, how-
ever, we have a solid solution of iron with,
let us say, 0.2 per cent, of carbon, then this
solid solution will begin to precipitate fer-
rite at a temperature of about 830° C. (1524°
F.). This ferrite will be, of course, in the
beta form and will be non-magnetic until
the mass cools below 760° (1390° F.), when
the previously precipitated ferrite will change
from the beta to the alpha form. Any fer-
rite which separates from solid solution after
that temperature is reached will separate in
the first instance in the alpha form. For
example, if we have a solid solution, con-
taining, say, 0.7 per cent, carbon, this will commence to precipi-
tate ferrite below 760° C. (about 720°), and the ferrite will be in
the alpha form.1

1 Some maintain that the ferrite always precipitates as gamma ferrite,
then immediately changes to beta and next to alpha ferrite. Others
maintain that when the solid solution is cooled near the line G O S, it changes
from a solid solution of gamma iron into a solid solution of beta iron, and then

FIG. 247. — COOLING
CURVE OF PURE
IRON.

320 THE METALLURGY OF IRON AND STEEL

Cementite. — The carbide of iron, Fe3C, is, next to ferrite,
the most important constituent of steel, and practically all of
the carbon is present in this form.1 Cementite is very hard and
brittle, scratching glass with ease and flying into pieces under a
blow. It crystallizes usually in thin flat plates, which are large
in size (sometimes up to J in. in diameter) when there is much
cementite present. It is attacked by reagents less than most
of the other constituents and is in this way distinguished under
the microscope. It is a little difficult to distinguish, by micro-
scopic evidence alone, between steel consisting of pearlite with a
slight excess of cementite and steel, consisting of pearlite with a
slight excess of ferrite. The practiced eye can usually tell; but
a chemical analysis readily distinguishes, since steel with less
than 0.9 per cent, carbon will have excess ferrite over pearlite,
and that with more than 0.9 per cent, will have excess cementite.

Cementite contains 6.6 per cent, of carbon, or, roughly, is
one-fifteenth carbon. We may therefore tell the amount of
cementite in any steel by multiplying the amount of carbon by
fifteen.2 Cementite may be separated from steel by electrolysis.3
It is magnetic at ordinary temperatures, but not above 700° C.
(1292° F.).

The carbon united with iron in cementite has been given
various names, such as ' cement carbon/ or ' carbon of cementa-
tion' (because of its prominent appearance in cemented steels),
and 'carbon of the normal carbide/ ' annealing carbon' (because
all the carbon of well-annealed steels will usually be present as
cementite) .

Manganiferous Cementite. — Manganese forms a carbide having
the formula Mn3C. This is isomorphous with Fe3C, and we often
find the two carbides together in one crystal. The name cementite
is still applied to this crystal, although it must be recognized that

into a solid solution of alpha iron, from which alpha iron then precipitates.
These questions are chiefly academic and of almost no practical importance ;
nor is it yet known which is the actual order, though many are inclined to
agree with Professor Sauveur (see No. Ill) upon the latter order.

1 There is not wanting evidence in favor of other carbides being discovered,
such as Fe2C and FeC.

2 In other words, steel containing 0.5 per cent, of carbon will contain 7.5
per cent, cementite; steel containing 1 per cent, carbon will contain 15 per
cent, cementite; etc.

3 See No. 112, page 332.

FIG. 248. — FERRITE, PURE IRON

(ELECTROLYTIC.)
1000 diameters. Etched with picric acid.

FIG. 250. — BIG CEMENTITE CRYSTAL

IN PEARLITE.

Magnified 250 diameters. Polished in
relief.

PIG. 252. — ELONGATED BUBBLE OF

MANGANESE SULPHIDE.
Magnified 250 diameters. Unetched.

FIG. 249. — PEARLITE CRYSTALS.

Surrounded by ferrite. Magnified 250

diameters. Etched with HNO3.

FIG. 251. — CRYSTALS OF MANGA-

NIFEROUS CEMENTITE.

Magnified 50 diameters. Etched with

nitric acid.

FIG. 253. — EUTECTIC OF Fe3P AND

IRON.

Magnified 1000 diameters. Etched with
picric acid.

322 THE METALLURGY OF IRON AND STEEL

a part of the iron has been replaced by manganese, and the formula
for the compound is usually written (FeMn)3C. The amount
of manganese in these crystals is very variable, running almost
all the way from nothing to 100 per cent. As manganese has an
atomic weight almost the same as that of iron (Mn 55, Fe 56), one
weight of manganese will replace almost exactly an equal weight
of iron in the crystal. The peculiarity of the manganiferous
cementite is that the crystals of free cementite are liable to be
larger, especially when the proportion of manganese is large,
and are seemingly harder and more difficult to machine.

Manganese Sulphide. — Manganese and sulphur unite to
form manganese sulphide, having the formula MnS, and this
compound is found in all steels. Indeed, all of the sulphur will
be found in this combination, provided there is enough manganese
in the steel to unite with it. It is necessary to have more than
the theoretical amount of manganese for this purpose, however,
because unless there is a surplus present, the attraction of the
manganese for the sulphur does not seem to be always sufficient
to catch it all. Steel should therefore always contain about four
times as much manganese as sulphur, because it is advantageous
to have the sulphur all in the form of manganese sulphide.

Manganese sulphide is seen under the microscope as a dove-
gray substance before the polished material is etched with any
reagents. It is usually collected together in round drops, which
are sometimes, if large in size, seen to be elongated by the rolling
or hammering of the material (see Fig. 152).

Manganese tends to make the crystals of steel smaller, which
is advantageous, but makes the metal more liable to crack in
heating, and still more so in cooling suddenly from a red heat.

Iron Sulphide. — The bulk of the sulphur not united with the
manganese will be found in the form of iron sulphide, FeS. This
iron sulphide is more brittle than manganese sulphide, and
instead of coalescing in drops, it spreads out in webs or sheets.
It is therefore very weakening to the steel, because the area of
weakness is more extensive than the tiny spots of manganese
sulphide. Steel containing iron sulphide is liable to show poor
tensile test and low ductility. It is at the rolling temperature,
however, that iron sulphide produces the greatest weakness,
because at this point it is in a liquid form and therefore has
practically no adhesion to the crystals of steel, which are liable

THE CONSTITUTION OF STEEL 323

to break along the planes or meshes of sulphide. The same is
true, to a less extent, of the effect of manganese sulphide, which
is also in a liquid or pasty condition at the rolling temperature;
but the extent of its damage is not so great on account of its
drawing together in drops.1 These facts explain the well-known
beneficial effect of manganese in counteracting the damage due to
sulphur in iron and steel.

Iron Phosphide. — Iron forms at least one phosphide, having
the formula FesP, and this phosphide forms with iron a series of
alloys, of which the eutectic contains 64 per cent, of Fe3P (10.2 per
cent, of phosphorus). This results in a considerable lowering ot
the melting-point of iron for each addition of phosphorus. Even
1 per cent, of phosphorus will make the melting-point of the
metal very much lower, and it is for this reason that foundry irons
are often desired with a high content of phosphorus. Even
where there is a smaller amount of phosphorus, there will be some
of the phosphorus eutectic formed, and this remains in a molten
condition for some time after the bulk of the steel has solidified.
This liquid eutectic tends to migrate to the spaces between the
crystals, where it remains after solidification and forms a very
brittle network, which naturally makes the whole mass more or
less brittle. For these reasons phosphorus is the greatest source
of brittleness in steel, and especially brittleness under shock,
and is thus a great enemy to the engineer and other users of
the material.

Besides forming the eutectic, as I have described, phosphorus
also tends to produce coarse crystallizations in steel, and this
makes it both weak and brittle. It is a fact observed many
times that the embrittling effect of phosphorus on steel is much
less when the steel is very low in carbon, and as the carbon rises
so the brittleness caused by phosphorus rises. This and other
effects of phosphorus have been explained by Prof. J. E. Stead
in two very able papers.2 He has shown that a little phosphorus
will dissolve in ferrite and that then the eutectic which produces
the brittleness will not form, but as the carbon in the steel in-
creases, it precipitates the phosphorus from the ferrite solid
solution and therefore causes the eutectic to form. Hence, the
more ferrite and the less cementite in steel, the less will be the
brittleness produced by phosphorus.
1 See No. 113, page 332. 2 See No. 184 on page 157 and No. 115 on page 332.

324 THE METALLURGY OF IRON AND STEEL

Iron Silicides. — There seem to be three or more silicides of
iron, but the one having the formula FeSi seems to be found
most commonly in steel. It appears to increase very slightly
the strength of steel, and also, to a limited extent, its hardness.
The chief importance of silicon, however, as already pointed out,
is in promoting soundness.

Iron Oxides. — Oxygen occurs in steel in the form of FeO
and Fe.203. In either form its presence is very harmful, producing
brittleness in both hot and cold steel, besides causing the liability
to blow-holes already discussed. There is probably no constit-
uent more harmful to steel than oxygen, and unfortunately the
chemists are far behind in respect of not yet having found a
satisfactory method of accurately determining small traces of
this gas. The effect of oxygen is somewhat similar to that of
sulphur and, in common parlance, makes the steel 'rotten/

Nitrogen and Hydrogen. — Both nitrogen and hydrogen occur
in steel, and one of the theories to explain the superiority of
crucible steel is based upon the relative freedom of this material
from these two gases. The amount of nitrogen and hydrogen
present is usually very small. Hydrogen dissolves very easily
in iron at a high temperature, but is evolved in part as the metal
cools. In order to obtain entire freedom, however, it is necessary
to heat and cool several times in vacuo.

THE STRENGTH OF STEEL

The properties of steel most commonly desired are strength
and ductility. Unfortunately there is more or less incompatibility
between these two. That is to say, as the strength of steel
increases, the ductility usually decreases; and, conversely, as the
ductility increases, the strength usually decreases. There are
other properties of steel which are likewise of importance, either
because they are desired, or the reverse. Among these we shall
especially discuss hardness, brittleness, electric conductivity, mag-
netic permeability, magnetic hysteresis, permanent magnetism
and weldability.

Pure iron has a tensile strength of about 45,000 Ibs. per square
inch and a compressive strength of about 80,000 Ibs. per square
inch. These are increased by means of several of the ordinary
impurities found in steel, but the most important strengthener is

THE CONSTITUTION OF STEEL

325

carbon, because this will give the maximum increase in strength
with the least decrease in ductility.

Carbon. — Each increase in carbon (cementite) gives an in-
crease in tensile strength until we reach a maximum of about
0.9 to 1 per cent, of carbon (13.5 to 15 per cent, cementite). With
further increase in cementite there is a decrease in tensile strength.
It is probable that the reason for this maximum of tensile strength
at approximately the eutectoid ratio of the steel is due largely

170,000

140,000

110,000

80,000

50,000

Percentage of Carbon
0.10 0.20 0.30 0.400.50 O.CO 0.70 0.80 0.90 l'00 1.10 1.20 1.30 1.40 1.50 1.60 1.70 1.80 1.90

2.00

RELATION BETWEEN THE PHYSICAL

CHARACTERISTICS OF STEEL AND THE

PROPORTIONS OF FERRITE AND Fe3C

00*

80*

<Fe300 5

J Ferrite 100 95

30 £Fe3C
70*Fenlte

FIG. 254. From Howe, "Iron, Steel and other Alloys."

to the small crystallization and the intimate mixture of the
crystalline constituents when the steel is at, or near, the eutectoid
proportions.1 With less cementite the grains of pearlite are sur-
rounded by a network of ferrite; with more cementite the grains
of pearlite are surrounded by a network of cementite, and both
of these networks have a weakening effect upon the material by
decreasing the adhesion of the crystals. The relation between
carbon and tensile strength is shown graphically in Fig. 254.2

1 The still greater increase in strength when the two constituents of the
pearlite (ferrite and cementite) are even more intimately associated is ex-
plained on page 392. 2 See page 438 of No. 1 on page 8 herein.

326 THE METALLURGY OF IRON AND STEEL

H. H. Campbell and W. R. Webster have studied exhaustively
the effect of the different impurities upon the tensile strength of
steel. The latest word on this subject has been said by Campbell
who gives 1 the following rule for the effect of each 0.01 per cent,
of carbon on acid and on basic open-hearth steel. Starting with
40,000 Ibs. per square inch for pure steel, each 0.01 per cent, of
carbon will increase the strength by 1000 Ibs. per square inch in
the case of acid open-hearth steel, and by 770 Ibs. per square
inch in the case of basic open-hearth steel. Because the color
method of determining the amount of carbon does not show all
the carbon present, the figures given above must be replaced by
1140 Ibs. and 820 Ibs. for each 0.01 per cent, of carbon, when the
color test is used. We have no data to determine the strengthen-
ing effect of carbon in Bessemer and crucible steel, but it is prob-
able that a little lower value than that given for basic open-hearth
steel would be used for Bessemer, and a little higher value than that
given for acid open-hearth steel would be used for crucible steel.

Silicon. — The effect of silicon on strength is probably very
small in the case of rolled steel. In the case of castings, however,
an important increase in tensile strength may be obtained by
increasing the silicon to 0.3 or 0.4 per cent. This practice results
in practically no decrease in ductility, but it is necessary to supply
larger risers on account of the deep piping that will be produced.
It is probable that the beneficial effect of silicon in this case is due
very largely to its producing soundness.

Sulphur. — H. H. Campbell says that the effect of sulphur on
the strength of acid and basic open-hearth steel is very small.
It is probable, however, that this statement is only true when
the sulphur is in the form of manganese sulphide, because the
effect of iron sulphide would be to lower the strength and the
ductility of the material. The worst effect of sulphur is un-
doubtedly its production of ' red-shortness ' and the liability to
cause checking during rolling, or, in the case of a casting, during
cooling.

Phosphorus. — Campbell states that each 0.01 per cent, of
phosphorus increases the strength of the steel by 1000 Ibs. per
square inch. It should be observed, however, that this increase
of strength is measured by the resistance of the material to
stresses slowly applied and that it ceases with 0.12 per cent.
1 Page 391 of No. 2, page 8.-

THE CONSTITUTION OF STEEL 327

phosphorus and is reversed. In the case of vibratory stresses and
sudden shocks, phosphorus is probably the most harmful of the
elements, so that it is undesirable to increase the strength of steel
by means of this element. This is the more true because of the
brittleness produced by phosphorus, for an increase of strength
obtained through this medium is accompanied by a much greater
decrease in ductility than when the same increase in strength is
obtained through the medium of carbon.

Manganese. — The beneficial effect of manganese on tensile
strength begins, according to the same authority, only when the
manganese is above 0.3 or 0.4 per cent. With less manganese
than this, in the case of open-hearth and Bessemer steels, the
presence of some other condition, probably iron oxide, masks the
effect of manganese. It will be remembered that when the man-
ganese is low, open-hearth and Bessemer steels are harmfully
charged with oxygen. Furthermore, the effect of manganese
is dependent upon the amount of carbon present. In acid open-
hearth steel each 0.01 per cent, of manganese (beginning at 0.4
per cent.) increases the strength 80 Ibs. per square inch when the
carbon is 0.1 per cent., but each increase of 0.01 per cent, of carbon
increases the strengthening effect of manganese by 8 Ibs. So
that, for example, if we have an acid open-hearth steel containing
0.4 per cent, of carbon, then each 0.01 per cent, of manganese
will increase the strength by 320 Ibs. per square inch. In the
case of basic steel each 0.01 per cent, of manganese (beginning
at 0.3 per cent.) increases the strength 130 Ibs. per square inch
when the carbon is 0.1 per cent., and so on to 250 Ibs. per square
inch when the carbon is 0.4 per cent., for each additional 0.01 per
cent, of carbon increases the strengthening effect of manganese
on basic steel by 4 Ibs. (See Table XXVI.)

Formula?. — Campbell gives the following formulae for the
strength of acid and basic open-hearth steels:

For acid steel: 40,000+ lOOOC+lOOOP + xMn+R = ultimate strength.
For basic steel: 41,500+ 770C + 1000P+yMn+R = ultimate strength.

In these formulae, 0^=0.01 per cent, carbon (determined by com-
bustion), P = 0.01 per cent, phosphorus, Mn = 0.01 per cent,
manganese, and R = a variable, depending upon the heat treat-
ment which the steel has received. For X and Y see Table XXVI.
Copper. — Copper, up to at least 1 per cent., does not appear

328

THE METALLURGY OF IRON AND STEEL

TABLE XXVL— EFFECT OF EACH 0.01 PER CENT. MANGANESE
ON OPEN-HEARTH STEEL

PERCENTAGE OF CARBON

On Acid Steel

On Basic Steel
Y

Lbs. per sq. in.

0 05

HO2

0 10

801

130

0 15

120

150

0 20

160

170

0 25

200

190

0 30

240

210

0 35

280

230

0.40

320

250

0 45

360

0 50

400

0 55 .

440

0.60

480

1 Beginning only with 0.4 per cent, of manganese.

2 Beginning only with 0.3 per cent, of manganese.

to have any important effect upon the strength or ductility of
low- and medium-carbon steels, but increases the brittleness of
steel containing 1 per cent, carbon. When the sulphur is more
than 0.05 per cent, copper appears to make the steel roll less easily
and above 0.5 per cent, of copper appears to make high-carbon
steel draw less easily into wire.

Arsenic. — Some steels contain arsenic, which does not appear
to have any effect when it is below 0.17 per cent., but any larger
quantity raises the tensile strength and decreases the ductility to
a very important extent.

Oxide of Iron. — All Bessemer and open-hearth steels contain
more or less oxide of iron, there probably being more in basic steel
than in acid, and more in Bessemer steel than in basic open-
hearth steel. It is probable that this oxide of iron has not any
very marked effect upon strength, as Campbell quotes some steels
containing quantities larger than usual whose strength is good.
No data are given as to the ductility, however, and it is probable
that oxide of iron has a deleterious effect upon this quality.

HARDNESS AND BRITTLENESS OF STEEL

As a general thing the hardness and brittleness of steel increase
together. The chief commercial application of this property is
in such articles as railroad rails and car- wheels, bevel- and spur-

THE CONSTITUTION OF STEEL 329

gears, axles and bearings, the wearing parts of crushing machinery,
etc.1 To produce hardness in these materials carbon is the chief
agent used, because it gives the maximum hardness with the least
brittleness. It is for this reason that railroad rails are now made
up to 0.7 per cent, carbon, and although this material is somewhat
brittle and breakages occasionally occur, the high carbon is de-
manded in order that the head of the rail may be durable. It is
probable that even higher carbon than this would be used if it
were not for the brittleness already in the ordinary railroad rails,
due to the fact that, being made by the acid Bessemer process, they
contain nearly 0.1 per cent, phosphorus. Phosphorus increases
both the hardness and the brittleness, especially under shock.
Manganese likewise increases hardness, and especially the kind
of hardness which makes it more difficult and more expensive to
machine the metal. With more than about 1 per cent, of man-
ganese the steel becomes somewhat brittle. When the content
rises above 2 per cent, the steel is so brittle as to be practically
worthless. In this connection a curious phenomenon is observed,
in that a still further increase in manganese produces a reversal
of its influence, and when we have more than 7 per cent, the
metal is not only extremely hard and practically impossible
to machine commercially, but becomes, after heat treatment,
very ductile and tough. This will be considered more fully in
Chapter XV.

Silicon. — Silicon makes the steel slightly harder, but ap-
parently without increasing its brittleness, unless we have more
than 0.5 or 0.6 per cent.

ELECTRIC CONDUCTIVITY OF STEEL

The purer the material and the nearer it is to ferrite, the better
will be its electric conductivity; therefore only wrought iron, or
the softest and purest forms of steel, are used generally for wire
for electric conduits.2 The case is somewhat complicated when

1 1 purposely omit here the consideration of such hardness as that pro-
duced in springs, cutting tools, etc., by heating the steel to a bright-red heat
and plunging into water, as this will be discussed in Chapter XIV.

2 Omitting, of course, the use of copper, which is not under consideration
in this book, but which is probably the most important material used for
electric conductors.

330 THE METALLURGY OF IRON AND STEEL

we come to third rails for electric railroads, which have now
become a very important industrial product, because so pure a
material will be very soft and will rapidly wear away under the
abrasion of the contact-shoes. To increase the hardness of these
rails with the least decrease in electric conductivity, it is best to
avoid nickel and manganese, which decrease electric conductivity
in greater proportion than the other elements, and to obtain the
hardness as much as possible from phosphorus, and not from
carbon, because phosphorus will give the greatest amount of hard-
ness with the least decrease in the purity of the iron. As high
phosphorus steels are difficult to roll, however, the section of the
rail chosen should be as free as possible from sharp corners and
thin flanges, in order that the tearing action in rolling may be as
slight as possible.

MAGNETIC PROPERTIES OF STEEL

Alpha ferrite is the magnetic constituent of iron and steel,
and therefore the greater the amount of this constituent present,
the greater will be the magnetic force and magnetic permeability
of the material and the less its magnetic hysteresis. On this
account the cores of electromagnets, the armatures of dynamos,
etc., are commonly made of Swedish wrought iron, which is the
purest commercial form of iron made.1

Ferrite has no permanent magnetism, but immediately loses
its magnetic force when it is out of contact with a magnet, or, in
the case of cores of electric magnets, when the electric current is
cut off. Permanent magnets are therefore made of a high-carbon
steel (1 per cent.), whose strength and permanency are increased
if about 5 per cent, of tungsten is present. This steel is heated
above the critical temperature and hardened in water, after which
it is magnetized by causing an electric current to flow around it for
a short time. Steel so treated will retain the magnetic force for
many years. Osmond has explained the permanent magnetism
of steel in the following very ingenious manner: Each molecule
of alpha ferrite is believed to have a north and a south magnetic
pole, which, in the ordinary unmagnetized condition of the iron,

1 A silicon alloy of iron with a double-heat treatment discovered by R. A.
Hadfield, and having a very high magnetic force and permeability, will be
discussed under the head of "Alloy Steels."

THE CONSTITUTION OF STEEL

331

will be oriented in many different directions, as shown in Fig. 255.
When this piece of iron is placed in the magnetic field, however,
the molecules arrange themselves in accordance with the lines of
magnetic force, with their north poles all in one direction and
their south poles all in the opposite direction, thus making the
whole piece of iron a magnet. As soon as the magnetic force is
removed, however, the molecules all return to their original
orientation, and the whole piece of iron loses its magnetism. We
have already learned that it is only the alpha molecules which have

FTG. 255.

C» Of 0 O &

£* c* <=»

FIG. 256.

north and south magnetic poles, and if the steel consists entirely
of beta or gamma molecules, it is not capable of becoming magnetic.
According to Osmond's theory, when steel is cooled rapidly from
above the critical temperature, the shortness of the time taken in
reaching the atmospheric temperature is such that only a part
of the molecules are able to change to the alpha allotropic form,
and the remainder are retained in the gamma form. This re-
tention is assisted by the 1 per cent, of carbon present, which
acts as a brake to make the change slower. When, now, this
hardened steel is subjected to the magnetic force, the alpha
molecules orient themselves with their north poles all in one
direction; but when the magnetic force is removed there is a
certain number of gamma molecules present to interfere with the
free movement of the alpha molecules and prevent them from

332 THE METALLURGY OF IRON AND STEEL

returning to the original unoriented position. This explanation
implies that the magnetic force is sufficient to overcome the
resistance of the gamma molecules and force the alpha molecules
into a magnetic position, but that the force of the alpha molecules
tending to return to their original position is not so great.

The welding of steel and the effect of different elements upon
it will be discussed in Chapter XIV.

REFERENCES ON THE CONSTITUTION OF STEEL

110. F. Osmond and G. Cartaud. "The Crystallography of Iron/'

Journal Iron and Steel Institute, No. Ill, 1906, pages 444-
492.

111. Albert Sauveur. "The Constitution of Iron-Carbon Alloys."

Same Journal, No. 4, 1906, pages 493-575.

112. J. O. Arnold and A. A. Read. "Chemical Relations of

Carbon and Iron." Journal of the Chemical Society, vol.
Ixv, page 788.

113. J. O. Arnold and G. B. Waterhouse. "The Influence of

Sulphur and Manganese on Steel." Journal Iron and Steel
Institute, No. 1, 1903, pages 136-160.

114. J. O. Arnold. "The Influence of Elements on Iron." Same

Journal, No. 1, 1894, page 107.

115. J. E. Stead. Same Journal, No. 11, 1900, pages 60 et seq.;

and also Cleveland Institution of Engineers, September 6,
1906.

116. Bradley Stoughton. " Notes on the Metallography of Steel."

International Engineering Congress, 1894. Transactions,
American Society of Civil Engineers, vol. liv, Part E, pages
357-421.

117. Proceedings of the American Society for Testing Materials,

vol. i, 1901; vol. vii, 1907. Philadelphia. This society is
composed of representative men from the great purchasers
and users of engineering and other kinds of materials in
America, from the manufacturers and of representative
scientific men. All sides of the question are usually repre-
sented in the discussions of this society.

118. Proceedings of the International Association for Testing

Materials. Vienna, Austria.

XII
THE CONSTITUTION OF CAST IRON

PRACTICALLY all the cast iron which is not purified is used
for making iron castings, so that a study of the constitution of
cast iron resolves itself into a study of iron castings. The difference
between cast iron and steel is that the former contains less iron
and more impurities, especially carbon, silicon, phosphorus, and
occasionally sulphur and manganese. The advantages of cast
iron, and the reason it is used as much as it is, are its fluidity,
lesser amount of shrinkage when cooling from the molten state,
relative freedom from checking in cooling, and the ease with which
very different properties are conferred upon it at will. Its dis-
advantages are its weakness and lack of ductility and malleability.
The last-named deficiency renders it practically impossible to
put any work upon cast iron ; hence it can never be wrought to
shape and must always be used in the form of castings. Its most
important advantage is probably its ready fusibility, which makes
it so easy to melt and cast, and its cheapness.

Graphite. — All of the characteristic qualities of cast iron are
due to the presence of the large amount of impurities in it. These
impurities are the same in kind as the impurities in steel, and
differ only in amount, with the single exception of graphite.
This constituent is almost never found in steel, or is found in
such a very small number of cases, and those cases being confined
to the high-carbon steels, the amount of which is very small in
comparison, that we may almost disregard it. In cast iron,
however, it is the largest and one of the most important con-
stituents. It occurs in thin flakes, in sizes varying from micro-
scopic proportions to an eighth of a square inch in area,1 dis-
seminated through the body of the metal and forming an intimate

1 In rare and unusual cases the flakes of graphite may be as much as an
inch and a half to two square inches in area, but practically never so large
in the commercial cast irons.

333

334 THE METALLURGY OF IRON AND STEEL

mechanical mixture, a magnified section of which is shown in
Fig. 259. Each flake of graphite is composed of smaller flakes,
built up somewhat like the sheets of mica with which all are
familiar, but with very little adhesion between the small com-
ponent flakes, so that the sheet of graphite may be split apart with
very little force. Graphite is very light in weight, having a specific
gravity of only about 2.25 as compared with a specific gravity of
7.86 for pure iron; consequently although the percentage of
graphite by weight is only 4 per cent, or less of the iron, its per-
centage by volume may be, in normal cases, as much as 14 per
cent. This may readily be seen by noting the amount of space
occupied by the graphite flakes in Fig. 263.

Combined Carbon. — We have already discussed the solidifica-
tion and cooling of cast iron, and it will be remembered that all
the carbon which does not precipitate as graphite forms first
as austenite, • which later decomposes into ferrite and cementite.
In short, all the carbon in cast iron will ultimately be found partly
in the form of graphite and partly in the form of cementite. The
carbon of cementite in cast iron commonly goes under the name
of ' combined carbon/ but it must be remembered that cementite
is the constituent which gives the observed effects.

White Cast Iron. — In white cast iron the carbon, amounting
often to 3 or 4 per cent., will all be in the form of cementite, which
will therefore form 45 to 60 per cent, of the material ; consequently
white cast iron will possess largely the properties of cementite.
It is very hard and brittle, being machined only with the greatest
difficulty and with special kinds of cutting tools, and resisting
wear by abrasion very effectively. It is so brittle as to be readily
broken by the blows of a hammer, and is weak because of the
presence of very large plates of cementite, which adhere but
slightly to one another. Consequently white cast iron has few
uses and is employed usually only as a hard surface on the outside
of gray-iron castings.

Gray Cast Iron. — Gray cast iron will have about the same
total amount of impurities present as white cast iron, the only
difference being that the carbon is now partly or wholly pre-
cipitated as graphite. The gray color of a freshly broken fracture,
from which this material receives its name, is due altogether to
the graphite present, for this constituent is so weak that the
iron breaks chiefly through its crystals, which are rent asunder,

THE CONSTITUTION OF CAST IRON 335

leaving one part sticking to each side of the fracture. The weak-
ness of gray cast iron as compared with steel is thus readily
understood, since there is but a small proportion of metallic
surface to be broken and the graphite splits so easily. An inter-
esting experiment is to take a freshly broken surface of gray pig
iron and brush one-half of it for some time with a stiff brush.
In this way the adhering crystals of graphite are partially removed
and we get a surface which is almost as white as the fracture of
white cast iron. This shows clearly that the gray color is due
altogether to the graphite and that the metallic part is as silvery
white as iron itself. The prevalence of the gray color also shows
how completely fracture takes place through the graphite crystals.
Gray cast-iron castings are by far the more important, and
the study of their constitution is the chief object of this chapter.
These castings usually contain 2 per cent, or more of graphite
and less than 1 J per cent, of combined carbon. It will be observed
that this limit of combined carbon is also the range found in steel.
Furthermore, it will be observed that the graphite is not a chemical
component of the metallic body, but is mechanically mingled with
it. In this sense, therefore, we may consider gray cast iron as a
very impure steel,1 mechanically mixed with graphite, and upon this
reasoning the study of its constitution becomes much simpler,2
as we may study first the properties of the metallic part, and
next that of the graphite, and so be able to foretell to some extent
the properties of the mixture. Indeed, the properties of the
metallic part are already understood pretty well from our dis-
cussion of the constitution of steel, and there is no new con-
stituent or new condition except the larger amounts of silicon and
phosphorus, which are of minor importance, because their effect
is collectively far less than the weakening and embrittling effect
of the graphite. Even though we had a very pure metallic

1 The silicon in gray cast iron is usually between 0.75 per cent, and 3 per
cent., or, let us say, ten times that in steel, while the phosphorus is usually
from 0.5 to 1.5 per cent., or, again, about ten or more times that in steel.
The sulphur varies greatly, but is not infrequently as high as 0.15 to 0.2
per cent. Manganese is an exception and is usually no higher in cast-iron
castings than in steel.

2 This theory of the constitution, which meets with very favorable ac-
ceptance in many quarters, was independently evolved by J. E. Johnson, Jr.
(American Machinist, 1900), and H. M. Howe (Trans. A. I. M. E., 1901,
vol. xxxi, pp. 318-339).

336 THE METALLURGY OF IRON AND STEEL

constituent, the strength and ductility of this portion would not
be sufficient to prevent the mass as a whole breaking at a small
load and without exhibiting any practical ductility, because of
the weakening effect of the crystals of graphite. In other words,
it is the carbon which is the great factor in determining the
properties of cast iron, for this may be either all graphitic, or all
combined, or part in both conditions.

Effect of Temperature. — By running the blast furnace very
hot, we may extend the saturation point of the iron for carbon
and thus get a slightly higher total carbon. This is not a very
potent influence, however, for we seldom have total carbon more
than 4.5 per cent., or less than 3.25 per cent. This control, such
as it is, may be exercised either during the manufacture of the
pig iron or during the remelting in the cupola, because in the
latter furnace the liquid iron is in contact with coke and will
absorb carbon up to its saturation point at the existing temperature.

Rate of Cooling. — A far more potent influence, however, is the
transfer of carbon from the graphitic into the combined form,
or vice versa, by rapid or by slow cooling from the molten con-
dition. It will be remembered that the carbon is always dissolved
in the iron when the mass is in a molten condition, that is, when
it is above the lines A B and B C in Fig. 246. As we cool from
the molten state, graphite precipitates, but this cooling must be
very slow indeed for this normal chemical change to take place
completely since it is a very sluggish change and requires several
seconds for its accomplishment. If, therefore, we cool with great
rapidity, as, for example, by pouring the iron into a metallic mold
which ' chills ' it, or by some other form of artificial rapid cooling,
we may prevent the precipitation of graphite by denying the time
necessary for the chemical reaction, and obtain a metal in which
all the carbon is in the combined form, i. e., white cast iron. It is
also evident that, by a rate of cooling intermediate between this
rapid rate and the slow rate which permits the precipitation of
the normal amount of graphite, we may obtain an intermediate
amount of carbon in the graphitic form. This variation in the
speed of solidification is a very important means of producing
combined carbon and is employed very largely in the 'chilling7
of the surfaces of gray-iron castings, whereby we may have a
relatively soft gray iron in the interior of each article and a hard
surface extending to almost any desired depth. For example,

THE CONSTITUTION OF CAST IRON 337

chilled-iron rolls are made in this way (see page 201), and also
American railroad car wheels l which are cast against an iron
chill (see page 337), giving nearly an inch depth of white iron
around the tread and flange where the metal is to suffer abrasion
in grinding over the rails, while the web and bore will be of gray
cast iron, because cooled more slowly in the sand part of the
mold, and thus will be less brittle and better able to withstand
the shocks of service and machining.

THE EFFECT OF CARBON ON CAST IRON

The nature and constitution of gray cast iron is far more
difficult to understand than that of steel, and even greater is the
difficulty of predicting the effect of any change in composition or
in constituents. The chief reason for this complexity is that a
change in any one of the constituents of gray cast iron is liable
to effect changes in several others as well. The simplest example
of this is in the case of the carbon; we have total carbon, graphite,
and combined carbon, and if we change any one of these three,
we must change either one or both of the other two, and it makes
a great deal of difference which. Indeed, we almost never change
the amount of graphite without making the reverse change in the
amount of combined carbon, and vice versa. Thus a very loose
system of speaking of these matters has come into vogue among
foundrymen. For instance, it is very common to hear a foundry-
man say: 'In order to soften your iron, increase the graphite';
but what he really means is : 'In order to soften your iron, decrease
the combined carbon.' He knows that the one change usually
follows from the other, and he speaks of it in this way, regardless
of the fact that graphite can be increased (i. e., by increasing
the total carbon and leaving the combined carbon the same or
a dittle higher) , and yet the iron will not be made any softer, but
may .even be harder.

Graphite and Shrinkage. — The most important effect of
graphite on cast iron, aside from causing weakness, is in decreasing

JIn other important railroad countries it is more usual to have the car
wheels made of steel, as it is believed that the iron wheels are not sufficiently
strong and ductile. The manufacture of pressed-steel car wheels is increasing
in America; nevertheless, the chilled cast-iron wheels seem to give very good
service.

338 THE METALLURGY OF IRON AND STEEL

the shrinkage. The reason for this will be understood when we
consider what happens when cast iron solidifies. It will be
remembered that when the eutectic forms, the cast iron breaks up
into alternate plates of graphite and austenite. This separation
of graphite from solution is the birth of a new constituent, and
this constituent occupies space, so that there is an expansion of
the mass as a whole in proportion to the amount of graphite that
separates. If, therefore, we pour liquid cast iron into a mold,
which is, of course, entirely filled at the moment when the iron
begins to solidify, the first action that takes place after the begin-
ning of solidification is an expansion, due to the separation of
graphite. The expansion continues for several moments until
the chemical precipitation is completed, after which the metal
begins to contract, as all metals do in cooling from a high tempera-
ture; but the preliminary expansion has been so great that the
ultimate shrinkage may be only about one-half what it otherwise
would have been. We can thus control this shrinkage by con-
trolling the amount of the expansion, through varying the graphite.
This point will be more readily understood by referring to Fig. 257r
which is taken from a recent article by Prof. Thomas Turner of
England.1

Explanation of Fig. 257. — The point 0 marks the position
occupied by the end of the bars at the moment of solidification.
It will be seen that in the case of copper the metal contracts con-
tinuously from this point, as shown by the continuous drop of the
curve. In the case of white cast iron, the metal contracts con-
tinuously until we reach a certain point (which is at a temperature
of about 665° C.) , when a momentary arrest of the shrinkage takes
place, after which the metal again contracts. This arrest is
common to all cast iron and steel and marks the decomposition of
austenite at the point S, Fig. 246, page 314. Now, see what a
difference there is in the case of gray cast iron, which does not
shrink immediately after freezing, but expands very appreciably,
as shown by he rise in the curve. This expansion is due to the
graphite that is being expelled from the metal and occupies space
between the particles of iron.

Again, in the case of the Northampton iron, which is high in
both graphite and phosphorus, the expansion is very long-con-
tinued, and the metal cools to almost a black heat before the
1 Journal of the Iron and Steel Institute, No. 1, 1906, page 57.

THE CONSTITUTION OF CAST IRON

339

bar has shrunk again to the size it had when first cast. This
expansion is again due to the separation of carbon, and is assisted
apparently by the phosphorus keeping the iron in a semifluid
condition for a long time and thus allowing the graphite more
easily to separate and make place for itself. Here, too, we have
an explanation why phosphoriferous irons fill every crevice of

FIG. 257. — SHRINKAGE CURVES.

the molds so perfectly. Being in a pasty condition for some
time, and continually expanding, the semifluid metal is forced
into the tiniest crevices of the molds, filling all the corners with
astonishing sharpness.

It is evident that any increase in graphite, whether caused
by an increase in total carbon or by a decrease in combined carbon,
will produce less shrinkage. The extent of this may be judged by
noting that gray cast iron expands so much in solidifying that no
contraction cavity or pipe is formed, such as occurs in the case of

340 THE METALLURGY OF IRON AND STEEL

steel. If the iron is only slightly gray, or if it is a very large
section of metal, then a slight spongy place may be formed in the
center, which is the nearest approach to a shrinkage cavity that
is normally found in most iron castings.

Graphite and Porosity. — It is also evident that an increase
in graphite, whether produced by an increase in total carbon or
by a decrease in combined carbon, will increase the porosity of
the casting, which is often disadvantageous, as in the case of
hydraulic cylinders or other receptacles for holding liquids under
pressure. The separation of much graphite usually is accompanied
by large-sized graphite crystals, and therefore the crystals of
the mass, as revealed by the fracture, appear large and the grain
is said to be 'open/

Graphite and Workability. — When we come to consider the
effect of graphite upon the softness or workability of the cast
iron, it is evident that we must consider it in relation to other
things; for if we increase the graphite by increasing the total
carbon, then we increase the workability of the metal only in so
far as the graphite acts as a lubricant for the tool that is doing
the cutting. Evidently the tool will have no difficulty in cutting
through the soft flakes of graphite; the chief resistance to it will
be given by the metallic part of the mixture. Though the lubricat-
ing effect of the graphite undoubtedly helps the tool, its presence
evidently cannot increase the actual softness of the metallic
body with which it is mixed. But, on the other hand, if we
increase the graphite by decreasing the combined carbon, then
we have not only increased the amount of lubricant, but we have,
in addition, increased the softness of the metallic part of the
mixture by reducing the proportion of cementite, which is the
hardener, in it.

Graphite and Strength. — Everything else being equal, it is
evident that the more graphite we have in cast iron, the weaker
it will be, for we have already shown that gray cast iron breaks
by the ready splitting apart of the flakes of graphite. Thus, if
we make no change in the combined carbon, but increase the total
carbon of our castings, and consequently the graphite, we should
expect a corresponding decrease in strength, and this is in fact
found to occur. When, however, we increase the graphite and at
the same time decrease the combined carbon, we may or may not
get an increase in strength, this depending altogether on how

THE CONSTITUTION OF CAST IRON 341

much combined carbon there was before and after the change.
For example, if we had 3 per cent, of combined carbon and 1 per
cent, of graphite in a casting that casting would be weak be-
cause of too high combined carbon. To decrease this and in-
crease the graphite would have a beneficial effect on strength.
On the other hand, if we had 1 per cent, of combined carbon
and 3 per cent, of graphite, to decrease the combined carbon
and increase the graphite would have a detrimental effect on
strength. We must therefore consider the question of strength
from a much larger viewpoint than by considering any one con-
stituent alone.

Combined Carbon and Shrinkage. — Combined carbon has very
little effect on the shrinkage of cast iron except in so far as it
changes the graphite. That is to say, if by increasing the com-
bined carbon we decrease the graphite, we will get an increase in
shrinkage, and vice versa.

THE EFFECT OF SILICON, SULPHUR, PHOSPHORUS, AND
MANGANESE ON PIG IRON

The constitution of cast iron is, furthermore, very complicated
because of the double influence of silicon, sulphur, phosphorus,
and manganese, Each of these elements has a direct influence
upon the properties of the material, which is in general similar to
its influence upon steel. For example, silicon produces freedom
from oxides and blow-holes and makes the iron more fluid ; man-
ganese counteracts the effect of sulphur and increases the difficulty
of machining the material; sulphur makes the metal very tender
at a red heat, and therefore liable to checking if put under strain
during this period. For example, if a casting in shrinking tends
to crush the sand, this strain will be more liable to break it in case
the sulphur is high. Sulphur also makes solidification take place
more rapidly, and causes blow-holes and dirty iron. Phosphorus
makes the metal very fluid and reduces its melting-point. It also
makes it more brittle under shock, especially when cold, and
produces a fusible eutectic, a photomicrograph of which is shown
in Fig. 253. Phosphorus and sulphur increase the tendency to
segregate.

Furthermore, the various compounds of the impurities with
iron and with each other, which we find in steel, are also found in

342 THE METALLURGY OF IRON AND STEEL

cast iron. Indeed, some of them are far more important in the
latter than in the former, because the amount of the impurities is
greater. This is especially true of manganese sulphide and iron
sulphide, for the sulphur in cast iron is often large in amount,
while the manganese is often intentionally small on account of the
difficulty which this element produces in the machining of the
casting. Therefore we are even more liable to find iron sulphide
in cast iron than in steel.

But the direct effect of these impurities is usually far less
important than their indirect effect, namely, their influence upon
the carbon. After all, it is the carbon which is the chief factor in
controlling the most important properties of the cast iron, and
we may vary this either by increasing or decreasing the total
amount, or else leaving the total amount the same, by increasing
the graphite and decreasing the combined carbon, or vice versa.
It is the ease with which we may vary the amount or the condition
of the carbon, and therefore the properties of the iron, that is one
of the most important advantages of the material. But strangely
enough, although it is easy to keep this control, it can never
be accomplished in a direct way. It will be remembered (see
pages 35 to 38), that the blast-furnace manager can vary the
amount of silicon and sulphur in his pig iron at will; that he
has only a small control over the manganese and practically
none over the phosphorus or carbon, but that the metal al-
ways saturates itself with this latter element; and it has also
been seen (see page 306) that the limit of this saturation is
small.

Silicon. — By means of his control over the silicon and sulphur,
the metallurgist exercises indirectly his most important control
over the condition of the carbon; for silicon acts as a precipitant
of carbon, driving it out of combination and into the graphitic form,
so that with about 3 per cent, of silicon, slow cooling and very
low sulphur and manganese, we may obtain a cast iron in which
almost none of the carbon is in the form of cementite. That is to
say, the presence of this amount of silicon acts so strongly that
it may partially prevent the formation of austenite during solidifica-
tion, and also cause graphite to precipitate instead of cementite
at 690° C. (1275° F.) when the eutectoid decomposes (point
S, Fig. 246, page 314), so that it decomposes into ferrite and
graphite instead of ferrite and cementite. The maximum pre-

THE CONSTITUTION OF CAST IRON

343

cipitation of graphite seems to occur with about 2.5 to 3.5 per
cent, of silicon. With each increase of silicon up to that point
(the amount of sulphur, the rate of cooling and other influential
conditions remaining the same), we get an increase in the amount
of graphite precipitation, but when the amount of silicon exceeds
about 3 per cent, it seems to reverse its effect, and each addition
of silicon thereafter causes an increase in the proportion, not of
graphite, but of combined carbon. At this point large amounts
of various iron silicides (Fe2 Si, Fe2 Sia, etc.), make their appearance.
Then the color of a freshly broken fracture begins to be bright
like a mirror, in contradistinction to the white color of ordinary
white cast iron, which has more nearly the appearance of frosted
silver.

Sulphur. — The influence of sulphur upon the formation of
graphite is almost the exact opposite of the influence of silicon.
That is to say, each increase in the amount of sulphur present
increases the amount of combined carbon in the iron. It is
usually considered that each 0.01 per cent, of sulphur will neu-
tralize fifteen times as much silicon (i.e., 0.15 per cent.) in its
effect upon the condition of the carbon in the iron. It is also
very important to note that when the sulphur is in the form
of MnS, it is not so potent in increasing the combined carbon
as when it is in the form of FeS. An interesting example of
this is shown in the analysis of the two railroad car wheels
given below:

Total
Carbon

Graph-
ite

C.C.

Good wheel
Poor wheel

94.79
95.00

3.84
3.52

0.69
0.65

0.13
0.12

0.43
0.52

0.12
0.19

3.30
2.35

0.54
1.17

Good wheel required 150 blows of 25-lb. sledge to break it. Poor wheel
required 8 blows of 25-lb. sledge to break it.

It will be observed that the poor wheel has more than twice as
much combined carbon as the good wheel, although the sulphur
in the poor wheel is only about 50 per cent, more than the sulphur
in the good wheel, the other impurities being nearly the same.
When we come to figure out the amount of MnS and FeS in the
two wheels, we find, however, the explanation of the large amount
of combined carbon in the poor wheel. We also have an ex-

344

THE METALLURGY OF IKON AND STEEL

planation of the poor quality of this wheel in the increased amount
of FeS present.

CONSTITUENT

Good Wheel
Per Cent.

Bad Wheel
Per Cent.

MnS .

0.206

0.195

FeS

0.121

0.315

FeSi

2 045

1 923

Fe3P . ...

2 755

3 335

Pearlite .

67 610

84 492

Ferrite

23.963

0 000

Cementite

0.000

7 390

Graphite

3 300

2 350

Totals

100 000

Manganese. — Manganese increases the total carbon in pig
iron. Manganese also increases the proportion of the carbon that
is in the combined form, but its influence in this respect is far
less than that of the sulphur; moreover, the statement requires
the following qualification: as much manganese as is combined
with sulphur in the form of MnS does not increase the proportion
of carbon in the combined form. Indeed, it has really the reverse
effect, because it takes the sulphur out of the form of FeS, in
which it is most powerful in increasing the combined carbon. In
this sense therefore, the manganese actually decreases the amount
of combined carbon.

The excess manganese over that necessary to form MnS (that
is, the manganese in the form of [FeMn]3C) increases the propor-
tion of carbon in the combined form, and also increases the amount
of total carbon even more potently than does the manganese,
which is in the form of MnS. We therefore have a strange con-
tradiction, in that when the manganese is high an increase in
sulphur will, by decreasing the amount of (FeMn)3C, actually
decrease the tendency of manganese to raise the total carbon as
well as the combined carbon. To sum up, manganese and sulphur
both tend to increase the total carbon and the combined carbon,
and yet they neutralize each other in this respect.

Phosphorus. — The effect of phosphorus upon the carbon is
somewhat self-contradictory: from a chemical standpoint it. tends
to increase the proportion of combined carbon, and this is es-
pecially true when the silicon is low and the phosphorus high
(say above 1.25 per cent,). But phosphorus also has the effect of

THE CONSTITUTION OF CAST IRON 345

lengthening the period of solidification. That is to say, it makes
the iron pass through a somewhat mushy stage of solidification,
and this mushy stage lasts for several minutes. This lengthening
of the solidification period gives a longer time in which graphite
can precipitate. Therefore, when the silicon is relatively high
(at least over 1 per cent.), and there is consequently a strong
tendency for graphite to precipitate during solidification, this
precipitation is actually aided by the phosphorus, and the graphite
occurs not only more abundantly, but in larger-sized flakes. When,
however, the amount of phosphorus is very large, its chemical
effect is great enough to retain the carbon in the combined form,
in spite of the long period of solidification. We may sum this up
by saying that if the chemical conditions are such that graphite
is bound to precipitate, then the physical effect of the phosphorus
makes this precipitation the more easy; but if there is enough
phosphorus present to produce a strong chemical effect of its own,
or if the other chemical influence is not very powerful (i.e., if
the silicon is low), then phosphorus tends to keep the carbon in the
combined form.

THE PROPERTIES OF CAST IRON

Let us now consider the properties of cast iron, and sum-
marize under the head of each the influence of the various elements
and conditions upon them.

Shrinkage. — The shrinkage of cast iron is of more importance
than might at first appear, because the greater it is the greater
will be the strains set up in the cooling of the casting, and con-
sequently the liability to check; also, the greater will be the
allowance necessary in order that the casting may be true to the
size called for by the drawings. Graphite is the most important
impurity in this connection, because of the expansion which its
separation causes. This separation should take place at the
moment of solidification, but is usually not complete then, and
therefore the precipitation continues during the fall of the temper-
ature to several degrees delow the freezing-point. Furthermore,
when the silicon is high, graphite instead of cementite separates
at the lower critical point (i.e., the line P S K in Fig. 246, page 314).
As silicon and the rate of cooling are the chief influences which
control the separation of graphite, they become the governing

346

THE METALLURGY OF IRON AND STEEL

factors in the shrinkage of the iron. Indeed, when sulphur is
practically normal and no other unusual conditions prevail, there
is such a close relation between the size of the castings 1 and the
percentage of silicon on the one hand, and the amount of shrinkage
on the other hand, that any one of the three may be calculated
when the other two are known (see Table XXVII) .

TABLE XXVII.— RELATION OF SHRINKAGE TO SIZE AND
PERCENTAGE OF SILICON

PER

CENT.

^in.

1 in.

2 in. X

2 in.

3 in.

4 in.

SILICON

square

lin.

square

Perpendicular readings
show decrease due to

increase in silicon.

1.00

.183

.158

.146

.130

.113

.102

1.50

.171

.145

.133

.117

.098

.087

2.00
2.50

.159
.147

.133
.121

.121

.108

.104
.092

.085
.073

.074
.060

Horizontal readings show
decrease of shrinkage

3.00

.135

.108

.095

.077

.059

.045

due to size.

3.50

.123

.095

.082

.065

.046

.032

Sulphur is important in this connection, and its effect is con-
trary to that of silicon, because of its tendency to retain the carbon
in the combined form. Manganese and phosphorus each has a
less important influence. Manganese, by increasing the total car-
bon, tends to increase graphite and therefore decrease shrink-
age. So far as it neutralizes sulphur, moreover, its effect is in the
same direction. Phosphorus decreases shrinkage, both because it
contributes to the fluidity, of the metal and therefore gives a better
opportunity for carbon to separate, and also because of the ex-
pansion caused when the phosphorus eutectic separates from
solution. A hotter casting temperature of the iron has the effect
of delaying solidification by heating up the mold, so that graphite
has a little more chance to separate. This effect is noticed but
slightly amidst the other conditions.

A table showing the relation between the size of the casting,
the amount of silicon, and the shrinkage is given above and is
taken from page 155 of No. 93. Slight changes must be made in
this table by each foundry for the conditions of sulphur, phos-
phorus, temperature, etc., obtaining there; but those given
herewith will be found sufficiently accurate for all ordinary pur-
poses where conditions are anywhere near normal.

1 Which is the chief influential feature in the rate of cooling.

THE CONSTITUTION OF CAST IRON

347

Shrinkage Tests. — At many foundries it is the custom to
make a shrinkage test of the iron from each cupola at least once
a day. The simplest way of making these tests is to pour into a
mold 12 in. long, with a sectional area approximately proportionate
to the size of the castings made, some of the iron from about the
middle of the cupola run. The casting must be poured flat, and
the difference between 12 in. and the length of the cold bar is the
shrinkage of the metal. This method is somewhat crude and,
although it gives valuable results, has been greatly improved by
W. J. Keep 1 and Prof. T. Turner,2 who have devised simple and
inexpensive pieces of apparatus whereby the iron, after it begins
its solidification, draws a curve showing first the expansion and
later the contraction. It is by means of Professor Turner's ap-
paratus that the curves shown in Fig. 257, page 339, were made.
With very little care these curves can be obtained to show with
sufficient accuracy for all ordinary purposes the percentage of
graphite and also (other conditions being normal, or nearly so),
the percentages of silicon and combined carbon, and the strength,
hardness, and porosity. Indeed the curves are more useful than
many single tests, because they show at a glance the net effect
of several varying conditions.

Density. — The maximum density of cast iron occurs with
about 1 per cent, of silicon. With less than that, the iron is liable
to contain spongy spots, due to high shrinkage on account of low
graphite. With more silicon the separation of graphite decreases
density. Above 2 per cent, of silicon, the grain of the iron becomes
so open as to be actually porous and the density falls off by 12
per cent.

TABLE OF DENSITIES

Specific
Gravity

Weight per
Cubic Foot

Pure iron

7.86

490

White east iron

7.60

474

Mottled cast iron

7 35

458

Light pray cast iron

7.20

450

Dark gray cast iron

6.80

425

Sample of gray cast iron when cold ... .

7 17

448

Same, when liquid

6.65

416

See Chapter XX of No. 93, page 291.

Reference on page 338.

348 THE METALLURGY OF IRON AND STEEL

To make a close-grained iron for hydraulic work the sulphur
should be from 0.03 to 0.055 per cent. If more than this, the iron
is liable to be dirty, to contain spongy spots on account of low
graphite, to be difficult to machine on account of high combined
carbon, and to be weak, because high sulphur, aside from its effect
on carbon, reduces the strength. Especially, if the phosphorus is
high must the sulphur be kept down to these limits, or the iron
will be hard, brittle, and weak.

There should be from 0.4 to 0.6 per cent, of manganese. We
do not want more manganese than this, or the casting will be
difficult to machine. We do not want less than I have indicated,
because manganese assists in counteracting the bad effects of
sulphur and phosphorus.

Phosphorus has a double-acting influence on the porosity of
cast iron: (1) It increases the size of the crystals, decreases shrink-
age and causes a large expansion after solidification, as explained
in connection with Fig. 257; but (2) it fills all the crevices between
the crystals and in the interior of the iron, which, by decreasing
the porosity, counteracts its first influence. When the phosphorus
is high, the phosphorus and iron form a eutectic, which remains
fluid for a long time and fills the tiniest crevices in the interior of
the metal. For this reason iron for hydraulic work may run
up to 0.7 per cent, phosphorus, but above that the iron is liable
to be weak and ' cold-short/ especially under impact. In fact,
where very strong iron is desired, the phosphorus should be kept
down to 0.4 per cent, at least.

With the various amounts of impurities mentioned above
the combined carbon will be in the neighborhood of 1 per cent,
and the graphite about 2.5 per cent., the exact amounts depending
upon the thickness of the castings and the rate at which they are
cooled. If we desire to keep the combined carbon the same and
reduce the graphite it will be necessary to reduce the total carbon.
This can be accomplished by mixing in steel scrap and melting
fast in the cupola, or by melting in an air-furnace instead of a
cupola. This reduction in graphite results in a closing of the
grain of the steel, with consequent increase in strength and
density.

Segregation. — A common cause of porosity in castings is
segregation, or the collection together of impurities in spots.
This segregation is the greater the greater the amounts of phos-

THE CONSTITUTION OF CAST IRON 349

phorus, sulphur, manganese, and silicon. Phosphorus increases
the segregation by making a fluid eutectic, which does not so-
lidify until after the remainder of the casting, but then runs
into that part of the metal having the loosest texture. This
part is usually in the middle of the larger sections of the cast-
ing, and when the silicon is high and there are shrinkage spots the
segregation will be excessive in the neighborhood of these spots.
Manganese and sulphur are also liable to collect in the same way
and place. These localities, where the segregation is high, and
which are known, when very bad, as 'hot spots/ are sometimes
porous or surrounded by porous parts of the casting. They are
sometimes so extremely hard that no tool will cut them. One
way of getting rid of them is to use very large risers, or headers,
which solidify last and serve as feeders for the remainder of the
metal. Under these circumstances the segregation occurs in the
riser, and is thus temporarily removed. This method is not ad-
visable as a regular practice, however, because these risers ulti-
mately find their way back into the cupola as scrap and result in
increasing the impurities in a subsequent set of castings.

Headers themselves increase the density of iron castings by
feeding the metal and so preventing the porous spots, and also by
keeping the metal under a pressure during solidification. This
latter is especially serviceable when the phosphorus is high, which
tends to make the metal expand during solidification, as I have
shown.

Checking. — The time when a casting usually checks is when
it is just above the black heat, when the metal is in a weak and
tender condition and, as shown by Fig. 257, page 339, is under
strain because it is contracting upon the sand. Sulphur greatly
increases weakness at this temperature, because both sulphide
of manganese and sulphide of iron are now in a pasty condition,
and therefore offer very little resistance to breaking. The sulphide
of iron is much worse, however, because this is spread out in thin
plates or membranes which offer much more extended planes of
weakness than the sulphide of manganese, which is in small spots
or bubbles, resembling blow-holes in its effect. Phosphorus, by
decreasing shrinkage, decreases the liability of checking, but phos-
phorus ha-s another influence, shown in the production of large-
sized crystals, and in this respect it increases the liability of the
metal to check.

350 THE METALLURGY OF IRON AND STEEL

Manganese, by decreasing the size of crystals, tends to counter-
act partially the effect of the phosphorus. The size of crystals
can also be decreased to some extent by chilling the weak
points and feeding them well under a head of metal. Feeding all
localities liable to check has the double advantage of lessening
shrinkage and segregation, both of which increase the liability to
checking.

Softness, Workability, and Strength. — It is the combined car-
bon which is the great hardener of cast iron, the other elements
producing hardness chiefly in proportion as they produce com-
bined carbon, except manganese, which not only produces com-
bined carbon, but also produces a compound having the formula
(FeMn)3C, which is very hard and difficult to machine.

Silicon, by decreasing combined carbon, decreases hardness.
When we get above 3 per cent, silicon, however, there begin to
form new compounds with silicon which make the iron hard.
Furthermore, silicon above 3 per cent, increases combined carbon,
instead of decreasing it.

The maximum softness of cast iron is obtained with about
2.5 to 3 per cent, of silicon, the sulphur being not above 0.1 per
cent, and the manganese not above 0.4 per cent. In large or
slowly cooled castings the silicon should be near the lower limit,
and in small or rapidly cooled castings near the upper limit, in
order that the combined carbon may be down below 0.15 per
cent, and the graphite more than 3 per cent. Such a cast iron
would correspond to a soft steel, mechanically mixed with crystals
of graphite. This soft steel would machine with great ease, and
the graphite would act as a lubricant for the cutting tool. The
mixture will have a transverse strength of about 2000 to 2200
lb., will be low in density and open in grain. To increase the
strength without increasing the hardness, the best way is to cut
the sulphur and phosphorus down to a low point, if possible,
because sulphur, and next to it phosphorus, are the impurities
which weaken iron most (aside from their influence on carbon).
Another way is to decrease the total carbon, and hence the graphite,
because graphite crystals, especially if large, are great weakeners
of cast iron.

The strength of steel is more than double the strength of
cast iron, the difference being due almost altogether to the graphite
in cast iron, because silicon in itself (aside from its influence on

THE CONSTITUTION OF CAST IRON 351

the carbon) is a strengthener of both iron and steel up to at least
4 per cent.

To a slight extent the total carbon may be reduced by melting
steel scrap with the iron, or by decreasing the amount of man-
ganese, provided that the manganese left be always at least twice
the sulphur, otherwise the iron will be weak and brittle.

The strength of cast iron may be increased by increasing
the combined carbon, but this is done at the expense of softness
and workability. Cast iron containing from 1.5 to 2 per cent,
of silicon (depending upon the size of the castings and rate of
cooling), 0.9 per cent, of combined carbon, 0.5 per cent, of
manganese and not more than 0.08 per cent, of sulphur and 0.3
per cent, of phosphorus, will work without difficulty in the machine
shop and have a tensile strength of over 28,000 Ib. per square
inch. In many cases foundries are unwilling to go to the expense
of such a low sulphur and phosphorus. In this case the strength
must be obtained by raising the manganese, which is not advisable,
as it decreases the softness more than any other element, causes
dull iron and high total carbon.

An important point in connection with the strength of cast
iron is the size of the crystals of graphite — the smaller these
crystals are the greater the strength, because the smaller are the
planes of easy rupture. A notable example of this is malleable
cast iron, which may have a tensile strength of 45,000 Ib. per
square inch, even when the percentage of graphitic or temper
carbon is as high as 3 per cent. The very small size of the flakes
of the temper carbon does not reduce the strength as much as the
same amount of the larger graphite crystals.

It is believed by many that smaller graphite crystals are
obtained by mixing different brands of iron in the cupola, even
though the analysis of the mixture may be the same. This,
however, is denied by others, and no reliable data exist upon
which we can base a definite statement. It is also believed by
many that when the silicon is added to the cast iron immediately
before pouring into the molds, the crystals of graphite are smaller
than those formed when high-silicon irons are melted in the cupola.
The practice of adding a small amount of ferrosilicon to the ladle
of cast iron after it is received from the cupola is thus said to be
doubly advantageous, because the silicon does not have to go
through the cupola, where it suffers some oxidation, and it pro-

352

THE METALLURGY OF IRON AND STEEL

duces the desired softness by precipitating the graphite, but in a
form which does not decrease strength so much.

To obtain high transverse strength, the silicon should be
about 0.2 per cent, lower and the combined carbon about 0.2
per cent, higher than the figures given for tensile strength. Other-
wise the effects are very similar.

Some estimates of the strength and workability of cast iron
are given in the following table,1 and on page 57 of No. 94.

TABLE OF CAST-IRON STRENGTH AND WORKABILITY

Silicon
Per
Cent.

Sulphur
Per
Cent.

Phos-
phorus
Per
Cent.

Man-
ganese
Per
Cent.

Tensile
Strength
Ib

Trans-
verse
Strength
Ib

Soft iron for pulleys, small j
castings, good tooling j

2.20
to
2.80

Not
over
0.085

Not
over
0.70

0.30

to
0.70

28,000

2200

Medium iron for engine j
cylinders, gears, etc . . j

1.40
to
2.00

Not
over
0.085

Not
over
0.70

0.30
to
0.70

30,000

2500

Hard-iron cylinders for (

1.20

ammonia, air-corn- -j

to'

pressors, etc (

•1.60
1.60J

Not
over

0.70
to

Not

over

25,000

2800

tQ2

0.095

0.40

0.60

1.90

If annealed.

2 If cooled fast.

Chill. — In the making of cast-iron rolls, railroad car wheels,
anvils, etc., at least one surface of the casting is desired to have
great hardness, to resist wear, and to be backed by metal which
shall be stronger and not so brittle. This is accomplished by
chilling the surface that is wanted in a hard condition, and so
producing white cast iron to varying depths, regulated at the
will of the foundrymen. The making of this kind of casting is
one of the most difficult problems of cast-iron metallurgy. The
metal must be very close to the given composition, and the
temperature of the mold, of the chill, and of the metal when
cast must be regulated with care. Therefore, air-furnaces are
often employed for melting in this class of work, or else uniform

1 Abstracted from page 197 of No. 120, page 355.

•§

FIG. 258. —WHITE PIG IRON.

0.75 per cent. Si. 0.120 per cent. S. 50

diameters. Unetched.

FIG. 259. — GRAY PIG IRON.

0.75 per cent. Si. 0.012 per cent. S.

50 diameters. Unetched.

FIG. 260. —GRAY PIG IRON.

1.75 per cent. Si. 0.025 per cent. S.

50 diameters. Unetched.

FIG. 261. — GRAY PIG IRON.

2.5 per cent. Si. 0.012 per cent. S. 50

diameters. Unetched.

FIG. 262. — GRAY PIG IRON.

3.5 per cent. Si. 0.025 per cent. S. 50

diameters. Unetched.

FIG. 263. — NO. 2 CHARCOAL PIG IRON

VERY SLOWLY COOLED.

50 diameters. HNO3.

354

THE METALLURGY OF IRON AND STEEL

conditions of cupola melting are maintained with great care, and
very little, if any, scrap, which must necessarily be of somewhat
uncertain analysis, is used, except the return scrap from the
foundry itself, that is, defective castings, sprues, gates, shot-iron
spillings, etc., and also scrap castings of like nature, such as
worn-out car wheels and broken rolls. The most important

FI6. 264. — METHOD OF MEASURING THE DEPTH OF CLEAR CHILL IN A
CAST-IRON ROLL.

factors in regulating the depth of the chill are the silicon and the
sulphur, and in the following table is given the depth of clear
chill from the surface for several different percentages of silicon
and sulphur. The figures here given must only be taken as ap-
proximations, as they will vary to an important extent with
different conditions in each foundry; but, starting with this as a
basis, one can quickly prepare a table for himself to suit the
practice in his foundry. Phosphorus has very little effect on the
depth of chill, and manganese is also relatively less effective,
although it increases the hardness of the chilled portion. The
hotter the iron when cast the deeper the chill.

THE CONSTITUTION OF CAST IRON

355

DEPTHS OF CLEAR CHILL FROM SURFACE IN INCHES

Silicon
Per Cent.

Sulphur
0.2
Per Cent.

Sulphur
0.15
Per Cent.

Sulphur
0.1
Per Cent.

Sulphur
0.075
Per Cent.

Sulphur
0.05
Per Cent.

1 25

0 625

0 250

0 125

0 000

1.00
0 75

1.000
1 500

0.625
1 000

0.250
0 625

0.125
0 250

0.000
0 125

0.50

1 500

1 000

0 625

0 250

0.40

1 250

1 000

0 625

0.30

1.500

1.000

REFERENCES ON THE CONSTITUTION OF CAST IRON

See Nos. 1, 32, 90, 91, 93, and

120. W. G. Scott. " Effect of Variations in the Constituents of

Cast Iron." Proceedings of the American Society for
Testing Materials, vol. ii, 1902, pages 181-206.

121. Transactions of the American Foundrymen's Association.

122. G. B. Upton. "The Iron-Carbon Equilibrium." The

Journal of Physical Chemistry, October, 1908, vol. xii,
pages 507-549. This is a very valuable study and resume
of the recent researches on the constitution of all the alloys
of iron and carbon, and gives the chief known facts of
scientific interest in a concise form.

XIII
MALLEABLE CAST IRON

MALLEABLE cast iron is iron which, when first made, is cast in
the condition of cast iron, and is made malleable by subsequent
treatment without fusion. Gray cast iron is weak and brittle, on
account of the flakes of graphite which destroy its continuity and
form planes of easy yielding. It will readily be understood that if
the amount of graphite were less, or if its flakes occurred in a very
finely pulverized form, or both, the material would be stronger and
would endure a slight degree of deformation without cracking;
that is, it would be malleable. These changes are indeed brought
about in the manufacture of malleable cast iron. The process,
which was invented by Reaumur in 1722, but has only been in
practical use about 100 years and of importance less than 50, is a
very ingenious operation. Malleable castings have two of the
greatest advantages of cast iron; namely, fluidity and a low melt-
ing-point, combined with about three-quarters the strength and
one-sixth the ductility of steel. They are very popular for rail-
road rolling-stock construction, especially for draw-bars, couplers
and knuckles, on account of their high resiliency, resistance to
shocks, and ability to be made into thin, light castings. Probably
about one-half of the malleable cast-iron production of the United
States goes into railroad work, although it is to be observed that at
present the use of steel castings for this purpose is increasing rela-
tively faster. The next most important use is for pipe-fittings,
where malleable cast iron is equally advantageous; also for small
machinery castings needing to be strong and light, household and
building hardware, etc.

Process. — Pig iron of the proper kind is first melted and cast
into molds of the desired size and shape. In these operations two
precautions are observed: first, the proportion of silicon is low,
and second, the castings are not allowed to cool too slowly. We
have already learned that the precipitation of graphite is a slow

356

MALLEABLE CAST IRON

357

action and does not occur unless ample time is allowed or there is
sufficient silicon to produce a strong chemical action. In the in-
tentional absence of these factors malleable castings, as first made,
are practically free from graphite and consist entirely of white cast
iron — hard, brittle and weak.

After cooling the castings are cleaned and packed in some pul-
verized material, as iron ore, mill scale, lime, sand, placed in an
annealing furnace and heated to a temperature of 675° to 725° C.
(1250° to 1350° F.), which is, roughly, 450° C. below their melting-
point, and at which temperature they are kept for many hours.
While under this heat there oc-
curs the precipitation of graph-
ite, which normally should have
occurred during solidification, or
shortly thereafter, and in the
majority of cases almost all the
combined carbon throughout the
body of the casting is changed
to graphite. But the graphite
does not here form in flakes, as
in ordinary gray cast iron, but in
a finely comminuted condition,
like a powder, to which the
name of ' temper carbon ' or
'temper graphite' is given (see
Fig. 265) . In this form it is not

nearly so weakening or embrittling to the casting as flakes of
graphite would be.1

Properties of Malleable Cast Iron. — Malleable cast iron con-
sists almost entirely of ferrite and temper carbon. It has a tensile
strength of 40,000 to 60,000 pounds per square inch, which is about
double that of gray cast iron, with an elongation of 2£ to 5J per
cent, in 2 in. and a reduction of area of 2 J to 8 per cent.2 A one-
inch square bar on supports 12 inches apart should bear a load at

1 We may liken this to two samples of putty, in one of which had been
embedded a large number of plates of mica, and in the other the same amount
of mica ground to powder.

2 In the case of iron very carefully melted and annealed in iron oxide, the
elongation may go as high as 8 per cent, and the reduction of area as high as
12 per cent.

FIG. 265. — ANNEALED MALLE-
ABLE CAST IRON.

Magnified 50 diameters. Unetched.

358

THE METALLURGY OF IRON AND STEEL

the center of at least 3500 pounds, and be deflected at least half an
inch before breaking. Thin sections should be capable of flatten-
ing out under a hammer and bending double without cracking.

Total Carbon in Malleable Cast Iron. — In melting the iron we
take pains to produce a low total carbon, and if we anneal in iron
ore or mill scale the carbon is still further reduced by a curious
reaction which takes place between it and the iron oxide —
3 C + Fe3 O3 = 3 CO + 2 Fe,

whereby it forms carbon monoxide and is eliminated. Sometimes,
when the sections of metal are thin, we may eliminate almost all
the carbon to the very center of the casting, which makes a more
ductile material. It must be observed, however, that this reduc-
tion of carbon is not an essential feature of annealing and that the
real function of this operation is to change the combined carbon to
temper carbon, for the malleable cast iron owes its superiority over

FIG. 266. — AIR-FURNACE.

gray cast iron chiefly to the finely pulverized form of its temper
carbon.

Melting in the Air-Furnace. — The commonest melting-furnace
for malleable cast iron is the air-furnace, because thereby we get a
better control of the metal than in the cupola and the ability to
produce castings with lower total carbon, lower sulphur, and any
desired amount of silicon. After the metal is melted it is retained
in this furnace for 15 minutes to an hour longer, and test samples
are taken at intervals, from the fracture of which and the tem-
perature of the iron we determine the correct moment for tapping.
The fracture of the test ingot sample should be a clear white
throughout, except when the castings are to be of very light sec-
tion, in which case the metal might be tapped when the test sample

MALLEABLE CAST IRON

359

shows a few specks of graphite in the center. The practice of
judging from test samples is different in each foundry, but there
must be some system which insures that the metal shall be of such

Modified

Pittsbucg Type Air Furnace
Cap. 1 5 Tons

a composition when tapped that the castings will have not more
than a trace of graphite if any at all (say, less than 0.15 per cent,
in small castings and a little more in larger ones). It is to be
remembered that any graphite is a detriment. The longer in the

360

THE METALLURGY OF IRON AND STEEL

furnace after melting the more silicon will be burned out, and
therefore the less the liability to graphite. Consequently, the
hotter we want to get the iron the higher must be the percentage of
silicon in it to start with. Also, the longer the metal is in the fur-
nace after melting the more carbon will be burned out of it. (See
also pp. 284-286.) During the operation the bath is rabbled at
intervals, and is skimmed so as to expose the metal directly to
oxidation by the gases.

Tapping the Air-Furnace. — When the time comes to tap the
furnace we may either allow a small stream to flow, which is caught

FIG. 268.

in ladles and immediately poured into the castings, whereby it
takes from 20 minutes to an hour to empty a furnace of 10 to 30
tons capacity, or else we may allow the metal to run out in a big
stream into a large ladle, from which it is repoured into smaller
ones for casting. The first method gives a less uniform product,
because the last part of the bath having been exposed longer to
the oxidizing influence of the furnace, is lower in silicon and total
carbon than the first part. This is not altogether a disadvantage,
however, because the first metal, having come from the top of the

MALLEABLE CAST IRON

361

bath nearer the flame, is hotter, and therefore very suitable to pour
into smaller castings; and these smaller castings, because they cool
more rapidly, can well contain more silicon without danger of
graphite precipitating. Moreover, the hotter the metal the greater
the tendency for it to be white. A disadvantage of pouring the
metal first into a big ladle is that it must be hotter when it comes
from the furnace, and moreover the first ladle must be preheated.
A middle course is possible : we may pour the top of the bath into
small castings by means of ladles receiving their metal direct from
the furnace tap-hole, and then enlarge the tap-hole and take all the

FIG. 269.

rest of the bath into a big ladle, whence it can be poured into other
ladles and go to the larger-sized castings.

Cupola Melting. — In cupola melting we get metal having
practically the same composition at all times of the heat,1 and also
about the same temperature. It is also cheaper in fuel, and
especially so when it is desired to get very fluid iron, because ob-
taining hot iron in the air-furnace requires a continuation of the
heating after the iron is melted, and this entails not only the use of
more fuel, but also a burning out of silicon and carbon, both of
which elements increase the fluidity of the metal.2 Cupola metal
is higher in total carbon and in sulphur, both of which decrease

1 Except for the slightly higher sulphur at the beginning and end.

2 When I say 'hot iron' here, I mean 'fluid iron/ i. e., the degree of heat
above the melting-point ; and in this sense hotness includes both the tempera-
ture and the state of impurity. An iron with 2 per cent, total carbon and
0.60 per cent, silicon at 1300° C. (2375° F.) will not be nearly as fluid or as
far above its melting-point as one with 3 per cent, total carbon and 0.75 per
cent, silicon at the same temperature.

362

THE METALLURGY OF IRON AND STEEL

strength and ductility. Because of the higher total carbon it is
more difficult to prevent graphite separating; therefore cupola
metal is used for light castings, which cool more quickly and do
not usually require so much strength.

Regenerative open-hearth furnaces are also used in melting
iron for malleable castings. (See also pages 285-286.)

Annealing-Boxes. — After the castings are cooled they are
carefully cleaned from all adhering sand by tumbling them around
in a tumbling-barrel (in which they are mixed with star-shaped

FIG. 270. — TUMBLING-BARREL.

pieces of metal something like children's jackstones), or by sand-
blast, or by some other suitable method. They are then packed in
the cast-iron ' saggers ' or annealing pots, or boxes, together with
the packing. Sometimes, though rarely, the tops of the saggers
are closed by means of an iron cover, sometimes by a thick layer
of the packing in the upper part, and sometimes by clay or wheel-
swarf. These pots last only from 4 to 20 heats before they are
largely oxidized away.

MALLEABLE CAST IRON

363

Annealing-Ovens. — The boxes are then placed in the anneal-
ing-ovens in such a way that the flame may play around them as
completely as possible. The general form of ovens is shown in

— v

»4

IJi

-H

•>

I o

Figs. 271 and 273. The flame usually comes in at the top and goes
out at the bottom along the side, and thence through flues under-
neath the oven. The fuel used may be coke, coal, oil or gas, the

364 THE METALLURGY OF IRON AND STEEL

latter being preferable on account of the better control of the tem-
perature, which should be increased at a very gradual and uniform
rate during the heating up, and kept as constant as possible during
the annealing period.

Annealing Practice. — It takes about six days for the1 anneal-
ing operation, including heating up and cooling down. Sometimes
this can be shortened a little by decreasing the time at the full
annealing heat, and by cooling rapidly or drawing the saggers out
of the oven and dumping them while the contents are still at a
dull-red heat. This practice is not conducive to a good quality of
castings and should never be permitted in important cases. The
time at the full heat should never be less than 60 hours, and pref-
erably it should be more than that. If less, the temperature of
annealing must be higher, and this decreases the strength and
ductility of the castings. Annealing should not occupy too long a
time, however, unless the temperature is quite low, ^because the
temper carbon tends to draw together to larger flakes; besides
which the metal may become oxidized between the grains, or
"burnt.' Air-furnace castings should be annealed at 675° to 760° C.
(1250° to 1400° F.) and cupola metal at 850° to 950° C. (1560° to
1750° F.).

Packing. — As originally planned, the castings were annealed
in a packing of iron oxide crushed to a size less than a quarter of
an inch in diameter. The packing must surround the castings at
every place, both inside and out, and no two castings must touch.
Iron ore, mill scale, 'bull-dog/ and similar forms of iron
oxide are used for this purpose. Usually two or three parts of old
packing are used with one part of new packing, because all new
packing is too energetic in its chemical action on the carbon, and
all old packing will not be energetic enough in decarburizing the
surface of the castings.

Annealing in iron oxide produces a white skin where the casting
has been deprived of its carbon, and a black interior, due to the
temper carbon; whence the name of ' black heart malleable' for
this material. Tests have shown that the casting with this white
skin upon it is much stronger than a similar one which has not been
decarburized on the surface, and therefore the packing in iron
oxide is advantageous, even though not an essential feature of the
operation. When the castings are packed in some nonoxidizing
material, such as sand, clay or lime, they may receive as perfect

MALLEABLE CAST IRON 365

an annealing, as far as the production of temper carbon is con-
cerned, but will be without the white skin and of lower strength.

Composition of Iron Used. — The pig iron employed in this
process is sold under the name of ' malleable coke iron' or ' malle-
able Bessemer/ As it is the composition of the metal poured
into the molds which determines the success of the annealing and
the quality of the product, we shall consider first the kind of
metal needed there, and from that calculate the composition
of the mixture necessary to charge into the air-furnace or cupola.
In the air-furnace we do not judge the iron by its chemical analysis,
but by the appearance of the test-ingot fracture when we are
ready to pour, but as the second quality depends upon the first,
it is the same thing in the end.

Silicon. — The proportion of silicon will depend upon the size
of the casting and the amount of total carbon, because the greater
each of these is the less will be the amount of silicon that will cause
a precipitation. It might at first appear that the less silicon the
better; but this is not altogether so, because temper carbon will
not come out during annealing unless a certain amount of silicon is
present ; and the more there is, the more quickly, easily, and com-
pletely will the precipitation occur. For castings one inch thick
the silicon may be as low as 0.35 per cent.; but this is unusual, as
three-quarters of an inch thickness is rarely exceeded. For half-
inch castings the silicon will be about 0.60 per cent., and for very
thin and light castings with low total carbon and high sulphur
(say 0.2 to 0.3 per cent.), the silicon may be up to 1 per cent. To
the percentage of silicon desired in the castings we must add the
amount which will he burned out in melting. In the cupola this
will be about 0.2 to 0.25 per cent., and in the air-furnace from
0.15 to 0.5 per cent., or more if desired, depending on the length of
time the metal is kept in the furnace after melting. The hotter
we want the iron or the more total carbon we desire to burn out,
the longer this time must be, and therefore the higher the silicon
in the original mixture charged.

Sulphur. — Sulphur increases the tendency of castings to
check, which is especially important in malleable work on account
of the shrinkage of white iron being nearly double that of gray
iron. Sulphur also reduces the strength and the ease of annealing.
For this reason over 0.06 per cent, should not be permitted in cast-
ings requiring strength, but it actually runs up to 0.2 and 0.3 per

366 THE METALLURGY OF IRON AND STEEL

cent, in inferior metal, both in America and England, and espe-
cially in small castings, which do not need strength so much, and
which, having less length for shrinkage, are not so liable to be
checked by cooling strains.

Manganese. — Low manganese is preferred by many foundries,
and one of the highest authorities in America 1 places the limit at
0.8 per cent. It should be remembered, however, that the man-
ganese should be at least twice the sulphur, and preferably three
times, though not when the sulphur is as high as 0.3 per cent.
Manganese of 0.5 per cent, tends to decrease checking. It also
protects silicon from oxidation, both during melting and anneal-
ing, and on this account hastens and makes more complete the
precipitation of temper carbon. It also protects the iron itself
from oxidation during annealing and thus prevents the formation
of 'scaled7 castings. More than 0.6 per cent, manganese makes
the iron hard and difficult to machine, which is disadvantageous,
especially for pipe-fittings, which must be threaded with great
economy in order to meet the trade competition.

Phosphorus. — Phosphorus makes the metal fluid, which is
especially desirable where total carbon and silicon are low, or
where sulphur and manganese are high. On the other hand, it
diminishes two of the most valuable properties of the material:
its resiliency and resistance to shocks. It also makes the metal
hard, difficult to machine and liable to check, and amounts over
0.225 per cent, should never be permitted by engineers where the
castings are subjected to strain.

Total Carbon. — Total carbon below 2.75 per cent, gives
trouble in annealing and therefore makes the castings weak. It
also makes the metal more sluggish. It is difficult to get as low
as this in cupola melting, although mixing in large percentages of
steel scrap and allowing the metal to run out of the cupola as fast
as melted will reduce the proportion appreciably. In air-furnace
practice the total carbon may be reduced as far as necessary.
Annealing in iron oxides also removes carbon from the outer layers
and even to the very center of thin castings. The lower the total
carbon in the annealed castings, the better.

Scrap Used. — Not more than about 20 per cent, of bought
scrap is used on the average in American practice, and a good deal
of this is steel, on account of the desirability of lower total carbon,
1 See No. 130, page 369.

MALLEABLE CAST IRON

367

and because iron scrap is too impure, too variable and too uncer-
tain in sampling and chemical analysis for castings requiring
strength, such as those for railroads and machinery. There is,
however, always a large amount of ' return scrap ' from the foun-
dry, consisting of defective castings, sprues, gates, etc., which, in
the case of small castings, may be greater in weight than the cast-
ings themselves. This return scrap is low in total carbon and
silicon as a result of having already suffered the melting changes.

Shrinkage. — The shrinkage of malleable iron from casting is
almost as great as that of steel, because almost no graphite forms.
The amount of silicon and the sectional area of the castings are
still the determining factors in this connection. Indeed, by means
of the measurement of the section and the percentage of silicon we
may estimate the shrinkage, or by means of the section and the
shrinkage we may estimate the silicon very closely, other condi-
tions and impurities being normal. The following table gives the
necessary data for these estimations :

SHRINKAGE IN INCHES PER FOOT OF LENGTH

PERCENTAGE
OP SILICON

1 Inch
Square

i Inch
Square

f Inch
Square

1 Inch
Square

0 35

0 225

0 200

0.190

0 175

0 50

0 220

0 195

0 183

0 170

0 75

0 215

0 190

0.176

0.162

1 00

0 211

0 183

0.137

0.102

Expansion due to Temper Carbon. — It is a very interesting fact
that when the malleable cast iron is annealed and the temper
carbon precipitates, the casting expands to an amount approxi-
mately equal to that which would have occurred if the graphite had
separated during solidification and gray cast iron had been pro-
duced in the first instance. In other words, the temper carbon,
although in a very finely powdered condition, occupies about the
same amount of space as an equal weight of graphite, and causes
about the same ultimate difference in size between the original
pattern and the annealed casting as when gray cast iron is made.
An interesting example of this expansion in annealing is shown in
Fig. 274, which is a swivel snap for hitching straps. Casting No. 1
is first poured, cooled and cleaned. It is then embedded in the

368

THE METALLURGY OF IRON AND STEEL

sand of a mold and casting No. 2 is poured around the shank of it,
as shown in No. 3 of Fig. 276. Casting No. 2 shrinks upon the
shank of No. 1 so as to make a close fit, and no swiveling is possi-
ble, but the combined casting is now sent to the annealing-ovens
and annealed. This causes the expansion referred to, and as cast-
ing No. 2 is larger in diameter, it expands the more, and now turns
very easily around the shank of No. 1.

Miscellaneous Iron Products. — In the form of small castings
malleable cast iron and similar products often masquerade under
the name of steel, because under that name the producer finds a

readier market for them.
On account of their
fluidity they may be
cast very cheaply in
small sizes, and there-
fore the temptation to
use them as a material
for 'cast-steel hammers/1
'hard-steel' bevel gears,
'semi-steel castings/ and
even automobile 'steel'
drop-forgings, is a strong

one. Engineers are warned to be on their guard against a de-
ception of this kind, for legal redress has been sought many
times in vain. A clever lawyer may easily confuse and outwit
a judge or jury with the involved definitions and technical de-
scriptions necessary to make the distinction clear. It is usual
for the manufacturer when putting material of this kind upon
the market to qualify the name 'steel' with some other letters
or name, such as ' P. Q. steel/ ' Smith steel/ etc. ; but they all differ
from true steel in that they were not " cast into an initially malle-
able mass." Some are made by melting a large proportion of
steel with cast iron, after which the cooled metal may or may not
be annealed in iron oxide. Others are made by a long or thorough
annealing of ordinary malleable castings in iron oxide, by means
of which the metal is decarburized to some depth, and is then car-
burized again by a cementation process. This makes a very good
material for some purposes, such as small bevel gears not requiring

1 The trade would ordinarily understand by this name hammers made of
crucible steel, so the use of this name is really a fraud.

FIGS. 274 TO 276.

MALLEABLE CAST IRON 369

strength or much ductility, but it ought not to be called ' steel/ If
the purpose for which it is to be used does not require any other
properties than malleable cast iron possesses, then it should be used
under its true name; but if it is to be used under circumstances
where it is liable to strain, calling it 'steel' will not enable it to
stand up under the work any better. The confusion is the more
easy because genuine steel is made by the cementation of wrought
iron, and wrought iron goes in England under the name of malle-
able iron. In America we seldom call wrought iron ' malleable
iron/ but we often abbreviate malleable cast iron to ' malleable
iron/ or even to ' malleable/

REFERENCES ON MALLEABLE CAST IRON

See 90, 91, 92, and 122.

130. Richard Moldenke. "Malleable Castings," Parts I, II, and
III. Instruction Papers Nos. 547A, 547B, and 547C,
of the International Correspondence Schools, Scranton, Pa.
This is the most thorough and valuable treatment of this
subject that I know of.

XIV
THE HEAT TREATMENT OF STEEL

WE have already discussed the heat treatment of cast iron
under other heads, namely: (1) the rapid cooling from the molten
state, or 'chilling/ and (2) the annealing of malleable cast iron.
Heat treatment is of much greater importance in connection with
steel, because nearly 99 per cent, of all the steel made is heated
either for the purpose of bringing it into the mobile condition in
which it can be readily wrought, or for annealing. Indeed, the
great majority of steel is heated several times, and some steel is
subjected to two or three different kinds of heat treatment.

IMPROPER HEATING OF STEEL

Overheating. — If steel be heated to a high temperature, say
1100° C. (2010° F.), and then cooled (either slowly or rapidly)
without being subjected to strain, it will be 'coarse-grained' as
it is called, that is, its crystals will be relatively large in size.
This can be readily seen by breaking it and examining the frac-
ture, which will be bright and sparkling if the crystals are coarse,
or dull-looking and fine-grained if they are small (see Fig. 284) .
The bright fracture is technically called 'crystalline' or 'fiery/
while the fine-grained one is called 'silky7 or 'sappy.' The
size of the crystals may also be learned with great accuracy by
means of the microscope (see Figs. 277 to 282 and 286) . Now, if the
steel which was coarse-grained after heating to 1100° be heated
instead to 1200°, the crystals will be still larger in size; if heated
to 1300° they will be larger still, and so on. The size of the crys-
tals will depend first upon which of these high temperatures it
was heated to, and second upon the amount of carbon it contains.
Low-carbon steel is normally larger in crystal-size than high-
carbon steel.

Even the best quality of steel, if rendered coarse-grained by

370

THE HEAT TREATMENT OF STEEL 371

' overheating/ will suffer in its valuable properties, and may be-
come quite unfit for use. Medium- and high-carbon steel will
lose both strength and ductility; low-carbon steel will lose strength
even up to 50 per cent, of the original, but does not seem to be
materially damaged in ductility unless the overheating is con-
tinued for a long time or at a very high temperature.

Cure for Overheating. — Let our first example be steel con-
taining 0.9 per cent, carbon, that is, steel consisting entirely of
pearlite. If this be heated from some point below the line P SK
in Fig. 246, page 314, to some point above that line, a new crys-
tallization will begin, and all traces of previous crystallization
will disappear. It seems as if dissolving the ferrite and cementite
in each other produces forces which obliterate almost all existing
crystalline forms. So, if this particular steel has been made
coarse-grained by overheating, we may make that grain fine
again by reheating the steel from below the line P SK to just
above it. This process is known as ' restoring/ or, by some writ-
ers, ' refining ' the steel. It is an operation which should be thor-
oughly understood by every metallurgist and engineer. When we
reheat the steel we must be careful not to go to a high tempera-
ture again, for a new crystal-size is born at the line P SK, and
the crystals grow with every increase in temperature. The re-
searches of Professors Howe and Sauveur 1 indicate that the size
of the crystals is almost directly proportional to the temperature
reached above the line PSK. If, therefore, we barely cross the
line, we will obtain the smallest grain-size that the steel is capable
of (see Fig. 281).

The cure for coarse crystallization in steel with less than 0.9
per cent, carbon is to reheat it from below the line PS K to above
the line GO S, at which the last of the ferrite goes into solution.
That is to say, the correct temperature for restoring the grain-
size will depend upon the amount of carbon in the steel; low-
carbon steel must be heated to nearly 900° C. (1650° F.) ; 0.4 per
cent, carbon steel must be heated to nearly 800° C. (1470° F.);
and so on.2 We can never get as small a grain-size in steel with

1 See page 246 of No. 1, page 8.

2 It is to be remembered that the changes indicated by the lines in Fig.
246 occur at a higher temperature on heating than on cooling (see page 312);
so it is well to heat the steel about 25° C. higher than the points on those
lines.

FIG. 277. — NO. 1A. STEEL OF 0.05

PER CENT. CARBON ROLLED.

Magnified 40 diameters.

FIG. 278. — NO. IB. STEEL OF 0.50

PER CENT. CARBON ROLLED.

Magnified 60 diameters.

FIG. 279. — NO. 2A. SAME AS NO. 1A

OVERHEATED TO 1420° C. (2588° F.)

Magnified 40 diameters.

FIG. 280. — NO. 2B. SAME AS NO. IB.

OVERHEATED TO 1420° C. (2588° F.)

Magnified 60 diameters.

FIG. 281. — NO. 3A. SAME AS NO. 2A.

REHEATED SLIGHTLY ABOVE AC8.

Magnified 40 diameters.

FIG. 282. — NO. 3B. SAME AS NO. 2B.

REHEATED SLIGHTLY ABOVE AC3-3.

Magnified 60 diameters.

Series A by F. C. Wallower in the Metallographic Laboratory of Columbia University,
Department of Metallurgy.

Series B by G. Rocour in the Metallographic Laboratory of Columbia University, De-
partment of Metallurgy.

THE HEAT TREATMENT OF STEEL 373

less than 0.9 per cent, carbon as we can in that which is exactly
0.9 per cent, carbon, because a new grain-size begins to grow
after we have crossed the line PS K, and yet we cannot entirely
eliminate the old grain-size until we cross the line G 0 S. Where
the lines GOS and PSK are near together (say, with 0.7 per
cent, carbon), the new grain-size does not have much chance to
grow before the restoration is complete, and therefore we may
obtain steel with a pretty small grain; but where they are far
apart (as in the low-carbon steels) the restoration can never be
very thorough, because we have to go so far above PSK to ob-
literate the old grain-size that the new grain-size will have at-
tained ample proportions. But the evidence seems to show that
the best net result is obtained by going just above the line GOS
in all cases.

In the case of steel with more than 0.9 per cent, carbon a some-
what similar condition exists: we must reheat the steel above
the line S a in order to produce complete elimination of the previ-
ous grain-size, but a new grain-size begins to grow from the cross-
ing of the line PSK. But here we disregard the line Sa, and
restore our steels in every case by reheating them over the line
PSK, just as in the case of pure pearlite. The reason for this is
that the lines Sa and PSK diverge so rapidly that we have to
heat very far above the line PSK before we cross S a, and there-
fore the new grain-size has grown greatly. Furthermore, the
only object of heating above the line Sa is to take the excess
cementite into solution; for the ferrite and cementite in the
pearlite all went into solution as soon as we crossed the line
PSK, but the amount of excess cementite is always small in
proportion, and therefore in its influence on restoration. Even
with steel containing 2 per cent, of carbon the excess cementite
is only 16 per cent. This is different from the low-carbon steels,
where the excess ferrite will be usually over 80 per cent.

Evidence of Overheating. — A piece of steel may be heated
many times above the line PSK and cooled again, but obviously
only the latest heating will leave its impression on the structure,
because each crossing of the line on the way up removes the effect
of previous heat treatment.1 The relation between the size of

is only true in a qualified sense, in that the previous overheating
must not have been very close to the melting-point. We shall discuss this
point under the head of " Burning." Indeed, even where burning has not

374 THE METALLURGY OF IRON AND STEEL

the crystals and the temperature above PS K is so constant that
we may determine what this temperature was from the analysis
of the steel and an examination of the grain. To do this it is
usually necessary to get a piece of steel of the same analysis,
heat different pieces of it to various temperatures, and compare
(see page 380). The analysis must be approximately the same
not only in carbon but also in phosphorus, sulphur, silicon, and
manganese, as well as in any alloying elements, if present, such as
nickel, chromium, tungsten, etc., because all of them have an
effect upon the size of grain and also upon the change in size of
grain by overheating. Generally, it is not important to know
the exact temperature of overheating, but only whether or not
overheating in some degree has occurred; and this is not difficult
to prove, because almost all who use steel are familiar with the
normal fracture of steels of different carbon and can tell at a
glance if the grain is large; those who are not so familiar with its
appearance may easily become so. The grain being large is proof
that overheating was the cause, provided chemical analysis shows
everything about normal, especially phosphorus and silicon.

Steel members of bridges or' other structures sometimes break
and disclose a crystalline fracture which is often attributed to the
effect of vibration. The same thing occurs with points or shanks
of rock drills and similar implements. It is the more general
opinion among metallurgists that the crystalline fracture in all
these cases is due to faulty heat treatment during manufacture,
and especially to finishing the forging or rolling while the tem-
perature is still too high. The manufacturers of steel like to
maintain the opposite opinion, for obvious reasons, but I do not
know of there ever having been any reliable proof offered that vi-
bration had caused, or is capable of causing, large-sized grain in
steel. It may be possible, but the more we learn about the sub-
ject the more we are inclined to believe that improper manufac-
ture is the cause, and that the grain was large before the steel was
put in service, although its nature was not disclosed until the
break occurred.

Mechanical Cure for Overheating. — When steel is to be rolled

occurred, a skillful microscopist may sometimes discern the effect of over-
heating after the steel has been restored by reheating; because, although the
crystals are all small, they are arranged in groups which show the form of the
previous large crystals.

THE HEAT TREATMENT OF STEEL 375

or forged it is frequently heated to a temperature of 1100° to
1350° C. (2010° to 2460° F.), and it might be thought that this
treatment would seriously damage it. So it would, but for the
fact that the subsequent mechanical pressure upon the metal
breaks down the crystals and reduces them again to a small size.
The result is that the final size of the crystals is dependent upon
the temperature of the material at the finish of the mechanical
operation. In other words, steel finished at 900° C. (1650° F.)
has a finer structure than the same steel if finished at 1100° C.
(2010° F.). We do not feel warranted in stating numerically the
exact relation between the finishing temperature and the grain-
size, as we have not yet sufficient evidence, but several rules af-
fecting the final size of grain seem to be virtually established: (1)
It is more advantageous to have the mechanical work applied
continuously from the highest temperature employed down to
the finishing temperature, rather than to have long waits during
which the steel cools; and especially is this true when the amount
of work put upon the metal at the lower temperatures is small.
In other words, if the steel is formed roughly to shape and size
at a high heat, is then allowed to cool, and a little work is done
upon it at the lower temperature, the grain will not be good.

(2) It is best for the metal to be worked by several passes through
the rolls, or many blows of the hammer, rather than to effect the
same amount of reduction by a lesser number of heavy drafts.

(3) The greater the amount of reduction the better; that is, to
work a large piece down to the desired article gives a better
structure. (4) The best temperature at which to finish the work
it probably upon, or slightly below, the lines G-O-S or S-K in Fig.
246, page 314.

Action in Rolling. — The exact crystalline action that takes
place under mechanical treatment is not definitely known. In
the case of rolling Professor Howe has tentatively assumed the
conditions graphically shown in Fig. 283,1 in which the line DG
represents the size of grain at the different temperatures. At
1400° C. (2550° F.) the grain-size is represented by the distance
of the line from the axis 0-0. On the first passage through the
rolls the grains are crushed to a very small size, but on emerg-
ing again they grow very rapidly. Meanwhile, however, the metal
has been cooled, and this fact, as well as the inability of the grains
1 Page 263 of No. 1, page 8.

376

THE METALLURGY OF IRON AND STEEL

to grow instantly, causes the new size of grain to be smaller than
before. Therefore, each passage through the rolls renders the
crystals smaller in size, the final size depending upon the tempera-
ture and the amount of pressure in the last pass. The only

1400 C

(2552 F)

(1274 F)

= Diameter of grain at this finishing temperature

32°F

18345

Diameter of Grain
FIG. 283. From Howe, "Iron. Steel and other Alloys."

abnormal assumption in this argument is that the crystals grow
rapidly after the crushing, whereas we know that when steel is
heated to any of these high temperatures, the growth is relatively
slow. This objection is not strong enough alone to refute the theory,
but other hypotheses may be advanced for those who require
further explanation. For example, it may be supposed that the
steel is so mobile at the very high temperatures that it yields to

THE HEAT TREATMENT OF STEEL 377

distortion, not altogether by the crushing of the crystals, but by
the sliding of the crystals past one another; as the temperature
becomes lower, however, the mobility of the mass becomes less,
and less sliding is possible, so that more crushing of the crystals
takes place.

Finishing Temperatures. — William Campbell has studied the
finishing temperature of steel containing 0.5 per cent, carbon
and finds that the very best qualities are produced in the steel
if mechanical work is ended just at the time when ferrite
begins to separate from solid solution, that is to say, just when
the steel is below the line G-O-S in Fig. 246, page 314. Work
below that temperature greatly increases the brittleness of the
material, while finishing the work at a higher heat results in lower
strength. Upon the evidence at hand, we may tentatively as-
sume like conditions for steels of any carbon, and expect the best
results if mechanical work is ended when the steel is at a tem-
perature which brings it exactly upon the line G-O-S or S-K, but
reserving, perhaps, the right to change this statement slightly
when more data are obtained.

Welding. — This brings us to the subject of welding, or the
joining of two pieces of wrought iron or steel by pressing or ham-
mering them together while at a very high temperature. In this
way a joint may be made which cannot be seen by the eye unless
the steel is polished and etched with acid, which usually develops
the junction line very clearly. The exact temperature of welding
is not known, but probably it is very near the melting-point, when
the steel is in a soft and almost pasty condition. Low-carbon steel
welds most easily; moreover, all impurities, especially silicon and
sulphur, reduce weldability. The procedure in welding is very
simple, and consists in heating the two pieces that are to be welded
to a high temperature, dissolving off the iron oxide, and then
pressing the two pieces together forcibly. The dissolving off of
the oxide is usually accomplished by rubbing the metal in some
flux, such as borax. At the present time various patented ' weld-
ing plates' are sold. These consist of thin plates of flux which
are put between the two pieces to be welded and so get rid of the
oxide, the pieces being hammered together with the plate between
them.

In the actual manipulation for welding the two pieces that are
to be joined together are usually 'upset/ or in some way en-

378 THE METALLURGY OF IRON AND STEEL

larged in size, so that after the junction the part of the bar right
at the weld is larger in size than the remainder. This part is
then hammered continuously until the metal is at a red heat,
the object being to break up the coarse crystals produced by
the high temperature, and, by having a low 'finishing tempera-
ture/ to obtain a small grain-size. With proper welding this
object will be attained so far as the metal immediately adja-
cent to the weld is concerned, but there is always a spot within
six inches or so of the weld which must necessarily have been over-
heated without subsequently receiving mechanical treatment, i.e.,
' hammer refining/ down to the proper finishing temperature. Thus
it is that most welded pieces break at a point not far from the
junction and under a strain much less than the original strength
of the bar. Blacksmiths and experienced welders are wont to
declare that if a welded bar does not break in the weld itself, then
it must be as strong as the original metal. As I have shown, how-
ever, this is by no means true. In a welding test carried on with
great care in this country by skilful and experienced welders who
were placed upon their mettle,1 the strength and elastic limit of
the welded bar was almost never as great as the original bar, and
in some cases was less than half. In ductility even worse results
were obtained. In a. similar test carried on at the Royal Prussian
Testing Institute the average strength of welded bars of medium
steel was only 58 per cent, of the original, that of softer steel only 71
per cent., and of puddled iron only 81 per cent., while the poorest
results were only 23, 33, and 62 per cent, respectively. It was
seen that bad crystallization adjacent to the weld was the cause
of ihe damage.

This evidence shows positively that welded steel and iron
bars should always be reheated to a temperature just above the
line G-O-S in order to restore by heating the grain-size of all parts.

Burning. — In the vernacular of the trade, all overheated
steel is termed 'burnt/ but this is not correct usage, because true
burning takes place only when the overheating is most abusive,
and, indeed, when the metal is heated almost to its melting-point.
It is probable that steel is burnt only when it is heated above the
line A-a in Fig. 246, page 314. Alfred Stansfield has studied
this question very ably,2 and distinguishes three stages of burn-
ing. The first stage is reached when the steel barely crosses the
1 See pages 401 to 406 of No. 2, page 8. 2 See No. 143, page 395.

THE HEAT TREATMENT OF STEEL

379

line A-a, that is, when the first drops of melted metal begin to
form in the interior of the mass: they segregate to the joints be-
tween the crystals and cause weakness. Stansfield thinks that
steel burned only to this stage may be restored by reheating
it first to a high temperature, cooling, and then heating again
to a temperature just above the lines G-O-S-K. The second
stage in burning is reached when these liquid drops segregate as
far as the exterior and leave behind a cavity filled with gas.

FIG. 284. — METCALF TEST. FRACTURES OF STEEL CONTAINING ONE PER
CENT. OF CARBON.

Stansfield thinks that steel burned to this stage might be re-
stored by combined reheating and forging. As a matter of
safety, however, I believe it would be well to remelt all such ma-
terial; in other words, send it to the scrap pile. The third and
last stage of burning is reached when gas collects in the interior
of the metal under sufficient pressure to break through the skin
and project liquid steel, which produces the well-known scintil-
lating effect at this temperature. Into the openings formed by
these minute explosions air enters and oxidizes the interior.
There can be no remedying of steel which has been burned to this
extent.

380 THE METALLURGY OF IRON AND STEEL

Metcalf Test. — A very interesting experiment is the " Met-
calf test," originated by William Metcalf.1 It is best performed
upon a bar of high-carbon steel, because this material shows the
differences in structure so readily to the eye. A bar of steel about
12 in. long is notched with a hacksaw or chisel at intervals of
an inch. One end is then placed in a fire and heated to a tem-
perature at which it scintillates, while the other end is at a black
heat. Then it is removed and cooled. It is immaterial whether
the cooling be rapid or slow, but time may be saved by plunging
it into water. It is then broken at every notch, and an exam-
ination of the fractures will show a very large size of crystals at
the end which is burnt, gradually decreasing until a fine and
silky appearance is presented where the metal was exactly at
the temperature of restoration, while beyond that point the
fracture will be the same as that of the original bar. In case
this test is made upon low-carbon .steel it should not be notched
before treatment, but afterwards it should be cut apart with a
hacksaw at intervals of an inch, and then polished and examined
under the microscope, because soft steel will not break with-
out bending, and this bending destroys the indications of the
fracture. This Metcalf test is very serviceable in case we desire
to compare steel that we suspect of being overheated with over-
heated steel of like analysis to determine the degree of over-
heating.

Castings do Not Burn. — It might be thought that every steel
casting would suffer the injuries due to burning because it is
cooled through the space between the lines A-B and A-a, and es-
pecially so in the case of high-carbon steel, which is very easy to
burn, on account of the low temperature at which this line A-a
occurs and of the long distance between the two lines. Such
injury, however, does not ordinarily take place, anol this fortu-
nate circumstance is explained partially by each of three differ-
ences between the heating and cooling of steel: (1) When steel is
heated into the area where burning takes place, it is subjected
longer to the burning temperature, because it generally takes
longer to heat steel than to cool it. (2) When steel is being
heated, the heat is traveling inward from the outside, and there-
fore all parts are expanding, and there is some opportunity for the
crystals to draw apart and form cavities. On the other hand,
1 See pp. 405, 406 (especially 406) of No. 116, page 332.

THE HEAT TREATMENT OF STEEL 381

when it is cooling from the molten state, the outside layers are
the cooler, and tend to contract upon the interior and hold the
crystals more firmly together. (3) When steel is cooling from
a molten state, it is constantly giving off from solution hydrogen
and other deoxidizing gases which are soluble in it while liquid,
and these gases prevent the oxidation of the crystal faces by the
percolation of air into the interior.

Ingotism. — I have already discussed ingotism and said that
the crystals in cast steel are larger than those of rolled steel, due
to growth while the metal is at a high temperature, and I have
stated that sometimes these crystals are very large, because the
conditions of casting cause the steel to occupy a longer time in
cooling from the liquid state down to a black heat. It is prob-
able that ingots and castings do not show the effects of overheat-
ing (ingotism) to any marked extent unless they are a long time
above 1100° C. (2010° F.). In case these coarse crystals do form,
they may be restored to some extent by reheating the casting
to a point just above the line G-O-S. Why ingotism is not com-
pletely remedied by the same treatment that cures the coarse
crystallization due to overheating, I am unable to say, unless it
be that ingotism is accompanied by burning to at least a slight
extent.

Stead's Brittleness. — In addition to the damage caused by
overheating, steel very low in carbon (say under 0.15 per cent.)
is subject to another and peculiar danger, for if this soft steel
be held for a very long time at temperatures between 500° and
750° C. (930° and 1380° F.), the crystals become enormous and
the steel loses a large part of its strength and ductility. Fortu-
nately it takes a very long time, in fact days, to produce this
effect to any alarming degree, so that it is not liable to occur,
even through carelessness, during manufacture or mechanical
treatment. But steel is sometimes placed in positions where it
may suffer this injury, for example, in the case of the tie-rods of
furnaces, supports for boilers, etc., so that the danger should be
borne in mind by all engineers and users of steel. I recall an
instance where the breaking of a piece of chain that supported
one side of a 50-ton open-hearth ladle caused a loss of life under
the most horrifying conditions, due to the fact that the wrought-
iron chain had been heated up many times to a temperature above
500° C. (930° F.), and had finally reached a condition of coarse

382 THE METALLURGY OF IRON AND STEEL

crystallization, so that it was unable to bear the strain upon it
when the ladle was full of metal.

This phenomenon of coarse crystallization in low-carbon steel
is known as "Stead's brittleness," after J. E. Stead, who has
explained its cause. The effect seems to begin at a temper-
ature of about 500° C., and proceeds more and more rapidly with
an increase in temperature until we reach 750° C., above which
no growth seems to take place. The damage may be repaired
completely by heating the steel just above the line G-0. In other
words, the remedy for coarse crystallization in this case is the
same as that for coarse crystallization due to overheating, and
all steel which is placed in positions where it is liable to reach
these temperatures frequently, should be restored at intervals
of a week or a month, or as often as may be necessary.

HARDENING OF STEEL

If steel be raised to a bright-red heat and then rapidly cooled,
as, for example, by plunging it into water, it becomes very much
harder and at the same time stronger and more brittle. One cir-
cumstance is absolutely necessary to produce the increase in
hardness, namely, that the temperature from which rapid cooling
takes place shall be above the critical temperature of the steel.
Take, for example, steel containing 0.9 per cent, carbon; we may
heat this ever so little below the point S in Fig. 246, page 314,
and no increase in hardness will take place, even though we cool
with extreme rapidity. On the other hand, if we cool the same
steel rapidly from ever so little above the point S, it will be hard
enough to scratch glass and brittle enough to fly into pieces under
a blow of the hammer. This is the maximum practical hardness
which can be obtained, for if we quench the steel at a still higher
temperature, the only result of importance is to do it damage by
increasing its grain-size. In case we have less than 0.9 per cent,
carbon in our steel, the best temperature for hardening is just
above the line G-O-S, because that gives the maximum hardness
and also the best grain-size. The best temperature from which
to harden steel with more than 0.9 per cent, carbon is just above
the line S-K, because that gives the best grain structure, although
it is true that greater hardness is obtained if we cool from above
the line S-a.

THE HEAT TREATMENT OF STEEL 383

Carbon and Hardness. — The hardness of steel increases with
every increase of carbon. This applies to the hardness of steel
in its natural state, and still more influentially to its hardness
after the treatment I have just described. Although iron free
from carbon is hardened by rapid cooling from above the point
Ac2 (760° C. = 1400° F.), and a little more so when rapidly cooled
from above AcB (900° C. = 1650° F.), yet this degree of hardness
is so slight as to be perceptible only by means of delicate labora-
tory tests. With 0.25 per cent, carbon the hardness begins to be
perceptible by crude tests, but it is only when we get above 0.75
per cent, carbon that ordinary steel acquires sufficient hardness
for the process to be used commercially, — for example, for springs,
saws, etc. Metal-cutting tools are usually made of steel con-
taining 1 per cent, or so of carbon, while very hard implements,
such as files, etc., will contain 1.5 per cent., or slightly more.

Rate of Cooling and Hardness. — The degree of hardness of
steel also varies with the speed of cooling from above the critical
range of temperature. When the cooling is very slow, as, for
example, when it takes several days to cool, the steel will be as
soft as it is possible to make it. When it is cooled by being taken
out of the furnace and suspended in the air, or thrown upon a sand
floor, it will still be relatively soft. When cooling is still more
rapid, as, for example, when it is taken out of the furnace at a
bright-red heat and plunged into a heavy oil with a low conduct-
ing power for heat, it becomes quite hard and springy, provided
its carbon is in the neighborhood of 0.8 per cent, or above.
Quenching in a thin oil from the same temperature makes it still
harder. Quenching in water makes it harder still ; and so on, the
degree of hardness increasing as we quench in liquids which take
the heat away from it faster and faster, such as ice-water, ice-
brine, ice sodium chloride solution, and mercury near its freezing-
point (-39° C. = -38°F.).

Theories of Hardening. — One essential feature of hardening
is that the steel must be heated to a temperature above the line
P-S-K, that is to say, to a point where at least some of the solid
solution of iron and carbon is formed. There are several different
theories to explain the hardness produced by rapid cooling from
this point, the two most important being the ' carbon theory ' and
the 'allotropic theory/ Both of these theories depend upon the
following line of reasoning: At temperatures above the critical

384 THE METALLURGY OF IRON AND STEEL

range the molecules of steel are in a hard condition.1 As they cool
from this point and cross the critical range of temperature, the
molecules change from the hard state to a soft state, but this
change is not rapid and requires time for its accomplishment;
rapid cooling does not afford the necessary time, and so perpetuates
the hard state of the molecules. This line of reasoning leaves only
one point in doubt, namely, what causes the molecules to be hard
when the steel is above the critical temperature, i.e., when iron
and carbon are dissolved in each other?

The Carbon Theory. — The carbon theory assumes that the
hardness of steel is due altogether to the carbon dissolved in it,
and in evidence its advocates point to the extreme hardness of
one form of carbon — the diamond. This theory has the ad-
vantage of simplicity, and has in its favor the fact that the hard-
ness varies almost directly with the amount of carbon. Against
the theory, it is urged that the amount of carbon is really too small
to produce such a great degree of hardness in the whole mass of
metal. Furthermore, although carbon is found in many metals,
it does not confer hardness on any of them except iron.

The Allotropic Theory. — We have already learned that in
the solid solution the iron is present in the gamma allotropic form,
and there is one school of metallurgists which attributes the hard-
ness to the allotropic form of iron alone and denies that carbon
has any direct influence. It has been shown, by very delicate
laboratory tests on iron practically free from carbon, that the
gamma and beta allotropic modifications are harder than the
alpha modification. He was not able to show how great was the
increase of hardness of one form over another, because he was
never able to cool the iron fast enough to prevent it changing
back in part to the alpha form. In the absence of carbon the
change from gamma to beta and from beta to alpha is very rapid,
so that cooling has to be almost instantaneous in order to prevent
it, and this, of course, is impossible. The ' allotropists ' explain the

1 It is difficult for some to understand how molecules of steel can be in a
hard condition at a temperature at which we know that the mass as a whole
is soft and mobile; but this can be explained by the following comparison:
A ball of wet sand and clay is soft and mobile as a whole, because the particles
move by each other readily and the mass changes its shape under pressure;
yet the individual particles composing this mass are many of them hard
enough to scratch glass with ease.

THE HEAT TREATMENT OF STEEL 385

greater hardness of high-carbon steel as compared with low-carbon
steel upon the basis that the presence of carbon makes the change
from gamma to beta and to alpha iron slower, and therefore enables
more of the iron to be retained in the gamma and beta forms by
the rapid cooling.

One additional argument in favor of the allotropic theory is
that when steel cools slowly through the critical range, it loses
its hardness slightly before the carbon comes out of solution.
This would indicate that the allotropic change took place before
the carbon change, and that the allotropic change was the cause
of hardness. What is true of the loss of hardness is also true of
the other physical changes which take place at the same time.
That is to say, the steel regains its magnetism, decreases in electric
resistance, and increases in thermo-electric power in a large part
before much carbon is separated from solution.

Influence of Carbon on Hardness. — I think no one to-day
denies that the carbon in steel has a very important influence
upon its hardness, even though it may not be the sole cause of it.
This influence is twofold: (1) We have stated that the changes
that take place when steel is cooled through the critical range
were not rapid changes, and for this reason fast cooling was able
in part to prevent their taking place. Carbon has the effect of
making these changes still slower, and so increasing the effect of
the rapid cooling. Howe calls this the 'brake action' of carbon.
(2) The more the carbon in solid solution, the harder will that
solid solution be.

The Compromise Theory. — Several theories have been ad-
vanced which are a compromise between that of 'the carbonists'
and that of 'the allotropists.; The simplest of these, and the
theory now most generally accepted, is that the hardness of the
molecules of steel above the critical range is due partly to the
allotropic form in which the iron exists, and partly to the fact that
we have a solid solution of iron and carbon.1 In other words,

1 Where I speak of a solution of carbon and iron, I intend to include under
this also the solution of iron and a carbide of iron. That is to say, we have
two substances, iron and carbon, and they are dissolved in each other. It
may be that the carbon is united with part of the iron to form a carbide, and
then that this carbide is dissolved in the rest of the iron; but I use the term
"solid solution of iron and carbon" to cover either this condition or that of
elemental carbon dissolved in iron.

386 THE METALLURGY OF IRON AND STEEL

the hardness is due to the fact that we have a solid solution of
carbon in an allotropic form of iron.

The Internal Stress Theory. — There is an entirely independent
theory of the hardness of steel which attributes it, not to the re-
tention of a hard molecule existing above the critical range, but
to stresses set up in the metal by rapid cooling through the crit-
ical range. That such stresses exist cannot be doubted, for some-
times the rapid cooling of high-carbon steel causes it to break into
pieces, or to open up a cavity in the middle from end to end,
but the theory itself does not seem sufficient to explain all the
facts.

Tempering. — Hardened steel is too brittle to be used without
some degree of tempering; except for a small variety of purposes,
such as the points of armor-piercing projectiles, the face of armor
plate, etc. In order to understand just what tempering does,
let us consider the exact condition of hardened steel: it is in a
hard and brittle condition which is not natural to it at atmos-
pheric temperatures, but which has been brought down with it
from a higher temperature by means of rapid cooling. Theo-
retically, when the temperature fell below 690 C°. (1272° F.),
the molecules of steel should have changed over to the soft form.
Their hard condition is not in equilibrium at the lower tempera-
ture, in the same sense that ice is not in equilibrium in hot weather.
Why, then, does not the steel change back into the soft form?
Ice, if given time enough, will all change into water when the
temperature is above 0° C. (32° F.). The reason the change does
not take place in the steel after we have cooled it to the atmos-
pheric temperature is that the mass as a whole becomes too rigid
and immobile at the lower temperature to permit any alteration
in its molecules to take place.

However, it is only necessary to decrease this rigidity in order
to permit a slight change. For example, if a piece of hardened
steel be kept in boiling water for some days it will lose a part of its
hardness; if it be heated a little more, it will lose more hardness
and lose it much more quickly. Each loss in hardness is accom-
panied by a loss in brittleness as well. If it be heated to about
200° C.*(392° F.), quite a little of the brittleness will be lost and
a part of the hardness.1 It is now in condition to be used for steel

1 After the heating it is immaterial whether cooling is fast or slow, as the
same result will be produced.

THE HEAT TREATMENT OF STEEL 387

engraving tools, lathe tools, and other implements to cut metals.
If we heat to 250° C. (480° F.), we again temper, and to a point
where the steel is still less brittle and will withstand greater
strains, but is at the same time sufficiently hard to be used for
rock drills, penknives, stone-cutting tools, and so forth. If tem-
pered to 275° C. (525° F.), the steel has not even yet lost so much
brittleness as to be able to withstand a great deal of shock, or
even bending, but is still hard enough to be suitable for dental
and surgical instruments, swords, needles, hacksaws, etc. Wood-
saws, the majority of springs, and other articles that must be
ductile even at the expense of hardness, will be tempered to a
temperature of about 300° C. (570° F.), which is the greatest de-
gree of tempering that is ordinarily employed.

It is interesting to note that when hardened steel is tempered,
the physical changes produced by the tempering — the decrease
in hardness and brittleness, increase in electric conductivity, etc.
— precede the separation of carbon from the solid solution. By
tempering we may lose 70 per cent, of the hardness, 93 per cent,
of the electric resistance, and nearly 100 per cent, of the thermo-
electric power produced in the steel by the hardening operation,
when only 13 per cent, of the carbon has been changed from the
dissolved form.

Temper Colors. — Nature has provided a ready means of de-
termining the temperature of steel between 200° and 300° C.
(390° and 570° F.) without the aid of thermometers or other in-
struments; and, since this is the range of temperatures in which
practically all of the tempering of hardened steel takes place, this
provision is a most fortunate one. It comes about through the
oxidation of the metal at those different points. At 200° C.
(390° F.) a thin film of oxide forms upon the steel, but is not suf-
ficient to entirely hide the white color underneath, so that the
combination produces a light lemon color. As the temperature
rises the film of oxide becomes thicker and the yellow color darker
until, at about 225° C. (437° F.), it has changed to orange. At
250° C. the orange has changed to a pink which is known as
' pigeon wing/ At 275° C. the pigeon wing has turned into a light
purple, which, at 300° C., becomes a blue.

Hardening, Tempering and Annealing. — Only quenching in
water, or in some other medium which takes the heat away as fast
or faster, goes under the name of hardening. Quenching in oil,

388 THE METALLURGY OF IRON AND STEEL

melted lead, etc., cools the steel less rapidly and makes it less hard
and less brittle than quenching in water, so to this operation
the name of 'tempering' is given. Cooling in the air, in sand,'
in the furnace, or by any other slow method, is called 'annealing.7

Combined Hardening and Tempering. — If steel be cooled from
above the critical range by quenching in some slow conducting
liquid in the first instance, as in cylinder oil, the same intermediate
hardening and embrittling effect will be produced upon it as if it
were first hardened in water and then tempered a certain amount.
Therefore the quenching in oil and similar mediums has come to
be called 'tempering.'1 Another method of combining hardening
and tempering after only one heating is used in the tempering of
the cutting edges of chisels and similar tools: the end of the tool
is first heated just above the critical range, and then the extreme
point only is quenched in water until it is black, after which it is
withdrawn and rubbed bright upon a piece of sandpaper, or upon
a brick. This is done merely to give a bright surface upon which
to observe the play of temper colors. The heat from the shank
now begins to creep down into the point, which takes the various
temper colors in order, beginning with the lemon. When the
desired degree of tempering is reached — say, the pigeon- wing
color — the whole tool is put into water. This is merely to ' put
out the fire' and stop more heat coming down into the tempered
point ; it has nothing to do with the tempering operation itself.

Annealing. — If hardened steel be heated to a temperature
of 600° C. (1100° F.) its pristine softness and ductility is returned
to it. This process is known as 'annealing.' As a general thing
we do not anneal at as low a temperature as this because we not
only desire to make the steel soft, but also to give it as small
a grain-size as possible, and this is done by heating just above
the lines G-O-S-K in Fig. 246, page 314. Therefore the temper-

1 Strictly speaking this is a misnomer, and very bad usage, but like so
many incorrect terms that have become current in the iron and steel trade, it
is now too firmly established to be displaced. The error in terms has gone
even further than this, and the hardness of steel is known as its 'temper,'
while making it softer is known as 'drawing its temper.' To temper means
literally to soften, to mollify; or else to mitigate one quality with another
(as justice is said to be 'tempered' with mercy). Either of these meanings is
quite correct when we 'temper' hardened steel by heating it slightly, but is
just the opposite when we call hardening steel in oil 'tempering,' or when we
speak of 'drawing a temper.'

THE HEAT TREATMENT OF STEEL 389

atures just above these lines are not only the hardening and re-
storing temperatures, but the annealing temperatures as well, if
we cool slowly. It is only when we have a steel with a structure
already fine that we use the low-temperature annealing, as, for
example, with cold-rolled steel, wire, etc. To heat such steel
above the line G-O-S would not increase its softness and would
undo some of the benefit of the cold work.

Magnetism and the Lines G-O-S-K. — It will be remembered
that iron is present in the solid solution in the gamma allotropic
form, and therefore the solid solution is non-magnetic. Therefore
all the steels to the right of the point 0 in Fig. 246, page 314, lose
the last of their magnetism at the same time as they cross the
lines 0-S-K. These steels comprise all containing 0.4 per cent,
of carbon and more. To harden, anneal or restore such steels we
may guide our work of heating by means of an ordinary horse-
shoe magnet, which makes a most accurate and simple tool.
Let the magnet hang outside of the furnace and take the steel out
at intervals to test it. When it no longer attracts the magnet,
begin to cool it. For steels with less than 0.4 per cent, carbon we
can use the magnet to tell us when the temperature corresponding
to the line M-0 is reached, for all iron loses its magnetism at that
point, and then it is a comparatively simple matter to judge by
eye the relatively short temperature intervals above that point
which it is necessary for the steel to traverse before it crosses the
line G-0. If this method is followed I think it will be found in
many works that annealing temperatures have been much too
high, and that better steel will be obtained in future. We do not
get the steel any softer by annealing it hotter, but only by slower
cooling.

THE CONSTITUENTS OF HARDENED AND TEMPERED STEELS

It is now pretty generally admitted by metallurgists that aus-
tenite is the solid solution of gamma iron and carbon. When
this cools slowly through the critical range it decomposes into
ferrite and cementite. It is also very generally admitted that the
decomposition does not take place spasmodically, but progresses
by stages, and many believe that the substances identified under
the microscope as martensite, troostite, and sorbite are the prod-
ucts of these stages. That is to say, martensite is the first stage

390 THE METALLURGY OF IRON AND STEEL

of decomposition of austenite, troostite is the second, sorbite is
the third, and pearlite is the consummation. If this is so, then
austenite will be found in the steels cooled with the greatest ra-
pidity, martensite will be found in those cooled with the next de-
gree of rapidity, troostite will be found in the next intermediate
steels, sorbite in the next, and pearlite in those slowly cooled.
To this extent, indeed, the facts agree with the argument, but
hardly any dare predicate further too positively from this evi-
dence alone. Our knowledge upon this whole subject is still very
new, and though a small army of workers is busy collecting evi-
dence and interpreting it as best they can, all our present state-
ments must necessarily be made tentatively, with the idea of pre-
senting the facts so that they may be of some practical benefit,
even though later information may oblige us to change slightly
the scientific basis upon which we found them.

Austenite. — Austenite can be obtained at atmospheric tem-
peratures in ordinary carbon steels only when three conditions
are collectively present: (1) The brake action of carbon must be
very strong, that is, there must be above 1.1 per cent, of carbon
present; (2) the steel must be cooled with the greatest rapidity,
as by quenching it in iced solutions at, or a few degrees below
zero C.; and (3) it must be cooled from above 1000° C. (1830° F.).
Even then austenite cannot be preserved throughout the whole
mass of the steel, but at least a part of it will be decomposed to
the stage represented by the martensite structure (see Fig. 288)
while cementite will separate also if the carbon is very high. Un-
der the microscope austenite may be differentiated from marten-
site by its white color after etching with a 10-per-cent. solution of
hydrochloric acid. Better results are obtained if the etching is
aided by electrolysis, the steel being made the anode, or positive
pole, and a piece of platinum being made the cathode. Austenite
is not a very important constitutent practically, because prob-
ably it almost never occurs in commercial steels. It is imme-
diately decomposed into later stages upon tempering.

Martensite. — Martensite is the chief constituent of ordinary
hardened steels, that is, of steels quenched from above the critical
range in water or in an iced solution. Its structure is shown in
Fig. 288. It is even harder than austenite, for a steel needle drawn
across the surface of a polished piece of steel will scratch the aus-
tenite plainly without making a mark upon the martensite. Its

FIG. 285. — NO. 1A. STEEL OF 0.34
PER CENT. CARBON OVERHEATED
TO ABOUT 1300° C. (2372° F.).
Magnified 250 diameters.

FIG. 286. — NO. 2A. SAME AS NO. 1A.

REHEATED SLIGHTLY ABOVE AC3.

Magnified 250 diameters.

FIG. 287. —STEEL OF 1 PER CENT.
CARBON BURNT.

Magnified 265 diameters.

FIG. 288. — MARTENSITE.
Magnified 250 diameters.

FIG. 289. — MARTENSITE (WHITE)

AND TROOSTITE (DARK).
Magnified 500 diameters. Etched lightly
with tincture of iodine. (H. C. Boynton.)

FIG. 290. — AUSTENITE (WHITE) AND

TROOSTITE (DARK).
Magnified 60 diameters. Steel of 1.41 per

cent, carbon. Etched with picric acid.

(William Campbell.)

392

THE METALLURGY OF IRON AND STEEL

structure is developed by the 'polish attack ' method of F. Os-
mond (see p. 453). Martensite is also found very largely in the
tempered steels, to which they doubtless owe their quality of
hardness. It was long thought to be the solid solution itself, and
is still spoken of as such in several books upon the subject; but
the microscope shows positively that it is not free from some
decomposition.

Hardenite. — Hardenite is the name sometimes given to ' sat-
urated martensite/ that is, martensite containing 0.9 per cent,
carbon. It is often found in German and French, and occasion-
ally in English and American books.

Troostite. — Troostite is obtained in either of three ways : (1)
By quenching steel in water when it has cooled just to the line
P-S-K in Fig. 246, page 314; (2) by quenching steel from a higher
temperature in boiling water, oil, or some other of the tempering
mediums; and (3) by hardening in the usual way and then tem-
pering by reheating. In other words, troostite is a product of the
usual tempering operations, and is found abundantly in tempered
steels. Under the microscope it may be distinguished by the

' polish attack' (page 453) or
by etching with tincture of
iodine. It appears as yellow,
brown, blue or black color-
ations on the borders of the
martensite, and often between
the martensite and sorbite.
The division between it and
the martensite is very sharp,
but it shades very gradually
into the sorbite (Figs. 289, 290).
Sorbite. — Sorbite is very
close to pearlite, and differs
from it chiefly in that the
crystals of ferrite and cemen-
tite are not quite perfectly
developed and segregated from

one another (see Fig. 291). It might at first seem almost like
splitting hairs to differentiate between the two, but this is not so.
because of the importance of sorbite, due to its having greater
strength than pearlite. Pearlite has a finer and more intimately

FIG. 291. — FERRITE AND SORBITE
IN RAILED STEEL.

Magnified 250 diameters.
Etched with picric acid.

THE HEAT TREATMENT OF STEEL 393

entangled structure than any other slowly cooled steel, and to
this we attribute the fact that pure pearlite steel (0.9 per cent,
carbon) is stronger than any other. But the structure of sorbite
is even finer than that of pearlite, and sorbitic steels are corre-
spondingly stronger than pearlitic steels.

Sorbite may be obtained by quenching the steel immediately
below, or just at the end, of cooling through the critical range,
or by cooling the steel pretty fast through the critical range with-
out actually quenching, or by rapid cooling and then reheating
to about 600° C. (1110° F.).

Osmondite. — Only very recently what is apparently another
constituent of hardened and tempered steels has been discovered,
and to this the name of 'osmondite' has been given. It would
seem that osmondite is a solid solution of carbon (or of a carbide
of iron) in the alpha allotropic modification of iron. It is there-
fore an anomaly, and can only exist in equilibrium momentarily,
so to speak, because the normal solid solution contains the iron in
the gamma allotropic form. It would seem, however, that when
this gamma solid solution decomposes, it passes through a phase
wherein the iron has changed from the gamma to the alpha form,
before the precipitation of the ferrite occurs. In other words,
then, osmondite is simply another stage in the decomposition of
austenite into pearlite. It differs from the other stages by having
a definite constitution and nature, whereas martensite, troostite
and sorbite are more or less indefinite and uncertain in compo-
sition, being in fact probably a mixture of two or more constitu-
ents rather than definite and individual components.

It is evident that osmondite can never compose the whole of
any piece of steel, because the earlier stages of the austenite de-
composition will grade into it on the one side and the later stages
on the other. That is to say, we would not expect that the entire
piece of steel would change simultaneously and instantaneously
from austenite into osmondite, and then likewise from osmondite
to pearlite. Osmondite may be obtained by tempering hardened
steel at 400° C. (750° F.), but its distinguishment under the micro-
scope is not known at present. Our chief means of recognizing
it are by the fact that it causes the steel to dissolve more rapidly
in dilute sulphuric acid and to be colored more deeply by alcoholic
hydrochloric acid, and by the fact that steel containing a good
deal of osmondite has lost a major part of all the qualities given

394

THE METALLURGY OF IRON AND STEEL

to it by hardening, although almost all the carbon is still in the
solid solution.

It is to be noted that the evidence proving the existence of
osmondite is still open to the possibility of doubt, although the
probabilities are greatly in favor of it.

Summary. — We may summarize the constituents of har-
dened and tempered steels in the following table, which is adapted
from one designed by Howe:

STAGES OF THE TRANSFORMATION FROM AUSTENITE INTO PEARLITE

CONSTITUTION

Austenite

Martensite, a transition substance.

Troostite, a transition substance . .

Osmondite

Sorbite, a transition substance. . .

Pearlite.

Solid solution of an iron carbide in gamma
iron.

The next step. Some of the gamma iron
is changed to beta and to alpha iron.
It is harder than austenite.

The third step. The quantity of gamma
and beta iron constantly decreasing,
that of alpha iron constantly increas-
ing.

Solid solution of iron carbide in alpha
iron.

A mixture of a constantly decreasing
quantity of osmondite with a con-
stantly increasing quantity of pearlite,
too fine to be resolved by the micro-
scope.

A conglomerate, or a mechanical mixture
of free alpha iron (alpha ferrite) with
the iron carbide, Fe3C, cementite.

In a very illuminating discussion of the constitution of iron
carbon alloys,1 Albert Sauveur predicates that when steel cools
slowly through the critical range and there takes place the de-
composition of austenite — the solid solution of carbon and
gamma iron — it first changes into a solid solutiorr of carbon and
beta iron, thence into a solid solution of carbon and alpha iron,
and thence into cementite and alpha iron. In other words,
Sauveur considers that, instead of the solid solution decom-
posing by the precipitation of its gamma ferrite, which im-
mediately changes to beta and then to alpha ferrite, the fer-^
rite in the solid solution is changed from gamma to beta and
then to alpha, after which the precipitation occurs. By the

Journ. Iron and Steel Inst., No. 4, 1906, pp. 493-575.

THE HEAT TREATMENT OF STEEL 395

light of recent researches this theory of Sauveur's is greatly
strengthened.

REFERENCES ON THE HEAT TREATMENT OF STEEL

140. Report of the Alloys Research Committee of the Institution of

Mechanical Engineers (England). Minutes of the Proceed-
ings. Report No. 1, October, 1891; No. 2, April, 1893;
No. 4, February, 1897; No. 5, February, 1899; No. 6,
January and February, 1904; No. 7, November and
December, 1905.

141. Joseph V. Woodworth. "Hardening, Tempering, Annealing,

and Forging of Steel." New York, 1903. This book tells
much of the practice and manipulation of these operations.

142. John Lord Bacon. "Forge Practice (Elementary)." New

York, 1904. This book also deals largely with practice
and manipulation.

143. Alfred Stansfield. "The Burning and Overheating of Steel."

Journal Iron and Steel Institute, No. 11, 1903, pages 433-
468.

144. Carl Benedicks. "Recherches Physiques et Physico-Chimi-

ques sur PAcier au Carbone." Upsala, 1904. With a
good bibliography.

145. Hanns Freiherr von Jliptner. "Siderology, the Science of

Iron." Translated by Charles Salter, London, 1902.

146. Hanns Freiherr von Jiiptner. " Grundzuege der Siderologie."

In three volumes. Vol. i, " The Constitution of Iron Alloys
and Slags," Leipzig, 1900; vol. ii, "Relation Between the
Thermal and Mechanical Changes, Constitution and Proper-
ties of Iron Alloys," 1902; vol. iii, "The Influence of
Impurities; the Manufacture of Iron and Steel," etc.

147. J. W. Mellor. "The Crystallization of Iron and Steel."

London, 1905. This is a very clear and readily intelligible
discussion of the heat treatment of steel, and an introduc-
tion to metallography.

148. George W. Ailing. "Points for Buyers and Users of Tool

Steel." New York, 1903. Being a general review of the
main sources of trouble and how to avoid them.

ALLOY STEELS

WE have already described ordinary steel (which to distinguish
it from the so-called alloy steels is often known as ' carbon steel ')
as an alloy of iron and carbon, and it is impossible to consider it
otherwise with equal accuracy. But there is another class of
materials to which the specific name of ' alloy steels' is applied.
This comprises steels to which a controlling amount of some alloy-
ing element in addition to the carbon is added.

Definitions. — The International Committee upon the nomen-
clature of iron and steel defines alloy steels as "those which owe
their properties chiefly to the presence of an element (or elements)
other than carbon." The distinction between an element added
merely to produce a slight benefit to ordinary carbon steel and the
same element added to produce an alloy steel is sometimes a very
delicate one. For example, manganese is added in amounts usu-
ally less than 1.50 per cent, to all Bessemer and open-hearth steels
for the sake of getting rid of oxygen and neutralizing the effect of
sulphur. Likewise silicon is sometimes added in amounts of 0.1
to 0.2 per cent, to get rid of blow-holes. But neither of these
additions produce what is known as an alloy steel. When we
make ' manganese steel ' containing 10 to 20 per cent, manganese,
the material has new properties quite different from the same
steel without manganese and we therefore get an alloy steel.
Similarly ' silicon steel ' containing 2 or 3 per cent, of silicon will
have an entirely new set of properties due to the silicon, and will
therefore become an alloy steel.

Ternary Alloys. — A ternary alloy, or three-part alloy, is an
alloy composed of iron, carbon, and one other influential element.
This class includes the alloy steels which are used most abundantly
by man and the most important of which are nickel steel, man-
ganese steel, chrome steel, tungsten steel, molybdenum steel,
.silicon steel, vanadium steel, and titanium steel. There are sev-

ALLOY STEELS 397

eral other ternary steels which have been investigated and used to
a small extent, such as boron steels, cobalt steels, etc. The field
of useful ternary steels has not yet been investigated except in
the most meager degree and a wide scope is left for future in-
ventors. There are many elements whose influence on steel has
not yet been studied, and even among those which are commonly
used, there are some of which only limited proportions have been
employed.

Quarternary Steels. — Quarternary, or four part, steels consist
of iron, carbon, and two other alloying elements. The commonest
and most important of these are nickel-chromium steels, tungsten-
manganese and tungsten-chromium steels, nickel-manganese,
manganese-silicon, tungsten-molybdenum, tungsten-nickel, nickel-
vanadium steels, etc., etc. The result produced by adding an
alloying element to ordinary carbon steel is astonishing and inca-
pable of being predicated, arid that obtained by a combination of
two alloying elements is far more so. New products result with
properties entirely different, and in some cases almost the opposite,
of those of its constituents, so that almost any combination at ran-
dom may lead to a surprise, even when the effect of different com-
binations of the same components is known. Therefore the possi-
bilities of quarternary steels seem to be very great and the field
has, as yet, hardly been touched.

Manufacture of Alloy Steels. — The manufacture of alloy steels
is usually very simple and calls for no special comment here. As a
general thing the alloying element is added like the recarburizer.
For example, in the manufacture of manganese steel the requisite
amount of f erro-manganese will be added at the end of the process ;
in the manufacture of nickel steel we may add ferro-nickel in the
same way, but it is more common here to add shotted-nickel during
the process and allow it to dissolve in the steel bath and remain
there until the metal is tapped; tungsten steels and tungsten-
chrome steels are often made by the crucible process and the
requisite amount of ferro-chrome and ferro-tungsten, or of me-
tallic chromium and metallic tungsten, is placed on top of the
charge when the crucible is filled.

The treatment of some of the alloy steels is not so simple:
Nickel steel may be heated, rolled and forged without any great
precaution, but these operations are performed upon manganese,
tungsten, and some of the other alloy steels only after great diffi-

398 THE METALLURGY OF IRON AND STEEL

culty and experience. Under the head of these different steels, I
shall describe the proper method of treatment.

NICKEL STEELS

Nickel steels are the most important of all the alloy steels
and are the most abundantly used. In the ordinary commercial
alloys the nickel ranges from 1.50 to 4.50 per cent, and usually
from 3.00 to 3.75 per cent, while the carbon is usually from 0.20
to 0.50 per cent. Not counting armor-plate, which is really
a quarternary steel, containing both nickel and chromium, the
most important uses of nickel steel are for structural work in
bridges, railroad rails, especially on curves, steel castings, ord-
nance, engine forgings, shafting, especially marine shafting,
frame and engine parts for automobiles, wire cables, axles, es-
pecially for automobiles and railroad cars, etc., etc. We can
best learn- the reasons for these particular uses by discussing
the distinctive properties conferred by the nickel and their use-
fulness.

Tensile Properties. — The chief distinction between nickel
steel and carbon steel is the higher elastic limit of the former, and
especially the fact that this higher elastic limit is obtained with
only a slight decrease in ductility. About 3.50 per cent, of nickel
added to carbon steel will increase the elastic limit nearly 50
per cent., while reducing the ductility only about 15 to 20 per
cent. It is this increase in elastic limit which is probably the
chief reason for the increased resistance of nickel steel to what
is known as "fatigue," that is to say, its resistance to repeated
stresses and alternating stresses1 under which all steel will ulti-
mately break down, even though the load is far less than that
it can bear indefinitely if constantly applied. It is probable
that the molecular structure of nickel steel is also advantageous
in this same connection. About 3.50 per cent, of nickel will
give steel approximately six times the life in resistance to fatigue.
The records of shipping show that the great majority of acci-

1 Repeated stresses are stresses put upon a body at intervals and relieved
meanwhile, while alternate stresses are stresses first in compression and then
in tension, such, for instance, as the stresses in a wire that is bent backwards
and forwards, or in a rotating shaft that is not absolutely in alignment.

ALLOY STEELS 399

dents to vessels at sea come from the breaking of the propeller
shafts which is doubtless due to alternate stresses because these
long shafts are put out of alignment by each passing wave, and
now practically all large vessels use hollow nickel steel shafts.
It is the higher elastic limit that is responsible also for the use of
nickel steel in bridges, ordnance, automobile parts, and wire cables,
for we may obtain equal strength with less weight or greater
strength with the same weight. Besides the elastic limit the ulti-
mate tensile strength of nickel steel is increased also by the addition
of nickel. The increase is not so great in this particular, and
consequently the elastic ratio, i.e., the ratio of the elastic limit
to the tensile strength, is increased still more greatly. The elastic
limit of ordinary rolled carbon steel should be at least one-half
of the tensile strength while that of 3.50 per cent, nickel steel
should be at least 60 per cent.

Crystalline Structure. — The crystalline structure of nickel steel
is more minute than ordinary carbon steel, and this is prob-
ably one of the chief causes for the toughness of nickel steel, and
also for the fact that cracks develop in it relatively slowly: in the
yielding of steel to fatigue the damage starts by the opening of a
crack of microscopic proportions through the cleavage planes of
the crystals, and this crack grows and spreads from crystal to
crystal until it is visible to the unaided eye, after which it proceeds
with still greater rapidity. As already stated, this development
is much slower in nickel steel than in carbon steel. Furthermore
if an armor plate is struck by a projectile, it does not crack so
easily, and the cracks do not extend so far if the plate is made of
nickel steel. This fact and the greater strength of nickel steel are
the chief reasons for the 3.25 per cent, of nickel in all modern
armor plate.

Modulus of Elasticity. — The modulus of elasticity of nickel
steel containing not more than 4 or 5 per cent, nickel is about the
same as that of carbon steel, namely, about 29,500,000 pounds
per square inch. With higher contents of nickel, however, and
especially with more than 20 per cent, of nickel, the modulus
of elasticity is lower. This results in the steel being much more
resilient or springy, and this is one of the important reasons why
not more than 4 per cent, of nickel is put into structural steels,
for a bridge built of steel which was resilient, even though strong,
would vibrate so much with the motion of a passing load as to

400 THE METALLURGY OF IRON AND STEEL

be unpleasant, and even unsafe on account of the repeated stresses
set up. The price of nickel, of course, enters into the limita-
tion of the amount used in structural material as well, and it is
found that 3.50 per cent, can be added without great expense
and with beneficial results.

Hardness. — Nickel steel is considerably harder than carbon
steel, though not so much so but that it can be machined without
difficulty. This is taken advantage of in the use of nickel-steel
railroad rails for curves and other locations where the steel soon
wears out. The additional strength of the nickel steel is also an
advantage in this connection and nickel-steel rails have been tried
with success upon the famous horseshoe curve of the Pennsylvania
Railroad and other places. The hardness of nickel steel is also
accompanied by a lower coefficient of friction, and those properties,
together with the additional strength, are taken advantage of in the
use of nickel steel in axles for automobiles, locomotives and rail-
road cars. Equal strength can be obtained in an axle of smaller
size which has, of course, less bearing surface, and therefore still
further reduced friction.

Soundness. — Nickel -steel castings are relatively free from
blowholes and this together with the strength is a reason for the
use of this material' for castings. They also have a lower melting
point and run more easily in the molds.

Expansibility. — The coefficient of expansion of nickel steel is
one of its most astonishing and unusual characteristics, for in dif-
ferent samples it varies all the way from practically zero up to the
ordinary figure for carbon steels.

Colby1 gives the following figures for the average coefficient
of expansion for each 1° C. temperature:

Carbon Steel 0.00001036 (Guillaume)

Carbon Steel (.25 per cent. C.) 0.00001150 (Charpy)

Soft Carbon Steel 0.00001078 (Browne)

5.00 per cent. Nickel Steel 0.00001053 (Guillaume)

Invar. — But the coefficient of, expansion with the ordinary
atmospheric changes of temperature becomes less and less as the
percentage of nickel increases until, when we reach 36 per cent,
of nickel, it is less than any metal or alloy known and amounts
practically to zero. This alloy is patented and sold under the

1 Page 79 of No. 152, page 420.

ALLOY STEELS 401

name of " Invar " and is used for scientific instruments, pendulums
of clocks, steel tape-measures for accurate survey work, etc. In a
paper read before the American Association for the Advancement
of Science in December, 1906, it was stated that tapes made of
invar used experimentally for United States Government survey
work showed a very great increase in accuracy over ordinary steel
tapes, and also in rapidity of use. These tapes varied an infini-
tesimal amount during the first few months, after which they be-
came practically constant in length. The cause of this peculiar
effect of nickel upon the dilation of steel with an increase in tem-
perature is a result of the effect of nickel upon the critical ranges of
the steel, which we shall describe later.

Platinite. — As the amount of nickel increases beyond 36
per cent., there is a slight increase in the coefficient of expan-
sion so that when we reach about 42 per cent, of nickel, the steel
has the same coefficient of ' expansion and contraction with the
atmospheric temperature as has glass. It can therefore be
used for the manufacture of 'armored glass/ i.e., a plate of glass
into which a network of steel wire has been rolled and which
is used for fire-proofing, etc., because even though the glass should
break, it is held together by the steel network. It can also be
used for the electric connections passing through the glass plugs
in the base of incandescent electric lights. Platinum has been
used altogether for this latter purpose, in spite of its very high
cost, because it was the only metal hitherto known that had
the correct coefficient of expansion and contraction, and there-
fore the name of ' platinite ' has been given to this patented nickel-
steel alloy. It is much more valuable for employment in the
electric light industry than for armored glass, because ordinary
steel, although its coefficient of expansion is too large, still suffices
for the majority of armored glass work.

Corrodibility. — Nickel steel corrodes less than carbon steel,
both in the presence of the atmosphere, fresh and salt water, the
ordinary acids, the smoke of locomotives, etc. Moreover, the de-
gree of corrodibility decreases with each increase in the amount of
nickel present. For this reason 30 per cent, nickel boiler tubes
have been used, especially in marine boilers. The great expense of
this material is, however, an obstacle to its common adoption.

Other Properties. — Ordinary nickel steel containing about 3.50
per cent, of nickel has several other properties which distinguish it

402 THE METALLURGY OF IRON AND STEEL

from carbon steel, among which we may mention its higher com-
pressive strength and greater toughness under impact. This
latter makes nickel steel especially resistant to shocks, for it not
only takes a greater blow to bend it but it will bend through more
of an angle before cracking. Nickel steel also has a greater shear-
ing strength which makes it advantageous for rivets, because
smaller rivets may be used and this means smaller holes in the
structural members that are being joined, and consequently a
greater area of these members left to support the strains upon
them. In this connection, however, it should be remembered that
nickel steel does not weld as well as carbon steel, and therefore
greater care is required in upsetting the rivets during the processes
of construction. Nickel segregates very little in iron and it also
has the advantageous property of hindering the other elements
from segregation, so that nickel steel is less liable to these irregu-
larities than carbon steel. In steel over 0.50 per cent, carbon,
nickel has a tendency to make the carbon come out as graphite.

Critical Changes. — Nickel has a very important effect upon the
critical changes of iron and steel. This fact will readily be believed
because it is known that many of the elements added to steel pro-
duce important changes in the critical points. G. B. Waterhouse,1
while investigating under my direction the effect of 3.80 per cent,
of nickel upon iron,, showed that this amount of nickel did .not
make any appreciable difference in the mode of occurrence of the
critical points on cooling, but it did reduce the temperature at
which these critical points came by about 75° C. (167° F.).

As the amount of nickel in the alloys increases the tempera-
tures at which the critical ranges occur become lower and lower
until we reach 25 per cent, of nickel, when the critical ranges occur
below the atmospheric temperature. That is to say, the steel does
not ordinarily cool to the point at which the solid solution is decom-
posed and the beta and alpha allotropic modifications are assumed.

Irreversible Transformations. — The great peculiarity of the
critical changes of the nickel-steel alloys with less than 25 per cent,
of nickel is that they are irreversible. By this we mean that the
change which takes place at one temperature on cooling is not re-
versed on heating at the same temperature, or anywhere near that
temperature. In other words when we cool a nickel steel contain-
ing 20 per cent, nickel the solid solution is not decomposed and the
1 No. 150, page 419.

FIG. 292. — 1.54 PER CENT. CARBON.

ROLLED.

Magnified 225 diameters.
Etched with picric acid.

FIG. 293. — 1.24 PER CENT. CARBON.

ROLLED.

Magnified 225 diameters.
Etched with picric acid.

FIG. 294. — 1.24 PER CENT. CARBON.

SLOWLY COOLED STEEL.

Magnified 225 diameters.

Etched with picric acid.

FIG. 295. — 1.24 PER CENT. CARBON.

COOLED EXTREMELY SLOWLY.

Magnified 265 diameters.

Etched with picric acid.

FIG. 296. — 1.24 PER CENT. CARBON. FIG. 297. — 0.41 PER CENT. CARBON
COOLED EXTREMELY SLOWLY. COOLED EXTREMELY SLOWLY.

Magnified 265 diameters. Magnified 265 diameters

Etched with picric acid. Etched with picric acid.

Nickel steels containing about 3.80 per cent, nickel, 0.12 per cent, silicon, 0.05 per cent,
manganese, 0.014 per cent, sulphur, 0.008 per cent, phosphorus, 0.01 per cent, aluminum.
(G. B. Waterhouse in the Author's Laboratory.)

404 THE METALLURGY OF IRON AND STEEL

alpha allotropic modification is not assumed until we get below 100°
C. (212° F.). But having cooled the steel to that point and decom-
posed the solution, we can now heat it nearly to 600° C. (1112° F.)
before the reverse change takes place and we again form the solid
solution and the gamma allotropic modification. In other words,
it is possible to have a sample of nickel steel between 100° and
600° C. which shall be in either condition we like. With 20 per cent,
of nickel, nearly 1 per cent, of carbon and 1.40 per cent, of man-
ganese, the transformation point on cooling is 188° below zero C.
(306° below zero F.), while the transformation point on heating is
well above the atmospheric temperature. Therefore at atmos-
pheric temperature we may have such a piece of steel in either
condition we like, and a very interesting experiment is formed
by having a bar of this steel one end of which has been cooled more
than 188° below zero C., while the other end has not. The end
that has been cooled will be magnetic and the other end non-
magnetic.

When we have more than 40 per cent, of nickel in our steels, the
critical transformations are reversible like ordinary steels. That
is to say, they occur at nearly the same temperature on heating
as the reverse change does on cooling. It is an interesting fact
that the steels in which the irreversibility of the transformation is
most marked, — that is to say, the steels from 12 to 25 per cent, of
nickel, — have the highest strength and elastic limit; at about 25 to
30 per cent, of nickel, where the irreversible transformation is most
erratic, and beyond that point, the strength is much lower. The
whole interesting question of reversible and irreversible transforma-
tion is discussed very fully in Dumas 's paper No. 153, referred to
at the end of this chapter.

Occurrence of Nickel. — Waterhouse tested his steel contain-
ing 3.80 per cent, of nickel and found that a part of the nickel
was dissolved in the cementite which had the formula (FeNi)3C.
The amount of nickel in the cementite was not, however, as great as
that in the ferrite. That is to say, the steel, as a whole, contained
3.80 per cent, nickel, while the cementite contained only 1.86 per
cent., showing that the nickel dissolves more easily in the ferrite
than it does in the cementite.

Micro-structure. — In Figs. 292 to 297 I show some photomicro-
graphs of nickel steels taken by Waterhouse, and in reference
No. 1522 will be found L. Quillet's researches upon this subject.

ALLOY STEELS 405

MANGANESE STEEL

We owe the discovery of manganese steel to the untiring in-
genuity of Robert A. Hadfield, of Sheffield, England, and its
story will be an inspiration to every inventor, for it resulted in a
material whose properties are not only the opposite of what we
might reasonably have expected on logical grounds, but whose
combination of great hardness and great ductility- was hitherto
unknown and might readily have been believed to be impossible.
Constant study and perseverance must have been the qualities
that led to this revolutionary invention, and it has established
beyond question the principle that because a given amount of any
element produces a given effect upon steels, it does not follow that
a different amount will give the same effect in a different degree.
Indeed a different amount may give an entirely different, and per-
haps an exactly contrary, effect, as is the case of the effect of man-
ganese upon steel.

When the manganese in steel is over 1 per cent, the metal
becomes hard and somewhat brittle, and these qualities increase in
intensity with every increase of manganese until, when we have
4 to 5.50 per cent, the steel can be powdered under the hammer.
But as the manganese is increased from this point, these properties
do not increase and when we reach 7 per cent., an entirely new
set of properties begin to appear. These are well marked at 10"
per cent, of manganese, and reach a maximum at 12 to 15 per cent.

Composition. — Manganese steel usually contains about 12 to 13
per cent, of manganese and 1 .25 to 2 per cent, of carbon. With this
amount of manganese the strength and ductility of the material
reaches its maximum. This high carbon has been necessary hither-
to, because ferro-manganese contains much carbon, which therefore
unavoidably finds its way into the steel. In recent years, however,
manganese metal, relatively free from carbon, has been made by
the Goldschmidt Thermit process and otherwise, and this enables
manganese steel low in carbon to be made, which is now in process
of development and is giving evidence of having new and useful
properties of its own and of being more easily treated and worked.

Treatment. — After manganese steel has been cast into an ingot
or casting, and slowly cooled, it is almost as brittle as glass. But
it is then reheated to a temperature of more than 1000° C. (1832°

406 THE METALLURGY OF IRON AND STEEL

F.) and rapidly cooled by plunging it into water. The tempera-
ture from which it is necessary to quench it can readily be deter-
mined, for it must be so high that when the steel is quenched little
blue flames of hydrogen will appear on the surface of the water.
These are due to the decomposition of water into hydrogen and
oxygen by the intense heat of the steel at the moment of touching
it. The steel which was very brittle before this treatment is after-
wards as ductile as soft carbon steel or wrought iron, while its
tensile strength is about three times as great. Thus the sudden
cooling, which produces brittleness in ordinary steel, produces
ductility in manganese steel. The hardness of the manganese
steel is about the same in the slowly cooled and in the quenched
condition, and is so great that it is not commercially practicable
to machine it and there is no method known of making it softer.

Manganese steel must be heated very slowly and uniformly
lest it crack. It is also very difficult to forge it, and this can only
be accomplished within a narrow range of temperature above a
red heat and by beginning with very light taps of the hammer.
After a little working it becomes so tough that it can be rolled,
although somewhat gingerly. The knowledge as to the proper
method of performing these manipulations is only in a few hands,
and it is only recently that any large amount of forging has been
possible. Even railroad rails for curves have generally been made
by casting the metal in a mold of the proper shape, including the
curvature, and this has, of course, involved a great deal of expense.

Uses. — Manganese steel is used chiefly for the jaws and
wearing parts of rock-crushing machinery and similar apparatus,
for railroad frogs and crossings, for railroad rails on curves, mine
car wheels, and burglar-proof safes. Its life in these classes of serv-
ice is very many times that of all other kinds of steel, because it
is not only extremely hard but is without brittleness. There is a
famous curve on the Boston elevated railroad where carbon-steel
rails were worn out in a very short time and the use of manganese-
steel rails has proved very advantageous and economical. The
use of the steel for burglar-proof safes is also very advantageous,
because there is no known method of making the steel soft enough
to be penetrated by a drill. The uses of manganese steel are limited
chiefly because the metal must ordinarily be formed by casting,
since machining and cutting to shape is practically out of the
question, and forging is difficult. For structural work the advan-

ALLOY STEELS 407

tages of its high combination of strength and ductility are some-
what offset by its low elastic limit, which is only about 35 per cent,
of its ultimate tensile strength. One peculiarity of manganese steel
is that when it yields to tensile stresses it is elongated more uniform-
ly over its whole length than carbon steel, which suffers its greatest
elongation near the point of final rupture where a certain amount
of "necking" takes place. It will be remembered that on page
65 we showed that wrought iron stretched more uniformly over
its whole length than steel; manganese steel has this property in a
still more marked degree even than wrought iron.

Critical Changes. — The hardness of manganese steel is due,
in part, to the hardness of manganese, but still more potently to
the fact that the steel is in the austenitic condition. That is to
say the manganese has reduced the temperature at which the
critical changes occur below that of the atmosphere, and there-
fore manganese steel consists entirely of austenite. It is, of course,
non-magnetic. Whether or not it is capable of irreversible trans-
formation like nickel steel is not known, for its nature and
manufacture is kept as secret as possible by those who know it,
for trade reasons.

CHROME STEEL

Chromium has the effect of making the critical changes of steel
take place much more slowly. Therefore chromium steels are
capable of greater hardness, because rapid cooling is able more
completely to prevent the decomposition of austenite. They
contain usually 1 to 2 per cent, of chromium and from 0.80 to 2 per
cent, of carbon and are used in the hardened state. They are par-
ticularly adapted for making armor-piercing projectiles, on account
of their hardness and also their very high elastic limit. They are
also used for armor plate for the same reason, for parts of crushing
machinery, and for very hard steel plate. This latter is not or-
dinarily used by itself, but is made into 3-ply and 5-ply plate for
plows and burglar-proof safes, as described on page 195, for if the
hardened chrome steel were used alone, its brittleness would
cause it to be shattered.

Armor Plate. — Krupp armor plate contains about 3.25 per
cent, of nickel, 1.50 per cent, of chromium and 0.25 per cent, of
carbon. Its further manufacture is described on page 228.

408 THE METALLURGY OF IRON AND STEEL

Automobile Steels. — Chromium up to about 2 per cent, is also
used for automobile steels where hardness is required, as for in-
stance in gears and other parts requiring great hardness or great
strength. For the latter purpose it is more common to use a nickel-
chrome steel, and this is often subjected to a double heat treatment
or a simple oil tempering. This treatment has the effect of greatly
increasing its strength and elastic limit, so that steels of this char-
acter will have properties similar to those shown in Table XXX.
There cannot be said to be any uniform composition or method of
treatment for automobile steels, and my inquiries among American
manufacturers seem to indicate that there is not very much of the
high-priced alloy steels used in American cars, except for the
frames which, as already stated, are frequently made of 3.50 per
cent, nickel steel. The next most important use is probably for
gears of chrome steel of nickel-chrome steel, first case-hardened by
cementing with carbon to a depth of an eighth of an inch or so, and
then heat-treated. The composition and treatment of alloy steels
used in French automobiles is shown more fully in Quillet's article,
No. 1514.

SELF-HARDENING AND HIGH-SPEED TOOL STEELS

Self -hardening Steel. — Self -hardening steel is steel which is
hard without being subjected to any heat treatment or other
process for making it so. It is steel which cannot be made soft, or
annealed, by any process known at present. It is often called ' air-
hardening steel ' because when it cools in the air from a red heat
or above, it is not soft like ordinary steel, but is hard and capable
of cutting other metals. Manganese steel is a typical self-harden-
ing steel and so obviously is any steel which is in the austenitic
condition at atmospheric temperatures, — that is to say, whose
critical temperature is below the atmospheric temperature. All
the self-hardening steels are therefore non-magnetic.

Mushet Steel. — The name 'self-hardening' steel was first
applied to an alloy steel invented by Robert Mushet and which
owed its self-hardening properties to the simultaneous presence
of both tungsten and manganese. The analyses varied greatly
but were probably limited to between 4 and 12 per cent, of tungsten
with 2 to 4 per cent, of manganese and 1.50 to 2.50 per cent, of
carbon. A typical sample and one having excellent qualities

ALLOY STEELS 409

contained about 9 per cent, of tungsten, 2.50 per cent, of manganese,
and 1.85 per cent, of carbon. This steel is incapable of being
made soft by any known process and is non-magnetic. It is
one of those curious phenomena met with in the metallurgy of
steel, where a combination of two elements will produce a result
entirely different from anything that might be predicated: Tung-
sten does not lower the temperature of the critical change in steel
and 2.50 per cent, of manganese has but a slight effect in that direc-
tion. Nevertheless the combination of these two reduces the
critical point below the atmospheric temperature.

Mushet steel has been, for many years, a famous tool steel be-
cause of its capacity for performing a large amount of heavy cutting
work. It is very hard and durable and will retain its cutting edge
for a long time and under very severe service. It, or its equivalent,
is used very largely at the present time for very heavy, or deep, cuts
and especially for cutting extra -hard metal, such as the roughing
cuts on armor plate and other hard alloys. The cutting speed of
which it is capable is not much, if any, greater than ordinary
carbon tool steel, but the economy of its use is due to the fact
that it will take such deep cuts and last so long without regrinding.

Other Self -hardening Tool Steels. — The 2.50 per cent, of man-
ganese in Mushet steel can be replaced by 1 or 2 per cent, of
chromium and again produce a self-hardening tool steel which
has the advantageous properties of Mushet steel. This result is
even more astonishing than the self-hardening properties of
Mushet steel, because chromium has a tendency to raise the
temperature at which the critical change comes, and yet the
addition of 1 or 2 per cent, of chromium to a tungsten steel, which
was not previously self-hardening and whose critical temperature
was about 600° C. (1112° F.), reduces the critical temperature
to below the atmosphere. We may also replace the 9 per cent,
of tungsten in Mushet steel with 4 to 6 per cent, of molybdenum,
and it is stated that this latter change produces a self-hardening
tool steel which is a little tougher than Mushet steel.

Taylor and White. — Frederick W. Taylor and Maunsel White
of the Bethlehem Steel Works experimented for a long period
of time with the self-hardening steels existing in 1899 and previ-
ously, for the purpose of improving them by heat treatment. The
full record of these and other researches were presented by Taylor
in his presidential address to the American Society of Mechanical

410 THE METALLURGY OF IRON AND STEEL

Engineers in 1906, and form one of the most interesting records
of the kind ever presented to the world. The result of these
experiments was to produce a wholly new kind of steel which
has fairly revolutionized the machine shop industry of the world.
Taylor and White found that by applying a new method of heat
treatment to the self-hardening tool steels, they gave them much
greater toughness at a red heat, so that they could do their cut-
ting work at a speed so fast that the point of the tool would be-
come red hot with the heat of friction and the great chips of steel,
which were thick and heavy on account of the depth of cut which
could be made, were raised to a temperature of nearly 300° C. (572°
F.). In other words, the steel tool never lost its temper nor its
toughness at a red heat. The heat treatment which Taylor and
White employed consisted in raising the steel almost to the melting
point and then plunging it in a bath of molten lead at a temperature
between 700° and 850° C. (1300° and 1550° F.), where it was kept
until it was of the same temperature as the bath, and then re-
moved and cooled by plunging into oil. They usually followed
this cooling by reheating the steel to a temperature between 370°
and 670° C. (700° and 1240° F.).

The first public exhibition of the Taylor and White steel was
made at the Paris Exposition in 1900 and created first incredulity
and then astonishment. The amount of work performed by a tool
was unheard of, as also was the speed at which the tool was made
to travel through the metal it was cutting, and the length of time
that elapsed before it was necessary to regrind it. It was realized
that a new epoch in the tool-steel industry had been inaugurated.
The fact that the method of heat treatment used by Taylor and
White was subsequently shown to be unnecessary and that there-
fore the manufacture of high-speed steel tools, having qualities
like theirs, was begun by everybody, in no way lessens the credit
due them for teaching the world how to produce a new kind of
metal and effecting a tremendous decrease in the price of machine
work.

High-speed Steels. — The name 'high-speed steels' was not
given by Taylor and White to their product, but has subsequently
been adopted for all steels capable of these rapid-cutting speeds
which theirs had. Soon after they had shown the world what could
be done, it was found that the only heat treatment necessary to
give the steel its peculiar hardness and toughness at a low red heat

ALLOY STEELS 411

was to raise it to a temperature very near its melting point and
then cool it witn moderate rapidity, as for example by holding it
in a blast of cold air until it was below a red heat. The essential
feature seems to be that the steel shall attain a high temperature
which, in many cases, is so great that melted oxide drops from it,
and it is almost ready to scintillate, — that is to say, it has almost
crossed the line Aa in Fig. 246, page 314. After this heating it
sometimes suffices to merely allow the steel to cool in air, but in
this case its hardness is not as great, and cooling in a stream of air
is more usual.

Composition. — It was also soon found that the composition
of the self-hardening steels was not the best one for high-speed
steels. Tungsten was the element which gave the steel the prop-
erties of hardness and toughness at a red heat. After the peculiar
heat treatment had been learned and the presence of manganese
or chromium in addition to the tungsten was shown to be unneces-
sary, it was found that more durable qualities could be obtained by
increasing the percentage of tungsten, and steels have been put
upon the market with even as high as 24 per cent, of this element.
At the same time the carbon was greatly reduced and at the
present usually varies from 0.40 to 0.80 per cent, in the best high-
speed steels.1

It was also found that molybdenum could replace tungsten
as far as producing high-speed qualities was concerned, and man}7
believe that the molybdenum steels are more tough and durable
than the tungsten steels. Some difficulty was met with at first in
working the molybdenum steels as they proved to be seamy ard
liable to cracks, but this was overcome with experience. The
molybdenum steels do not require so high a temperature for heating
previous to cooling down in the air blast as the tungsten steels.

1 It is commonly stated in the trade that tungsten will take the place of
carbon in producing hardness, but this is not true. It is far more correct to
say that tungsten will assist carbon in producing hardness and therefore with
high tungsten steels we may have lower carbon. This distinction may appear
merely academic, but it is well worth recognition by those who expect to make
a study of these steels. No amount of tungsten or any other element will make
steel hard in the absence of carbon, or even when the carbon is low. The tung-
sten produces hardness by its effect upon the condition of the carbon, — that is
by helping to retain the carbon in its solid solution, — and not by any effect of
its own. It is for this reason that a lesser amount of carbon will produce
hardness in the presence of tungsten or some similar agent.

412 THE METALLURGY OF IRON AND STEEL

According to the researches of Carpenter,1 the molybdenum
steels should be heated between 1000° arid 1100° C. (1832° and
2012° F.), while the tungsten steels must be heated in the neigh-
borhood of 1200° C. (2192° F.). Furthermore it only takes about
one-half as much molybdenum as tungsten to produce the desired
result, which means that there is more iron in the molybdenum
steels than in the tungsten steels. In other respects the analysis
of the molybdenum and the tungsten steels is about the same,
containing usually 0.60 to 0.80 per cent, of carbon and anywhere
from zero chromium up to sometimes as much as 4 per cent.
Indeed chromium is sometimes recommended as high as 6 per
cent, and over, because it gives hardness, but it also reduces
toughness. The more durable qualities of the molybdenum steels
than the tungsten steels are believed to be due to the .larger
amount of iron in them and the lower temperature necessary for
tempering them.

In some cases molybdenum and tungsten have been used to-
gether in high-speed steels. In fact there are at the present time
scores of brands and analyses of high-speed steels on the market,
made both in America, England, and Europe and the art of manu-
facturing them is constantly advancing so that no very general
results can be quoted. In America the most advantageous per-
centages of molybdenum, 6 to 15 per cent., are patented and al-
though at one time a great many tons of this kind of high-speed
steel were manufactured and gave very good results (containing
usually about 10 per cent, of molybdenum) it can now be made
only by one company, so that tungsten steels are more common.

Forging. — High-speed steels can only be forged at tempera-
tures above a bright red heat, that is to say from 1050° to 1150° C.
(1922° to 2100° F.) and higher.

Annealing. — The heating and annealing of high-speed steels
requires a great deal of care. They must be heated up to the an-
nealing temperature (say about 800° C. = 1472° F.) with extreme
slowness, and cooled down in lime or ashes or in the furnace. They
are then soft enough to be machined easily, but not as soft as car-
bon steel.

Tempering. — The reason tungsten and molybdenum produce
in steel the high-speed quality of not losing its temper at a red heat
is because of their effect upon the critical temperatures. Their

1 See page 460 of No. 1517, page 421.

ALLOY STEELS 413

effect seems to be to prolong the critical range of temperatures of
the steel on slow cooling; that is to say, instead of the critical range
coming in the neighborhood of 690° C. (1285° F.), as with the
carbon steels, it begins, when the cooling is slow, at about 700° C.
and spreads out all the way down to 300° or 400° C. (572° or 752° F.),
or even lower. Molybdenum is more active than tungsten in
causing this prolongation of the critical range. But if the steels
are first heated to a very high temperature (1000° to 1100° C. for
molybdenum steel and 1200° C. for tungsten = 1832° to 2012° and
2192° F. respectively) and then cooled moderately fast, this treat-
ment suffices to prevent the critical changes altogether and pre-
serves the steel in the austenitic condition. We know that this
austenitic condition is one of hardness and toughness and its
peculiarity under these circumstances is that it is not transformed
into the pearlitic condition until the steels are heated to 650° C.
(1202° F.) or thereabouts.

Magnetic Steels. — Strictly speaking, the steels used for per-
manent magnets are not high-speed steels, because, of course, they
are never used for cutting work, but as their composition is so
similar it seems well to introduce them here. A permanent magnet
is made by putting a piece of hardened steel in a magnetic field for
a few moments, as, for instance, by winding an insulated wire
around it and passing an electric current through. The magnetic
force which it obtains in this way will remain in it for a very long
period of time. It is found that a steel containing about 4 to 5
per cent, of tungsten and 0.50 to 0.70 per cent, of carbon, if heated
to a red heat (say 800° C. = 1472° F.) and quenched in water,
will retain its magnetism better than ordinary hardened carbon
steel. Sometimes about 0.50 per cent, of chromium is added to
the alloy also.

SILICON STEELS

The genius of Hadfield has also given us a silicon steel alloy
of importance and usefulness. In 1888 1 Hadfield investigated
many alloys of iron and silicon and although these showed some
remarkable properties, especially in the matter of tempering and
cutting qualities, they did not lead to any alloy steels that were
produced in abundance. At a later period, however,2 Hadfield

1 See No. 157, page, 420. 2 U. S. Patent 12,691. September 3, 1907.

414 THE METALLURGY OF IRON AND STEEL

developed a silicon steel which, after a double-heat treatment,
showed some truly remarkable magnetic qualities. It had always
been believed that pure iron had the highest magnetic force and
permeability of any known substance or of any combination that
could be produced, but Hadfield's new silicon steel, whose com-
position and treatment is shown below, had not only a greater
magnetic permeability than the purest iron but also, charac-
teristic of silicon steels, it had a high electrical resistance. Its
hysteresis is, of course, low, this property always accompanying a
high permeability. It is therefore a very valuable material for use
in electromagnets, and in electrical generating machinery is the
most efficient material known. Its high magnetic permeability
gives high motor efficiency, and in addition its high electrical
resistance reduces the "eddy currents" which are a source of
waste.

Composition and Treatment. — Hadfield's patent covers silicon
from 1 to 5 per cent., but the alloy which he recommends con-
tains 2.75 per cent, of silicon and the smallest possible amounts of
carbon, manganese, and other impurities. Before the steel is ready
for use, it is subjected to a double-heat treatment by first heating it
to between 900° and 1100° C. (1652° and 2012° F.) and cooling
quickly, and then reheating to between 700° and 850° C. (1292° and
1562° F.) and allowing to cool very slowly. In some cases his
second cooling has been extended over several days. He finds the
best results by heating first to 1070° C. (1958° F.) and cooling
quickly to atmospheric temperature and then heating to 750° C.
(1382° F.) and cooling slowly, after which he sometimes again
reheats to 800° C. (1472° F.) and then cools slowly.

VANADIUM STEELS

Vanadium steels are still in their infancy, for although the
element has been used in steel metallurgy for many years, it is only
recently that any important development work has been carried
on. The results, however, are so very encouraging that we may
expect great extension of their employment and important progress
in their metallurgy. With the single exception of carbon no ele-
ment has such a powerful effect upon steel as vanadium, for it is
only necessary to use from 0.10 to 0.15 per cent, in order to ob-

ALLOY STEELS 415

tain very powerful results, while 0.30 per cent, should probably
not be exceeded so far as present knowledge indicates. In addi-
tion to acting as a very great strengthener of steel, especially
against dynamic strains, vanadium also serves as a scavenger in
getting rid of oxygen and possibly nitrogen. It is also said to
decrease segregation, which we may readily believe, as most of
the elements which quiet the steel have this effect. Vanadium
steel also has the advantage of welding readily.

The effect of vanadium is shown very well by Tables XXX and
XXXII,1 and it will be seen how efficiently this material resists al-
ternating stresses and the other forces producing fatigue. It is to be
observed that vanadium is especially advantageous when added to
nickel and to chromium steels, greatly increasing their strength,
toughness, and temper. In this connection it is important to note
that the nickel-vanadium steels have better quality when the car-
bon is low, especially if they are to be subjected to heat treat-
ment.

It would seem that vanadium should have especial advantage
in high-speed tool steels, but, strange to say, the results have not
always been favorable. Nevertheless experiments in this direction
continue.

Manufacture. — Pure vanadium has a very high melting point
and the element is therefore added to steel in the form of ferro-
vanadium. This alloy should be added to the bath about two or
three minutes after the manganese used for recarburizing; because
if the vanadium is added before the manganese, it is wasted by
oxidation. It therefore should be added always under reducing
conditions. Beyond this there is no special difficulty in manu-
facture, as the amount of alloy used is so small and as it distributes
itself readily through the metal and does not segregate.

Treatment. — Vanadium steels must be heated gradually but
are forged without difficulty although they must be worked a little
tenderly at first. Like all steels, they must not be forged at too
high a temperature, and like all alloy steels, they become even more
brittle when forged below a black heat than carbon steel does.

Uses. — It is advantageous to use vanadium steel for prac-
tically every purpose that will stand the additional price, which is
about the same as that of 3.50 per cent, nickel steel. It is, of course,
especially useful for all purposes where strength and lightness are
1 For which I am indebted to the American Vanadium Co.

to 00 I-H • • <N • • to • • • O
CO C^ CO • • ^ ^ • • • to

spuaq 3
-uv *t{3i

S-.t.IJ-

jo uon

to • o

,- -CO •

SAvojq

spunod -CJJ

'^otjdim
raninpuaj

ijo aad 'Baa
jo uop

-OBJ^UOQ

•^o aad

•m -bs

•m -
•eqi '

SI> QO <M • i— i O5 lO CO GO (N to 1> T-H O5 CO rt< CO Oi O5 O
rJH iO CO • CO 1C CO 1C 1C CO CO O »O CO »O ^ CO CO CO CO

Oi TJH CO • <N to r-t ^ O5
rf CO CO • CO CO l> (M CO

QO t» 1C X O X

CO CO (N rH CO O b»

O5 O5 O5O5 O5»O COt^Oi X 1>

CO OS Is* Is" CO Is* C^ to t*** ^> X Is* C^ C^ 1>» C^ ^ to to CO

tOcOOOI> O5 X O5 C^ O? CO CO "* i— i CO X CO i— i CO to O

T-H i-H T-( i— I i— t C^ (M rH rH <M 1— I

So
00 :
co_o

*~^ 00 Oi t^ . CO ^ ^H
CO CO '

o o o^os o_o t^ co_

O"^"l^'rH>(N"i-r'^' OS

(N (M i— i OJ

t severe crushing, bending, bulging, and expanding tests. 2 Equal to best Swedish mild steel. a Easily stands
* Bends, etc., as forging steel. 6 .75" round bends double cold D = 2T. 6 y%" sq. and 1" round bars bend close double.
up; %» rod bends double. After tempering a spring of it coefficient of safe working load = 40,000 with excellent
cient of safe working load = 20,000. 8 Bends close double cold. • Nash torsion test; length 6", diameter . 75".

he mo
= 2T.
t tight
the coe

1 Fulfils
plate bend,
7 Twists in fl
carbon steel

:|||

I 111

3 o3_c! c3

" 11 10

I I'll

^^ CO
C<N -^ 00 1-1
g . . . .

*jHl

1111

•Sill

|8^.SS

^1 1 1

• s

I'll

S S
'Sea

3 g^

C'n

k> ?

OHO

|!
PS

llllll

S 08 -g

418

THE METALLURGY OF IRON AND STEEL

•S .2

Q «3

i :,/

~OQ Q

M g

s 1

|| n6

o o

-f^

£ §b

^11 §1"

^ ^ i s ^ ""

•< **

§ ® S

c^ c^

03
^ •*»

E 2

>£ H

VANADIt

ng of the m<

-|{

^H 00 t1^ ^5 00 • ^^ *H *O t*^*

2 (N r-T
rH rH

•

H] 'S

< -°
00

^in

S I

— »H

JH ^

H «S
fe 3

o g

>• c8 ft

O O N.O v55 v<3 <N X 1C O OS C
Is* x to^ o^ ox • T- ( • O co C
tO CO CO CO ^ ^ CO 1s* ^
rf 0- CO CO CO -• N- N

co os cc

>mparati\

OQ |

a 1

1 1

H £

(N^S" _

O OxOKOxOC^ C^ ^o CO C

H} 3

o|"§| •

oT t>T *& co to ^

v
'
*

<•; t»C

"^ *

T^ x '-

Q S

S °

-§

o a

*H g^

^ ^ o^o^c^^ 2 ^ § £ g

3 J

o"£||

§ - § |l « ^ S ^

[

§D

fe §

^ §

o s

• aj ^ ' i

n . . a> cp . 3

s 1-

1—1 ' 1 * 8 H * 'S

1 1

1 • j ^ § £

s s

•a 1*0^^

PtH |

B § s"0 s
.s -s 8,11 1

h-J S

— "~

^^ ill i

X *

fH °° «S ~ "S ^

"^ >— b 73 "ooSrtfl
to of (N^M C =SB<0.£

£ S c 'SS ^ &.§ ^ ^

1 ,

^|.d °g^ g) lyfgl
.s. o---^-< 3S>^>
oJcr "43 -g JS g^^g^
^s3® ^ceS "S 3 ^7=! c^ K

I 11I?!J

ALLOY STEELS 419

desired, such as springs, axles, frames and other parts of railroad
rolling stock and automobiles. It would also seem to have special
advantages for shafts and other rotating parts. It also does very
well for case-hardened gears, on account of the good combination
of hardness and toughness and has been used to some extent for
steel castings where resistance to vibration has been demanded.

Vanadium has also been added to cast iron 011 account of its
ability to remove oxygen and possibly nitrogen (if any is ever
present), and because it makes the metal more fluid and tougher.
It is said also to make chilled cast-iron rolls more durable and more
ductile.

TITANIUM STEELS

Titanium has not been used much for making alloy steels, but
it has one advantage which has caused its use to be recommended,
namely, its high affinity for nitrogen. There is no element whose
addition to iron or steel would be expected to have a greater effect
in ridding it of nitrogen, but beyond these logical arguments the
experiments so far made are not absolutely conclusive, although
they are sufficient to cause a great deal of interest to be taken in
this subject.

titanium has often been found in pig iron in very small quan-
tities, but its presence in the blast furnace is usually objected to,
because it causes the slags to be so infusible and sticky. Ferro-
titanium is now being made in electric furnaces where the tem-
perature available overcomes the difficulty of infusible slags. This
alloy is added to iron castings in amounts such that the proportion
of titanium in the metal will not exceed 0.10 to OJ^per cent.
This results in an increase in strength of the iron of nearly 20 to 40
per cent, and also of greater hardness and durability against wear.

REFERENCES ON ALLOY STEELS

1 50. G. B. Waterhouse. " The Influence of Nickel and Carbon on

Iron.;' Journal of the Iron and Steel Institute, vol. ii,
1905. pages 376-407. This includes a bibliography of a
part of the subject.

151. G. B. Waterhouse. "The Burning, Overheating, and Re-

storing of Nickel Steel." Proceedings of the American

420 THE METALLURGY OF IRON AND STEEL

Society for Testing Materials, vol. vi, 1906, pages 247-
258.

152. A. L. Colby. " A Comparison of Certain Physical Properties

of Nickel Steel and Carbon Steel." Published by the
Bethlehem Steel Company, 1903. A very complete com-
pilation.

153. L. Dumas. "The Reversible and Irreversible Transforma-

tions of Nickel Steel." Journal of the Iron and Steel In-
stitute, No. 11, 1905, pages 255-300.

154. D. H. Browne. "Nickel Steel; A Synopsis of Experiment

and Opinion." Transactions of the American Institute of
Mining Engineers, vol. xxix, 1899, pages 569-648. Very
complete, with a good bibliography.

155. R. A. Hadfield. "Alloys of Iron and Manganese." Pro-

ceedings of the Institution of Civil Engineers (England),
. 1888.

156. R. A. Hadfield. "Manganese Steel." Journal of the Iron

and Steel Institute, 1888.

157. R. A. Hadfield. " Alloys of Iron and Silicon." Ibid., 1889.

158. R. A. Hadfield. "Alloys of Iron and Aluminum." Ibid.,

1890.

159. R. A. Hadfield. "Alloys of Iron and Chromium." Ibid.,

1892.

1510. R. A. Hadfield. "Alloys of Iron and Nickel." Proceedings

of the Institution of Civil Engineers (England), 1899.

1511. R. A. Hadfield. "Alloys of Iron and Tungsten." Journal

of the Iron and Steel Institute, No. 11, 1903.

1512. Henry M. Howe. "Manganese Steel." Proceedings of the

American Society of Mechanical Engineers, 1891.

1513. C. E. Guillaume. His articles will be found listed in some

of the cross-references cited.

1514. L. Guillet. "Steel Used for Motor Car Construction in

France." Journal of the Iron and Steel Institute, No. 11,
1905, pages 166-203.

1515. L. Guillet. "The Use of Vanadium in Metallurgy." Same

journal, pages 118-165.

For Guillet's other memoirs on alloy steels consult the
Revue de Metallurgie, and the index of the Journal of
the Iron and Steel Institute from 1904 to date.

ALLOY STEELS 421

1516. J. M. Gledhill. "The Development and Use of High-speed

Tool Steel." Journal of the Iron and Steel Institute, No.
11, 1904, pages 127-182.

1517. H. C. H. Carpenter. " High-speed Tool Steels." Same jour-

nal, No. 1, 1905, pages 433-473, and No. Ill, 1906,
pages 377-396.

1518. J. T. Nicolson. "Experiments with Rapid-cutting Steel

Tools." Manchester Municipal School of Technology,
Manchester, England, 1903.

1519. J. Kent Smith. On vanadium steels; see Transactions of

the American Institute of Mining Engineers, vol. xxxvii,
1907, pages 727-732. Also The Iron Trade Review,
November 7, 1907, page 729. Vol. xli.

1520. (For Vanadium steels see also the index of the Journal of the

Iron and Steel Institute, especially 1905 to date.

1521. L. Guillet. " Quaternary "Steels." Journal of the Iron and

Steel Institute, No. 11, 1906, pages 1-141.

1522. For all the alloy steels one should look in the four volumes

of Revue de Metallurgie, for which see No. 10, at end of
Chapter I.

1523. De Mozay. "Etude de la Durete dans les Aciers a Outil de

Tour." Revue de Metallurgie, vol. iv, 1907, pages 885-
900.

1524. Leon Guillet. "Nouvelles Recherches sur les Aciers au

Vanadium Ternaires et Quarternaires." Revue de Metal-
lurgie, vol. iv, 1907, pages 775-783.

1525. P. Nicolardot. "Le Vanadium." Paris. About 1906.

1526. A.J.Rossi. "The Metallurgy of Titanium." Transactions

of the American Institute of Mining Engineers, vol. xxxiii,
1903, pages 179-197, with cross-references.

1527. A. J. Rossi. " Titanif erous Iron Ores." Engineering and

Mining Journal, vol. Ixxviii, 1904, pages 501 and 502.

1528. J. Hoerhager. "Ueber titanhaltiges Holzkohlen-Roheisen

von Turrach in Obersteiermark." Oesterreichische Zeit-
schrift fur Berg- und Huettenwesen, vol. Hi, 1904, pages
571-577.

XVI
THE CORROSION OF IRON AND STEEL

IRON offers so little resistance to rusting or corrosion that there
are almost no circumstances of service in which it can safely be
placed without some means of protection from the elements. Cer-
tain parts of machinery, in situations where rusting is not very
rapid and where the metal will receive constant cleaning and over-
sight, are used without any protective coating, but structural
work, in and outside of buildings, tin roofs,1 wire fences, pipes, and
other metallic structures, all require to be protected by some
coating, such as paint, galvanizing,2 tinning, nickel plating,
oxidizing, etc. Boiler tubes and the inside of boilers, tanks, and
pipes it is usually impossible to protect by paint, galvanizing,
etc., and there is an annual loss of doubtless many thousands of
tons of iron and steel from the decay of these classes of articles
alone.

THE CAUSE AND OPERATION OF CORROSION

The brown powder with which we are all far too familiar under
the name of rust is a hydroxide of iron — ferric hydroxide, FeOsHs.
It is formed wherever iron is exposed to the action of water and air.
Neither dry air nor water free from oxygen has any effect upon it
alone, but as air is always moist and ordinary water contains some
oxygen in solution, the conditions for corrosion practically always
prevail against the iron in service. The alternate attack of oxygen
and water within a brief period is far more destructive than the
attack of either one.. For example, heavy rains, the splashing of
water intermittently upon piers and columns, the rise and fall of
the tide, etc., corrode the metal much faster than exposure all the

1 Tinplate consists of a thin sheet of steel or wrought iron covered with
metallic tin.

2 I.e., coating with metallic zinc.

422

THE CORROSION OF IRON AND STEEL 423

time to either damp air or oxygenated water alone. Acids, while
not essential to corrosion, greatly hasten its action, so that the
presence of carbonic acid in the air, sulphurous, sulphuric, and
hydrochloric acids in the smoke from locomotives and other fires,
all greatly increase the speed of rusting. It seems evident also
that at least some weak electrolysis is essential for any corrosion to
occur. I shall show later that this electrolytic action will always
take place where iron and water are in contact with each other, but
it is very small in amount under these conditions. Where a greater
electric force exists, as for example where pipe lines run close to
trolley tracks and receive a leak of electricity so as to carry a part
of the current, the electrolytic action is increased and corrosion
much hastened. As the use of electricity to-day is far more gen-
eral than ever before and the amount of coal burned and conse-
quently the amount of corrosive gases in our atmosphere is much
larger, it follows that the question of corrosion is of constantly in-
creasing importance. Affairs have indeed reached such a state
that the subject is now occupying a great deal of attention from
metallurgists and engineers in the hope of getting better means of
protection. Pipe lines embedded in the earth, which are alter-
nately wet and dry, supports in tunnels, subways, and other damp
places, wire fences and tin roofs are all suffering severely from
decay.

Theories of Corrosion. — In an able research,1 Allerton S.
Cushman has studied the current theories that have been advanced
to account for the corrosion of steel. He shows that the theory
that carbonic acid, or some other acid, is necessary is wrong, and
that corrosion will take place even in weak alkaline solutions. He
also shows that hydrogen peroxide is not the agency producing the
rust. By means of a series of extremely careful and accurate ex-
periments he demonstrated the probability, if not, indeed, the cer-
tainty, that the two factors without which the corrosion of iron is
impossible are electrolysis and the presence of hydrogen in the
electrolyzed or 'ionic' condition. In brief, it is the ions of
hydrogen which first cause the metal to pass in solution:

Fe + (4H + 20)2=Fe02H2 + H2

Ions of hydrogen are electropositive to iron, and when the reaction
takes place, they transfer their electric charge to the metal. This

1 See No. 160, page 436. 2 4H+20 =2HaO in the electrolyzed condition.

424 THE METALLURGY OF IRON AND STEEL

is, of course, an electrolytic action and the hydrogen which takes
part in it is converted from the electrolyzed (or ionic) condition to
the atomic (or gaseous) condition. Therefore it is evident that the
action must take place in the presence of oxygen, or some other
oxidizing agent that will complete the electrolysis, else the forma-
tion of ferrous hydroxide will soon cease. This explains why the
presence of oxygen so greatly increases the corrosion of iron,
although it is not the oxygen itself which, after all, is the cause of
the primary attack.

Rust. — Ferrous hydroxide, FeC^H^, is soluble in water and its
formation and solution is the first step in the production of rust.
Because of its solubility, however, it does not ordinarily make its
presence known until a further reaction occurs : —

2Fe02H2 + H20 + O =2Fe03H3.

The rust, FeOsHs, so formed precipitates from solution.

Unfortunately traces of ionic hydrogen are always present even
in the purest water, and larger amounts in ordinary water. Sub-
stances which increased the hydrogen ions, such as oxides, pro-
mote the rusting, and the same may be said of anything which
increases the electrolytic action, while substances which restrain the
formation of hydrogen ions will decrease corrosion. Indeed, dip-
ping a piece of bright iron into a solution of potassium bichromate
and then wiping it off will put it into an apparently oxidized con-
dition in which it will resist corrosion for days or even weeks.
Cushman, therefore, suggests dissolving a small amount of chro-
mic acid or potassium bichromate in boiler waters in order to
restrain corrosion of the metal.

Segregation. — Evidently anything that increases the electro-
lytic activity will increase the attack by hydrogen and therefore
the formation of rust. Unfortunately even the purest piece of iron
will show differences of electric potential at different parts and
therefore produce an electrolytic effect. When the metal is impure
or is badly segregated, these differences in potential will be quite
large, and when several pieces of steel are joined together, as in a
bridge or other structure, the difference in potential between the
different parts may be great. It is probable that each of the dif-
ferent microscopic constituents of iron and steel has a different
electric potential and therefore either assists or retards the progress
of rusting. Also scale, the slag in wrought iron, etc.

THE CORROSION OF IRON AND STEEL

425

Self-protection from Corrosion. — It is generally believed that
certain constituents in iron and steel assist in protecting underly-
ing layers of the metal from attack. For example, the graphite
which forms such a large proportion (say 10% or more) of the
volume of cast iron, slag which forms nearly 4% of the volume of
wrought iron, and cementite
in steel, all corrode less
rapidly than the pure metal,
and it is probable that they
are beneficial in protecting
it. It must not be for-
gotten, however, that they
likewise cause a difference
in potential and to that ex-
tent probably tend to hasten
corrosion. Their net effect
can be learned only by ex-
periment. Scale or foreign
substances on the surface of
the metal would also pro-
duce large differences in
potential.

Relative Corrosion of Iron
and Steel. — It is generally
believed that cast iron cor-
rodes less rapidly than either
wrought iron or steel and,
for this reason, cast-iron pipe
is greatly preferred for city
water mains and like uses
where great strength is not
required. The belief is a

reasonable one, since it might well be expected that the pres-
ence of graphite would be a protection, but it is only right to
remark that the theory rests upon no experimental evidence and
that there are other circumstances connected with cast-iron pipe
which may not have given rise to the belief but which are not in-
herent in the material itself. These conditions are (1) the common
practice of dipping cast-iron pipe in asphaltum, paint, or some sim-
ilar protecting material, before it is sold and thus putting it in a

FIG. 298. — CORRODED STEEL PLATE.

426 THE METALLURGY OF IRON AND STEEL

condition to be in service for a long time before the metal itself is
subjected to the corrosive influences. (2) When the iron is cast
in the sand there seems to take place some union between the
metal and the inside surface of the mold, whereby a very resisting
siliceous coating or 'skin' is formed on the casting. Some be-
lieve that after this skin is worn away, cast iron corrodes as rapidly
as wrought iron or steel. (3) Cast-iron pipe is thicker than the
same diameter of wrought iron and steel pipe, because of the
difficulty of making castings with thin sections of metal ; therefore
such pipe would remain in service much longer than steel or
wrought iron, even though the rate of corrosion were the same.
I have cited the above facts not to argue against the belief in the
slow corrosion of cast iron, but merely to explain the situation as
it exists, for we have as yet no scientific data upon which to form
an opinion either way.

Wrought Iron vs. Steel. — There is also a very prevalent and
widespread conviction that steel corrodes much more rapidly than
wrought iron. This opinion too rests upon no very exact experi-
mental evidence, although there are not incidents lacking which
make in favor of it, while others make in the opposite direction, but
it seems to be based principally upon the fact that corrosion is very
much more rapid to-day, when steel is the world's great metal,
than it was some years ago when wrought iron was chiefly used!
But I have already shown that conditions to-day are more con-
ducive to rapid corrosion than they were at any previous time.1

Opposed to this popular belief is the result of a great many
scientific tests which have shown, in almost every case, that the
difference in the speed of corrosion between wrought iron and steel
is very small, although favorable to wrought iron in the case of
sea water and alkaline water and to steel in the case of acids and

1 Indeed I hardly think that this argument ought to have anything like the
weight that is popularly given to it. It is not at all uncommon to hear the state-
ment that wrought iron made 20 years ago is still in service in places and con-
ditions alongside of steel which has been replaced several times. But I am
informed by a testing engineer of much experience and reputation that he has
a steel fence upon his property which has been in service many years and has
outlasted several other steel fences under similar conditions, but formed of steel
made recently. From this he argues that the steel made to-day is more sub-
ject to corrosion than the steel made many years ago. If this is so, we are
permitted to ask: Is the wrought iron made to-day more subject to corrosion
than the wrought iron made several years ago?

THE CORROSION OF IRON AND STEEL

427

acidulated water. As has been pointed out, however, these
scientific tests are not altogether reliable as a basis for commer-
cial comparison, because they have not usually been carried to
the point where either one material or the other becomes unfit for
service, but merely allow corrosion to proceed for several months
and then show the rela-
tive loss in weight.
Neither have these ex-
periments taken into
account sufficiently the
localized, i.e., "pitting,"
corrosion to which badly
made material is especi-
ally subject. It is im-
material whether or not
the metal has lost but
little weight, provided
it has pitted in any
one spot sufficiently to
fail, or to have become
dangerously thin. This
pitting is believed to be
due chiefly to blowholes
and possibly to segrega-
tion resulting in a local
increase in electric po-
tential.

Manganese and Cor-
rosion. — It has been
suggested that the pres-
ence of manganese in
steel causes an increase
in the rate of corrosion, but this assertion is based upon no reliable
evidence so far as I am aware. It was brought forward to ex-
plain the supposed rapid corrosion of steel as compared with
wrought iron, but if it is a true influence in this direction, then steel
should corrode in acids faster than wrought iron, which it appar-
ently does not if the steel has been made with due care and is free
from blowholes and much segregation.

Badly Made Material. — There can be no doubt that badly

FIG. 299. — CORRODED WROUGHT-IRON
PLATE.

428 THE METALLURGY OF IRON AND STEEL

made steel is much more liable to corrosion and to pitting than well-
made steel, and it may be from this cause that the bad name which
steel is popularly given comes. There can also be no doubt that
badly made wrought iron is extraordinarily subject to rusting,
and of this kind of material we are to-day getting a good deal. As
I have noted in Chapter III, probably more than half of the wrought
iron produced in America is made by ' busheling ' scrap into a pile,
rolling it down and marketing it as wrought iron. This material
is of good quality so long as the scrap from which it is made is
good, but when the scrap is collected from almost any source,
and especially when it contains steel, as it sometimes does, we
should expect great differences in potential and therefore rapid
corrosion.

Coating. — Wrought iron has one advantage over steel in the
case of articles which are to be coated, because its rough surface
gives a better opportunity for the paint to adhere than the com-
paratively smooth and even surface of steel.

Summary. — Badly made steel and badly made wrought iron
corrode faster than any other material; next in order come well-
made steel and well-made wrought iron, between which two classes
the difference is probably very slight, and has not been determined
by any sufficiently lengthy or convincing series of tests; the next in
order is probably cast iron, although as yet we cannot be quite
certain that this corrodes more slowly than wrought iron and
steel, except in so far as it is protected by its natural or artificial
coating or both. Steel and wrought iron are both liable to pitting,
which may greatly shorten their life in service even though the
average rate at which they corrode is slow. Several causes may
produce this pitting, such as, blowholes, segregation, bad welding
in places, particles of oxide, scale or dirt, etc. When pits or holes
are found with a smooth, hollow surface, it is altogether probable
that they are due to blowholes and sometimes these pits may be an
inch or more in diameter and extend an eighth of an inch into the
plate, even before the remainder of the surface has become severely
attacked. Numerous efforts are being made at the present time to
improve the quality of steel and to turn out a more uniform grade
of product. It is quite certain that painting is not as good to-day
as it was in years past and that the quality of paint used is also
worse, on the average, as shown by the fact that paint does not
last as well on wooden structures. The better adherence of paint

THE CORROSION OF IRON AND STEEL 429

gives an advantage to wrought iron, which -however, only applies
where work is so placed as to be capable of being painted.

PRESERVATIVE COATINGS FOR IRON AND STEEL

The oxidized surface which all steel retains after hot working
is in itself a certain protection against corrosion, but this is only
limited in effect, because the oxide is more or less porous and so
allows the corrosive agencies to penetrate it and attack the surface
underneath. Moreover, the scale does not adhere firmly in some
places, and as its coefficient of expansion and contraction is dif-
ferent from that of the metal, it is liable to loosen and fall off in
places and so expose the iron or steel.

Preparation of Surfaces for Coating. — It is not advisable to
coat wrought iron or steel until its surface has been carefully pre-
pared, for any rust or other product of corrosion, scale, grease,
dirt, or moisture underneath the coat will either start corrosion or
else, by becoming loosened, will cause the paint or galvanizing to
fall off and so expose the metal underneath. The opposite is the
case with cast iron, because when this metal is poured liquid into
the mold a skin is formed, consisting of a chemical union of silica
and oxide of iron, very firmly united to the metal and capable of
being relied upon to adhere beneath the paint or other coating and
serve as additional protection from corrosion. So much mill scale
as adheres to wrought iron and steel with great force is permitted
by some engineers to remain underneath the protective coating,
but others insist upon the removal even of this upon the ground
that it may become loosened later by expansion and contraction,
and then spall off.

Priming Coat. — A difference of opinion exists as to the ad-
visability of having wrought-iron and steel surfaces prepared at
the mill and there given a priming coat, or else having the priming
coat applied under the direction of the engineer of construction, or
else of omitting the priming coat altogether and not painting the
structure at all (except such parts of it as will be inaccessible after
erection) until the metal has been exposed so long that the weather
has loosened all the scale. This will occupy perhaps six months to
a year, depending upon the corrosive conditions. During that time
no great damage can be done by corrosion although the structure
will, of course, look very shabby. After that period, the scale is

430 THE METALLURGY OF IRON AND STEEL

removed with sand blast, wire brushes, pneumatic hammers, or
chisels, and after the surface is perfectly clean and dry, it is given a
priming coat and at least two other good coats of paint, each one
being allowed to dry thoroughly before the next is spread. For
indoor work one coat of paint! upon the priming coat is often con-
sidered sufficient.

Shop vs. Field Painting. — The advantage of painting the steel
at the shop is that it can be done inside of some building and there
is therefore less liability of hygroscopic moisture under the coating.
If the shop coat is put on with care and skill it unquestionably has
certain advantages, but it is too true that the shop painting and the
preparation of surfaces at the shop is often carelessly done, for the
manufacturer has not the interest in preserving the metal from
decay that the consumer feels. Furthermore it undoubtedly
saves expense to allow the structure to stand from six months to a
year, provided that it is then very thoroughly cleaned and well
painted when absolutely dry. If otherwise the whole work may
have to be done over again at the end of a short interval.1

Pickling. — To remove scale it is customary in many cases to
pickle the steel or wrought iron, i.e., to immerse it in dilute sul-
phuric acid (say 10 per cent.) preferably heated to boiling so as to
act more quickly. After a few minutes the scale is removed, when
the metal is washed once in boiling water, once in cold water, and
finally in lime water to neutralize the last traces of the acid. It
should remain in the lime water until it is ready for the application
of the coating, when it should first be washed free of lime, and then
heated slightly above 100° C. (212° F.) to drive off all the moisture.
Pickling is therefore applicable only when the metal is to be coated
at the shop, either with the priming coat of paint or with zinc, tin,
etc.

Comparison of Methods. — Pickling costs less than the other
methods of removing scale and accomplishes the work very thor-
oughly. Sand blast is the next cheapest method. This latter
does not get off all the scale unless it is very thorough, while, if it is
too thorough, it leaves the surface in a smooth condition so that the
paint does not stick so well. On the other hand, pickling must be
done with great care or it may leave hydrogen upon the surface of
the metal which will greatly hasten corrosion, so that pickled

1 In which event it is customary for all interested parties to blame the
paint, except the paint-maker who is in the minority.

THE CORROSION OF IRON AND STEEL 431

surfaces sometimes corrode more rapidly than those which have
been cleaned in any other way. Cleaning with wire brushes is more
expensive than sand blasting, but if performed with great care is
more effective and leaves the surface in a rougher condition which
assists the adherence of the paint.

Kinds of Paint Used. — There is a great difference of opinion as
to the best paint for preserving iron and steel, but some few things
seem certain: (1) That no one kind of paint is suitable protection
against all corrosive influences. For example, the best paint to
withstand the action of the open air may fail when exposed to the
elements in a damp tunnel, or when used on the parts of piers
under sea water, while a good protection against this latter influ-
ence might be inefficient when exposed to the oxidizing gases in
locomotive smoke, etc. ; (2) that whatever paint is used must be
sufficiently elastic to expand and contract with the changes in
temperature of the metal without 'cracking; and (3) that it must
contain nothing which will attack the metal and so commence cor-
rosion. In this last connection we must especially avoid all oxidiz-
ing influences. The reader will find in the Annual Proceedings of
the American Society for Testing Materials a very full interchange
of opinions between experts in paint manufacture and engineering
which will doubtless help him to form an opinion as to the best
paint to use in each case.

There are two parts to every paint: (1) The vehicle which car-
ries the pigment and undergoes a change to the solid state when the
paint dries, and (2) the pigment or originally solid part of the pre-
servative coating. These two must form a firm impervious coat-
ing upon the surface of the metal, but must not be so solid as to be
inelastic or brittle.

Linseed Oil. — Linseed oil is a very good and common vehicle.1
It is what is known as a ' drying oil.' That is to say, an oil which
when exposed to the atmosphere will change from a liquid to an
elastic or leathery consistency. This action takes place not by
evaporation but by a process of oxidation, whereby the oil absorbs
oxygen to the extent of from 10 to 18 per cent, of its weight and ex-
pands in volume, so that a coat of linseed oil spread upon glass will
wrinkle up upon drying. It is because linseed is the best of all the
drying oils that it is so much preferred as a paint vehicle, but when
allowed to dry in the raw state it requires too long a time and the
1 See page 336 of No. 163 ; page 41 of No. 164, page 436'.

432 THE METALLURGY OF IRON AND STEEL

drying is therefore hastened by boiling it and adding some oxidizing
agent, known as a ' drier/ of which the best, as far as iron and steel
preservation is concerned, are the salts of lead or manganese, used
without rosin. It will readily be understood how dangerous it is to
indiscriminately use driers (i.e., oxidizing agents) in steel paints,
because so many of them will oxidize the steel and so cause the very
corrosion which it is the object of the paint to prevent.

Purity of Linseed.' — This brings us to the question of the purity
of linseed oil, for the usual adulterants are all harmful to steel work
and cause in the end more painting and more of the expensive
cleaning of structures to receive the coating, out of all proportion
to their lesser first cost. Freedom from adulterants probably
cannot be obtained except by constant watchfulness and frequent
chemical analysis on the part of the consumer. Some impurity
arises from the presence of a few per cent, of foreign seeds with the
linseed, which is not always avoidable, but the greater harm comes
from the fact that the oil is obtained from the linseed by pressing it
while hot, in which way a larger amount of product is extracted
than if cold pressure is applied, because some of the solid part of the
seeds are thereby extracted together with the oil. Cold pressed
linseed oil has a golden yellow color and remains clear in cold
weather as distinguished from the yellowish-brown color of the hot
pressed oil which also has a more acrid taste, is not so fluid, and con-
tains more solid fats, solid organic matter, and fatty acids, all of
which are harmful either because they attack the metal or else
because they make a pervious paint.

Pigments. — The pigment is not as important as the vehicle
and many different ones can be chosen, provided they are chemic-
ally inert to the steel. Red lead has been very much used and is
very good especially for the priming coat, for it seems to form with
the linseed oil a very dense, impervious coating. For the outer
coats, however, it is generally well to mix the red lead with some
substance that shall reduce its weight, such as graphite. Ferric
oxide, Fe2O3, and other oxides of iron in the form of iron ore are
very cheap and withstand the action of sulphur gas better than the
red-lead paints. They are very good for outer coats where loco-
motive smoke and similar gases are liable to be present. Sulphate
of lead, white lead (a mixture of oxide and sulphate of lead with
often some sulphate of zinc) and sulphate of zinc are all good white
paints although expensive. Pulverized asphaltum and other

THE CORROSION OF IRON AND STEEL 433

hydrocarbons are also used with success as pigments, especially
where the metal is exposed in damp ground or under water.

Other Paints. — Pipe is often coated very cheaply by dipping
it in melted asphaltum or pitch. The objection to this coat is that
it is very hard and brittle when cold and in time it forms a network
of myriads of cracks through which the atmosphere attacks the
metal. For cast-iron pipe it is very useful, however, because this
is protected by its natural skin. Dipping in tar would form an
elastic coating, but unfortunately tar contains certain acids and
oxidizing agents which attack the metal. There is a paint made
by distilling off the creosote and other volatile components of tar
until the solid asphaltum is left. This is then redissolved in
two of the distillates, neither of which will attack iron work,
and thus a paint is obtained which is said to be practically tar
without any of its harmful constituents. It forms a very elastic
coating which does not crack after years of exposure nor does it
disintegrate under the action of the sun as the linseed oil paints
sometimes do.

Galvanizing. — Galvanizing is the process of coating with
metallic zinc and where this coating adheres firmly, it undoubtedly
forms a very efficient means of protecting iron from corrosion. As
zinc is electrically positive towards iron, whatever electrolysis
exists would tend to corrode the zinc and protect the iron. In-
deed this fact is taken advantage of by some engineers who hang
pieces of zinc in their boilers, by means of a wire connected to the
steel work, so that the electrolytic action shall corrode the zinc and
protect the wrought iron or steel.

Galvanizing is usually applied to wire and wire products, thin
sheets, especially corrugated sheets used for the outside of build-
ings, etc., tubes, hollow ware, and a great variety of articles, after
the surfaces have been cleaned by pickling. There are three
methods by which the galvanizing is effected, known respectively
as cold galvanizing, hot galvanizing, and dry galvanizing.

Cold Galvanizing. — In the cold galvanizing process zinc is de-
posited electrolytically upon the surface of metallic articles which
are made the cathode of an electro-plating cell. The zinc is first
dissolved in sulphuric acid and water and this solution is made the
electrolyte. The anode is a piece of zinc, so that as fast as the elec-
tricity deposits zinc upon the surface of the article being galvan-
ized, it replenishes the electrolyte by dissolving zinc from the anode.

434 THE METALLURGY OF IRON AND STEEL

The coating is about 0.0003 to 0.0005 inch thick, equivalent to
about 0.2 to 0.3 ounce of zinc per square foot of surface.

Hot Galvanizing. — In the hot galvanizing process, which is the
commonest one used, the articles to be galvanized are dipped into
a bath of molten zinc at a temperature of 425° to 460° C. (800° to
860° F.), i.e., slightly above the melting point (419° C. =786° F.).
The metal is exposed to the zinc bath usually about 1J to 7J
minutes, depending upon the thickness of coating desired, which
will vary between 0.0003 and 0.0010 inch or about 0.2 to 0.6 ounce
of zinc per square foot of surface covered, or about 0.3 to 0.6 ounce
per pound of wire. In the case of wire the iron or steel is drawn
slowly through the bath of melted zinc and usually passes over a
wiper as it comes out, which removes the still molten zinc and
causes the zinc remaining to stick a little more firmly and have a
more uniform thickness. The coating on this wiped wire is not
so liable to crack and break off when the wire is bent and twisted
as the coat of unwiped wire but, on the other hand, it is thinner
and gives but little protection against corrosion. Sometimes arti-
cles to be galvanized are first dipped in a bath of melted lead and
then in the melted zinc. This gives a cheaper coating.

Dry Galvanizing. — The process of dry galvanizing is a recent
invention and consists in heating the articles to be galvanized in-
side a closed vessel and while they are covered with what is known
as 'blue powder/ which is a zinc dust containing some oxide of
zinc and relatively cheap in price because it is a by-product in the
metallurgy of zinc. The temperature is about 300° C. (575° F.),
and", although this is below the melting point of zinc and of iron, it is
sufficiently high to produce an alloy between the two, forming, it
is said, a very resisting coating which is more thoroughly attached
to the surface of the metal and therefore much more durable
against cracking off.

Comparison of Galvanizing Methods. — Cold galvanizing depos-
its a thinner coating of zinc which, if improperly performed, is
liable to be porous or spongy, but it gives a better connection be-
tween iron and zinc and therefore a more durable coating. Hot
galvanizing necessitates the use of a flux on the bath of melted
zinc in order that the zinc may not be oxidized by air, and these
fluxes probably have the effect of sometimes beginning the corrosion
of the iron underneath the layer of zinc. The process of dry gal-
vanizing is too new yet for any comparison to be drawn.

THE CORROSION OF IRON AND STEEL 435

Tinning. — A large amount of metal is coated with tin in order
to give protection against organic acids, such as those present in
cooked foods, and also in order to give a more effective resistance
to the elements. Thus cooking utensils, roofing sheets, tin cans
for preserves, and many such articles are coated with tin in pref-
erence to zinc, either because zinc would not withstand so long
the corrosive influence, or else would not resist it at all. In the tin-
ning operations the metal sheets are usually drawn through a bath
of liquid tin by four to six pairs of rolls which are immersed in it.
Each pair of rolls presses the tin which has solidified on the surface
of the iron firmly upon the metal and the result is a smooth, bright,
adhering coat which protects the metal very successfully. Tin
plating is more expensive than galvanizing, chiefly on account of
the additional cost of the tin.

Terne Plate. — Sometimes sheet" metal is coated with a mixture
of two-thirds lead and one-third tin and then goes under the name
of terne plate, which is used very largely for roofing and outdoor
purposes. It is applied by the same method as tinning but is
less expensive.

Nickel Plating. — Articles requiring a very high polish and
which are to be subjected to handling, etc., are often plated with
nickel. This is an electrolytic process, similar to the general oper-
ation described under the electrolytic galvanizing. Nickel plating
is more expensive than galvanizing or tinning, but gives a more
highly resisting surface.

Oxidized Coating. — There are one or two processes by which a
black oxidized surface can be given to iron and steel which will re-
sist rust for years and form what are known as ' black iron ' objects.
It is used chiefly for fancy iron work in house decorations, etc.

Enameling. — A number of articles, such as bath tubs, wash-
bowls, cooking utensils, are made of cast iron or steel and then
coated with a white or variously colored film known as enamel.
Enameling processes are more or less secret, but usually consist in
powdering the enamel upon the surface of the metallic article which
has been heated to a red heat. At this temperature the mixtures
forming the enamel melt and spread themselves uniformly over
the surface where they chill and harden. Enamel must be in-
soluble in water and in chemicals with which they are liable to come
in contact, and must also be sufficiently elastic to expand and con-
tract with the metal without breaking off.

436 THE METALLURGY OF IRON AND STEEL

REFERENCES ON CORROSION

See especially Nos. 117, 118, and

160. Allerton S. Cushman. "The Corrosion of Iron." Proceed-

ings of the American Society for Testing Materials, vol. vii,
1907, pages 211-228. Same published as Bulletin No. 30,
Department of Agriculture, Washington, D. C. With many
cross-references.

161. Henry M. Howe. "Relative Corrosion of Wrought Iron and

Steel." See Proceedings, International Association for
Testing Materials, 1900, vol. xi, Part I, pages 229-266.
Paris, 1901. Abstracted in No. 162, and in Journal Iron
and Steel Institute, No. 11, 1900, pages 567, 568.

162. Frank N. Speller. "Steel and Iron Wrought Pipe." The

Iron Age, March 2, 1905, vol. Ixxv, pages 741-745. With
several cross-references on relative corrosion.
See also the indices of No. 8 for abundant references.

163. J. Lewkowitsch. "Chemical Analysis of Oils, Fats and

Waxes." London, 1898.

164. A. H. Church. "The Chemistry of Paints and Painting."

London, 1901.

XVII
THE ELECTRO-METALLURGY OF IRON AND STEEL

IN the electric smelting and refining of iron and steel, four
modifications in practice are produced:

1. Practically any desired temperature in reason may be
obtained.

2. The impurities introduced with the usual fuels are avoided.

3. The temperature is regulated with much greater accuracy.

4. The cost of the processes is increased.

For many years the first three modifications have been taken ad-
vantage of in the production of ferro-alloys — e.g., ferro-tungsten
ferro-chrome, ferro-molybdenum, etc. — that is, alloys of pig iron
and some other metal which is used for recarburizing in the
manufacture of alloy steels. The high temperature necessary
for the production of these alloys gives electric smelting especial
advantages while the high price at which they can be sold enables
their manufacturers to stand the additional expense with profit.
But in the year 1900 a number of important electro-metallurgists in
Europe and America began to use electric processes for the pro-
duction of pig iron and steel on a commercial scale, and from this
time the industry dates. Because the electric processes have
apparently secured for themselves a permanent place in the
metallurgy of iron and steel, and because of the great interest
which they have evoked, I have decided to devote some space to
them here, although the amount of metal produced by them is
still very, very small in comparison with the older methods.

Iron and steel electro-metallurgical processes naturally divide
themselves into three classes:

(1) Ore smelting for the production of pig iron.

(2) Refining of pig iron to produce steel, and

(3) Electrolytic refining of steel or wrought iron to produce
almost chemically pure iron.

The first two classes are electro-thermic; that is to say, they
use electricity for conversion into the heat necessary for smelting;

437

438

THE METALLURGY OF IRON AND STEEL

the third class is electrolytic, that is to say, the electric current
serves to produce chemical changes.

ELECTRO-THERMIC ORE SMELTING

The two most successful ore-smelting processes are those of
Heroult and Keller, and in each of those the furnace is filled with

the charge, consisting of ore,
flux, and coke which resembles
the ordinary blast - furnace
charge, except that the amount
of coke is much smaller since
the electricity is relied upon
for the production of heat.
No blast is driven into those
furnaces, but a powerful cur-
rent of electricity is passed
through the charge, the high
resistance of which, on account
of its poor conductivity, pro-
duces a great deal of heat.
The degree of heat can be reg-
ulated both by the intensity
of the current and the charac-
ter of the charge, in which the
coke is the best conductor.
The process will be better un-
derstood by reference to Figs.
300 and 301, and the descrip-
tions thereunder. When a
certain temperature is reached
the carbon reduces the ore,
and the iron thus produced
absorbs from the coke carbon,
silicon, sulphur and phos-
phorus,1 and collects in a

conducted mainly by the coke, which is heated ^ -j j t ^ bottom of

on amount, nf its fiWtrioa.1 resistance, and to T.

the furnace. The gangue
unites with the flux to form a
1 Some phosphorus is absorbed
from the ore also.

FIG. 300. — EARLY HEROULT ORE-
SMELTING FURNACE.

The ore enters the furnace through A, and
meets the coke which enters through H. The
electric current enters at I, to the positive elec-
trode B, which is a block of solid carbon; it
passes from B up through the charge, being

on account of its electrical resistance, and to
the negative electrodes, F and G, which are also
blocks of solid carbon. The negative electric
connection is a J. The charge begins to become
heated at the electrodes F and G, and con-
stantly gains in temperature as it falls lower
and lower in the furnace.

THE ELECTRO-METALLURGY OF IRON AND STEEL 439

FIG. 301. — KELLER ORE-SMELTING FURNACE.

The two shafts are connected below by means of the channel AAA. The charge, con-
sisting of ore, coke and flux, enters the shafts at BB, BB, and passes downward in the furnace
as smelting progresses. C is the positive, and D the negative electrode, each consisting of
a block of carbon. E and F are also blocks of carbon, electrically connected by the cable,
G. The current enters at C and passes through the charge to E; thence it passes through
G to F and thence through the charge in the other shaft to D. In this way the charges in
both shafts are heated; liquid iron is formed and collects in the channel, AAA, while the
slag floats on top of it. When the liquid pool extends all through AAA the current passes
through it instead of through the cable G, and thus the melted iron is kept hot. If it should
become too cold, the auxiliary electrode, H, is brought into action and the current enters
at this point until the desired temperature is obtained.

slag and, on account of the very high temperature available,
the slag is made intentionally very rich in lime, wherefor it has

440

THE METALLURGY OF IRON AND STEEL

a great solvent power for phosphorus and sulphur. On this
account pig iron, very low in these two harmful elements, can
be produced from relatively impure ores, and this is one of the
greatest advantages of this type of ore smelting. The pig iron
and slag are tapped from the bottom of the furnace through
tap holes, as in other processes.

Electric vs. Coke Smelting. — The cost of electric smelting under
diverse conditions has been frequently estimated and figures

FIG. 302. — SECTION OF KJELLIN INDUCTION FURNACE.

may be found in several of the references given at the end of this
chapter. It seems unwise to quote figures except where the full
circumstances and conditions are described, although we may
estimate that the cost of smelting may be as low as $10 per ton
of pig iron produced under the most favorable circumstances and
up to $30 per ton where ore and water power are not so cheap.

THE ELECTRO-METALLURGY OF IRON AND STEEL 441

The expense depends chiefly upon the cost for the production of
electricity, and this is so large that electric smelting is practically
never advisable except where impure ore and water power for

the production of electricity are very cheap, and where pig iron,
coke and pure ore are expensive. Besides the advantages of
being able to produce a very pure pig iron and to employ very
impure ores, we can use electric smelting even where coke is not
available by the use of charcoal as a reducing agent.

442 THE METALLURGY OF IRON AND STEEL

ELECTRO-THERMIC MANUFACTURE OF STEEL

There are three important types of electric furnace now in use
for the refining of pig iron, known respectively as the Kjellin-
Colby or induction type; the Heroult, and the Keller types.

Induction Furnace. — The induction furnace is based upon
the principle of the ordinary static transformer, whereby alter-
nating electric current is transformed to lower voltage. It was
independently developed by E. A. Colby, of the United States,
and F. A. Kjellin, of Sweden, whose American rights have been
joined under one management. A sectional elevation of the
furnace is shown in Fig. 302, in which CCCC is the core of an
electro-magnet, around one leg of which a coil of wire, AA, passes.
When an alternating current goes through the coil, A A, it sets
up an alternating magnetic field in the core, CCCC, and this in
turn sets up a secondary current in the circle, BB, parallel to
the coil, A A. In other words, an alternating current passing
through the coil, A A, induces an alternating current in the coil,
BB, without there being any metallic connection between the two.
This is a well-recognized phenomenon in electrical engineering
and requires no further comment here. In the furnace operation,
the circle, BB, is a hollow ring in the brickwork into which melted
metal is poured. The resistance offered by this melted metal
to the passage of the induced current generates heat which will
maintain the temperature or raise it to any desired point. The
slot, BB, is really an annular crucible into which pig iron, steel
scrap, iron ore and flux may be charged as if it were an open-
hearth furnace, and the operation of steel making is practically
the same in principle except that electric heat is employed instead
of regenerated gas and air. We may charge solid metal if desired,
but, in such a case, it is well to leave a shallow circle of metal in
the bottom of the slot, BB, after each operation is ended, to serve
to carry the induced current during the beginning of the next
operation, until the solid charge is melted.

There is a tap hole for slag at an upper level into the slot,
and one or two tap holes for metal at a lower level; or sometimes
the whole furnace is supported so as to tip forward and pour its
contents out of a spout. The great advantages of the furnace
are the freedom of the charge from contamination by impurities

THE ELECTRO-METALLURGY OF IRON AND STEEL 443

either in fuel or in electrodes or other connections, the excellent
control and wide range of temperature, and the ease and simplicity

FIG. 304.— COLBY FURNACE POURING.

of operation. Its disadvantages are the high cost for electricity
on account of the low efficiency of the induction process, and

444

THE METALLURGY OF IRON AND STEEL

the expense for replacing the annular crucible when it is badly
corroded by slag and oxides. Indeed, the cost of this type of
operation is so great as to practically preclude it from competi-
tion with the open-hearth process under ordinary circumstances
and conditions, but it does enter into direct competition with
the crucible-steel process, because of the purity obtainable when

Longitudinal Sections-A B &-C D
FIG. 305 — HEROULT REFINING FURNACE.

pure charges are used, and because of the high cost of the latter
process for crucibles, labor and fuel, so that altogether more than
a dozen furnaces of the induction type have been installed for
commercial production of steel in Europe, England and America.
Heroult Furnace. — The design and operation of the Heroult
steel furnace is even more like an open-hearth than the induction

THE ELECTRO-METALLURGY OF IRON AND STEEL

445

furnace, although again electric heat is substituted for regenerated
gas and air. The general form of the furnace is shown in Figs.
305 to 306. There are two electrodes of carbon, one of which
leads the current to the charge and the other conducts it away.
The electrodes do not touch the charge, but the current arcs
from one electrode to the charge, through which it passes, and
thence arcs to the negative electrode. Thus combined arc and

Transfers Section E F
FIG. 306. — HEROULT REFINING FURNACE.

resistance heating is used, and the impurities of the electrode
do not come in contact with the charge except in so far as the
electrode burns up and deposits its ash. When desired the
electrodes are lowered until they dip into the slag, but never
touch the metal. The charges consist of solid or liquid pig iron,
steel scrap and flux, regulated according to the principle of the
basic open-hearth process, for the usual Heroult furnace is lined

446 THE METALLURGY OF IRON AND STEEL

with basic material. On account of the high temperature obtain-
able a liquid slag very high in basic components may be made,
and thus steel low in phosphorus and sulphur may be produced
from impure raw materials. The furnace is mounted so as to tip
for discharging its contents. There are holes for the admission
of air blast to the interior of the furnace and the operations of
adding ore and lime and of working the charge are similar to
those of the basic open-hearth process. There are four Heroult
steel furnaces in Europe, and one in the United States.

ELECTROLYTIC REFINING OF IRON

The greatest amount of work on the electrolytic refining of
iron has been done by C. F. Burgess, of the University of Wis-
consin, who produces an iron that is almost chemically pure,
the chief foreign element being hydrogen, whose presence renders
the metal very hard and brittle. The hydrogen is driven off
by heating the metal to a high temperature, but this is accom-
plished with the result of vitiating the iron with traces of carbon
and sometimes other impurities, for iron has such a great affinity
for carbon that it will absorb it at a red heat from coke, charcoal,
and even from gases and oil vapors. Indeed, if a piece of electro-
lytic iron, or other iron very low in carbon, be heated in contact
with steel or wrought iron higher in carbon, there will be a small
amount of transfer of carbon from the low- to the high-carbon
metal. The same difficulty is met with in melting the metal,
and as its melting point is 1507° C. (2745° F.), there are not
many kinds of crucibles that will stand the heat necessary and
yet fail to yield carbon, silicon, or some other impurity to the
iron. As I understand it, it is this difficulty which has been the
chief obstacle in the electrolytic process, for the electrolysis
itself seems to be accomplished by Burgess with success and
economy. If the cost of the electrolytic product could be brought
somewhere near that of Swedish iron and the other very pure
forms, it is probable that it would become a commercial commodity
on account of its high magnetic permeability, electric conductility
and softness.

Electrolytic Process. — In Burgess's process the electrolyte is
a mixture of ferrous sulphate and ammonium sulphate and it

THE ELECTRO-METALLURGY OF IRON AND STEEL 447

is kept at a temperature of 30° C. (86° F.). The anode is ordinary
wrought iron or steel and the primary cathode upon which the
first metal is deposited is a thin strip of sheet iron. The deposited
metal sticks so slightly to this that no difficulty is met with in
separating them after the operation is finished. The electric current,
with a density of from 6 to 10 amperes per square of cathode
surface, deposits the dissolved iron upon the cathode with an
efficiency of nearly 100 per cent., and cathode plates averaging
about three quarters of an inch thickness are produced in a
four weeks' run.

REFERENCES ON THE ELECTRO-METALLURGY OF
IRON AND STEEL

171. Eugene Haanel, Superintendent of Mines. "Reports of the

Commission Appointed to Investigate the Different Elec-
tro-Thermic Processes for the Smelting of Iron Ores and
the Making of Steel." Department of the Interior, Ottawa,
Canada, 1904 and 1907.

172. A number of valuable articles and references in vols. i to v,

1902 to 1907 inclusive. The Electro-Chemical and Metal-
lurgical Industry.

173. C. F. Burgess and Carl Hambuechen. "Electrolytic Iron."

American Electro - Chemical Society, April, 1904. Ab-
stracted in The Electro - Chemical and Metallurgical In-
dustry, vol. ii, 1904, pages 183-185.

174. Geo. P. Scholl. "The Manufacture of Ferro-Alloys in the

Electric Furnace." The Electro - Chemical and Metal-
lurgical Industry, vol. ii, 1904, pages 349-351, 395-396
and 449-452.

175. A number of valuable articles in W. Borcher's Metallurgie,

vols. i to iv, 1904 to 1907 inclusive.

176. Several valuable articles and abstracts in Henry Le Chatelier's

Revue de Metallurgie, vols. i to iv, 1904 to 1907 inclusive.

177. John B. C. Kershaw. " Electric Furnace Methods of Iron

and Steel Production," vols. xxxix and xl, The Iron Trade
Review, 1906 and 1907.

178. G. K. Burgess. "Melting Points of the Iron-Group Ele-

ments," Reprint 62. Bureau of Standards, Washington,
D. C. See this Bulletin for best references on pyrometers.

XVIII
THE METALLOGRAPHY OF IRON AND STEEL

METALLOGRAPHY in its larger sense is the description or study
of the structure of metals. That branch of the subject which
comes under the head of microscopic metallography is, however,
the most important because the structure of most metals, espe-
cially iron and steel, is discernible only when magnified. We shall
see, however, that the observation of structure by eye — known as
magroscopic metallography — is not without great value.

Microscopic metallography has nowr reached that stage of im-
portance where it is viewed almost on a par with chemical analysis
and physical testing. In the United States practically every
large steel works is well equipped for the microscopic analysis of
its product, and, too, important laboratories of the universities and
of consulting metallurgists devote much attention to the study.
Although only a little more than twenty years have elapsed since
the art first received public attention, it has advanced so far as to
have become by now another and a very serviceable tool in the
hands of the expert. I take this opportunity, however, of offering
a word of warning: reputations have more than once suffered
severely, because of erroneous deduction made from microscopic
evidence, and history has shown that those who "rush in where
angels fear to tread " are sure to be caught sooner or later. The
wise man is he who never bases an opinion upon a sample whose
chemical analysis is unknown to him: who never bases an opinion
upon a microphotographic negative or print, and who polishes
and etches his own specimens, or has these operations performed
by some one well known to him and working under his immediate
direction. With these precautions the microscope is a very relia-
ble index to an experienced mind.

448

THE METALLOGRAPHY OF IRON AND STEEL 449

PREPARATION OF SAMPLES FOR MICROSCOPIC EXAMINATION

Samples of iron and steel for microscopic analysis can be cut out
of soft samples by means of a hacksaw, lathe, or other machine,
and broken out of hard and brittle samples by means of a hammer.
A good rule to follow is to have the surface that is to be polished
about f to i of an inch on a side. If larger than that, it requires
excessive labor for polishing. It requires about sixteen times as
much labor to polish a sample an inch square as to polish one J
inch square, and it requires about sixty-four times the labor to
polish one 2 inches square. On the other hand it is not advisable to
polish too small an area, because the surface will be liable to con-
vexity and therefore very difficult to get into focus, especially for
high powers.

Rough Polishing. — After the proper size of a specimen is ob-
tained the next step is to give its surface a bright mirrorlike polish,
free from any scratches discernible even with high powers of the
microscope (1,000 diameters or so), and this result is achieved
with the greatest economy in labor by proceeding by gradual
steps :

If the surface contains deep gouges, or marks produced by cut-
ting or breaking it out, it should first be brought to a plane surface
by rubbing across a rough file, a grindstone, or an emery wheel.
For my own work I much prefer to rub the specimen across a file
rather tnan to put it in a vice and draw a file across it, for I believe
the former method produces a more even, — i.e. less rounded, —
surface and therefore conduces to economy of labor in the later
operations. It is best to hold the specimen lightly in the fingers
and draw it back and forth across the file in a straight line, avoiding
any circular motion and therefore having the polishing marks all
parallel and straight across, the specimen.

When a plane surface has been produced in this way the speci-
men should be rubbed on a very smooth file. In this operation
the specimen should again be held in the fingers and rubbed in a
straight line back and forth, and should be turned 90° from the
first rubbing so that the marks now made will cut vertically across
the first scratches. In this way it is very easy to tell when the
scratches made by the first file are entirely eliminated and the oper-
ation on the second file should be continued to at least this point
no matter how short and faint the old scratches may prove to be.

450 THE METALLURGY OF IRON AND STEEL

It only takes a minute or two to remove the last scratches on this
smooth file, but it would take several minutes to remove them by
means of one of the later polishing mediums, and the greatest
economy is obtained by having each stage of the operation abso-
lutely complete.

After coming off the smooth file, the specimen is again turned
90°, so that the marks now to be made run in the same direction as
those made on the rough file, and rubbed across a sheet of ordinary
00 emery paper cut to about 3| inches wide by 9 inches long and
pinned with thumb tacks upon a piece of planed, J-inch board.
This operation is continued until the last marks from the smooth
file are removed.

Fine Polishing. — The polishing then proceeds in the same
manner by steps upon French emery paper of gradually increasing
fineness, each piece being cut to about 3f inches by 9 inches and
mounted on a smooth board. Of the Hubert brand the grades are
designated 0, 00, 000, and 0000.

After the 0000 French emery, the surface is very smooth and
bright and is given a final burnishing by rubbing across a piece of
broadcloth or baize stretched over a piece of wood, moistened with
water and covered with a very thin liquid paste of water and best
washed rouge. This should leave the specimen polished as bright
as a mirror and free from all scratches. It may be, however, that
the eye or a hand magnifying glass would discover microscopic
rounded furrows in the specimen. These are due to scratches
made in the early stages of polishing, which have not been elimi-
nated, but whose corners have been rounded off by the finer grades
of polishing mediums. I never permit a specimen containing these
furrows to be used as the basis of any opinion.

Preparation of Rouge. — The best jeweler's rouge purchasable
is not good enough for polishing, as it contains very fine particles of
dirt and grit which produce scratches in the surface of the specimen
during the final stages of the process. It is best washed by the
metallographer himself. This is best done upon samples of not
more than one teaspoonful of rouge at a time stirred in about a
glassful of water in a flat pan or dish until thoroughly wetted.
After allowing to settle for about five minutes the wrater is poured
into an ordinary chemist's wash bottle where it is kept until ready
for use. The last dregs of the water, containing a good deal of
coarse rouge and grit, is thrown away. A little experience soon

THE METALLOGRAPHY OF IRON AND STEEL 451

teaches one to get the maximum amount of good rouge from a
sample, without any particles that would produce scratches. The
rouge in the wash bottle is protected from dust and dirt and can be
poured out of the glass tube on to the broadcloth polishing board
as needed. For the preparation of special powders for the very
finest grades of polishing, the reader is referred to the references
at the end of this chapter. I have described here the methods
which I prefer and use, but there are many others which are doubt-
less equally as good.

Precautions as to Polishing. — Do not rub the specimen too
hard on the polishing mediums. This does not produce the de-
sired effect any more rapidly and may distort the structure of the
metal so as to lead to erroneous conclusions.

Do not allow the specimen to become heated by the polishing.
This is especially true of hardened steel and other heat-treated
specimens which may become tempered and so altered even upon
gentle heating.

Rub a piece of hard steel over each piece of polishing paper and
broadcloth before using it for polishing your specimen. This is
to get rid of grit.

Do not lay polishing boards down where dust will get on them,
but let them stand with the full height upward inside a closed box
or a small closet.

Never form an opinion upon a specimen that retains ^scratches
or polishing marks.

Mechanical Polishing. — Instead of rubbing the specimen across
different mediums by hand, we can press them against the emery
papers and rouge-cloth mounted upon wooden discs about 8 inches
in diameter, revolving at speeds of about 600 r.p.m. This method
saves time and the necessary apparatus is very simple to make, or
can be purchased complete.1 For my own use I prefer hand-
polishing except upon the rouge, because of the liability to heating
the specimens with the higher surface-speed, and to damage of the
specimen by having it snatched out of one's fingers.

1 Consult references 180, 184 and 185, page 457.

452 THE METALLURGY OF IRON AND STEEL

DEVELOPING THE STRUCTURE FOR EXAMINATION

To develop the structure of iron and steel so as to differentiate
between the constituents, four methods are available :

(1) Polishing in bas-relief .

(2) Etching with chemicals.

(3) 'Polish attack/ and

(4) Heat tinting.

Polishing in Bas-relief. — Where some of the constituents are
less durable than others the method of polishing that I have de-
scribed, upon a soft background, produces a bas-relief, since the
softer constituents are worn to a greater depth than the harder
ones. The parts thus worn down appear darker than the higher
places which reflect the light better. It is by this method that
graphite is best distinguished in pig iron and slag in wrought iron,
because the attack by acids is liable to produce other dark spots
which may not be readily differentiated. It is by the bas-relief
method that F. Osmond developed so beautifully the structure of
pearlite shown in Fig. 241. This method has the disadvantage,
however, of rounding off the edges of the harder constituents and
so causing the softer ones to appear larger than they really are.

Etching with Chemicals — Nitric Acid. — This method is prob-
ably the commonest one of developing the structure of steel. Many
different strengths of acid are used by different metallographers
from 0.1 per cent, up to 20 per cent., and the length of time that
is necessary to expose the specimen will depend upon this factor
and upon the amount of carbon present. High-carbon steel will
require a longer time than soft steel and may take as much as
2J minutes, although this is very rare. The nitric-acid solutions
are usually made with alcohol instead of with water, in order
to dry more quickly. Some metallographers prefer to immerse
their specimens in the acid for a given length of time, and others
prefer to hold the specimen in the hand with the polished sur-
face upward, and then deposit a few drops of the etching fluid
upon it with a piece of rubber tubing on the end of a glass rod.
In any event it is usually better to etch a shorter time than is
estimated as suitable, examine under the microscope, and then
etch again if necessary. After every etching it is necessary
to wash the specimen off with alcohol and dry as quickly as
possible. This drying is best accomplished by absorbing all

THE METALLOGRAPHY OF IRON AND STEEL 453

the liquid left on the surface after the alcohol wash with a piece
of soft cloth and then holding in a stream of air. Some metal-
lographers wash the specimen after etching with alkali, and
then with water and then with alcohol. Sauveur immerses his
specimen in the strongest nitric acid for a few seconds and then
places it in a stream of running water, washes with alcohol and
dries. As the strong nitric acid puts the steel in a passive state,
the length of attack by this method is only momentary, so that
it is necessary to repeat it a few times. It results in a very even
etching of the surface.

Iodine Etching. — Some investigators use iodine for etching,
by rubbing a few drops over the surface until a film covers it, and
allowing it to remain until the iodine color has disappeared,
after which the iodine is washed off with alcohol and dried. A
good iodine solution for this purpose is the ordinary tincture of
iodine diluted with an equal voluhie of alcohol. It is also well to
have an auxiliary solution of about J- this strength for lighter
etching work.

Picric Acid Etching. — I have found the 5 per cent, of picric
acid in alcohol, which Igevsky used for hardened and annealed
steels, very useful also for wrought iron, very low carbon steels, and
pearlite.

Polish Attack. — The method of ' polish attack ' advised and
used by Osmond, consists in rubbing the specimen upon a piece
of parchment stretched over a piece of soft wood and moistened
with a 2 per cent, solution of ammonium nitrate. This method
gives very beautiful results in Osmond's hands and is especially
valuable for developing the structure of martensite, troostite,
pearlite, and sorbite.

Heat Tinting. — Heat tinting consists in 'warming the steel
until it becomes oxidized. The different constituents are oxi-
dized at a different rate and so may be distinguished from one an-
other. It is most serviceable in distinguishing the phosphide of
iron from the carbide, for by heating until the carbide is red, the
phosphide (FesP) will be yellow. The phosphorus eutectic may
be distinguished from pearlite in the same way, because the former
will be yellow when the latter is blue. The heat tinting must be
accomplished in such a way as not to expose the metal directly to a
flame. The simplest method is to put the specimen upon an iron
plate heated from beneath.

454 THE METALLURGY OF IRON AND STEEL

MICROSCOPE AND ACCESSORIES

It is hardly desirable to describe here the different forms of
microscope and accessories used for metallographic work, as the
manipulations cannot well be learned, except by practice, and the
catalogues of the manufacturers of scientific instruments of all
iron-producing countries now give full data. A photograph of a
common illuminating device, microscope and camera, is shown in
Fig. 307. Those who expect to take up the subject should pro-
cure a book upon it and study it much more fully than we have
space for here. The chief difference between the microscopic ex-
amination and photography of iron and steel, and of biological
specimens, botanical specimens, thin rock sections, etc., arises
from the non-transparency of the metal. It is necessary to illu-
minate them from above, since we cannot cause the light to pass
through the specimen and into the instrument. This necessitates
special forms of illuminators and powerful lights. For magnifi-
cations of about 500 diameters, the Welsbach mantle gives suffi-
cient illumination. The Nernst lamp is also very useful, but for
very high powers it is necessary to use the electric arc lamp. This
introduces especial difficulties, because the focus which gives good
definition to the eye by means of the arc light will be blurred upon
the photographic plate. The use of a light-yellow screen in front
of the light assists in this difficulty, but a good deal of experience is
required to get good results. A mercury light avoids this difficulty.

MACROS COPIC METALLOGRAPHY

An experienced eye may get a very good idea of the size of
crystals in iron and steel by examining a freshly broken fracture,
and this is one of the branches of magroscopic metallography.
For the most accurate results, however, it is not as good as mi-
croscopic examination.

It is also possible to get other information by etching a polished
surface for many hours with dilute hydrochloric acid. For this
purpose the polishing need only go as far as the commercial 00
emery paper, following the smoother file. In this way the center
of ingots, because of their looser texture, will be eaten away much
more rapidly, and this will be evident to the unaided eye. Also
the interior of sections of large area will be attacked more than the

456

THE METALLURGY OF IRON AND STEEL

outside, which has become harder through the work of rolling.
Some blowholes, which cannot be seen by eye or have been par-
tially welded up, may often be discovered, because the etching
action is more severe in their neighborhood, and the same is true

FIGS. 308 TO 311. — RAILS ETCHED FOR SEVERAL HOURS WITH DILUTE

HYDROCHLORIC ACID.
FIG. 308. — RAIL WITH SOFT INTERIOR; ROLLED FROM TOO NEAR

TOP OF INGOT.
FIG. 309. — RAIL WITH BLOWHOLES AND SOFT INTERIOR.

of spots where segregation has occurred. After etching and exam-
ination, permanent records may be kept of the indications by
covering the etched surface with printer's ink and then pressing it
with a letter file on to a piece of cardboard or heavy paper. A few
examples of records made in this way are shown in Figs. 308
to 311.

THE METALLOGRAPHY OF IRON AND STEEL 457

REFERENCES ON THE METALLOGRAPHY OF IRON AND STEEL

For preparation of microscopic specimens:

180. W. Campbell. " Notes on Metallography." Columbia School

of Mines Quarterly, vol. xxv, No. 4 and vol. xxvii, No. 4.

181. F. Osmond and J. E. Stead. "Microscopic Analysis of

Metals." London, 1904.

182. Henry Le Chatelier. " Sur la Metallographie Microscopique."

Bulletin de la Societe d' Encouragement pour I'lndustrie
Nationals, April, 1896. (In English, see) The Metallograph-
ist, 1901, vol. iv. Also: Congres International de Liege,
Section de Metallurgie, vol. i, pages 255-284.

183. E. Heyn. See especially, Verhandlungen des Vereins zur

Befoerderung des Geiverbfleisses, 1904, pages 355—397.
Also: Stahl und Eisen, vol. xxvi, pages 8-16 and vol.
xxvi, pages 580-596.

184. J. E. Stead. Journal of the Iron and Steel Institute, 1894,

No. 1, page 292; and the Proceedings of the Cleveland In-
stitution of Engineers (England), 1900.

185. The Metallographist, vols. i to vi, 1898 to 1903. Edited by

A. Sauveur, Boston, Mass. Succeeded by The Iron and
Steel Magazine to 1906.

186. Paul Goerens. "Einfuhrung in die Metallographie." Halle

a. S., 1906.
For micro-photography:

187. Edward Bausch. "Manipulation of the Microscope." Third

edition. Rochester, N. Y., 1897.

188. Walter Bagshaw. "Elementary Photo-Micrography." Lon-

don, 1902.

189. Louis Derr. "Photography for Students of Physics and

Chemistry." New York, 1906.
1800. H. Behrens. " Das Mikroskopische Gefuege der Metalle und

Legierungen." Hamburg, 1894.
See also bibliography given at end of No. 116.

XIX

CHEMISTRY AND PHYSICS INTRODUCTORY TO
METALLURGY

Chemical Changes. — If a piece of coal be burned, it ceases to
exist as such. It disappears from sight, except for its slight res-
idue of ash, and apparently has been wiped out of existence for-
ever. Likewise, if a piece of steel be attacked by some acid, it
disappears as such, and only a coloration of the acid gives evidence
to the eye of the metal previously present. Lastly, if a piece of
bright iron or steel be exposed to the weather, it is soon converted
into reddish-brown rust, which bears but little resemblance to the
original metal. In the first example the solid coal has been com-
bined with oxygen of the air and converted to the form of an
invisible gas; in the second case, the iron has been combined with
the acid and water and converted into a liquid; in the third case
the iron has been combined with oxygen and converted into a
powder. In no case has there been any loss in total amounts or
weights, but only a difference in composition or substance. These
changes in composition are chemical changes.

Physical Changes. — Changes may occur in form or properties
without any change in composition: For instance, water may be
converted into ice by mere cooling, and no change in composition
will take place. Or it may be converted into steam by heating
and will still be composed of the same elemental constituents as
when it was in the form of ice or water. Iron may be liquid or
solid; it may be cold or hot; it may be magnetic or non-magnetic,
and all without change in substance. These changes in proper-
ties are known as physical changes. They may consist of changes
in form, strength, heat, light, magnetism, electricity — in fact,
everything but composition.

Relation between Chemical and Physical Changes. — Every
chemical change produces one or more physical changes. Thus,
chemical changes are always accompanied by a loss or gain of heat,

458

CHEMISTRY AND PHYSICS 459

and in the examples cited in paragraph 1, there were also observed
changes in form, in color, etc. Conversely, we shall see too that
physical changes are often the cause of starting chemical changes:
heat is necessary to start the chemical action of the burning of
fuel; pressure produces the explosion of dynamite; electricity
breaks up many chemical compounds; light produces the chemical
changes that make the photograph.

Chemical Compounds and Mechanical Mixtures. — When coal
is burned it is chemically united to the oxygen of the air, and the
gas formed contains properties entirely different from anything
belonging to either of the substances that compose it. Likewise
when steel is dissolved in an acid, or is converted to rust. This
is the essential characteristic of chemical action: that the sub-
stances acting upon each other lose their individual properties
and produce a new substance with different properties. Not so
when the substances are merely mixed together, no matter how
intimate that mixture may be. If finely ground sulphur be
mixed with finely ground iron no new properties are produced,
but if the mixture be heated until a chemical action takes place
and the two substances unite, a new compound is formed, with
new and different properties. Moreover, once the chemical union
has taken place the iron and sulphur cannot be separated except
by chemical means, whereas the mixture of the two could be sep-
arated by mechanical means, such as blowing the sulphur away
with a slow blast of air, or picking up the particles of iron with a
magnet.

Chemical Affinity. — The power that causes substances to unite,
and that holds them together afterward is known as 'chem-
ical affinity.7 Like gravity, magnetism and some other great
forces, its nature is not understood but its influence is very evi-
dent. Some substances have seemingly no affinity for each
other (for instance, mercury and iron will not form a compound),
while others have tremendous affinity — such as sodium and
oxygen, which cannot be brought into each other's presence with-
out uniting violently and generating much heat. Some sub-
stances have almost universal affinities, like oxygen which unites
with every other elemental substance known except one, while
others are relatively inert and form few compounds.

Conditions under which Chemical Action will Occur. — To start
chemical action it is sometimes only necessary to mix the sub-

460 THE METALLURGY OF IRON AND STEEL

stances, as, for instance, sodium and water; in other cases we must
apply heat, or pressure, or electricity, etc., even though the sub-
stances have great affinity for each other and unite vigorously
after the action is once started. In other cases chemical action
is very slow, no matter how started as, for instance, the rusting
of iron, the dissolving of rocks in water, etc. As a general thing
an increase in temperature increases all chemical affinities, but it
increases some faster than others.

The Elements. — In the universe there are millions of chemical
compounds, and these are mixed together to produce animal and
plant forms, the earth, sea, etc. All these compounds are different
combinations of only about eighty elemental substances, which
are known as 'the elements/ The compounds can all be sep-
arated into their component parts by chemical means (perhaps
aided by electricity), but the elements have so far resisted every
attempt to break them down into simpler substances, and they
are therefore considered as the simple substances and the basis
of all matter. These eighty elemental substances are therefore
of great importance, but some much more so than others, for only
eleven of them form the great bulk of the earth as we know it,
and the remainder are less abundant. The crust of the earth is
made up of the following elements:

Oxygen 47 . 29 per cent.

Silicon 27 . 21 per cent.

Aluminum 7.81 per cent.

Iron 5 . 46 per cent.

Calcium 3 . 77 per cent.

Magnesium 2 . 68 per cent.

Sodium . 2 . 36 per cent.

Potassium 2 . 40 per cent.

All others 1 .02 per cent.

The air is 77 per cent, nitrogen and 23 per cent, oxygen.
Pure water is 89 per cent, oxygen and 11 per cent, hydrogen.
Plant and animal forms are composed chiefly of combinations of
carbon, hydrogen, oxygen and nitrogen.

The air is the only one of the foregoing bodies in which the
elements are not severally united in the form of compounds of a
more or less complicated nature, and it is because the oxygen of
the air is in the free, or elemental, state that it is capable of per-
forming the chemical work of supporting life and burning fuels.

Summary. — So far we have learned that the whole universe,

CHEMISTRY AND PHYSICS 461

so far as we know it, is built up of only about eighty different sim-
ple substances, which we call elements, and which are sometimes
mixed together and sometimes chemically united, forming many
millions of different combinations. We have also learned that a
chemical compound has properties different from those of any of
the substances composing it, and that the elements of the com-
pound are held together by what is known as ' chemical affinity.'
We have learned that chemical action does not always take place
when two or more substances are mixed together, but we have
often to start it by heat, electricity or some other means. Lastly
we have learned that all chemical action is accompanied either
by a production or consumption of heat.

Synthesis. — W^hen two or more elements are chemically united
to form a compound, or when two or more compounds are chem-
ically united to form a further compound, the building-up process
is called a 'synthesis/

Analysis. — On the other hand, when a compound is separated
into the elements which compose it, the breaking-down process
is called an 'analysis'; or, another name for it is a 'decom-
position/

Definition of Metallurgy. — Metallurgy is the art of extracting
metals from their ores and adapting them to their intended service.
As iron occurs in the earth combined with one or more other ele-
ments, the process of extracting it consists in decomposing the
compounds and obtaining the metal from them. But this is not
all of the metallurgy of iron for, after it is extracted, it must be
adapted to the service for which it is intended, and for this pur-
pose various other elements are combined with it in proportions
depending upon the uses to which the metal is to be put. Even
one part of some of the added elements in ten thousand parts of
iron will make a great difference in its properties, so that these
syntheses must be performed with great care.

Qualitative and Quantitative Chemical Analysis. — Chemical
analysis is important in metallurgy in another connection, because
every substance put into the furnaces for smelting, and every
substance produced must be carefully ' analyzed ' in order that we
may know exactly what is in them and guide our operations ac-
cordingly. A 'qualitative analysis' will show what elements are
in any substance, while a 'quantitative analysis' will show how
much of each is present.

462 THE METALLURGY OF IRON AND STEEL

OXYGEN

Occurrence. — Oxygen occurs free in the air; combined with
hydrogen in water, and combined with silicon, with metals and
with many other elements in the earth's crust. It is the most
abundant element known to us.

Uses. — When breathed into our lungs oxygen performs cer-
tain chemical reactions which produce the heat that keeps us
alive, and which purify the blood. The only other abundant
constituent , of the air is nitrogen, and this is so inactive chem-
ically that it serves chiefly to dilute the oxygen. Oxygen also
reacts chemically with coal, coke, oil, gas and other fuels to
produce the heat for our fires, among which we must include
prominently the fires so necessary in metallurgy.

Preparation. — Oxygen is very easily prepared in a concen-
trated form in a great variety of ways. If, by means of combined
pressure and cold, air be converted into a liquid, its two com-
ponents may be separated by centrifugal force, or else the nitrogen
may be allowed to evaporate, leaving the liquid oxygen behind.
No chemical processes are necessary for this separation because
the elements are not combined. Where a compound exists, other
means must be employed: For example, water may.be decomposed
into oxygen and hydrogen by an electric current, and from 100
parts by weight of water we can always obtain 89 parts of oxygen
and 11 parts of hydrogen, nothing being lost in the change.

Chemical Action. — Oxygen is an odorless, colorless, tasteless
gas. At the ordinary temperature it forms few chemical reac-
tions, but, when heated, is one of the most active of the elements,
vigorously attacking, for example, hydrogen and carbon, as well
as their mutual compounds in the form of gases, when once a
chemical reaction is started with a match, electric spark or similar
means. It also attacks almost all the metals at a red heat, and
some of them at lower temperatures, — iron for instance, which
is coated with an ' oxide ' upon being heated about twice as hot
as boiling water. All the simple compounds of the elements with
oxygen go under the name of 'oxides/ and their formation is
accompanied with the production of heat.

Phlogiston Theory. — Centuries ago it was observed that lead,
when melted and exposed to the air, became an apparently new

CHEMISTRY AND PHYSICS 463

substance, and at the same time gained in weight. Starting with
the same weight of lead and allowing the action to go on until
complete, always resulted in the same gain in weight. The an-
cients believed that this action was the transfer from the fire to
the metal of a certain indefinable substance which they called
1 phlogiston/ They learned that if the lead and ' phlogiston ' were
later heated with charcoal, metallic lead was again produced,
and they said that the 'phlogiston' was driven out of it. We
now know that it was oxygen from the air that attacked the lead
when melted and formed 'lead oxide/ and that when the lead
oxide was heated with charcoal (carbon) the charcoal robbed the
lead of its oxygen and formed ' carbonic oxide ' leaving metallic
lead again.

Oxidation and Reduction. — When oxygen attacks an element
and forms an oxide the process is said to be an 'oxidation/ and
when an oxide is deprived of its oxygen and reduced to a metal
the process is said to be a 'reduction/ Iron is 'reduced' from
its ores, which are usually oxides. Oxidation and reduction are
therefore opposite actions in chemistry, the first adding something
to a substance and the second taking something away. At first
the terms were used in connection with oxygen alone, but are now
applied to adding or taking away anything. In metallurgy oxida-
tion and reduction are all-important, for everything that is re-
duced goes with the metals, and everything that is oxidized passes
away with the impurities. In the example cited where melted
lead was oxidized, the oxide separated itself from the metal just
as fast as formed, floating upon the top of it, and when the reduc-
tion with charcoal was effected, the lead separated itself from the
mass and dropped down into the bottom, of the furnace as fast as
it was reduced. This latter was then a metallurgical operation.

Combustion. — Combustion is a form of oxidation, in which a
' combustible ' is chemically united with oxygen, and, as we know,
this combustion, or burning, is our chief means of obtaining heat.
When there is just the right amount of both oxygen and com-
bustible to enter into combination, we have 'perfect combus-
tion/ but if the compound is formed and there is either oxygen
or combustible left over, we have 'incomplete combustion/ In-
complete combustion always means waste of heat. Thus, we
know that too much air passed through a fire-bed will carry waste
heat up the chimney, or if we have too little oxygen and carry a

464 THE METALLURGY OF IRON AND STEEL

combustible gas up the chimney, or leave unburned fuel on the
grate, we again waste heat.

THERMOCHEMISTRY

Chemical Energy. — If two or more substances combine and
produce heat, they have chemical energy, and they transform
this chemical energy into heat energy, which can be transformed
in turn into other forms and into work. Not all chemical actions
produce heat, but some are accompanied by a consumption of
heat, and therefore use up energy, or rather, they transform en-
ergy into chemical work. But the heat energy so transformed
into chemical work is not lost, for we can get it back again by
reversing the action. For instance, if lead oxide be reduced, a
consumption of heat occurs, and the same amount of heat will be
produced if we burn the lead and produce the oxide again. So
it is in every case: the heat produced by the formation of a com-
pound is the same in amount as the heat consumed when the
compound is broken up, and the heat produced by any chemical
reaction is the same in amount as that consumed when the action
is reversed. The science that treats of the heat changes accom-
panying chemical changes is called 'thermo-chemistry.' As a
general thing syntheses and oxidations are accompanied by a pro-
duction of heat, while decompositions and reductions are accom-
panied by an absorption of heat. Thermo-chemistry is very im-
portant in metallurgy because metallurgy is chemistry carried on
at high temperatures, and the metallurgist must know how much
heat is required for all his reactions, and by what reactions he may
obtain it.

Maximum Affinity. — If iron is heated in air the oxide of iron
is formed, and if this be mixed with powdered aluminum and the
action started with a fuse, the aluminum will rob the iron of its
oxygen and unite with it instead. The reason for this is that
aluminum has a greater affinity for oxygen than iron has. This
process of selection is a common^ one in chemistry and any sub-
stance will decompose a compound provided it can form a new
compound with greater chemical affinities. Likewise, if we have
a limited amount of a substance in the presence of two others, it
will combine with the one for which it has the greatest affinity to
the exclusion of the other. For example, if we have liquid iron

CHEMISTRY AND PHYSICS 465

and aluminum in the presence of oxygen, none of the iron will be
oxidized until all the aluminum has been oxidized.

Net Heat of Chemical Reactions. — When iron oxide is formed,
195,600 units of heat are evolved; when aluminum oxide is formed,
392,600 units of heat are evolved. Therefore, when aluminum
decomposes iron oxide and forms aluminum oxide instead, the
net heat effect of the reaction is to evolve (392,600-195,600 = )
197,000 units of heat. But suppose a reaction occurs in which
the decomposition brought about consumes more heat than the
compound formed produces? For example, if iron oxide is at-
tacked by carbon and deprived of its oxygen there will be a net
loss of heat, instead of a gain, because carbonic oxide generates
only 29,160 heat units in its formation, while 195,600 heat units
are consumed in the decomposition of iron oxide. Such a reaction
would not go on unless we constantly supplied heat to the bodies.
This is important in metallurgy because it means that when
we smelt iron oxide with coke (carbon) we cannot reduce the
iron unless we continually heat the bodies. In the case where
aluminum reduced the iron it was only necessary to start the
reaction, but with carbon smelting it is not only necessary to
start the action with heat, but also to continually supply the
(195,600-29,160 = ) 166,440 units of heat that are absorbed.

Temperatures. — Temperature is the degree of heat. There
are two scales by which it is commonly measured, known respec-
tively as the Fahrenheit and the Centigrade scale. In both of
these the freezing and boiling points of water are taken as the
standards. In the Fahrenheit scale, 32° is the freezing point of
water, and 212° is the boiling point. Each degree is therefore
y^-g- of this interval. In the Centigrade scale 0° is the freezing
point and 100° is the boiling point of water. Each degree is there-
fore y^-g- of this interval. One degree Centigrade equals 1|°
Fahrenheit. A table of comparison is shown in Table XXXIV,
page 487.

Heat Units. — The amount of heat in a body is different from
its temperature ; it takes much more heat to raise a pound of water
to 200° F. than it does to raise a pound of iron, and more to raise
iron than copper, lead or gold. There are two standards by
which amounts of heat are measured: A British Thermal Unit
(known as B. T. U.) is the amount of heat required to raise one
pound of water one degree Fahrenheit; a calorie is the amount

466 THE METALLURGY OF IRON AND STEEL

of heat required to raise one gram of water one degree Centigrade.
In both cases the water must start at its maximum density, which
is at 39.1° F. ( = 4° C.). One B. T. U. equals 252 calories.1 In
this book I shall generally use calories, as that is the ordinary
system in scientific work. One Calorie = 1,000 calories.

Summary. — We have now learned that metallurgy is chemis-
try at high temperatures, and that when iron is reduced from its
ores, we must decompose the ores, which consumes a great deal of
heat. We have also learned how heat is obtained from chemical
reactions, and chiefly from combustion. We have learned that
oxygen forms oxides with many of the elements, and that some
of the elements have a greater affinity for oxygen than others,
so that they will keep, or even take, the oxygen away from them.
Lastly we have learned that any reduced substances will join with
the metal in our furnaces, and any oxidized ones will join with
the impurities, and that the reduced substances will Jiot ordinarily
mix with the oxidized ones.

CHEMICAL EQUATIONS

Combining Weights. — When metals unite in compounds they
always do so in certain definite porportions. The compound of
oxygen and iron contains 16 parts of oxygen and 56 of iron; that of
oxygen and calcium, 16 parts of oxygen and 40 of calcium; that
of iron and sulphur, 56 parts of iron and 32 of sulphur; that of
calcium and sulphur, 40 parts of calcium and 32 of sulphur. These
characteristic combining weights are known as ' atomic weights'
for a reason that will be evident shortly. A list of about one-half
of the known elements with their atomic weights is given in the
table on the next page, those which are of the least importance
in metallurgy of iron and steel being omitted, and those which
are of greatest importance being printed in small capitals.

The Atomic Theory. — The atomic theory supposes that all of
the elements are made up of a myriad of tiny particles called atoms.
The atoms are the smallest particles of matter that can exist; too
small to be even conceived of by the imagination, and yet all
having a definite size and weight and incapable of being divided
into finer particles. All the atoms in any one element are alike
in composition, size, and weight, but differ in these three proper-
ties from the atoms of all of the other elements. When two or

1 See page 171. One pound avoirdupois =453.59 grams; one ounce
avoirdupois =28.3495 grams.

CHEMISTRY AND PHYSICS

467

TABLE XXXIII

MAGNESIUM Mg 24

MANGANESE Mn 55

Molybdenum Mo 96

NICKEL Ni 59

Nitrogen N 14

OXYGEN O 16

PHOSPHORUS P 31

Potassium K 39

SILICON Si 28.4

Silver Ag 108

Sodium Na 23

SULPHUR S 32

Tin Sn 118.5

Titanium Ti 48

TUNGSTEN Wo 184

VANADIUM V 51

Zinc.. Zn 65.4

ALUMINUM Al 27

Antimony Sb 120

Arsenic As 75

Barium Ba 137.4

Bismuth Bi 208

Boron B 11

CALCIUM Ca 40

CARBON C 12

Chlorine Cl 35.5

CHROMIUM Cr 52

Cobalt Co 59

Copper Cu 63.6

Fluorine F 19

Gold Au 197

HYDROGEN H 1

Iodine I 127

IRON Fe 56

Lead Pb 207

These combining weights are used in the laboratories of chemical analysis,
in calculating furnace burdens and similar work. A list should be kept for
convenient reference.

more elements combine chemically the atoms of one are locked
with bonds of chemical affinity to the atoms of the others, and
thus the weights of each element entering the compound is in
proportion to the weight of its atoms.

Chemical Symbols. — In writing the elements it is customary
to represent each by one or two initial letters, instead of writing
the name out in full. This is a sort of shorthand of chemistry.
The representative letters are taken from the Latin name of the
elements, and the same symbols are employed in every civilized
country of the world. The symbol for iron is Fe, because the
Latin name for iron is ferrum. That for aluminum is Al; for cal-
cium, Ca; for carbon, C; for hydrogen, H; for magnesium, Mg;
for manganese, Mn; for oxygen, O; for phosphorus, P; for silicon,
Si; and for sulphur, S. These eleven symbols should be learned,
as they will be used frequently in the body of the work. Others
are shown in Table XXXIII.

Multiple Proportions of Atomic Weights. — It does not require
much thought to see that atoms may combine in more than one
way. For instance, one atom of carbon may combine with one
atom of oxygen, and again one atom of carbon may combine with
two atoms of oxygen. In the first compound there will be 12
weights of carbon and 16 of oxygen; in the second, 12 weights of
carbon and 32 of oxygen. Likewise, one atom of iron may com-

468 THE METALLURGY OF IRON AND STEEL

bine with one atom of oxygen, or two atoms of iron may combine
with three atoms of oxygen. Each of these compounds will have
different properties. It is possible to represent these compounds
in a very simple way by using the symbols for the elements, for
each symbol designates one atom of the element. To represent
the first compound of carbon and oxygen we write their symbolic
letters together — thus, CO. To represent the compound con-
taining one atom of carbon and two of oxygen we write, COO, or,
CO2. To represent the first iron oxide we write — FeO. To rep-
resent the second one, Fe2O3. Then the formulae for these com-
pounds tell us not only what elements make up the compound, but
also how much of each is present. For example, in the first iron
oxide we have 56 parts of iron and 16 parts of oxygen; in the sec-
ond we have 112 parts of iron and 48 of oxygen.

Molecules. — When two or more atoms are held together by
chemical affinity the particle formed is known as a. molecule. The
symbol, CO, represents a molecule of carbonic oxide; Fe2O3 rep-
resents a molecule of iron oxide.

Chemical Equations. — Chemical shorthand may be used to
represent chemical reactions, and will indicate at a glance what
is taking place. A synthesis will be shown as follows:

C -f O CO;

Carbon and oxygen produce carbonic oxide.

The decomposition of water would be written:

H2O 2H + O;

Water gives hydrogen and oxygen.

The reduction of iron oxide by carbon:

FeO + C Fe + CO;

Iron oxide and carbon give iron and carbonic oxide.

The reduction of iron oxide by .aluminum:

Fe2O3 + 2A1 = 2Fe + A12O3;

Iron oxide and aluminum give iron and aluminum oxide.

Indestructibility of Matter. — In the equations written above
it will be noticed that there are always as many atoms of each ele-
ment on the left-hand side of the equation mark as on the right.
This is in accordance with the fundamental law of chemistry that
matter can neither be destroyed nor created. If we combine car-

CHEMISTRY AND PHYSICS 469

bon with oxygen the weight of carbonic oxide formed is exactly
equal to the weight of carbon and oxygen together. Also, if we
decompose water the total weight of hydrogen and oxygen will be
equal to the weight of water from which it came.

HYDROGEN

Occurrence. — Hydrogen forms 11 per cent, of water, which is
a compound of hydrogen and oxygen, whose molecules contain
two atoms of hydrogen and one of oxygen, so that they have the
formula, H2O. Hydrogen also occurs in all living forms.

Properties. — Hydrogen is a colorless, tasteless, odorless gas,
and the lightest substance known, so that it would be very useful
for filling balloons except for its cost. It has a high chemical
affinity for oxygen and a stream of it when ignited will burn read-
ily in the air and produce water vapor with the evolution of 58,060

calories :

2 H + O = H2O (+ 58,062 cals.).

It is therefore a good combustible, and an impure form of it is
indeed one of our important fuel gases, going under the name of
'water gas/ It is also a good 'reducing agent'; that is, it will
reduce substances by taking their oxygen away.

Hydrocarbons. — Hydrogen has a strong affinity for carbon,
and forms with it a long series of compounds known as 'hydro-
carbons/ of which there are about two hundred different com-
binations. These form the basis of mineral oil, or petroleum,
from which we get kerosene, gasoline, naphtha, benzene, lubricating
oils, vaseline, paraffine, etc. The ' light hydrocarbons ' are found
in kerosene, gasoline, etc., while the 'heavy hydrocarbons' are
found in the less volatile oils. Some of the more important com-
pounds are as follows: Methane, whose molecule has the formula,
CH4, is the chief constituent of natural gas; when it burns the fol-
lowing reaction takes place: —

Ethylene, C2H4, is a heavier hydrocarbon than methane because
its molecule contains more atoms of carbon, while acetylene,
C2H2, is heavier still and is the most powerful illuminating gas
known. Benzene, C6H6, has the same relation between the atoms
of hydrogen and carbon as acetylene, but a different number of

470 THE METALLURGY OF IRON AND STEEL

t'hem in the molecule, so that it is an entirely different substance
with different properties.

Thermo-chemistry of the Hydrocarbons. — When methane burns
we get the following reaction: —

CH4+4O = 2H20 + CO,. (+19 1,270 cals.)
-22,250 cals.+ 116,320 cals. + 97,200 cals.

If we put the heat of combination under each of the compounds
then we can readily calculate the net heat produced or consumed
by the reaction, because all the compounds on the left of the
equation mark are decomposed and all those on the right are
formed. Therefore the sum of the heats on the left is to be com-
pared with the sum of the heats on the right. If the right-hand
sum is greater, heat is produced ; if the left-hand sum is greater,
heat is destroyed. In the burning of methane 191,270 calories
are produced and we therefore place ( + 191,270 cals.) at the end
of the equation.

Let us now consider for comparison the reduction of tin oxide
by carbonic oxide, as follows :

SnO2 + CO = SnO + CO2 (-2,560 cals.)
-141,300 cals. - 29,160 + 70,700 cals. + 97,200 cals.

In this case we find that the sum on the left is greater, and 2,560
calories are consumed. We therefore place (—2,560 cals.) next
to the equation.

Preparation of Hydrogen. — The cheapest method of obtaining
hydrogen is by decomposing water. This may be done as follows:

Na + H2O = H + NaOH(+ 33,700 cals.)
(sodium) -69,000 ' + 102,700

We see that this is a heat-producing reaction. Another way:

C + H2O = 2 H + CO (-28,900 cals.)
-58,0631 +29,160

We do this by passing water vapor over red-hot carbon, but the
reaction consumes heat so we must frequently heat the carbon
or the reaction will not go on.

Still another way is to pass an electric current through a body
of water. Hydrogen gas appears at one electric connection and
oxygen gas at the other. This process is known as the 'electrol-
ysis' of water, and it is an operation in ' electro-chemistry/ If

1 The heat of formation of water is 69,000 cals. in liquid, form and 58,060
in gaseous form. For other heats of reactions, see page 124, No. 53, page 125.

CHEMISTRY AND PHYSICS 471

it is carried out perfectly the amount of electric energy passed
into the water will be equal to the amount of heat energy required
for the decomposition, — that is, 69,000 1 calories.

Summary. — Now we have learned that each element is con-
stituted of infinitesimal particles called atoms which are all identical
in weight and composition, and that when elements form com-
pounds the atoms of one are joined to those of the other by bonds
of chemical affinity. We have also learned that the atoms of ele-
ments are represented by letter symbols, and that we can express
reactions between them by putting the symbols together in mole-
cules and then showing how they break up and change places to
form other molecules, and that there are no atoms and no weight
lost or gained in any of these changes, but there is a gain or loss
in heat energy to correspond exactly with each loss or gain of
chemical energy. And we have seen how the loss or gain of
heat may be determined by reckoning the total heat of com-
pounds decomposed as heat lost, and the total heat of com-
pounds formed as heat produced. From this point we shall go
on to consider the important elements more in detail as to their
chemical behavior.

ELEMENTS, COMPOUNDS, AND RADICALS

Metallic and Non-metallic Elements. — All the metals are ele-
ments, but some of the elements are not metals; for instance, we
know without being told that oxygen is not a metal. The dis-
tinction between metallic and non-metallic elements is not very
clear. It once was considered that all elements which looked like
metals should be classified as such; they were said to have 'me-
tallic luster/ And all others were classified as non-metals. But
this classification has been shown to be deceiving, and a chemical
one has taken its place : Now all the elements that form ' bases '
are classified as metals, and all those that form ' acids ' are classi-
fied as non-metals.

Acids and Bases. — Acids are generally sharp to the taste and
have certain other characteristic chemical properties of which one
of the most distinctive is their ability to turn litmus a red color.
Bases have certain other characteristic properties of which one
of the most distinctive is their ability to turn litmus a blue color.
1 See footnote, page 470.

472 THE METALLURGY OF IRON AND STEEL

Acids will destroy the characteristic properties of bases and neu-
tralize them, and conversely, bases will neutralize acids. Acids
and bases have strong affinity for each other and either one will
attack the other if opportunity offers. That is why a basic slag
will attack an acid furnace lining, or an acid slag will attack a
basic lining. Metallurgical acids are ' anhydrous ' (water-free).

Salts. — When an acid and a base just neutralize each other
they form what is known as a salt. Let us dissolve 36.5 grains of
hydrochloric acid, HC1, in water, and put a piece of paper soaked
in litmus in it; the paper will at once turn a brilliant red. Now
let us dissolve 40 grains of caustic soda, NaOH, in another vessel.
Caustic soda is a strong base ; if we put one end of our piece of red
litmus paper in the solution it will turn blue. Now let us pour
the acid solution into the basic solution, and we will get the
following reaction: —

HC1 + NaOH = NAC1 + H2O.
36.5 40 58.5 18

Now, 36.5 is the molecular weight of HC1 (because one atom of
hydrogen weighs 1, and one of chlorine weighs 35.5), and 40 is the
molecular weight of NaOH (one atom of Na=23; one of O = 16,
and one of H = 1) ; therefore there must be as many molecules of
HC1 present as of NaOH, and a complete neutralization will occur.
Moreover it will be seen that the total weight of atoms at the right
of the equation mark is the same as that at the left. This neu-
tralization forms • ' sodium chloride/ which is our common table
salt. The salt will be dissolved in the water used in the experi-
ment. If now we put in this salt solution the piece of litmus
paper, one end of which is red and the other blue, it will not
change its colors at all.

Radicals. — Let us consider the neutralization of caustic soda
by sulphuric acid, H2S04*.

2NaOH + H2SO4 = Na2SO4 + 2H2O.
80 98 142 36

In this reaction the S04 has changed places with the OH. When
two or more atoms are joined together and travel around in com-
pany in this way, acting as if they were inseparable, they are
called 'radicals/ In this reaction, the S04 is called an 'acid
radical/ and the OH is called the 'hydroxide radical.' For the
time being these radicals act as if they were elementary substances.

CHEMISTRY AND PHYSICS 473

Valence. — Let us consider four compounds with hydrogen,

as follows:

C1H OH2 NH3 CH4

Hydrochloric acid, water, ammonia, methane.

One atom of hydrogen can hold one of chlorine, but it takes two
to hold one atom of oxygen, three to hold one of nitrogen, and
four to hold one of carbon. Conversely, one of chlorine can hold
one of hydrogen, one of oxygen can hold two of hydrogen, one of
nitrogen, three, and one of carbon, four. This capacity for hold-
ing numbers of atoms is called 'valence/ In the compounds
shown above, chlorine is uni-valent, oxygen is bi-valent, nitrogen,
tri-valent, and carbon, quadri-valent. In each case hydrogen is
uni-valent; indeed hydrogen is established as the standard of
valency, with a holding power of one. We can determine the
valence of other elements by learning how many atoms of hydro-
gen they will hold, or, if they do not form a compound with hy-
drogen, we can compare them with some other element that does.
For example, calcium forms a very common oxide, CaO, known
as lime. In lime calcium holds one atom of oxygen; but it takes
two atoms of hydrogen to hold one atom of oxygen ; therefore cal-
cium is bi-valent.

Chemical Stability. — We have already seen enough com-
pounds to know that the valence of several of the elements is not
a constant quantity. For example, carbon and hydrogen atoms
unite in nearly two hundred different combinations. Likewise,
iron forms FeO and Fe203. In the first it has a valence of two;
in the second, of three. But there is a difference in the stability
of these compounds; the oxide, FeO, can only exist under strong
reducing conditions, and will take on more oxygen with the least
opportunity. Iron forms two sulphides, designated as FeS and
FeS2. In the second compound it has a valence of four, but this
sulphide is not as strong a one as the other, and the second atom
of sulphur may be driven off by heating it slightly: FeS2 = FeS + S.

Ferrous and Ferric Compounds. — The oxide, FeO, is called
' ferrous oxide'; while FeS is called ' ferrous sulphide.' The ox-
ide, Fe203, is called 'ferric oxide/ while FeS2 is called 'ferric sul-
phide/ Manganese forms two oxides: MnO is called 'manga-
nous oxide/ and Mn02 is called 'manganic oxide.' So with all
compounds; that having the lower valence is given the suffix
-ous, and that with the higher valence, -ic.

474 THE METALLURGY OF IRON AND STEEL

Mono-, Bi-, Tri-, etc. — FeS is also called ' iron mono-sul-
phide/ from the Latin, meaning one; FeS>2 is sometimes called
'iron bi-sulphide.' MnO is called 'manganese monoxide/ and
MnO2, ' manganese bi-oxide ' (di-oxide is sometimes used instead
of bi-oxide). H20 is 'hydrogen monoxide'; H2O2 is 'hydrogen
di-oxide. ' Fe20s is called 'iron sesqui-oxide/ from the Latin
meaning three halves.

Sub- and Per-. — When an element has a very low valence
it is given the prefix 'sub-/ and when it has an unusually high
valence it has the prefix 'per-.' For example, Fe2O (if such a
compound were capable of forming) would be called 'iron sub-
oxide'; while FeO2 (if possible) would be called 'iron peroxide.'
H2O2 is often known as 'hydrogen per-oxide.' In the same way
we may have sub-sulphides (Fe7S8 is called ' iron sub-sulphide ') ,
sub-carbides, etc.

Oxidizing Agents. — When additional atoms are put in the
molecules of a compound it is said to be oxidized. For example,
FeO will be oxidized to Fe2O3 (2 FeO + O = Fe203) ; FeS will be
oxidized to FeS^ Oxidation can only be produced by means
of some 'oxidizing agent.' The commonest oxidizing agent in
metallurgy is the oxygen of the air, and the next most important
in iron and steel processes is Fe2O3, and slags very rich in Fe2O3:

3 Si + 2 Fe2O3 = 3 SiOa + 4 Fe
6 P + 5 Fe2O3 - 3 P2O5 + 10 Fe

Another important one is carbon di-oxide :

Redwing Agents. — When atoms are taken out of the mole-
cule of a compound it is said to be reduced. Reduction can only
go on in the presence of some ' reducing agent/ The commonest
reducing agent in metallurgy is carbon in the form of coke, char-
coal, etc.

Another one is carbon monoxide, and another is hydrogen:

Manganese is also a reducing agent for iron:

= Fe + MnO.

CHEMISTRY AND PHYSICS 475

CHEMICAL REACTIONS AND COMPOUNDS

Organic and Inorganic Chemistry. — The chemistry of living
organisms, such as plants, animals, etc., is a very complex sub-
ject, and quite distinct from inorganic chemistry. Because car-
bon enters into all organisms we may describe organic chemistry
as the chemistry of the carbon compounds. Inorganic chemistry
is the chemistry of the metals and of compounds in which carbon
enters in relatively small proportions. Inorganic chemistry is
the only one that concerns metallurgists especially.

Wet and Dry Chemistry. — In the analytical laboratories they
perform their chemical reactions by dissolving everything in
water and so getting them in the liquid form, because solids do not
unite with each other rapidly, and gases are not easily controlled.
This branch of chemistry is known as 'wet chemistry/ In iron
and steel metallurgy, however, we get everything in liquid form by
melting it. This is known as 'dry chemistry/ The reactions that
take place in dry chemistry are the same in principle as those of
wet chemistry. The chief difference is that we cause substances
to react directly instead of dissolving them all in water.

Carbon. — Carbon occurs in the earth in the crystallized form
as graphite and as diamonds. Of these the diamond is the purer
variety, but both may be considered as pure carbon in different
forms. The element may be obtained in a massive, or uncrys-
tallized, form by burning organic matter, such as wood, when a
black residue of carbon (charcoal) will be left. The most abundant
occurrence of carbon is, however, in combination with other ele-
ments in the various forms of living matter, and also in inorganic
compounds with metals, known as carbonates, such as the car-
bonate of lime, CaCOs, called limestone, or, when in the crystallized
form, marble.

When bituminous coal is burned in a smothered sort of way,
that is, in the absence of much air, a silvery-gray residue is left
which is an impure form of carbon, called coke. Crystals of
graphite often are present on the surface of coke. This coke is
one of the most important of all metallurgical reducing agents,
as well as fuels, Carbon also forms a number of hydro-car-
bons which are used in the form of gases as reducing agents and
fuels, because both their carbon and hydrogen will unite with
oxygen.

476 THE METALLURGY OF IRON AND STEEL

Carbon forms two oxides, — CO and C02. The first combina-
tion is accompanied with the production of 29,160 calories, and
the second, 97,200 calories. The formation of CO is not complete
combustion, because it will itself be further oxidized : CO + O = C02,
with the evolution of 68,040 calories. The heat of formation of
C + O together with that of CO + O is just equal to that of the

reaction :

C + 2 O = CO2 (+ 97,200 cals.) ;
29,160 + 68,040 = 97,200.

Carbon di-oxide, CO2, is an acid radical and unites with many
bases to form 'carbonates/ of which the commonest are those
of calcium, CaCO3, magnesium, MgCO3, and sodium, Na2CO3.
As CO2 is very volatile, the carbonates may be decomposed by
heat, which drives the CO2 off as a gas:

Na2CO3 = Na2O + CO2.

Chemical Behavior of Iron. — Iron is attacked by many of
the wet acids, — sulphuric, nitric, hydrochloric, acetic, etc.
When heated it is attacked by oxygen, and also when cold pro-
vided the air is damp. At a red heat, iron decomposes water
vapor (2Fe + 3H2O = Fe2O3 + 6H). It has a high affinity for
oxygen, and also for small amounts of carbon, silicon, sulphur,
phosphorus and hydrogen. The last-named gas will penetrate
solid iron very readily at a red heat and form a compound with it.
Iron practically never occurs in the earth except combined with
oxygen or some other elements.

Chemical Behavior of Silicon. — Silicon has a high affinity for
oxygen, with which it forms a very common oxide, SiO2, which
is known as silica. This compound is decomposed with great
difficulty; the following reaction takes place only when we get
to the very high temperature in the hearth of the iron blast fur-
nace :

SiO2 + 2C = Si + 2 CO (-121,680 cals.)
- 180,000 + 2 X 29, 160 = 58,320.

We must remember the difference between silicon and its oxide,
silica. Silicon never occurs uncombined in the earth, but silica
is the most abundant constituent known to us. Quartz is a crys-
tallized form of pure silica, while flint, jasper, agate, etc., are
uncrystallized forms. Opal is silica combined with water.

CHEMISTRY AND PHYSICS 477

Silicates. — Silica is the great acid of dry chemistry, and when
in the melted condition will neutralize every base with which it
comes in contact, forming a series of salts known as ' silicates.'
The great bulk of the earth's rocks are either pure silica or silicates
of the different metals, and all metallurgical slags are silicates.
The mono-silicate of iron has the formula, — Fe2SiO4. But it is
more commonly written, — (FeO)2Si02, which is the same as
Fe2O2Si02. The series of commonest iron silicates are given

below:

FeO.SiO* Sub-silicate.
(FeO)^SiOa Mono-silicate.
(FeO)2(SiO2)3 Sesqui-silicate.
FeO(SiO2)2 Bi-silicate.
FeO(SiO2)3 Tri -silicate.

With lime, CaO, and magnesia, MgO, a similar series is formed,
but the silicates of alumina, A1203, are more complicated in com-
position. The different metallic silicates have the property of
dissolving in each other when melted, and of dissolving the oxides
of metals, and various other oxidized substances, but not of dis-
solving metals or reduced substances.

Feldspar. — With potassium and aluminum silica forms a
series of silicates known as the feldspars, which are common con-
stituents of the earth's crust. The feldspars are chiefly important
because when reduced to powdered form they become clay, which
has the peculiar property of becoming plastic when moistened.
The purer clays melt at a very high temperature and are therefore
used as the bond to hold together the material for the linings of
furnaces, but the clays that contain much potassium or sodium
melt relatively easily, and are not so ' refractory.' Clays contain
a certain amount of water of crystallization, that is, water chem-
ically combined with the molecule of the silicates. If they are
heated so hot that this water of crystallization is driven out of
them, they will not again become plastic.

Chemical Behavior of Aluminum. — Aluminum has great affin-
ity for oxygen and is therefore used, like silicon, for the purpose
of de-oxidizing steel:

3 FeO + 2 Al = 3 Fe + A12O3.

Indeed aluminum retains its oxygen more tenaciously than silicon,
and even the highest temperature of our fuel furnaces does not

478 THE METALLURGY OF IRON AND STEEL

effect its reduction. The oxide of aluminum, A12O3, called alu-
mina, is a common constituent of rocks, and when nearly pure is
used as an ore of the metal, its reduction being effected in electric
furnaces. Alumina is very refractory, that is, it will stand a high
temperature without melting, and is neutral in character, that is,
it is attacked neither by acid nor basic slags. It is therefore used
as a neutral lining for some furnaces. Alumina is also useful in
blast-furnace slags; in acid slags it acts as a base, and in basic
slags as an acid, rendering the slags more fluid.

Chemical Behavior of Manganese. — Manganese has a higher
affinity for both oxygen and sulphur than iron has, and is there-
fore used as a de-oxidizer and de-sulphurizer of iron and steel. If
sufficient manganese is present, and the metal bath kept liquid a
sufficient time, neither oxygen nor sulphur will be found com-
bined with iron:

n = MnO+Fe;
n = MnS+Fe.

Chemical Behavior of Sulphur. — Sulphur is found in the earth
native (that is, free from combination), especially in volcanic re-
gions, and also combined with metals as sulphides. Iron bi-
sulphide, called 'iron pyrites/ FeS^ is very abundant and is the
chief source of sulphuric acid manufacture, while the sulphides
of copper, lead, and zinc are the principal commercial ores of
those metals. Iron sulphides are not used so much as ores on
account of the expense of ridding the iron of sulphur, which is
very harmful to it. Sulphur readily combines with oxygen at a
slightly elevated temperature to form S02 and SOs, and these com-
bine with water to form sulphurous and sulphuric acids (H2O-f
SO2 = H2SO3; and H2O + SO3 = H2SO4) . In wet chemistry sul-
phuric acid attacks metals to form sulphates:

2Fe + 3H2SO4 = 6 H + Fe2SsO12 or Fe2(SO4)3.
Ca + H2SO4 = 2 H + CaSO4.

Phosphorus. — Phosphorus occurs in nature usually as metallic
phosphates, and chiefly as phosphate of lime, Ca2(PO4)2, a natural
mineral to which the name of apatite is given. It is in this form
that it ordinarily gets into the blast furnace with the iron ores
which it accompanies in the earth. Phosphate is necessary to
animal and vegetable life and a good part of bones and living

CHEMISTRY AND PHYSICS 479

organisms are composed of it. The phosphates that will dissolve
easily are therefore valuable fertilizers. Consequently certain
slags which are used to remove the phosphorus from steel can be
sold for fertilizing purposes.

Phosphorus acts the part of an acid-forming element, and the
phosphate radical will form salts with many metallic oxides, but
especially with iron oxides, magnesium oxide and lime, for the
latter of which it has great affinity. But it is a weaker acid
than silica, and silica will drive the phosphate radical away from
all the basic radicals until the silica has completely satisfied itself.
For this reason phosphorus cannot be combined in slags unless
there is a superfluity of bases present over the amount necessary
to practically surfeit the silica. In our iron slags this means
usually at least 40 per cent, of lime plus magnesia plus iron oxide.

Calcium and Magnesium. — Calcium forms a very common
oxide, CaO, known as lime, and magnesia forms a similar one,
MgO, called magnesia. These occur in nature chiefly combined
with carbonic acid to form carbonates; CaCOa is called limestone,
and MgCOs, magnesite. The two carbonates often occur com-
bined together in a compound having the formula, — (Ca.MgJCOs.
This type of formula is used to indicate that the calcium and
magnesium replace each other in the carbonate in almost any
relative proportion. The natural rock, (Ca.Mg)CO3 has the miner-
alogical name of ' dolomite/

Limestone is used as a material to add to the charge of the
iron blast furnace because the carbonic acid is driven off in the
upper levels of the furnace as soon as it begins to become hot
(CaCOs + heat = CaO + C02) and the lime so produced serves as a
base in the blast-furnace slag. Burnt limestone, that is, limestone
from which the carbonic acid has been driven off by heat, is also
added to the slags made in some of the steel furnaces, in order to
increase their basicity. Lime has the peculiarity of absorbing
moisture from the air and forming a hydrate [CaO + H2O =
Ca(OH)2]. This is known as 'slacking/ and it causes the lime
to lose its coherence. For this reason furnace linings cannot be
made of it.

Magnesia is made by burning magnesite (MgCOa + heat = MgO
+ C02), and this is much used for making the basic linings of
furnaces. Burnt dolomite is used for patching basic furnace lin-
ings, but it is not as durable as magnesia for the original lining.

480 THE METALLURGY OF IRON AND STEEL

CHEMICAL SOLUTIONS

Chemical Compounds, Mechanical Mixtures, and Chemical
Solutions. — We have learned that the differences between me-
chanical mixtures and chemical compounds are: (1) The proper-
ties of compounds are different from those of its components ; (2)
the formation of a compound is attended with the production
of heat ; (3) the components of a compound are held together with
bonds of chemical affinity, and (4) the components always form
the compound in the same definite proportions. There is another
class of combinations different from both compounds and mix-
tures, and known as solutions. These have some of the charac-
teristics of compounds and also of mixtures. (1) The properties
of a solution are different from those of its components, but not
to as marked a degree as is the case with compounds; (2) its for-
mation is attended with the production or consumption of heat;
(3) the components of the solution are held together by bonds of
chemical affinity, and can only be separated by chemical means,
or by electricity, but (4) unlike compounds, and to a limited de-
gree like mixtures, the components of a solution may vary widely
in proportions. In some cases the variation is infinite, as with
melted gold and silver, which will dissolve in each other in any
proportion; likewise, with melted copper and silver. In other
cases the limit of solubility is very narrow, as in the case of
melted iron, which will dissolve only about 5 per cent, of car-
bon, while carbon will dissolve apparently only 1 per cent, or so
of iron.

Just as some substances resist all our efforts to make them
combine chemically, so others refuse to dissolve. For instance,
several salts and liquids will not dissolve in water, the best solvent
known, or rather, dissolve to such a slight degree that, for practical
purposes, it may be neglected. The same is true of iron and mer-
cury, melted lead, and zinc, etc.

Essence oj Solubility. — Just what the nature of the state of
solution is cannot be told at present, but the atoms, or molecules,
of the dissolved substance, known as the ' solute/ seem to be
held by the molecules of the solvent. One striking difference ex-
isting between a solution and a mixture is that the solute seems
to occupy no space. If we mix with hot water one-quarter of its

CHEMISTRY AND PHYSICS 481

weight of table salt, the level of the water will rise in the contain-
ing vessel an amount equivalent to the bulk of the salt, but, as
soon as the water dissolves the salt, it will fall back to its original
volume. In short, we now have a quarter more weight of ma-
terial in the same bulk, so that the specific weight of the mass
increases 25 per cent. In several ways which we have not space
here to discuss we can know of the presence of a greater number
of molecules than ordinary in the same space when two or more
substances are dissolved in each other; for instance, 'osmotic
pressure/ surface tension. Metals dissolved in each other are
heavier than the same bulk of any of the metals alone.

Precipitation. — If we dissolve 27 per cent, of table salt in hot
water and then allow the water to cool, some of the salt will fall
out of solution again and crystallize. This action is called t pre-
cipitation.' It is one of the most important actions in chemis-
try. We may cause a precipitation by another means : If we have
as much of any salt dissolved in water as it will take, and then
add to the solution a more soluble one, the water will dissolve the
new salt and precipitate the old one in corresponding amount.
The same applies in metallic solutions: If we have 5 per cent, of
carbon dissolved in iron and then add some metallic silicon, the
iron will precipitate graphite in flakes of 'kish/ and dissolve sili-
con. The commonest method of precipitating elements in wet
chemistry is by producing chemical change: Suppose we have
some table salt (NaCl) dissolved in water and add just enough
silver nitrate (AgNOs) to react with it; we form silver chloride
which is insoluble, and which precipitates almost instantaneously :

NaCl + AgNO, - AgCl + NaNO,.

This is only a partial precipitation, because sodium nitrate is still
left in solution, but is often of great service.

Solubility and Temperature. — As a general thing the higher
the temperature the greater amount of solute can be dissolved in
any given solvent. This rule is not universal, but usually applies
in practical metallurgical chemistry. Five per cent., or even
more, carbon will dissolve in iron at a high temperature, but some
of this precipitates as the metal cools near its solidification tem-
perature.

Alloys. — Metallic alloys are dependent upon solution for their
formation. Two metals which will not dissolve cannot be made

482 THE METALLURGY OF IRON AND STEEL

to form alloys, and, in practice, all alloys are made by dissolving
melted metals in each other. The only exception is certain al-
loys made by dissolving solid metals in each other under great
pressure.

Nature of Slags. — Slags are molten solutions, and as a gen-
eral rule, they will dissolve all the oxidized substances with which
they come in contact in the furnace (except of course the furnace
lining itself), and will precipitate all the reduced substances. For
example, metals will be precipitated as fast as reduced from their
combinations in the ores: phosphorus, if oxidized by being com-
bined with some base as a phosphate, will be dissolved, but, if
silica takes the base away from it, the reduced phosphorus will be
precipitated again.

SOME PRINCIPLES OF PHYSICS

Dalton and Gay Lussac Law. — When elements or compounds
are in the gaseous form they all expand and contract at the same
rate. In brief, every gas expands -^j-j of its volume for every
degree Centigrade that the temperature is increased, and ^T for
every degree Fahrenheit. Air or steam at 273° C. (491° F.) has
twice the volume of the same weight of air or steam at 0° C.
(32° F.). The amount of increase or decrease in volume can be
calculated by multiplying its volume at 0° C. by the number of
degrees increase or decrease in temperature and then dividing
the product so obtained by 273, if Centigrade units are used, or
by 491.4 if Fahrenheit.

Boyle's Law. — The volume of gases also increases or de-
creases in proportion to lessening or increasing the pressure. Or-
dinarily gases are under the atmospheric pressure, which is 15
pounds per square inch. If we increase the pressure upon them
to 30 pounds, their volume becomes one-half; if we increase it to
45 pounds, it becomes one-third, etc.

Specific Gravity. — The specific gravity of bodies is the rela-
tive weights of a unit volume. For example, a cubic inch of iron
weighs nearly 8 times as much as a cubic inch of ice, and a cubic
inch of lead or platinum weighs more still. The specific gravity
of solids and liquids are usually compared with water as a standard,
and water at its temperature of maximum density — 4° C. (39.2°
F.) — is given an arbitrary value of 1. Gases are usually com-

CHEMISTRY AND PHYSICS 483

pared with air as a standard, and air at 0° C. (32° F.) and
under a pressure of 760 millimeters of mercury ( = practically 15
pounds per square inch) is given an arbitrary value of 1. (See
page 27.)

Avogadro's Hypothesis. — Equal volumes of all gases contain
the same number of molecules.1 This is an important observa-
tion and leads us to another: that the specific gravities of gases
bear the same relation to each other as their molecular weights.
In other words, nitrogen is 14 times as heavy as hydrogen, and
carbon monoxide, (CO) 14 times.

Heat. — The atoms and molecules of all bodies are never in
a state of Test, even though the body itself appears to be quiet.
It is this constant and violent motion of the molecules which
we know under the name of heat. To raise the temperature of a
substance increases the motion and vice versa. In the case of
solids the vibration of each molecule is of course confined to a
very small space indeed, but in the case of gases, the molecules
travel until they strike against some other body with force enough
to resist them, as, for instance, some other molecule, or the walls
of the vessel in which they are contained, when they move with
equal velocity in another direction. It is, in fact, the constant
impact of molecules upon the walls of the containing vessel that
causes gases to exert pressure. This explains why it takes twice
as much pressure to confine the same weight of a gas into one-
half the volume, because now the molecules have only one-half
as far to travel between containing walls and therefore they strike
them twice as often. It also explains why gases expand when
their temperature is raised, provided the pressure under which
they are confined remains constant, because if their motion is
more rapid they exert greater pressure against the containing
walls.

Conservation of Energy. — The law of conservation of energy
tells us that energy can be neither created nor destroyed. We can
•convert chemical energy into heat, or heat into motion, but we
cannot get energy out of anything into which we do not put an
equivalent amount in some form or another. We may waste
energy, such as energy lost in heat from friction which is useless
to us, but it does not cease to exist.

1 It being understood of course that the conditions of pressure and tem-
perature are identical.

484 THE METALLURGY OF IRON AND STEEL

PHYSICAL PROPERTIES OF METALS

Tensile Strength. — The tensile strength of a body is its re-
sistance to being pulled asunder. It is usually measured in
pounds per square inch; that is to say, a bar of wrought iron for
example, with one square inch cross-sectional area1 will support
about 50,000 pounds weight.

Stress and Strain. — A stress is a force put upon a body, and
a strain is the deformation of the body produced by a stress.
For instance, if a bar of wrought iron one square inch in cross-
sectional area and 2 inches long be made to support a weight of
10,000 pounds, it will stretch about 0.0007 inch; the 10,000 weight
is the stress, and the 0.0007 inch is the corresponding strain.

Elastic Limit. — In the case just mentioned, if the 10,000
weight be removed the strain will be removed. That is, the bar
will return to its original length of 2 inches. Now if the same bar
be loaded with 20,000 pounds it will stretch 0.0014 inch, and again
this elongation will be lost when the weight is removed. If, how-
ever, we load the bar with 30,000, it will stretch a little more than
0.0021 inch, and now it will not return to its original 2-inch length
when the weight is removed, but will be permanently elongated.
It has taken a 'permanent set/ as it is called. The ' elastic
limit' of a body is the force necessary to produce the first per-
manent set. It is usually measured in pounds per square inch.
Another way of expressing the elastic limit is to say it is the force
beyond which the strain is not proportional to the stress.

Modulus of Elasticity. — The modulus of elasticity tells of the
resilience, or springiness, of a body, that is to say, how much it
will yield under any stress up to the elastic limit. The modulus
of elasticity is obtained by dividing any stress up to the elastic
limit by the strain produced per inch of length. For example,
the wrought iron mentioned in the last paragraph stretched
0.0014 inch in a length of 2 inches, =0.0007 per inch of length,
with a stress of 20,000 pounds per square inch; its modulus of
elasticity will then be:

20,000

00007

1 Say a round bar about 1^-inch diameter, or a square bar one inch on a
side.

CHEMISTRY AND PHYSICS 485

Percentage Elongation. — After a bar under tensile stress has
passed its elastic limit it begins to be permanently elongated in
the direction of the pull. A soft metal, like copper or mild steel,
will stretch out somewhat like molasses candy before finally break-
ing, and may be almost twice as long as it was originally. The
increase in length, divided by the original length, is the percentage
elongation. It is usually measured on a length of two inches,
or of eight inches.

Reduction of Area. — When a bar is elongated it of course
shrinks in cross-section; finally, just before the bar breaks, it
usually ' necks down' directly at the point on either side of the
fracture. This type of fracture occurs with soft metals. The
original area, minus the area of smallest cross-section after frac-
ture is called the ' reduction of area/ and this divided by the
original area is the 'percentage reduction of area.'

Ductility. — The percentage elongation and the percentage
reduction of area are usually taken together as the measure of the
ductility of a metal.

Compressive Strength. — The compressive strength of a body
is its resistance to crushing. It also is usually measured in pounds
per square inch.1 The terms ' stress and strain/ 'elastic limit/
and ' modulus of elasticity ' all have the same meaning when re-
ferred to compressive as to tensile stresses.

Transverse Strength. — If a bar one inch square be supported
on thin edges placed twelve inches apart its resistance to a force
applied half way between the supports is called its 'transverse
strength/ Here again we have the same terms, 'stress and
strain/ etc.

Impact. — If a bar be supported on thin edges placed a certain
distance apart and then a falling weight be allowed to strike upon
it at a point midway between the supports, its resistance to this
force will give an indication of its strength under impact, while
the amount that it will bend before breaking will indicate its
ductility under impact, or under 'shock/

Shearing Strength. — The resistance of a body to being cut
in two by a pair of knife edges, is called its shearing strength..
Rivets are sometimes tested in this way, because they are sub-
jected to this kind of stress in service.

1 In Great Britain they often use long tons (2,240 Ibs.) per square inch,
instead of pounds, both for tensile and compressive stresses.

486 THE METALLURGY OF IRON AND STEEL

Torsion. — The resistance of a bar to being twisted like a cork-
screw is called its torsional strength. The number of twists it
will endure before breaking gives an indication of its ductility
under this stress.

Repeated Stress. — If a certain kind of stress be applied to a
body, then relieved, applied again, and so on alternately, this
class of test is called ' repeated stress/ A metal will break under
many applications of a repeated stress much less in amount than
that required to fcreak it if constantly applied. It is to be under-
stood, however, that the interval between the applications of the
stress must be very short so the metal will have no opportunity
to rest between applications.

Alternate Stresses. — If we place a body first under tension,
then under compression, and so on alternately, it produces what
is known as ' alternate stresses.' It is like in nature to bending
a wire back and forth, and metals will break under alternate
stress even less than their elastic limits under either tension or
compression alone.

Toughness. — The toughness of a metal is its resistance to
breaking after its elastic limit is passed. It is the direct opposite
of brittleness.

Brittleness. — The brittleness of a metal is the ease with which
it breaks after its elastic limit is exceeded. A very brittle steel
will have an elastic limit exactly equal to its ultimate tensile or
compressive strength; that is to say, it will take no permanent
elongation or reduction or area; its ductility will be zero. Some
metals are more brittle under shock than under constantly applied
stress, and vice versa.

Malleability. — Malleability is the quality of being deformed
under a hammer. Gold is the most malleable of metals and
can be hammered into sheets of extreme thinness without
cracking.

Resilience. — We have already described resilience under the
head of modulus of elasticity. A very resilient metal, that is, one
with a small modulus, would be unsuitable for a bridge even
though strong, because its vibration under a moving load would
be so great.

Hardness. — The hardness of a metal is its resistance to being
scratched, or to wearing away under friction. In steel metallurgy
hardness is often used to mean brittleness, but this is no longer

CHEMISTRY AND PHYSICS 487

advisable, because we are now making hard steels that are also
tough.

AUotropy. — Allotropy is the capacity that certain elements
have of changing their properties without changing their com-
position or purity. For example, we may have pure carbon in
the form of diamond, graphite or charcoal; we may have iron in
a magnetic or non-magnetic condition; we may have sulphur in
a brittle or in a pasty state, etc. What the nature of allotropy is
we cannot at present tell. It may perhaps have to do with the
relations of the different atoms in the molecules of the element.
When elements form compounds the atoms of one are joined to
the atoms of the others, and even when elements are in the pure
state their atoms are often joined together to form molecules.
Allotropy may be a difference in the number of atoms that are in
each molecule, or perhaps in the form in which they are joined
together. An allotropic change -is always accompanied by a loss
or gain of heat.

Crystallization. — The tendency of most elements and com-
pounds to arrange themselves into regular forms called crystals
is really a powerful force of Nature, and one of the most wonderful
and charming studies imaginable. The crystalline forms of each
particular substance are usually the same, or very similar, but
different from almost all other substances. Each crystal is built
up of smaller crystals, and these in turn of still smaller ones.
The tendency to produce a regular form is well illustrated in the
case of alum: If a piece of an alum crystal be broken off and the
main part be immersed in a saturated alum solution, the crystal
will slowly repair itself and renew the lost part until it is again
perfect. Moreover, if the alum solution is impure, the crystal will
take to itself only the pure salt, and leave the impurity.

488

THE METALLURGY OF IRON AND STEEL

TABLE XXXIV.— COMPARISON OF DEGREES CENTIGRADE AND

FAHRENHEIT

Below zero

Above zero

Equivalents

-200° =

F.
-328°

+ 525°

= + 977°

+ 1,250° -

2,282°

1° =

F.
1.8

150 -

238

550

- 1,022

1,275 -

2,327

2 =

3.6

100 -

148

575

- 1,067

1,300 -

2,372

3 =

5.4

50 -

600

- 1,112

1,325 -

2,417

4 =

7.2

625

- 1,157

1,350 -

2,462

5 =

9.0

Above zero

650

- 1,202

1,375 =

2,507

6 =

10.8

675
700

- 1,247

= 1,292

1,400 =
1,425 -

2,552
2,597

7 =
8 =

12.6
14.4

25 -

725

- 1,337

1,450 =

2,642

9 =

16.2

50 -

122

750

- 1,382

1,475 -

2,687

10 =

18.0

75 -

167

775

- 1,427

1,500 -

2,732

11 =

19.8

100 -

212

800

= 1,472

1,525 -

2,777

12 =

21.6

125 -

257

825

- 1,517

1,550 -

2,822

13 =

23.4

150 -

302

850

- 1,562

1,575 -

2,867

14 =

25.2

175 =

347

875

- 1,627

1,600 =

2,912

15 =

27.0

200 -

392

900

- 1,652

1,625 -

2,957

16 =

28.8

225 -

437

925

= 1,697

1,650 =

3,002

17 =

30.6

250 -

482

950

- 1,742

1,675 -

3,047

18 =

32.4

275 =

527

1,000

- 1,832

1,700 =

3,092

19 =

34.2

300 =

572

1,025

= 1,877

1,725 =

3,137

20 =

36.0

325 -

617

1,050

- 1,922

1,750 =

3,182

21 =

37.8

350 -

662

1,075

- 1,967

1,775 =

3,227

22 =

39.6

375 -

707

1,100

- 2,012

1,800 =

3,272

23 =

41.4

400 -

752

1,125

= 2,057

1,825 -

3,317

24 =

43.2

425 -

797

1,150

- 2,102

1,850 =

3,362

25 =

45.0

450 -

842

1,175

= 2,147

1,875 -

3,407

475 -

887

1,200

= 2,192

1,900 =

3,452

500 =

932

1,225

= 2,237

2,000 =

3,632

INDEX TO AUTHORITIES CITED

ALLING, GEORGE W., 395.
AKERMAN, RICHARD, 125, 183.
ATHA, HERBERT B., 245.
AUSTEN. See ROBERTS-AUSTEN.
ARNOLD, J. O., 9, 332.

BACON, JOHN LORD, 395.
BAGSHAW, WALTER, 457.
BALE, GEO. R.', 291.
BAKHUIS-ROOZEBOOM, H. W., 312,

315.

BAUERMAN, H., 93.
BAUSCH, EDWARD, 457.
BEHRENS, H., 457.
BELL, SIR I. LOWTHIAN, 67, 73.
BENEDICKS, CARL, 395.
BLAIR, A. A., 8, 144.
BORCHERS W., 447.
BRINELL, J. A., 174.
BROWNE, D. H., 420.
BURGESS, C. F., 446, 447.

CAMPBELL, H. H., 8, 117, 152, 155,

169, 326, 327.

CAMPBELL, WILLIAM, 377, 457.
CARPENTER, H. C. H., 412, 421.
CARTAUD, C., 332.
CASPERSSON, C. A., 183.
CHARPY, G., 400.
CHURCH, A. H., 436.
COLBY, ALBERT LADD, 400, 420.
COLBY, E. A., 442.
CORT, HENRY, 57.
CUSHMAN, ALLERTON S., 423, 436.

DAELEN, R. M., 197.
DEER, Louis, 457.
DE MOZAY, 421.

DORMAN, W. H., 94.
DUMAS, L., 404, 420.

FRITZ, JOHN, 195.

GAGES, LEON, 73.
GLEDHILL, J. M., 421.
GOERENS, PAUL, 457.
GUILLAUME, C. E., 400, 420.
GUILLET, L., 404, 408, 420, 421.

HAANEL, EUGENE, 447.
HADFIELD, R. A., 330, 405, 413, 420.
HAMBUECHEN, CARL, 447.
HARBORD, F. W., 72.
HEYN, E., 457.

HOERHAGER, J., 421.

HOWE, HENRY M., 8, 72, 86, 91, 125,
177, 178, 182, 183, 184, 310,
311, 335, 371, 375, 384, 385,
394, 420, 427, 436.

JOHNSON, JR., J. E., 39, 174, 335.
JUPTNER, HANNS FREIHERR VON, 395.

KEEP, WILLIAM J., 291, 347.
KERSHAW, JOHN B. C., 447.
KJELLIN, F. A., 442.

LE CHATELIER, H., 447, 455, 451.
LEDEBUR, A., 72.
LEWKOWITSCH, 436.

LlLIENBERG, N., 184.

MclLHiNEY, PARKER C., 245.
MCQUILLAN, W. S., 279.
MELLOR, J. W., 395.
METCALF, WILLIAM, 379, 380.

489

490

INDEX TO AUTHORITIES CITED

MOLDENKE, RICHARD, 369.
MONELL, AMBROSE, 158.
MUELLER, FRIEDRICH C. G., 125.
MUSHET, ROBERT, 408.

NICOLARDOT, P., 421.
NICOLSON, J. T., 421.
NOBLE, H., 73.

OSMOND, F., 317, 330-2, 392, 453,

457.
OVERMAN, FREDERICK, 94.

PAVLOV, M. A., 93.
PERCY, JOHN, 73.
PHILLIPS, J. ARTHUR, 94.
PHILLIPS, W. B., 126.
PROCHASKA, ERNEST, 126.

READ, A. A., 332.

RICHARDS, J. W., 122, 125, 136.

ROBERTS- AUSTEN, SIR WILLIAM, 312,

314-15.

ROE, JAMES P., 77.
ROOZEBOOM. See BAKHUIS-ROOZE-

BOOM.

Rossi, A. J., 421.
RYLAND'S DIRECTORY, 8.

SANITER, E. H., 149.

SAUVEUR, ALBERT, 320, 332, 371, 394,

395, 453, 455, 457.
SCROLL, GEO. P., 447.
SCOTT, W. G., 355.
SMITH, J. KENT, 421.
SPELLER, F. N., 436.
STANSFIELD, ALFRED, 378, 379, 395.
STEAD, J. E., 182, 323, 332, 382,

457.
SWANK, JAMES M., 8, 10.

TALBOT, BENJAMIN, 182, 184.
TAYLOR, FREDERICK W., 409.
TRURAN, W., 94.

TURNER, THOMAS, 93, 291, 338,
347.

VATHAIRE, A. DE, 93.

WATERHOUSE, G. B., 332, 402, 403,

404, 419.

WEBSTER, W. R., 326.
WEDDING, HERMANN, 72, 126.
WEST, THOMAS D., 291.
WHITE, MAUNSEL, 409.
WOOD WORTH, JOSEPH V., 395.

INDEX

Acetylene, 469.

Acid, compared with basic Bessemer,
125.

compared with basic steel, 59.

distinguished from basic steel, 67.

furnaces for steel castings, 286-

92.
Acids, definition of, 471-72.

effect on corrosion, 423.
Acid steel, oxygen in, 328.

vs. basic furnaces for castings,

287.

Adulterants in linseed oil, 432.
After-blow in basic Bessemer, 123.
Agate, 476.
Air, amount of moisture in, 38.

and moisture producing rust, 422.

composition of, 460.
Air-furnace, 284-85, 286, 348, 352,
358-62.

compared with cupola, 361-62.
Air-hardening steel, 408.

liquid, 462.

necessary to burn coke, 269.
Alloys, 292, 481.
Alloy steels, 396-421.

definition of, 396.
Allotropic modifications of iron,

317-19, 383, 384-86.
Allotropy, 486.

Alpha iron, 318, 384-85, 393, 394.
Alternate stresses, 398, 415, 416, 418,

485-86.

Aluminum, chemistry of, 477-78.
Aluminum in steel, 174.
American iron and steel manufacture,
scheme of, 51-53.

Lancashire process, 71.

Analysis, 461.
Anhydrous acids, 472.
Animal forms, composition of, 460.
Annealing boxes for malleable cast-
ings, 362.

carbon, 320.

malleable castings, 52, 357, 358,
364, 365, 366, 367, 370.

ovens, 363.

pots for malleable castings, 362.

steel, 224-25, 229, 370, 378, 382,

383, 387-89, 412.
Anthracite, 12.
Anvils, chilling of, 352.
Apatite, 478.

Armor plate, 227, 386, 398, 407.
Arsenic, effect on steel, 328.
Atmospheric pressure, 482.
Atomic weights, 466, 467.
Atoms, 466, 471.
Austenite, 305, 334, 338, 342, 389-95,

407, 408, 413.
Automobile steels, 190-91, 408, 417,

418, 419.

Avagadro's hypothesis, 482.
Axles, properties of, 329.

Baby Bessemer converters, 286-92.
Badly made material and corrosion,

427-29.

Balling in puddling process, 81.
Bands, 65.
Bar iron, 70.
Bars, forging of, 189.

rolling of, 196.
Bases, definition of, 471-72.
Basic Bessemer compared with acid,

125.

491

492

INDEX

Basic Bessemer compared with acid

steel, 59-60.
distinguished from acid steel,

66-7.

furnace lining, 472.
pig iron, 42, 104, 146.
process, 51, 122-25.
slag, 471-72.

steel for railroad rails, 62-63.
steel in U. S., 53.
steel, oxygen in, 328.
steel, strength of, 327.
vs. acid furnaces for steel cast-
ings, 287.

Bas-relief polishing, 452.
Bath in Bessemer converter, depth of,

102.

Bearings, properties of, 329.
Bed of cupola, 266, 269, 270, 274,

280-84.

Beehive coke oven, 12, 14, 16.
Bell, 224.

Bell-Krupp process, 67-68.
Bell of blast furnace, 23.
Bench, 226.

molding, 243.

Bertrand-Thiel process, 159.
Bessemer blow, 96, 105.
Bessemer boil, 118.

compared with open-hearth steel,

60-63, 287-88.
converter, 98-102, 286-92.
distinguished from open-hearth

steel, 67.

flames, 96, 105, 118, 119.
fume, 118.
gases, 116-17.
ingots, weight of, 197.
iron, analyses of, 95, 104.
ores defined, 61.
pig iron, 95.
process, 53, 61, 95-126.

recarburizing in, 54, 60, 103-5,

120.

removal of impurities in, 112-
13, 114-15, 116, 122.
j, 95, 100, 110, 113, 114,
117-18.

Bessemer steel in U. S., 53.

steel, strength of, 326.

steel, uses of, 61-63, 224.
Beta iron, 318, 384-85, 394.
Billets, 65, 196.
Bi-prefix, 473.
Bisilicate, 477.
Biting of piece by rolls, 203.
Bi-valent, 473.

Black heart malleable castings, 364.
Black iron, 435.
Blast furnace, 2, 11-50.

chemistry of, 30-91, 465.

cinder. See Slags.

dimensions and parts, 23, 24-28.

fuels and fluxes, 11.

limestone used in, 13.

lining of, 24, 25-26.

ore used in, amount of, 13.

slags, 37, 39-40.

smelting practice, 30.

stoves. See Stoves.

vs. electric furnace, 440-41.
Blast furnaces in United States, 16.
Blast, in Bessemer process, 110.

pressures in cupola, 264, 266,
268, 269, 270, 271, 272, 273,
280-3, 284.

volumes in cupola. See Blast

pressure.
Blister steel, 87.
Blooming rolls, 203, 220.
Blooms, 65, 219.
Blower for cupola, 268, 269.
Blow-holes, 54, 59, 173, 174, 175, 185,
287, 324, 341, 427, 428, 456.
Blowing engines for blast furnace, 26,

27.

Blowpipe, 44, 168.
Blue powder, 434.
'Body' in bar iron, 70-71.
Boil in puddling process, 79.
Boiler, supports, 381.

tubes, 224.

Boilers over heating furnaces, 231.
Boiling linseed oil, 432.
Boilings, 85.
Boron steels, 397.

INDEX

493

Bosh of blast furnace, 24, 25-26.

Bottom casting, 178.

Bottoms for heating furnaces, 233.

Boyle's law, 482.

Brake action of carbon, 385, 390.

Breast of, blast furnace, 25.

cupola, 270.
Breeze, denned, 19.
British thermal unit, 162, 171, 465.
Brittleness, defined, 486.

of Stead, 381-82.

of steel, 224, 327, 328-29, 386-87.

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "The metallurgy of iron and steel"

See other formats