Full text of "The corrosion and preservation of iron and steel"

See other formats

BOUGHT WITH THE INCOME
FROM THE

SAGE ENDOWMENT FUND

THE GIFT OF

ilenrg W. Sage

$&&h£.6f!.*r. i aajfeykd..

6896-1

The date shows when this volume was taken.

To renew this book copy the call No. and give to
the librarian.

HOME USE RULES.

i kI \ \950 instruction or research

A\j»^ are returna bl e within 4

All Books subject to
Recall.

Books not used for

U&^^fT"

weeks.

, A o *g£7 Volumes of periodi-

■MrV> &-" * caIs and of pamphlets

are held' in the library

■" I .1 f ■ * '' For special purposes

they are given out for
a limited time.
, Borrowers should
not use their library
privileges for the bene-
fit of other persons.

Books not needed
during recess periods
should be returned to
/•i the library or arrange-

(J ments made for their

return during borrow-
er's absence, if wanted.

Books needed by
more than one person
are'held on the reserve.
list

Books of special
value and gift books,
when the giver wishes
it, are not allowed to
circulate.

Readers are asked to
report alkases of books
marked or mutilated.

Do not deface books by marks and writing.

Cornell University Library
TA 467.C98

The corrosion and preservation of iron a

3 1924 004 109 124

The original of this book is in
the Cornell University Library.

There are no known copyright restrictions in
the United States on the use of the text.

http://www.archive.org/details/cu31924004109124

THE CORROSION AND PRESERVATION
OF IRON AND STEEL

■IVKVVVVVVVWVVVVVVVVVVVVVVVWWWVWVVVVVV

Published by the

McGraw-Hill Book. Company

New Yoirk.

fiucce^sors to theBoolcDepartments or tne

McGraw Publishing Company Hill Publishing" Company

Publishers of Books for

Electrical World The Engineering" and Mining Journal

The Engineering Record rower and The Engineer

ngii
Electric Railway Journal

American Machinist

THE

Corrosion and Preservation

IRON AND STEEL

ALLERTON S. CUSHMAN, A.M., Ph.D. (Harvard)

Assistant Director and Chemist in charge of Physical and Chemical Investigations, Office of Public

Roads, U. S. Department of Agriculture; Chairman, Committee U on Corrosion of Iron, and

member of Committee E on Preservative Coatings, American Society for Testing

Materials; Member of Iron and Steel Institute of Great Britain ; and Associate,

American Society of Civil Engineers

AND

HENRY A. GARDNER

Director, Scientific Section, Paint Manufacturers' Association of the United States; Member of

Committee E on Preservative Coatings, American Society for Testing Materials;

Life Member, Master House Painters' and Decorators' Association of

Pennsylvania; Honorary Member, Master Car and Locomotive

Painters' Association of the United States and Canada

McGRAW-HILL BOOK COMPANY

239 WEST 39TH STREET, NEW YORK
6 BOUVERIE STREET, LONDON, E.C.

1910

- &

Copzjright, 1910, by the McGraw-Hill Book Company

^T-b°\e

T&e Plimpton Press Norwood Mass. U.S.A*

TO THE MEMORY

CHARLES B. DUDLEY

WHO, DURING A LIFE OF SPLENDID SERVICE, WAS ALWAYS READY

TO ENCOURAGE INVESTIGATION, AND WHO EARNESTLY

BELIEVED THAT THERE WAS SOMETHING NEW TO BE

LEARNED IN EVERY BRANCH OF KNOWLEDGE,

THIS VOLUME IS DEDICATED, WITH

PROFOUND RESPECT, BY

THE AUTHORS

PREFACE

It has been found a difficult matter to bring a work of this
kind up to date. New material, which touches directly, or in-
directly, on the corrosion and preservation of iron and steel,
is appearing almost daily in the technical press of the world.
The authors have made the effort to include or mention the results
of all recent investigations and original researches touching on
the subject which have appeared up to the time of going to press,
even if to some extent the systematic arrangement of the work
has had to suffer by frequent insertions. Although a consider-
able space is given to the description and results of the researches
of one of the authors working in a Government laboratory, no
material is included from this source that has not already been
printed in official bulletins for free distribution to all persons
desiring it. While the book is written mainly to elucidate the
electrolytic theory of corrosion (which some one has humorously
referred to as "The New Thought" in the protection problem),
it is nevertheless hoped that it will be found useful and suggestive
even to those to whom the theory does not especially appeal.

PREFACE

It has been found a difficult matter to bring a work of this
kind up to date. New material, which touches directly, or in-
directly, on the corrosion and preservation of iron and steel,
is appearing almost daily in the technical press of the world.
The authors have made the effort to include or mention the results
of all recent investigations and original researches touching on
the subject which have appeared up to the time of going to press,
even if to some extent the systematic arrangement of the work
has had to surfer by frequent insertions. Although a consider-
able space is given to the description and results of the researches
of one of the authors working in a Government laboratory, no
material is included from this source that has not already been
printed in official bulletins for free distribution to all persons
desiring it. While the book is written mainly to elucidate the
electrolytic theory of corrosion (which some one has humorously
referred to as "The New Thought" in the protection problem),
it is nevertheless hoped that it will be found useful and suggestive
even to those to whom the theory does not especially appeal.

CONTENTS

PAGE

Introduction. — The Object of the Work xvii

CHAPTER I. — THE PROBLEM OF CORROSION

The Importance of the Corrosion Problem . . 1

Contending Statements as to Rust Resistance of Different Types

of Metal . ... . 2

Corrosion and Conservation . . . 4

Three Phases in the Problem of Preservation . 6

Discussion Aroused by the Electrolytic Theory of Corrosion 6
Influence of Various other Elements Contained in Manufac-
tured Iron . . . .9

Problems Confronting Manufacturers of Structural Material . 10

Local Conditions Affecting Corrosion . 11
Working Knowledge Necessary to the Understanding of the

Corrosion Problem . . ... 12

CHAPTER II. — THEORY OF SOLUTION

Water as an Universal Solvent . . . 14

Solution Pressure . . 14

Osmotic Pressure 15

Relations between Osmotic Pressure and Gas Pressure . 17

The Theory of Electrolytic Dissociation .... 18

Hydrolysis . ... 20

Electrolysis . . 21

Modes of Ion Formation . ... 22

Oxidation and Reduction . 24
Electrolysis and Polarization . . 26
The Passive State of the Metals 28
The Electro-Chemical Series of the Metals ... 30
The Theory of Indicators . 31
Phenol-phthalein as an Indicator of the Presence of Elec-
trolysis . . . .... 32

Colloids and Crystalloids . . 33

Electro-Chemical Properties of Colloids .... 33

Conclusion of the Chapter . 34

CONTENTS

CHAPTER III.— THEORY OF CORROSION page

The Three Theories which have been Advanced to Explain

Corrosion . 35

The Carbonic-Acid Theory .... .... 35

Corrosion Occurs when Carbonic Acid is Absent ... 36

The Peroxide Theory op Corrosion . 39

The Electrolytic Theory ... .40

Pure Water as a Solvent of Iron ... .41

Rusting of Iron Primarily Due to Attack of Hydrogen Ions 44
Brief Explanation of Corrosion of Iron from Standpoint of

Electrolytic Theory ... 44

Development and Use of Ferroxyl Reagent 48
Ferroxyl Indicator Shows that Solution Tension of Iron Varies

at Different Points on the Surface 49

Preparation of Ferroxyl Mounts . . 50

Ferroxyl Indicator Rational Proof of Electrolytic Theory 51

Varied Rust Formation in Different Samples . 51

CHAPTER IV.— APPLICATION OF ELECTROLYTIC
THEORY

Two Distinct Effects in Rust Formation . . 56

Practical Application of Electrolytic Theory . . 58

Modern Steels Seem more Corrodible than the Old . 65

Chemical Purity an Aid in Resisting Corrosion 67

Difficulties Introduced by Improper Heat Treatment . 68

The Carbon Constituents of Steel . . . 72

Influence of Iron-Carbon Constituents on Corrosion . . 75

Possible Effect of Assembling Metals of Various Carbon Types 76

Effect on Corrosion of other Constituents in Steels 78

Various Factors which Modify the Effect of Impurities . . 80
Problem of Confining Structural Metals of Different Chemical

Constitution . ..... 80

Conclusions Based on Measurements of Potential Differences . 82

Effect of Heat Treatment, Tempering, Stresses, Strains, etc. 83
Relative Stability and Solubility of Strained and Annealed

Metals . . 85
The Influence of Quenching and Reheating Soft, Mild Steel on

its Solubility . . . .88

Influences of Wire-Drawing and Subsequent Annealing 91

Cold- Working Increases Solubility of Mild Steel . 92

Relation Between the Acid Test and Resistance to Corrosion 97

Probably no Reliable Acceleration Test Possible . . 99

Efficient Annealing Essential to a Rust Resistant Metal . 99
Effect of the Rusting Medium, Electrolytes, Natural Waters,

etc . 99

CONTENTS xi

PAGE

Highly Alkaline Solutions Prohibit Corrosion .... 100

Natural Waters may Contain Inhibitive Impurites 100

Dissolved Oxygen a Stimulator ... .... 101

Relative Corrosion in Fresh Water and in Sewage . . . 102

Non-Corrosion of Deeply Immersed or Buried Iron 104

"Busy" Iron Does not Rust 107

Application of Electrolytic Theory to Special Corrosion Prob-
lems . ... ... 109

CHAPTER V. — THE INHIBITION AND STIMULATION OF

CORROSION

The Use of the Word "Inhibition" ... .110

Hydroxyl Ions as Inhibitors . . 110

Passivity of Chromated Iron . Ill

Bichromate Solutions do not Attack Iron Free from Manganese . 112

Experiment with Bichromate to Induce Passivity 113
Inhibitive Power of Pigments Containing Certain Oxidizing

Agents . 113
Solution System with Two Contending Forces, Inhibitive and

Stimulative . 113
Practical Bearing of Alkaline Solution on Corrosion 115
Stimulative Action of Certain Paint Films Acting as Depolar-
izers .... . 116
Wounds on Steel Surface Stimulate Rusting 120
Mill-Scale, a Stimulator . 121
Various Inhibitive Expedients which have been Tried 121
Passivity Maintained by Plunging Electrodes into Neutral or

Alkaline Solution . . 122

Rusting Caused by Escaped Currents 123

Factors Causing Stimulation or Inhibition . 123

Factors which Inhibit Corrosion 124

CHAPTER VI. — THE TECHNICAL PROTECTION OF IRON
AND STEEL

Different Phases of the Protection Problem 125

Protection with Zinc 126

Methods of Applying Zinc Coating 127

Hot Dip Method of Galvanizing 128

The Weight of the Zinc Coating . . . 130

The Preece or Copper Suplhate Dip Test . 130

Standard Solution 131

Cleaning of Samples . 131

Apparatus . 132

Test 132

xii CONTENTS

Specified Dips. . 133

(a) Coating 133

(6) Cleaning . . . 133

{d) Quantity of Solution ... ... 134

(e) Samples . . .... 134

(/) Test . . . ... 134

(g) Rejection . ... 135

Objections to the Standard Copper Sulphate Test . . . 135

Walker's Contribution to the Testing of Galvanized Coatings . 137
Walker's Desiderata for a New Test for Determining the

Durability or Resistance to Corrosion. . . . 138

Walker's Test for Detecting Pin Holes, Cracks. . 139
Walker & Campbell's Test for Determining the Combined Effect

of Imperfections . . .... 139

First, Thickness of Coating . .... 140

Second, Purity of Zinc . . . ... 141

Third, Flexibility of the Zinc Coating 141
Circumstances Affecting the Weight of the Zinc Coating and the

Corrosion op Fence Wire . 142

The Electrolytic or Cold Method of Galvanizing 144

The Vapor Deposition Process . . . 146

Tests of Galvanized Wire . . 148

Preservation with Tin . 152

Preservation with Copper, Lead, and other Metals and

Alloys . 155

Processing after Manufacture . . . 157

Unprotected Steel 160

Corrosion of Boilers 160

Special Cases Met with in the Preservation of Iron . . 162

CHAPTER VII. — RELATION OF PIGMENTS TO THE
CORROSION OF IRON

Testing Pigments in Water Suspension . ... 164

Effect of Impurities and Other Factors Governing Nature of

Pigments . ... 166

Electrical Conductivity of Paint Films . . . 169

Water-Shedding and Excluding Properties of Pigments . 171

Moisture Penetration Tests of Paint Films .... 172

Industrial Application of the Theory of Inhibitors . . . 174

Thompson's Work on Solubility of Paint Films 178

CHAPTER VIII. — RECENT FIELD TESTS ON PROTECTIVE
COATINGS FOR IRON AND STEEL

Steel Test-Panels at Atlantic City . 180

CONTENTS xiii

PAGE

Pickling and Preparation op Plates .... . 180

Fence Erection and Preparation for Work 181

Method Followed in Painting Plates . . . . 181

Vehicles Used and Reasons for Avoidance of Japan Driers 182

Testing Effect of Various Prime Coats . ... 186

Combination Formulas Tested .... . 186

Drawing Conclusions from Results of Field Tests . . . 187

Inspection of Painted Surfaces . . 188

Other Field Tests and Their Value . 196

Tests on Preservative Coatings Carried on under the Supervision

of Committee "E," American Society for Testing Material 199

Tests of Paint Coatings Disigned to Resist Sea Air 201

CHAPTER IX. — PAINTS FOR VARIOUS PURPOSES

The Importance of the Special Design of Paint Formulae . . 206
Galvanized Iron and Modern Methods Used on its Decoration

and Preservation . . . 206

Painting of Tinned Surfaces . 207

Bituminous Coatings and their Application for the Protection of

Iron . . ... . 208

Regarding the Use of Coal Tar 209

Tar Paints . ... .211

Paper Paints and Paper Preservations .... 215

Painting Metals Subject to Marine Growths 215

Painting Steel Cars . . .... . 216

Paints for Locomotives and Fenders ... . 217

Protection of Iron in Tunnel Work . . 218"

Painting Train Sheds ... 218

Paint Protection for Water Tanks 219

Painting Steel Railroad Ties . . 219

Inhibitive Pigments for Railways Equipment in General . 221

The Preservation of Steel Mine Timbers . 221

Painting of Steel Mine Timbers ... . 224

Painting of Field Wire Fence . . 227

Prime Coatings for Structural Metal . 229

The Use of Oil as a Shop or Prime Coating . . . 231

Paint Coatings Used at Panama . . . 235

Painting Various Municipal Accessories, etc. . . 236

Painting Refrigerating Machinery . 237

Ornamental Iron Work Protection . .... 238

CHAPTER X. — THE TESTING AND DESIGN OF PROTECT-
IVE PAINTS

The Testing and Design of Paints . 239

General Directions in Regard to Testing 239

Laboratory Acceleration Tests . . ... 241

XIV

CONTENTS

Design of Protective and Inhibitive Paint Coatings

Inhibitive Paint Formulas

Wire Fence Paint

Condenser Paint

Paint for Iron Piping

Formula Labeling

page
245
247
249
249
250
250

CHAPTER XI. — PROPERTIES OF PIGMENTS

The Requisites of Protective Coatings

Meaning and Cause of Hiding Power

Whiteness

Stable and Chemically Active Pigments

Spreading Value

Effect of Gases on Paints

Paint Coat Strengtheners and Preventatives to Settling

Excluding Properties and Elasticity

Working Properties of Paints ....

Description of Pigments in General Use for Painting Iron

Basic Carbonate (White Lead)

Zinc Oxide

Basic Sulphate — White Lead (Sublimed White Lead)

Sublimed Blue Lead

Lithopone .

Zinc Lead White

Leaded Zinc .

Barium Sulphate (Barytes)

Gypsum (Calcium Sulphate) .

Mangesium Silicate (Asbestine and Talcose)

Whiting (Calcium Carbonate)

Aluminium Silicate (China Clay)

Silica

Litharge

Red Lead

Orange Mineral

Artificial Iron Oxides

Venetians Reds

Metallic Brown

Indian Red

Ochre, Sienna and Umber

Graphitic Pigments ....

Bone-Black

Lampblack

Carbon Black

Vine Black and Willow Charcoal

Mineral Black Pigments

Orange Chrome Yellow

251
251
251
252
252
252
252
253
253
254
254
256
257
258
258
259
260
260
261
261
262
262
262
262
262
263
263
264
264
264
264
265
265
266
266
266
266
266

CONTENTS XV

PAGE

Medium Chrome Yellow . . 267

American Vermilion . 267

Lemon Chrome Yellow 267

Barium Chromate . ... .... ... 267

Zinc Chromate .... . 267

Zinc and Barium Chromate . . 268

Chrome Green (Blue Tone) . 268

Chrome Green (Yellow Tone) . . ... 268

Chrome Green (Oxide) . 268

Prussian Blue . . . 268

Ultramarine Blue . 269

CHAPTER XII. —THE PROPERTIES OF PAINT VEHICLES

Linseed Oil . 271

Chemical Characteristics of, .Linseed Oil 272

Boiled Oil . . 273

Chinese Wood Oil . 273

Soya Bean Oil . . 274

The Use op Driers . . . 275

Turpentine . . . 275

Paraffins Spirits . . 276

Benzol . . . 276

Lacquers and their Application . . 277

Varnishes . .... . 278

APPENDIX A. — PRESENTING A DISCUSSION BEFORE THE
AMERICAN INSTITUTE OF MINING ENGINEERS ON THE
CORROSION OF WATER JACKETS OF COPPER BLAST
FURNACES.

The Corrosion of Water Jackets of Copper Blast Furnaces . 279

Discussion . . ... . . . . 280

The Corrosion of Water Jackets of Copper Blast Furnaces . 288

Appendix B. — Bibliography . 301

INTRODUCTION

THE OBJECT OF THE WORK

In the preparation of this work it has been the aim of the
authors to present as simply as possible the results of the more
recent researches on the corrosion and preservation of iron and
steel. The results of these researches have led to the electro-
lytic theory of corrosion, which it is safe to say is now provision-
ally accepted by almost all of the authorities on the subject.
Friend, 1 who has appeared in several papers in opposition to the
electrolytic theory, now very recently states that he is not an
adherent of the carbonic acid theory, but rather of the general
"acid theory" of corrosion. He agrees, however, that the acid
theory is in harmony with the electrolytic theory of ionization.
The question at issue, therefore, involves a distinction and not a
difference.

The electrolytic action which can be demonstrated taking
place when iron rusts in contact with water should not be con-
fused with the rapid destruction of steel in the neighborhood of
escaped currents from high potential circuits used for electric
lighting or power lines. As a matter of fact, the phenomena do
not differ except that in one case there is an external source of
energy at work, whereas in the other the energy of electro-chem-
ical action is due to the slow combustion or oxidation of iron.
In order to avoid confusion of terms and ideas, it has been pro-
posed that we should speak of the underlying electro-chemical
cause of corrosion as autogenous electrolysis, while another sug-
gestion is that the shorter term, autoelectrolysis, should be adopted.
There is no reason why these terms should not come into general
use, but the fact remains that they have not yet done so; the
authors have therefore held to the electrolytic explanation of
corrosion, without the introduction of these or other new terms.

1 See Electro-Chem. and Met. Industry VII., 11, 489, Nov., 1909. Also
Jour. Iron and Steel Inst. 77, 5, 1908.

xviii INTRODUCTION

In the authors' opinion corrosion must be considered an electro-
chemical phenomenon, and therefore a special effort has been
made to treat the subject in such a manner as to render it intelli-
gible to technologists, even if they are not thoroughly grounded
in the underlying principles of physical chemistry.

In the chapter on the theory of solutions an explanation is
given of the fundamental electro-chemical principles, without
an understanding of which no discussion of the corrosion problem
from this point of view is possible. It is hoped that this exposi-
tion of the subject will be found helpful by manufacturers,
engineers, metallurgists, and, in short, by all persons to whom
the rapid rusting and decay of iron is a matter of anxious solici-
tude. Wherever citations from other authors have seemed to
strengthen the development of the subject both as to theory and
as to fact, these have been freely made with the fullest credit to
the authority quoted. It has been remarked that the omission
by some of the prominent workers in this field, to mention in
their publications the contributions of others, has been the cause
of confusion in the minds of investigators who have had occasion
to review and discuss the problem. A considerable body of
material has been included which has been already previously
published by the authors from time to time in bulletins and
papers before technical societies. The object of the insertion of
this matter here is to better support the general argument and to
preserve it in a more permanent and concrete form. So much
has been claimed in regard to the relative rust resistance of iron
and steel that it has seemed best to include names of both types
of metal in the general title of the book. There are, however,
many places in a general discussion in which the word "iron"
can be used in its general sense, and whenever it has been
possible without leading to confusion this has been done.

The purely metallurgical problem of the manufacture of
highly rust-resistant metal has been touched on lightly, as the
subject is one on which much difference of opinion exists and
little consistent information is available. It is hoped that neither
the defenders of Bessemer and open-hearth steels, nor those who
believe in puddled and charcoal irons, will find any biased opin-
ions set forth in the succeeding pages.

The tendency to rust is a characteristic inherent in the element
known as iron, and will in all probability never be entirely over-

INTRODUCTION xix

come. Nevertheless, it is perfectly well known that there is
the greatest variability both in the manner and rapidity with
which different specimens and types of this metal and its alloys
suffer from corrosion. Before substantial advance in the manu-
facture of resistant types and in the protection of all types can
be made, a thorough understanding of the mechanism of corro-
sion and the function of protective coatings must be obtained.

The demand of the engineer and architect for the most
advanced knowledge regarding protective coatings for iron comes
as a result of the failure of many types of paint now in use.
Technical literature is replete with information regarding the
protection of iron, but probably the most remarkable advance in
knowledge of this subject has taken place during the last few
years, and almost wholly as a result of a well-planned series of
investigations recently undertaken by a number of independent
workers. The aim of the authors is to present the latest de-
velopments on this phase of the subject in a general way so that
the underlying principles which govern corrosion can be applied
by each investigator to his own particular problem.

So diverse a subject as the corrosion and preservation of
iron would become encyclopedic, if the attempt were made to
take up every phase in detail, or to abstract the literature of the
subject from the technical and scientific journals of the world.
The protection of boiler tubes presents a very different problem
from the protection of bridge structures, and yet the same main
principles can be applied to the consideration of each case. It
is this point that the authors have had in mind, and while a num-
ber of specific cases of corrosion are presented and discussed the
main object has been to treat the subject in a general way. For
this reason no separate chapters are included on special phases
of the problem, such for instance as the corrosion of boilers, the
corrosion of fence wire, or the corrosion of ships' bottoms. All
these subjects, as well as many other special cases of corrosion,
are, however, brought up and treated as special instances to
which the general discussion applies. On this account those
readers who propose to use this work as a reference book in special
cases of inquiry will depend upon the index to the volume rather
than upon chapter headings to guide them.

That much of the work which has been done along different
lines in the past in the effort to supply adequate protection for

xx INTRODUCTION

iron and steel has been unsatisfactory requires no proof. Unfor-
tunately, much of this work has been based on purely empirical
knowledge of the subject, that is to say, it has depended on indi-
vidual experience or observation without due regard to science
and theory. Thus, for instance, we find one engineer experi-
menting with red lead as a prime coating material, and reaching
a conclusion favorable to its use, while another engineer as a
result of his own observation arrives at a quite contrary opinion.
The case comes up for discussion before an engineering or tech-
nical society and adherents to the two sides of the argument are
not wanting to carry it on to indefinite lengths. The old ques-
tion in regard to who is to decide when doctors disagree applies
quite as well to such a case as this, and less learned people are
perplexed by the violent controversies between those who are
supposed to be expert. A more scientific acquaintance with
the technology of protective coatings from the standpoint of
recent investigations would at once suggest that red lead, like
other protective agents, is not a standard substance, and that
material from different sources might or might not contain stimu-
lative or accelerative impurities or properties. This same empiri-
cism may be noted in numerous other phases of the protection
problem, as in the steel versus iron controversy, or in the rival
processes for galvanizing and tinning. A theory serves the pur-
pose of a well-articulated skeleton, to which may be attached
the sinews and muscles represented by observed facts. If most
of the results of experiment and observation find their natural
and prepared points of attachment the theory is, to say the least,
useful as a working mechanism, even if some observations are
found difficult to fit in. As a matter of fact the recent electro-
chemical theory of corrosion presents few if any difficulties which
require to be explained away, while on the other hand, it furnishes
a rational explanation for many previously unexplained occur-
rences and phenomena.

In presenting and developing the electro-chemical explanation
of corrosion the authors' hope is to put into the hands of tech-
nologists generally a working theory which will be both suggestive
and practically useful.

THE CORROSION AND PRESERVATION OF
IRON AND STEEL

CHAPTER I

THE PROBLEM OF CORROSION

Importance of the Corrosion Problem. — The importance to
the human race of arresting the rapid decay of structural iron
has never been denied. It is almost impossible to find a volume
of transactions of an engineering or learned technical society for
years past that does not include papers and discussions on this
all-important subject. The different theories that have been
formed to account for the effects produced have been numerous,
and the suggested treatment and methods for protection, without
end. As Sang in a recent resume of the whole subject aptly puts
it: 1 "The decay of iron and steel by corrosion, if natural agencies
are allowed to act on them, is far more rapid than that of wood
and other materials of construction. Steel is being used more
and more every day for buildings and other permanent struc-
tures and therefore on the prevention of this decay depends the
permanency of these works and the safety of future generations.
Were it not for iron and steel, the erection of large works of engi-
neering would be impossible and their very size and consequent
high cost, representing as it does a large sum of human energy —
which is after all the only true foundation for wealth — make
it a duty to preserve them from decay.

"On a structure like the Forth bridge a number of men are
kept at work, cleaning rust-spots and repainting. The wise
course of preserving such structures for the use of our descend-
ants is not generally followed, and it is only when accidents like
the one at Charing Cross Station in London take place, that
interest is revived, for a time, in the question. Wood, in referring
to the roof of a gas-works in New York City which collapsed for
lack of attention, forecasted a similar fate, sooner or later, for

1 Proc. Eng. Soc. West. Pa., Vol. XXIV, 10, 493.
1

2 CORROSION AND PRESERVATION OF IRON AND STEEL

structures like the viaducts of the elevated railway of the same
city, which almost casual observation will show are repainted
over the rust without cleaning.

"On account of this necessity of combating corrosion, it is
imperative that engineers arrange the design so that every part
of structural works be readily accessible for frequent inspection.
It has been truly said that ' wrought iron is not only a bad but a
dangerous material if neglected'; this is equally true of steel."

Contending Statements as to Rust Resistance of Different Types
of Metal. — No attempt will be made in this work to decide
between contending statements as to the relative rust resistance
of wrought iron, charcoal iron, and the various types of steel.
The authors are satisfied to reiterate their previously published
opinion that there is great variation in each type of metal in rust
resistance and that in this respect there are good and bad irons
just as there are good and bad steels. If this is true, and a mass
of evidence seems to support the statement, it is useless for manu-
facturers of the different types, through their agents, to engage
in endless polemical debates over the comparative excellence
of their respective products. The proper control of unavoidable
impurities, their homogeneous distribution and a careful heat
treatment, particularly as far as the annealing processes are
concerned, will improve the rust-resistant quality of metal, whether
it is called iron or steel, and regardless of the method by which
it is made.

It is now very generally admitted as the result of recent
researches by a number of investigators that a state of stress or
strain in a metal invites rapid corrosion, as will also a burnt
condition of the metal leading to high porosity and the presence
of occluded gas and blow holes. Sang says in regard to this
point: 1 " Carelessness of manufacture, which tends to heterogene-
ousness, is an invitation to corrosion and in itself goes far to
explain why modern steel, which is tortured into shape at such
a high speed that the molecules are not permitted to readjust
themselves, is said to be more corrodible than the metals pro-
duced a generation ago; in those days iron and steel were pro-
duced in small quantities, without the addition of other metals,
and were rolled slowly and allowed to cool naturally. The inter-
nal strains due to mechanical treatment are not to be confounded
iProc. Eng. Soc. West. Pa., Vol. XXIV, 10, 511.

THE PROBLEM OF CORROSION 3

with the unevennesses in the distribution of the impurities due
to segregation in cooling; these mechanically induced strains are
really equivalent to straining the metal beyond the elastic limit
which, as will be seen later, makes it more corrodible. More-
over, the tonnage-craze, from which the quality of product in so
many industries is to-day suffering, is causing to be placed on the
market a great mass of material, only a small proportion of which
is properly inspected, which is not in proper condition to do its
work: — rails and axles which fail in service and steel skeletons
for high buildings which may carry in them the germs of destruc-
tion and death." With this summing up the authors are in
complete accord and no further space in this work will be given
to the discussion of the comparative merits of iron versus steel in
the resistance to corrosion. 1

That the old, largely hand-worked metal of about thirty
years ago is superior in rust-resisting quality to the usual
modern steel and iron is attested by the recorded evidence of a
large number of observers. Many citations could be given to
prove this and the examination of the discussions of numerous
papers before engineering and technical societies seldom fail to
bring out evidence of the very general belief in the superiority
of the older product. A more detailed discussion of this subject
must be postponed to a later chapter. The accompanying illus-
tration (Fig. 1) showing a contrast between two nails is interest-
ing though not unusual. Sample A is a forged nail that was
used in the old Masonic Hall in Richmond, Va., and was prob-
ably driven in 1807. It was in service about one hundred years
and for a large portion of that time was freely exposed to the
weather, as the old clapboarding rotted away. The illustration
shows that even the thin edges of the forged head are still sharp
and uncorroded. Sample B is an ordinary modern nail after six
months' exposure in a wooden gutter on a roof at Washington, D. C.
No thoughtful person supposes that it would be practically
possible to return to the earlier laborious methods of iron manu-
facture, in order to produce metal highly resistant to corrosion.
The modern problem must be solved by modern economic methods,

1 For a full discussion of this subject by advocates of various types of metal
the reader is referred to the Transactions of The American Society for Testing
Materials, for years 1906-7-8, also to Transactions American Institute of
Mining Engineers for 1905.

4 CORROSION AND PRESERVATION OF IRON AND STEEL

and there is no reason to suppose that great improvement will
not be made in the quality of both iron and steel, as soon as the
principles governing the rate and kind of corrosion which takes
place on different types and kinds of iron are thoroughly estab-
lished. These principles will be brought out in detail in suc-
ceeding chapters.

Fig 1. — Showing the relative cor-
rosion of a modern steel nail and
an old forged nail.

Corrosion and Conservation. — No discussion of the corrosion
problem is complete without reference to its bearing upon the
conservation of natural resources. In the introduction to an
address before a recent meeting of the Iron and Steel Institute
of Great Britain,- one of the authors called attention to the im-
portance of this phase of the problem in the following words:
"It is significant of the present age that the tendency of thought
among all nations is toward conservation of natural resources.
Increasing consumption of the world's supplies and constant
decay of materials menace the future of the human race. It is
evident that carbon, from which we derive energy, and iron,
which provides the principal means of applying energy, are the
materials which should particularly engage the attention of those

THE PROBLEM OF CORROSION 5

who are studying the problems of conservation. The annual
production of pig iron in the United States alone grew from
about 14,000,000 tons in 1900 to about 26,000,000 tons in 1907,
and while for the year 1908, owing to industrial conditions, it
decreased to about 16,000,000 tons, it seems sure to again show
an increase in the near future. How much of the enormous and
constantly increasing world's production of iron and steel is
wasted for lack of adequate preservation? Where will the growth
of demand stop, and how many years will the world's ore supply
stand the drain upon it? These are questions of vast importance,
the answers to which can only be vaguely guessed. One thing
seems certain, namely, that civilization must learn to conserve
more efficiently its stores of iron and steel already manufactured,
and seek methods to prevent the almost resistless tendency of
iron to return to its lethargic union with oxygen, from which it
was won only by tjie consumption of vast quantities of the ever
dwindling coal supply. It is not generally realized that about
four tons of coal or its equivalent is used in preparing one ton of
finished steel from the ore. The world's store of gold is not
subject to loss by corrosion, but suffers to some extent from
attrition, owing to the softness of the metal. If steel could by
any means whatsoever be ennobled and thus protected from the
inevitable decay due to corrosion, future conditions for all pos-
sible years to come could be viewed with complacency. There
exists at Delhi, India, an iron monument that, since the dim
beginning of history, has been exposed to the weather without
rust or decay, and yet this column has been provided with no
protective coating, other than that which the atmosphere has
itself formed upon it. Could we to-day, with all our boasted
knowledge and our great pneumatic processes, build its like?
It is probable that we could not, and yet an art is not necessarily
lost forever. It is simply a question whether or not it is worth
our while to rediscover it."

The President of the Iron and Steel Institute in discussing this
paper stated in his remarks that no doubt this was a phase of
the subject which was most prominently before the world at the
present time, and that nothing could be more valuable to the
iron industry at large than that the question of the preservation
of the metal which they were engaged in manufacturing should
be investigated and solved.

6 CORROSION AND PRESERVATION OF IRON AND STEEL

These citations serve to bring out the great importance of
the corrosion problem and show the earnest attention that is
being given to it by metallurgists and learned bodies.

Three Phases in the Problem of Preservation. — One of the
writers has in previous papers 1 referred to the fact that the
problem of the preservation of iron has three distinct phases.
The first phase has to do with the manufacture of a metal highly
resistant to corrosion, such as is represented by the iron column of
Delhi, Fig. 2, or the splendid examples of hand-forged metal that
have come down to us from past centuries, and with which we may
contrast the pitted, lacelike condition of specimens of modern
steel after only a few years' exposure to the atmosphere. The
illustration, Fig. 3, shows the condition of certain members taken
from a signal bridge on an American railroad. 2 The structures
were about twelve years old and exhibited great differences in
the rate of corrosion of different members. This point will be
discussed more fully later on, but the photographic reproduction
serves to illustrate very forcibly the lacelike condition assumed by
metal undergoing rapid corrosion.

The second phase of the subject concerns the general subject
of protective coatings, which may consist of other metals, such
as zinc, tin, copper, and lead, or of oil paints, varnishes, lacquers,
and bituminous materials, or, finally, the production of a higher
oxide on the surface, as in the Bower-Barf, Wells, and Speller
processes. The third phase includes the interesting study of the
passive condition which iron is capable of assuming, and the
possibility of maintaining the surface in such an ennobled con-
dition, either by the use of inhibitive pigments in paint com-
pounds, or by the use of electric currents as described later on.

Discussion Aroused by the Electrolytic Theory of Corrosion. —
Many discussions 3 have arisen in regard to the several rival

1 U. S. Dept. Agr., Office of Public Roads, Bui. 30, 1907; Trans. Am. Soc. for
Testing Materials, 1907, 7, 209; Trans. Am. Electro-Chem. Soc, 1907, 12, 403.

2 Samples and information furnished by Mr. J. P. Snow, Bridge Engineer,
Boston and Maine R. R.

J Am. Soc. Mech. Eng., Trans. 1894, 15, 998; ibid., 1895, 16, 350, 663; Zts.
Elektrochemie, 1903, 9, 442; Treadwell and Hall, Analytical Chem., 1907,
p. 92; Manchester Lit. Phil. Mem., 1871, 5, 104; Jour. Iron and Steel Inst.,
1888, 129-131; Proc. Chem. Soc. (Lond.), 1906, 22, 101; Jour. Chem. Soc.
(Lond.), 1905, 87, pt. 2, 1548; Jour. Am. Chem. Soc, 1903, 25, 394; The
Analyst, 1905, 30, 232; Trans. Chem. Soc. (Lond.), 1906, 98, 1356.

THE PROBLEM OF CORROSION

theories of corrosion, but the evidence which has been collected
in numerous recent researches appears to the authors to prove

Fig. 2. — The Iron Column of Kutab Minar Delhi, India, erected 900 b.c.

beyond argument that the corrosion of iron, like that of other
metals, is an electro-chemical phenomenon. Some confusion
has arisen in regard to this contention which the authors take

8 CORROSION AND PRESERVATION OF IRON AND STEEL

THE PROBLEM OF CORROSION 9

this occasion to explain. H. M. Howe, an eminent authority,
has stated: 1 "Electrolytic action surely hastens corrosion very
greatly, and in practice it may well be true that nearly the whole
of corrosion is electrolytic. But imagine the case of absolutely
pure iron, without stress or slip planes, immersed in pure water
containing dissolved oxygen. Will not the iron go into solution
till the solution tension is reached, and will not the iron so dis-
solved oxidize, precipitate, and leave room for more iron to dis-
solve? It seems to me in this and like ways corrosion may go
on without electrolysis. If so, then electrolysis is an aggravator
and hastener of corrosion, but not essential to it."

Every one interested in the subject who will read the succeed-
ing chapter on the Theory of Solutions will understand that iron
which is pushed into solution by its solution tension immediately
ionizes and assumes an electric charge and that probably iron is
never in such a quiescent condition that some polarization does
not take place. The further discussion of this point as well as
the consideration of the effect of outside or extraneous currents
of electricity in effecting corrosion must be postponed to a later
chapter.

Influence of Various other Elements Contained in Manufac-
tured Iron. — In considering the corrosion of iron it is important
to remember that iron is a metal which readily combines with or
dissolves nearly all the other elements. With possibly one or
two exceptions, there are no elements that do not either dissolve
in or combine readily with iron. It is also unique in the fact
that very small quantities of impurities suffice to entirely change
its physical characteristics. On account of this fact metallurgists
scrutinize the hundredth of a per cent, of some of the principal
impurities that are generally associated with this metal. This
is particularly true, for instance, of the element phosphorus.
It is so important to modern metallurgy that the amount of
phosphorus should be controlled in certain forms of steel that
an animated discussion is going on at the present time between
certain interests as to the control of the amount of phosphorus
that steel shall carry, and the question at issue amounts to no
more than a few hundredths of one per cent.

Manganese is also an element which is nearly always asso-
ciated in modern metallurgy with iron and steel. Manganese
1 Trans. Am. Soc. for Testing Materials, VIII, 278, 1908.

10 CORROSION AND PRESERVATION OF IRON AND STEEL

decreases the electrical conductivity of iron, and as the percentage
of manganese, starting from zero, rises, the electrical resistance
increases up to a certain specific maximum. It will be seen that
if the presence of manganese in iron raises the electrical resistance,
any variation in the distribution of the manganese means that
there will not be a constant electrical conductivity throughout
its mass, or on any given surface. There is abundant evidence
to show that manganese associates itself to a considerable extent
with sulphur when both these impurities are present in steel. 1
That manganese sulphide shows a difference of electrical potential
against iron is also well known. One who is familiar with the
methods of modern metallurgy knows that the manganese is
added for certain specific purposes, not as a rule quantitatively,
but in accordance with the views of the iron master who has con-
trol of the mill or furnace. Moreover, the manganese is usually
added by throwing lumps of ferromanganese into the molten
metal, either in the furnace itself or in the ladle into which it
has been poured. Chemists know the extreme care that has to
be taken in order to get uniform mixtures of substances in the
course of chemical operations. On the large scale on which
metallurgical processes are conducted, even if it were possible
to take great care in the mixing, it still happens that when iron
is cooled from the molten state segregation takes place — that is to
say, the impurities, although they may have been thoroughly
mixed in the molten mass, do not remain homogeneously dis-
tributed after the metal is cooled.

For these reasons we must remember that in studying iron
and steel from the standpoint of their stability, under the con-
ditions of service, we are not dealing with homogeneous pure
metal.

Problems Confronting Manufacturers of Structural Materials. —
If we consider the results of recent experimentation along the
lines of the three phases as indicated above, we find that con-
siderable progress has been made in the course of the last few
years. Many leading manufacturers of iron and steel have been
paying special attention to the careful control of impurities, and
the heat treatment of their products, as well as to the equally
important problem of evolving rust-resistant metallic coatings
in the various processes of galvanizing with zinc and special
iFay, Proc. Am. Soc. Test. Mat., VIII, 74 (1908).

THE PROBLEM OF CORROSION 11

alloys. 1 The condition of maintained passivity which has come
to be known as "inhibition" is especially a problem for the paint
manufacturer and will be dealt with in detail in later chapters
of this book.

One phase of the corrosion problem which is of the utmost
importance and which has been discussed with anxious foreboding
is the possible corrosion of steel embedded in concrete. The
consensus of opinion among engineers and investigators of the
corrosion problem appears to be that concrete furnishes ample
protection to steel embedded in it, except in certain cases in
which infiltrating or percolating waters find a way through the
concrete, washing away the free alkali present in the form of
lime or calcium hydroxide. The reasonableness of this explana-
tion will appear later on.

Local Conditions Affecting Corrosion. — It is quite apparent
that the corrosion problem has increased steadily with the growth
of civilization, for the waste gases of combustion pollute the atmos-
phere and the water with which iron may come in contact,
and besides this there are many stray electric currents escaped
from high potential circuits which undoubtedly aid in the work
of destruction. H. M. Howe, in making a plea for modern steel,
has said: 2 "The fact that steel has come into wide use simul-
taneously with a great increase in the sulphurous acid in our city
air and of strong electric currents in our city ground may well
lead the practical man, be he hasty or cautious, into inferring
that the rapid corrosion of to-day is certainly due to the new
material of to-day, steel, whereas, in fact, it may be wholly due
to the new conditions of to-day, sulphurous acid and electrolysis."

While the above statements are perfectly true they do not
furnish a complete and satisfactory explanation of the whole
problem. Take the interesting case of the steel light-ship, U. S.
Light-Vessel No. 71, built in 1897, which was anchored at Dia-
mond Shoal off Cape Hatteras for eleven years, and was towed
in at the expiration of that time so that 8400 four-inch bolts

1 Since the writing of this paragraph, patents have been granted for a
process of producing exceptionally pure iron in basic open-hearth furnaces
which is claimed to be slagless, degasified and easily workable. Since the
effort to produce this type of metal was originally suggested by one of the
authors it is proper to record here its successful accomplishment on a com-
mercial scale.

2 Proc. Am. Soc. Testing Materials, 1906, VII, 155.

12 CORROSION AND PRESERVATION OF IRON AND STEEL

which were destroyed by accelerated electrolytic action could be
replaced. 1 In such a case as this serious damage occurred after
eleven years, due to electrolytic corrosion, although the material
was subjected to no stray currents or sulphurous acid gases.
Probably not all the bolts were in actual contact with salt water,
although the general saltiness of the atmosphere undoubtedly
stimulated the action. The condition of the destroyed bolts is
well shown in the illustration,. Fig. 4. The discussion of this and
similar cases of corrosion will be postponed to a later chapter.

Fig. 4. — Showing the corroded condition of bolts taken from the U. S.
Light-Vessel, No. 71.

The Working Knowledge Necessary to the Understanding of the
Corrosion Problem. — It is not possible to make an intelligent
study of the problem of corrosion without at least a working
acquaintance with the principles of the new knowledge in physics
and chemistry which has gone so far to elucidate and explain
many heretofore obscure phenomena. It is, however, the opinion

1 Information and specimens furnished by A. B. Johnson, Supt. Light-
house Establishment, 5th District, Dept. Commerce and Labor, also by
Robt. L. Russell, Commander, U. S. N.

THE PROBLEM OF CORROSION 13

of the authors that the groundwork for a proper understanding
can be acquired by any person with some technical knowledge
and experience, even if chemistry, and more particularly physi-
cal chemistry, has never before engaged his attention. To this
end the following chapter on the theory of solutions has been
included. In the treatment of this chapter the effort has been
made to present in the simplest possible way the underlying
principles which enter into the corrosion problems in their most
recent and modern aspects. The new theories which attempt
to explain corrosion have hitherto been presented mainly before
learned technical bodies, and in the columns of the highly scien-
tific and technical press. It is the earnest hope of the authors
that a careful perusal and study of the following chapter will
enable any intelligent person who may be interested in the prob-
lem to understand the discussions and data presented in the
succeeding chapters.

CHAPTER II

THE THEORY OF SOLUTIONS

Water a Universal Solvent. — Water is the universal solvent,
and it is doubtful if there is a form of matter in existence which
is not to some extent dissolved, disintegrated, or otherwise changed
by it in the course of time. The action of water is not, however,
to be considered as universally destructive, for it is the medium
by means of which, and in which, synthesis as well as decomposi-
tion takes place in nature. Whether the action of water is to
be considered as destructive or constructive depends therefore
on the point of view. For instance, the immense bodies of iron
ore on which we must always depend for our supplies of iron
were undoubtedly originally formed under the action of water,
and now once more water carries on the attack on our finished
iron and steel, which has been laboriously smelted and reduced
from the ore at the expense of vast quantities of free carbon.
To the layman iron is an insoluble substance as far as water
alone is concerned, but as has been pointed out, when we speak
of a substance as being insoluble, we may be merely employing
a relative term. If the water is impure, that is to say, if it already
contains slight amounts of dissolved substances, its action on any
given body is modified. Whether the solvent action is stimu-
lated or inhibited depends upon the nature of the case.

Solution Pressure. — Leaving aside substances usually classed
as insoluble, we will for the present consider only those which
have a decided and appreciable solubility in pure water. Such a
body, when immersed or brought into contact with water, tends
to pass into solution; that is to say, its molecules or atoms (de-
pending upon whether the substance in question is a compound
or a simple element) tend to distribute themselves equi-spatially
among the molecules of the solvent. The driving force which
produces this tendency, and which is known as solution tension
or solution pressure, is exactly analogous to the pressure exerted
by a gas or vapor confined in a vessel, as for instance, in the

THE THEORY OF SOLUTIONS 15

case of steam in a boiler. But just as the rising steam pressure
in the boiler will tend to resist the evaporation of the remainder
of the water, the growing number of free particles entering the
solution will produce a back pressure which tends to resist the
entrance of more. This back pressure acting against solution
tension is called osmotic pressure, and the class of phenomena
which it produces is known as osmosis. It is at once apparent that
for any given substance at a given temperature, its maximum
solubility would be reached just as soon as the solution pres-
sure and the osmotic pressure were equal. It is also apparent
that any outside cause operating on the system, which tended
constantly to lower the osmotic pressure, would allow more and
more of the substance to enter into solution so that the action
would become continuous, provided the supply of solvent and
solute was maintained. This point is an important one in this
discussion, inasmuch as it has a direct bearing on the corrosion
of iron.

Osmotic 'Pressure. — The existence of osmotic pressure was
first recognized by the Abbe Nollet about the middle of the eight-
eenth century. He used animal membranes for demonstrating
the pressure. A glass tube closed at one end with animal parch-
ment was filled with alcohol and plunged into a vessel of water.
The alcohol tended to dissolve or diffuse into the water and the
water into the alcohol, but alcohol cannot pass rapidly through a
parchment diaphragm, whereas water can. The consequence was
that the level of liquid rose in the tube, revealing the existence of
osmotic pressure. Diaphragms which allow the passage of water
molecules but resist the passage of substances soluble in water
are known as semi-permeable membranes. Traube in 1867 dis-
covered that by depositing a film of a gelatinous preparation of
copper ferrocyanide in a suitable manner very good artificial
semi-permeable membranes could be formed. This method of
measurement has been improved and developed by Pfeffer in
Germany and later by Morse in the United States. The copper
ferrocyanide is now by an ingenious process deposited as a film
inside the walls of a clay cylinder to the top of which the
manometer tube for measuring the pressure is luted. It is only
necessary here to bring out the fact that osmotic pressures and
therefore solution pressures represent values of very considerable
magnitude. This is best shown by results obtained by Pfeffer

16 CORROSION AND PRESERVATION OF IRON AND STEEL

working with dilute solutions of cane sugar. Here again the
water passes through the semi-permeable film while the sugar
molecules are stopped so that the pressure can be recorded in
the manometer tube in barometric readings.

Pfeffer's results, which were taken at about 14° Centigrade,
are as follows:

C = Concentration of

Solution in Per Cent.

P = Osmotic Pressure
in Centimeters of mercury

P'= Osmotic Pressure
in Inches of mercury

P"= Osmotic Pressure
in Atmospheres

1 per cent.

53.5

21.0

0.70

2 per cent.

101.6

40.0

1.33

4 per cent.

208.2

82.0

2.73

6 per cent.

307.5

121.0

4.03

It will be seen from this that the osmotic pressure exerted by a
six per cent, solution of sugar is equal to about four atmospheres,
or reduced to ordinary terms sixty pounds to the square inch.
Sugar, of course, is a very soluble substance and its solution
pressure is high, but that even the more insoluble substances
have solution pressures of definite and appreciable magnitude
is incontestable.

If a strip or rod of polished steel is plunged into a clear glass
cylinder containing distilled water and placed so that it can be
observed in a strong light, it will be seen that in the course of a
few minutes delicate festoons of a whitish brown substance are
clouding the sharp edge or outline of the steel specimen and fall-
ing away from the surface. This phenomenon furnishes evidence
to the eye that the iron has a direct solution pressure, for the
whitish brown substance which is forming is a hydrated oxide
or so-called hydroxide of iron, and cannot be formed unless iron
first passes into solution. Careful observation of this simple
experiment will bring out another point which has an important
bearing on our subject. It will be noted that the festoons of
hydroxide do not appear at all points of the polished surface,
but seem to segregate at certain points or areas, leaving others
still bright and unattacked. This observation furnishes the first
indication that the solution tension of iron is not the same at all
points on the surface. This fact is important to the considera-
tion of the corrosion of iron, and will be taken up again after the
principles governing solutions in general have been presented.

THE THEORY OF SOLUTIONS

It has been found possible to illustrate the uneven solution ten-
sion of steel rods immersed in water and very dilute electrolytes.
A photographic representation of the appearance of three polished
steel rods after a few hours' immersion is shown in Fig. 5.

Fig. 5. — Showing the unequal solution tension and rusting of three steel
rods immersed in fresh water, sea water, and dilute sodium nitrate.

Relations between Osmotic Pressure and Gas Pressure. — Pfef-
fer's object in carrying out his researches was to study the role
played by osmosis in certain physiological processes, such for
instance as the swelling and bursting of seeds, and the movement
of sap in plants. It was reserved for Van't HofT to make the
wonderful observations and deductions which, published in an
epoch-making paper in 1887 * was destined to become the foun-
iZeit. Phys. Chem., 1, 481.

18 CORROSION AND PRESERVATION OF IRON AND STEEL

dation of a new chemistry. The law of Boyle for gases points
out that the pressure of a gas varies directly with its concentra-
tion, and the general gas law is simply expressed by the formula,
PV = RT, where P is the pressure, V the volume, T the tempera-
ture, reckoned from the absolute zero ( — 273° Centigrade), and
R a constant. Using Pfeffer's results, Van't Hoff pointed out
that the osmotic pressure of a solution of cane sugar is exactly equal
to the gas pressure of a gas having the same number of molecules in
a given volume, when the temperature is the same in both cases.
This is a surprising generalization, and especially so when we con-
sider the different conditions under which gas molecules and the
molecules of a dissolved substance are acting. Nevertheless, still
greater surprises were in store, for Van't Hoff further pointed
out that while one class of substances which were chiefly organic
in their nature conformed with the gas law, another class in which
could be included the acids, bases, and salts of inorganic chemistry
gave osmotic pressures in dilute solutions, in many instances about
twice as high as they ought to if they were in conformation with the
generalization. It was at this point that Arrhenius, a Swedish
physicist, in 1881 made his great contribution to the dawning science
of physical chemistry. Arrhenius developed the discussion of
Van't Hoff's deductions as follows l : If a gas shows a deviation from
the gas law, as many of them do at high temperatures, we explain
this by supposing that the molecules of the gas which are com-
posed of two associated atoms (as, for instance, hydrochloric acid,
which would be expressed molecularly by the symbol H — CI),
break down at certain critical temperatures into simple atoms, thus
providing twice as many free particles in a unit space. Arrhe-
nius then proposed to apply the theory of dissociation to the mole-
cules of inorganic substances in solution and inquire whether
they were not split up on entering solution into constituent par-
ticles, or, as he called them, ions.

The Theory of Electrolytic Dissociation. — The theory of elec-
trolytic dissociation, as it is stated at the present time, holds
that when acids, bases, and salts are dissolved in water to form
dilute solution, they break down or dissociate into ions. Ions
are atoms or groups of atoms carrying in relation to their masses
enormous charges of static electricity. The reason why these
charges are not apparent in the molecule is because they are of
^eit. Phys. Chem., 1, 631.

THE THEORY OF SOLUTIONS 19

necessity always equal and of opposite sign. These dissociation
reactions are simply expressed as follows:

Acid: hydrochloric HC1 = H + CI

+
Base: sodium hydroxide NaOH — Na + OH

+
Salt: sodium chloride NaCl = Na + CI

or to give more complex examples:

Acid: sulphuric H 2 S0 4 = H • H + S 4

+ + — —

Base: calcic hydroxide Ca(OH) 2 = Ca + OH • OH

+ +

Salt: calcic sulphate CaS0 4 = Ca + S0 4

The positive hydrogen ion is the distinctive characteristic of all
acids in solution, just as the negative hydroxyl ion (OH) is the
characteristic of all bases in solution. Hydrogen, in spite of the
fact that it occurs and is generally known as a gas, belongs to
the type of elements which are called metals.

This subject can be more clearly set forth by quoting from a
recent excellent text-book on physical chemistry: 1

"Each compound dissociates into a positively charged part
called a cat-ion, and a negatively charged part, an an-ion. These
ions may be charged atoms as the above cations, or groups of
atoms as the anion OH. The cations are usually simple atoms
charged with positive electricity. The cation of all acids is hydro-
gen; the nature of the anion varies with the nature of the acid.
It may be chlorine, bromine, the N0 3 group, S0 4 , etc. The
anion of bases is the group (OH); the cation varies with the
nature of the base. It may be potassium, barium, ammonium,
etc. The anions and cations of salts both vary with the nature
of the salt. They depend upon the nature of the acid and the
base which have combined to form the salt.

"It was stated that hydrogen is the cation into which all
acids dissociate. It may be added that this is the characteristic
ion of all acids, and whenever it is present we have acid proper-
ties. Further, we never have acid properties unless there are
hydrogen ions present. The same may be said of the hydroxyl
ions into which bases dissociate. This is the characteristic ion
of bases.

1 Jones, Elements of Physical Chemistry, MacMillan (1909), pp. 211-212.

20 CORROSION AND PRESERVATION OF IRON AND STEEL

"It has been repeatedly urged that the theory claims that a
compound like potassium chloride dissociates into potassium and
chlorine, and since neither potassium nor chlorine can remain in
the presence of water under ordinary conditions without acting
upon it, the theory is self-evidently wrong. This objection, like
so many others, is based upon an imperfect understanding of the
theory. No one has ever claimed that a compound like potas-
sium chloride dissociates in the presence of water yielding atomic
or molecular potassium, having the properties of ordinary potas-
sium. The products of dissociation are a potassium ion and a
chlorine ion, and the potassium ion is a potassium atom charged
with a unit of positive electricity. There is no reason whatever for
supposing any close agreement between the general properties of
a potassium atom and those of a potassium atom charged with
electricity. About the only property which we would expect to
remain unchanged is that of mass, and the mass of an atom is
not changed by charging it. The properties of atoms are doubt-
less very closely connected with the energy relations which obtain
in or upon the atom. When we change these fundamentally,
as by adding an electrical charge, we would expect fundamental
changes in properties; and such are the facts. It can be safely
stated that whatever may be the ultimate fate of the theory of
electrolytic dissociation, it will never suffer seriously from any
such objection as that just referred to."

Hydrolysis. — We have now familiarized ourselves with the
principles of electrolytic dissociation in which certain substances
in solution are broken down into constituent ions. We shall
now have to consider a different form of dissociation in which ions
and electrical charges are only indirectly concerned, but which
has a bearing on many of the problems of the corrosion and pres-
ervation of iron. It has long been known that compounds which
are formed by association of a weak acid with a strong base-
forming element have an alkaline reaction in solution, while
conversely a compound made up of a strong acid with a weak
base will tend to show an acid reaction. Thus, nearly all car-
bonates of strong bases, such as potassium, sodium, calcium, and
barium, are alkaline, whereas nearly all the salts of very weak
bases, such as aluminum, are acid in reaction, if they exist at all. 1

1 Aluminum carbonate cannot exist at all. Both the acid and base being
weak, hydrolysis takes place with the formation of aluminum hydroxide.

THE THEORY OF SOLUTIONS 21

The explanation of the hydrolysis is best given by writing
the chemical reactions as follows:

Sodium Carbonate Water Sodic Hydrate Carbonic Acid

(1) Na 2 C0 3 + 2HOH = 2NaOH + H 2 C0 3

Aluminum Sulphate Water Aluminum Hydrate Sulphuric Acid

(2) Al,(SCg 3 + 6HOH = 2A1(0H) 3 + 3H 2 S0 4

In the first reaction sodium hydrate is highly dissociated, car-
bonic acid very slightly dissociated into ions, so that the alkaline
reaction due to hydroxyl ions predominates.

In the second reaction aluminum hydroxide is not dissociated,
while sulphuric acid is highly dissociated, so that the acid reaction
due to the hydrogen ions is produced.

Between these two extremes we meet with a very great num-
ber of cases of hydrolysis, the principles of which should be under-
stood by all manufacturers of protective paint compounds. In
many cases salts and compounds which do not immediately
hydrolyze when dissolved in water, will undergo slow decom-
position under the action of acid gases and moisture in the
atmosphere, and hydrolytic products will be formed. An under-
standing of these points will frequently explain the stimulative
corrosion effects produced hy certain types of pigments.

Electrolysis. — The phenomenon known as electrolysis takes
place whenever a current of electricity passes through a solution
capable of conducting the current. Such a solution is known as
a conductor of the second class to distinguish it from an ordinary
conductor like a metallic wire, which is of the first class. A
solution of sugar will not conduct a current of electricity while
a 1 solution of salt will readily do so. This difference in behavior
is accounted for by the fact that the salt is dissociated into ions
while the sugar is not. A substance which in solution will con-
duct electricity is known as an electrolyte. The phenomenon of
electrolysis shows that when a current is passed through a solution
of an electrolyte, there is a mechanical movement of the ions
towards the electrodes. Thus, if a current is passed through a
solution of hydrochloric acid the positive hydrogen ions will
proceed to the negative electrode where they will plate out after
giving up their electrical charges. Having now assumed the
atomic or gaseous condition, the hydrogen escapes from the sys-
tem in the form of minute bubbles. While this action is occur-

22 CORROSION AND PRESERVATION OF IRON AND STEEL

ring at the negative pole an equivalent amount of chlorine is
being plated out and disengaged at the positive pole.

It is not necessary that an outside or external source of elec-
tricity should be at work before electrolysis can take place. If
two strips of dissimilar metal are plunged part way into a solution
and connected by a wire, or by any other means, across the top
a current will flow around the circuit. This current is generated
at the expense of the more electro-positive metal in the couple.
The electro-positive element rapidly shoots off positive ions into
the solution, thereby leaving itself negatively charged so that it
invariably appears as the negative pole in the circuit. Even
two steel needles from the same package are sufficiently dissimilar
to show a slight difference of potential when coupled in such a
way, and one will be protected while the other suffers accelerated
corrosion. From the standpoint of the electrolytic theory all iron
and steel must be thought of as a composite structure, as though,
indeed, it was compounded of more or less well consolidated bundles
of more or less homogeneous needles or units. This statement
contains the crux of the whole electrolytic explanation of corro-
sion and will be further developed in another chapter.

Modes of Ion Formation. — Jones describes the various modes
of ion formation very clearly in the following paragraphs: 1

"From what has been said thus far, the impression might be
gained that ions can be formed from molecules in only one way
■ — the molecules breaking down directly into an equivalent num-
ber of anions and cations. This is one way in which ions are
formed, and the way with which we are most familiar, since it
occurs most frequently. The following examples illustrate this
mode of ion formation:

(1) HC1 = H 4- CI, H 2 S0 4 = H + H + S0 4

NaOH = Na + OH, Ba(OH) 2 = Ba + OH + OH
KN0 3 = K + N0 3 , K,S0 4 = K + K + S0 4

Another method by which an ion can be formed, is for an atom
to take the charge from an ion, converting it into an atom, the
original atom becoming an ion. Thus, when a bar of zinc is
dipped into a solution of copper salt, the copper which was present

1 Jones, Elements of Physical Chemistry, loc. cit., pp. 419-421.

THE THEORY OF SOLUTIONS 23

in the solution as an ion gives up its two charges to an atom of
zinc, becoming an atom; while the zinc, having received the
charges, becomes an ion. This is the well-known precipitation of
copper from a solution, by zinc. We will call this the second
mode of ion formation.

(2) Cu + S0 4 + Zn = Zn + SO, + Cu.

"All that occurs is a transference of electricity from the copper
to the zinc. This is exactly analogous to what takes place when-
ever one metal replaces another, as it is said, from its salts.

"The replacement of the hydrogen ion from acids by a metal
like zinc is an illustration of the same mode of ion formation.

2HC1 + Zn = Zn + CI + CI + H 2 .

What takes place here is simply the transference of the electrical
charge from the hydrogen which does not hold its charge firmly,
to the zinc which holds its charge much more firmly than the
hydrogen, and, therefore, takes the charge from the hydrogen.

"This is typical of the reaction of acids on metals in general;
and is probably typical of substitution in general. The work of
Thomson makes it highly probable that when substitution takes
place in organic compounds, the entering atom or group takes
the charge away from the atom or group displaced. The sub-
stituting atom or group always has the same charge that the
substituted atom or group had when in the compound. The
entire act of substitution is essentially an electrical act, and not a
chemical act } as that term is usually understood. This is true
whether we are dealing with the substitution of the hydrogen
ions of an acid by a metal, or with substitution in organic com-
pounds.

"Another method of ion formation is where an atom of one
substance passes over into a cation, at the same time that an
atom of another substance passes over into an anion. When
gold is dipped into chlorine water, both the gold and chlorine
are in the atomic or molecular condition. But under these con-
ditions the gold can become a cation, and the chlorine can form
anions. This we will term the third method of ion formation.

(3) Au + CI + CI + CI = + Au + + CI + CI + CI.

24 CORROSION AND PRESERVATION OF IRON AND STEEL

This is usually expressed by saying that gold dissolves in chlorine
water.

"The fourth and last method by which ions are formed is
where an atom passes over into an ion, at the same time convert-
ing an ion already present into one with a different quantity of
electricity upon it. Thus, chlorine brought in contact with
ferrous chloride in solution forms an anion, at the same time
converting the ferrous ion into a ferric ion.

(4) Fe + C1 + CI + CI = + Fe+ CI + CI + CI.

This is an example of what has so frequently been called in chem-
istry, oxidation. The reverse phenomenon is, of course, what
has been termed reduction. In this sense oxidation is simply
increasing the number of charges carried by an ion, and reduction
is diminishing the number of such charges.

"These four methods of ion formation, which have been so
clearly pointed out by Ostwald, 1 include all the cases which are
known. If we study them carefully and apply them to chemical
reactions, we shall see that they throw much light on many
problems in chemistry, the meaning of which has hitherto been
concealed in darkness.' 1

Oxidation and Reduction. — The true nature of the processes
known as oxidation and reduction, as set forth above, should be
thoroughly understood by every student of the problem of corro-
sion. By valence the chemist means the combining power of a
given atom for other atoms, though as has been shown it may
be considered as an expression for the measure of the electrical
state of the ion. Many of the elements have only one known
valence state or combining value, as for instance, hydrogen,
which is always univalent and unites with one other univalent
atom to form a compound as in hydrochloric acid (HC1). Biv-
alent elements like calcium unite with two univalent atoms, as
in calcic chloride (CaCl 2 ). Iron i§ typical of a number of other
elements in that it can occur in, two states of valence, viz., the
ferrous condition in which it is bivalent as in FeCl 2 , and the ferric
condition in which it is trivalent, as in FeCl 3 . Every change from
the lower to the higher state is known as oxidation, and every
change from the higher to the lower is conversely a reduction.

1 See Lehrb. d. Allg. Chem., II, 7S6.

THE THEORY OF SOLUTIONS 25

A neutral solution of ferrous chloride exposed to the air rusts
just as surely as does a moist piece of metallic iron, and it is
probable that both actions are due to ionic changes dependent
upon the electrical states of the reacting ions.

It is characteristic of ferrous ions that they are in an unstable
condition, and in the presence of the oxygen in the atmosphere
are changed to the ferric state. A solution of ferrous chloride
contains ferrous ions and chlorine ions and may be written thus:

FeCl 2 ^ Fe + CI CI

If we admit no more chlorine ions to the system, as the change
from the ferrous to ferric states goes on hydrolysis takes place
and finally in the course of time a reddish rust consisting of
ferric hydroxide will make its appearance.

Fe + CI CI + 2 + 3HOH —

Fe(OH) 3 + HH + CI CI + 2H 2

Ferric 1 hydroxide is an insoluble compound so that when it is
formed it drops out of solution. This action of dropping out
of solution is known to chemists as precipitation, and the result
of such action is called a precipitate. It is at once evident that
such a precipitation removes a certain amount of material from
the solution system, and thereby alters the equilibrium between
solution pressure and osmotic pressure, as has- been previously
explained.

Now if we apply these same principles to the discussion of
the simple system represented by the immersion of a piece of
iron in pure water, the rust which forms can be easily and simply
accounted for. Iron has a definite solution tension in pure water,
especially at certain points on its surface, which are electro-posi-
tive to other points or areas. 3 The iron here enters the solution
as positive ferrous ions, the corresponding negative charges being
assumed by hydroxyls. But immediately the oxygen of the air

1 As a matter of fact ferric hydroxide cannot be properly expressed by a
chemical formula such as Fe(OH)3. It appears as a colloidal precipitate
of indefinite composition as far as the amount of water associated with it
is concerned. It would be more appropriately written as Fe 2 3 xH 2 but, for
the purpose of the present discussion this point may be overlooked.

2 Compare p. 17 and illustration, Fig. 5.

26 CORROSION AND PRESERVATION OF IRON AND STEEL

changes the ferrous ions to the ferric condition resulting in the
hydrolytic formation of the insoluble ferric hydroxide or rust.
The more detailed discussion of the electrolysis which takes place
will be taken up in a later chapter.

Electrolysis and Polarization. — When a metal electrode is
immersed in an electrolyte it tends to dissolve, or in other words,
to pass from the atomic to the ionic condition. This is known
as electrolytic solution pressure. As the atom of metal passes
into solution it assumes a positive charge of electricity which it
must take from some source. If the electrolyte contains positive
hydrogen ions, that is to say, if the electrolyte is acid, the enter-
ing metallic ions will acquire their charges by exchange with
hydrogen, which immediately leaves the system in the form of
minute gaseous bubbles. If the electrolyte is neutral the metallic
ions take their positive charges from the electrode which is thereby
left negatively charged. In almost neutral media a combination
of the two results may take place, the hydrogen being disengaged
too slowly to be visible.

Such a state of affairs leads to a polarization effect first
described by Helmholtz 1 and called by him an electrical double
layer. The positively charged metallic ions cannot escape or
free themselves from the negatively charged surface, although
there is sufficient pressure to prevent the reneutralization of the
charges. This effect polarizes the surface, which for practical
purposes may be considered as plated with positive ions. Walker 2
describes this phenomenon in the following words:

1 Wied. Ann., 7, 337 (1879).

2 Jour. Iron & Steel Inst., I, 70 (1909).

THE THEORY OF SOLUTIONS 27

tively; this property or escaping tendency of the metal is termed
its solution pressure. It is obvious, however, that this action
dan continue for only a short time; owing to the fact that the
mass of metal and solution are of opposite polarity, the electro-
lytic tension becomes so great that no more atoms can escape
to the ionic state, and the solvent action ceases. This condition
was first described by Helmholtz, and called by him an electro-
lytic double layer. If now there be in the water ions of another
metal which has a smaller solution pressure than the one under
consideration, the action as above described will be reversed
and the ion with the less solution pressure will pass back to the
metallic state, plating out on the first metal and giving up its
charge of electricity. At this point the first metal will be charged
positively, and the solution in the immediate vicinity negative,
and there will tend to be set up a second electrolytic double layer
opposite in polarity to the first. The result is, a current of elec-
tricity flows from the metal to the solution at the point where
the metal passes into solution, through the solution to the metal
at the point where the ions of the second metal are plating out,
and back through the first metal to the starting-point again.
The electrolytic double layers are thus destroyed, an electric
current passes, and the solvent action of the water on the first
metal continues.

"This phenomenon and its relation to the corrosion of iron
are clearly exemplified in the well-known Daniel or gravity cell.

"In the case of pure iron in water a perfectly analogous con-
dition is found to exist. Water itself is dissociated to a small
but perfectly definite extent into its ions, hydrogen (H), and
hydroxyl (OH). When a strip of pure iron comes into contact
with water, it sends into the water iron atoms in the form of
positively charged ions. Hydrogen as a metal has a much smaller
solution pressure than iron, and hence an equivalent number of
hydrogen ions plate out on the iron strip (leaving the free hydroxyl
ions with their negative charges to balance the iron ions with
their positive charges), and an electric current flows from the
iron by means of the iron ions to the solution, and by means of
the hydrogens from the solution back to the iron again, thus
completing the circuit. But here comes an important break in
the analogy of the film of copper in the Daniel cell. Deposited
copper is a good conductor of the current, and offers no resistance

28 CORROSION AND PRESERVATION OF IRON AND STEEL

to its flow from the solution to the iron on which it is attached.
The reverse is true of the deposited hydrogen; here we have a
high insulator — a film of gas which offers a great resistance to
the flow of the current. Hence, although in the case of the iron
strip in water all the conditions for continuous solution are pres-
ent, owing to the resistance offered by the deposited hydrogen
film (called polarization) the action must cease.

"Just as in the case of iron in a copper sulphate solution the
rapidity of the action depended upon the number of copper ions
present in the solution, so here the solution of the iron, in the
first instance, depends upon the number of the hydrogen ions
present. This number of hydrogen ions, or the concentration
of these ions, is increased by the addition of any acid. So weak
an acid as carbonic increases the number, but to a relatively
small amount; while a strong acid, like hydrochloric or sulphuric,
adds to the number to such an extent that the solvent action
becomes violent, and the deposited hydrogen comes off as a stream
of gas."

The Passive State of the Metals. 1 — It is well known that
under certain conditions iron and some other metals assume a
condition in which they no longer have all the properties they
usually possess. Iron which has been dipped in certain strong
oxidizing agents no longer exhibits a solution pressure when dipped
into dilute electrolytes, and comports itself as though it were a
noble metal. This was first noticed by Keir, a British chemist,
as early as 1790. 2 He found that specimens of bright steel which
had been dipped into strong fuming nitric acid became passive
and were no longer soluble even in dilute acid, or able to exchange
places with copper when dipped into a dilute solution of copper
sulphate. Other strong oxidizing agents, especially chromic
acid and its soluble salts, have been found to induce this condi-
tion of passivity on the surface of iron. This fact has been made
the subject of special study by one of the authors, the results of
which will be detailed in a later chapter. Since the phenomenon
of passivity is produced only by strong oxidizing agents or by
galvanic contact which causes oxygen to separate on the sur-
face of the iron, it was explained by Faraday, Wiederman, and

1 For the bibliography and an excellent discussion of the passive state see
Byers Jour. Am. Chem. Soc, 1908, 30, 1718.

2 Phil. Trans., London, 1790, 359.

THE THEORY OF SOLUTIONS 29

others 1 as due to a thin oxide film. Another explanation is that
the passivity of iron is due to a polarization effect produced by
the separation and retention of oxygen on the surface of the metal.

Keir 2 observed that polished iron which had been immersed
in red fuming nitric acid was altered in some manner so that its
power of precipitating silver and copper from their solutions was
inhibited, and this occurred, in the discoverer's own words, "with-
out the least diminution of metallic splendour or change of color."
Mugdan 3 discussed the passivity acquired by iron which was
immersed in fuming nitric or sulphuric acids and concluded that
it was not due to the formation of an oxide film, but was a true
passivity in the sense of an ennobling (Veredlung) of the metal,
accompanied by a low electrical potential. Jones 4 has discussed
the subject in the following words: "A number of attempts have
been made to explain the passivity of the metals. Faraday 5
and Schonbein explained the passivity in the case of iron, as due
to the formation of a layer of oxide on the surface of the metal.
This was natural when we consider that iron is rendered passive
by strong oxidizing agents, and loses its passivity when heated
in a reducing gas.

"The oxide layer theory of passivity is now regarded as
untenable, since the passive state has been brought about under
conditions where oxidation is impossible; and further, has been
destroyed under conditions where any layer of oxide if formed
would not be disturbed.

"The same fate has befallen the theory that passivity is due
to the formation of a protective layer of gas over the surface
of the metal. The two views of passivity that have acquired
the greatest prominence are those of Finkelstein 6 and Hittorf. 7
According to the former, active iron is bivalent and passive iron
trivalent. This conclusion was based upon the difference in
potential between iron electrodes and the iron salt in which they
were immersed. The potential difference depends upon whether
the iron salt is in the ferrous or in the ferric conditions.

1 Dammer's Anorg. Chem., 1893, V. 3, p. 294.

2 Ibid.

3 Ztschr. Elektrochemie, 1903, 9, 454.

4 Jones, EL Phys. Chem., p. 441.

6 Phil. Mag. (3), 9, 53 (1836); 10, 175 (1837).
c Ztschr. Phys. Chem., 39, 91 (1901).

7 Ibid., 30, 481 (1899); 34, 385 (1900).

30 CORROSION AND PRESERVATION OF IRON AND STEEL

"Hittorf also points out that in the case of chromium the
passive condition corresponds to the highest valence, and the
active to the lower valence. He thinks that we have to do with
two allotropic modifications of the elements, one of which is
active and the other not."

In spite of these various explanations of the interesting and
extraordinary phenomenon of passivity, its fundamental cause
is not thoroughly understood and is still an open question. It
is possible that plating out of oxygen and a change of valence
of the surface ions takes place simultaneously. Whatever the
cause, however, the condition of passivity has an important
bearing on the problem of the protection of iron and steel, and
will be taken up for further discussion later on.

The Electro-chemical Series of the Metals. — If a strip of iron
is plunged into a solution of a salt of copper, iron will go into
solution and copper will plate out. This is because iron is electro-
positive to copper, or, in other words, iron has a higher solution
pressure than copper. The copper ions in solution give up their
positive charges of electricity to the iron, which in turn assumes
the ionic condition. If now a strip of iron is plunged into a
solution of a salt of zinc no such phenomenon takes place because
iron is electro-negative to zinc, and it is only when a strip of zinc
is plunged into a solution of an iron salt that exchange can take
place. If two metals are connected together in metallic contact
to form a couple and are then plunged into a dilute solution of
an electrolyte, the more electro-positive metal will dissolve and
oxidize (rust), while the more electro-negative will be practically
speaking unacted on. In the meantime current will flow around
the circuit. This is the underlying principle of all galvanic action
and explains the action of all primary batteries. In the following
list a number of the more familiar metals are put down in the order
of their diminishing solution tensions, and this may be termed a

tension series.

Magnesium Nickel

Zinc Lead

Aluminum Copper

Cadmium Tin

Iron Antimony

A metal anywhere in the above series will tend to precipitate,
from a solution of its salts, a metal lower in the series. Thus

THE THEORY OF SOLUTIONS 31

zinc and iron will precipitate copper. A metal at any point
when coupled with another lower in the series and plunged into
a corroding medium will throw off positive ions into solution,
and thereby become the negative pole. Zinc is the negative
pole in almost all the common types of primary batteries. It
is apparent from this that if metals which stand below iron
are successfully used as protective coatings it is because they have
very low solution tensions. It is equally apparent why pinholes
or cracks in a tin or copper coating will result in more rapid cor-
rosion of the exposed portions of the iron. These points must
await further discussion in a later chapter; it should at once be
stated, however, that the electro-chemical state of a metal may
be considerably modified by the presence of other metals which
may be alloyed with it. This is an important point which is
often overlooked in connection with the problem of protecting
iron from corrosion. Thus zinc which is strongly electro-positive
to iron is much used as a protective coating. There are, however,
certain alloys of zinc with iron, which are said to be electro-
negative to iron, and as these alloys may be formed in the spelter
baths of the hot dip process, much of the zinc coating which is
turned out in metallurgical processes may have quite the opposite
effect from the one desired.

The Theory of Indicators. 1 — Certain substances usually of an
organic type when dissolved in water are observed to change
color when the reaction changes. Thus, a solution of litmus,
which shows red in an acid solution, will change to blue in an
alkaline solution. The color change is dependent upon the degree
and state of the electrolytic dissociation of the compounds. Sub-
stances which undergo these delicate changes are known as
indicators. The only one of these which will be referred to in this
work is phenolphthalein, and for the benefit of many persons
who are interested in the corrosion problem, but who are not
familiar with the theory of indicators, the following explanation
is included. Phenolphthalein shows the presence of hydroxyl
ions in a solution by the formation of a pink color, thus indicating

1 It is fair to state that the explanation of the action of indicators based on
the theory of electrolytic dissociation is not universally accepted by chemists.
The principle, however, remains the same, even if a better explanation can be
advanced, for the pink color developed by phenolphthalein will always furnish
a sensitive indication of the negative pole in an electrolytic circuit.

32 CORROSION AND PRESERVATION OF IRON AND STEEL

an alkaline reaction. Phthalic acid was first prepared by Laurent
in 1836, by the oxidation of naphthalene, and was first called
naphthalinic acid. It was afterwards shown that the compound
was not directly related to the naphthalene structure, and Laurent
changed the name to phthalic acid, the derivatives of which
became known later as phthaleins. Phenolphthalein is a prod-
uct which is formed by the condensation of two molecules of
phenol or carbolic acid with the anhydride of phthalic acid. It
is in its nature so weak an acid that it is not dissociated in solu-
tion, and as the molecule is colorless no color is seen when it is
added to a perfectly neutral solution. If, however, an alkali is
added the corresponding salt of the weak acid is formed, which
immediately dissociates with the formation of a colorless metallic
cation, and the strongly rose-colored organic anion. Thus all
hydroxides of basic elements will show the pink color in solution,
even when present in only the slightest excess. On this account
phenolphthalein is an exceedingly delicate indicator of the pres-
ence of hydroxyl ions.

Phenolphthalein as an Indicator of the Presence of Electrolysis.
— As hydroxyl ions are always found in more or less excess around
the negative pole of a galvanic circuit after the positive hydrogen
ions have neutralized their charges and disappeared, phenol-
phthalein can be used as an indicator of the existence of a negative
pole. In order to show the sensibility of this substance for indi-
cating an excess of hydroxyl ions by the development of a distinct
pink color, the following simple experiment has been described: 1

Five hundred and fifty c.c. of distilled water containing 1 c.c.
of phenolphthalein indicator was boiled down in a Jena flask to
500 c.c. One one-hundredth normal potassium hydroxide solu-
tion was then run into the quickly cooled water from a burette.
It was thus found that about 1 c.c. of one one-hundredth normal
potassium hydroxide was the limit of the quantity necessary to
produce a distinctly visible pink color; 1 c.c. of one one-hundredth
normal potassium hydroxide contains 0.00017 gram of hydroxyl.
This quantity in 500 c.c. of water represents a concentration of
about 0.35 part of hydroxyl per million.

This experiment shows how extremely delicate the phenol-
phthalein test is as an indicator of the presence of hydroxyl
ions, and hence of the negative pole in an electrolytic circuit.
1 Cushman, Bull. 30, Office Public Roads, U. S. Dept. Agr.

THE THEORY OF SOLUTIONS 33

Colloids and Crystalloids. — Before closing this chapter on
solutions, a brief reference should be made to the condition of
substances as they precipitate out of solution, as a general knowl-
edge of the subject will be of assistance in the design and prepara-
tion of inhibitive compounds. The action of colloids under the
effects of electrolysis will also be found later to have a bearing
on the corrosion problem.

Electro-Chemical Properties of Colloids. — Whenever substances
suddenly precipitate out of solution they appear either in a finely
crystalline condition in which they are denominated crystalloids,
or in a gummy, gelatinous, more or less coagulated condition, as
the case may be, in which we define them as colloids, or colloidal
precipitates. It will not be necessary here to give a full descrip-
tion of the properties of colloids in general, as this information
can be obtained from any complete text-book on physical chem-
istry. The points that particularly interest us in this work are
the peculiar absorbent powers of colloidal precipitates and their
electro-chemical properties. Many colloidal precipitates are
-dried and used as pigments, and such substances not only absorb
considerable quantities of water, but also absorb and carry down
with them impurities from the mother liquors or solutions from
which they were precipitated.

There seems to be satisfactory evidence that colloidal pre-
cipitates in suspension in an electrolyte carry electrical charges
just as ions do. This is shown by the fact that the colloidal
particles when suspended in a solution migrate to the electrodes
under the influence of electrolysis. Colloidal ferric hydroxide,
which although not exactly is essentially similar to all freshly
formed iron rust, moves with the current to the cathode, this
peculiarity it shares with all basic hydroxides. Acid hydroxides
like precipitated silicic acid move in the opposite direction. The
ferric hydroxide is therefore charged positively and the silicic
acid negatively. This property appears to be a general one for
all colloidal suspensions, and seems, under certain conditions, to
give them the tendency of absorbing from the solution ions of
opposite sign. According to Jones, two theories have been pro-
posed to account for the electrification of the colloid particles.
According to one of these theories, the particles of the colloidal
precipitates assume the charge of one sign while the surrounding
water takes on the charge of the other sign. This explanation

34 CORROSION AND PRESERVATION OF IRON AND STEEL

does not seem to be in accord with the observed facts, and the
more probable theory is that from every colloidal aggregate there
split off either positive or negative ions, leaving the residue of
the aggregate carrying the opposite charge. Thus, basic hydrox-
ides would split off hydroxyl ions, and the residue be charged
positively; in the same way silicic acid would split off hydrogen
ions, leaving the colloidal residue charged negatively. 1

This absorbent tendency of colloidal precipitates, both for
water and salts, as well as their electro-chemical properties,
should be thoroughly understood by paint technologists who are
studying the protection of iron and steel. The explanation of
the stimulating effects produced by certain pigments which are
theoretically inhibitors is usually furnished by these colloidal
absorptions which take place in the course of manufacture of
the pigments.

Conclusion of the Chapter. — It is hoped that a careful reading
and study of this chapter by metallurgists and others interested
in the special problems with which this book deals will lead to a
clearer understanding of the discussion and data which is included
in succeeding chapters. It is not necessary that the reader should
be a chemist in order to understand the explanations which have
been presented, and it is hoped that those who have never studied
even the rudiments of chemistry and electrical action will have
been introduced to theories which are further developed in suc-
ceeding chapters. In conclusion, the authors desire to state
that they are conversant with the interesting generalizations
of Kahlenberg 2 and his school, who do not accept the theory of
solutions, and who have brought strong arguments, based on ex-
perimental evidence, to bear against it. Nevertheless, the theory
has been and still is a useful one whatever its ultimate fate may
be; the authors are strong in the conviction that the theory of
solutions expresses fundamental truth even if it must eventually
be modified in form.

1 Jones, El. Phys. Chem., p. 283.

2 See Science, XXXI, 785, 41.

CHAPTER III

THE THEORY OF CORROSION

The Three Theories which have been Advanced to Explain Cor-
rosion. — Three separate theories which, though they all more
or less overlap, nevertheless involve distinctly different reactions,
have been advanced and strenuously defended in the effort to
furnish an explanation for the rusting of iron. These may be
stated as the carbonic-acid, the hydrogen-peroxide, and the
electrolytic theories.

Before any distinct progress can be made in the manufacture
of metal that shall be more than ordinarily resistant to corrosion,
it is of great importance that the underlying causes of oxidation
should be clearly understood. It is the object of this chapter to
discuss the different theories and to present certain evidence
which bears directly upon the subject.

The Carbonic-acid Theory. 1 — The carbonic-acid theory is the
one which until recently was most generally held. The theory
presumed that without the interaction of carbonic or some other
acid the oxidation, or better, the hydroxylation, of iron cannot
take place. The theory is best set forth in the words of a text-
book recently published. 2

"The process of rusting is a cyclical one, and three factors
play an important part: (1) An acid, (2) water, (3) oxygen.
The process of rusting is always started by an acid (even the weak
carbonic acid suffices); the acid changes the metal to a ferrous
salt with evolution of hydrogen:

2Fe + 2H 2 C0 3 = 2FeC0 3 + 2H 2 .

"Water and oxygen now act upon the ferrous salt, causing the
iron in this salt to separate out as ferric hydroxide, setting free the
same amount of acid which was used in forming the ferrous salt :
2FeC0 3 + 5H 2 + = 2Fe(OH) 3 + 2H 2 C0 3 .

1 The best presentment of the general acid theory is due to Friend, Jour.
Iron and Steel Inst., 77, 5, 1908.

2 Treadwell and Hall, Analytical Chemistry, 1907, p. 92.

36 CORROSION AND PRESERVATION OF IRON AND STEEL

"The acid which is set free again acts upon the metal, form-
ing more ferrous salt, which is again decomposed, forming more
rust. A very small amount of acid, therefore, suffices to rust
a large amount of iron. If the acid is lacking, the iron will not
rust. If we desire to prevent this rusting, we must neutralize the
acid, e.g. } add milk of lime. Iron remains bright under an alkali."

Corrosion Occurs when Carbonic Acid is Absent. — Although
the above explanation is sufficiently plausible, and in spite of the
fact that carbonic acid, as well as other acids, does act a part in
the ordinary rusting of iron, it will presently be shown that iron
readily oxidises, not only when carbonic acid is entirely absent,
but also in dilute alkaline solutions. It is only when the hydroxyl
ions supplied by an alkaline solution have reached a certain con-
centration that rusting is entirely prohibited.

The carbonic-acid theory was founded originally on the inves-
tigations of Crace Calvert, 1 as interpreted by Crum Brown. 2 It
has also more recently been vigorously defended by Moody, 3
who insists that with water and oxygen quite free from carbonic
acid iron cannot rust. This view is, however, not shared by
Dunstan, Jowatt, and Goulding, 4 or by Whitney 5 or Cribb, 6 all
of whom give experimental evidence to show that rusting takes
place rapidly in the absence of carbonic acid, provided liquid
water and oxygen ar6 present. The experiment of Dunstan and
his coworkers was so carefully carried out that there seems to be
no doubt that if carbonic acid plays any role whatever it is an
unimportant one, and that rusting can go on with extreme rapidity
in its absence.

In order to confirm this conclusion the following experiment
was made by one of the writers :

The Jena glass flasks A, B, and the beaker C, shown in Fig. 6,
were nearly filled with freshly distilled water and boiled vigor-
ously for one-half hour. While the boiling was still proceeding
bright polished strips of charcoal iron and steel were slipped into
flasks A and B, and the rubber stoppers, which had been previously

1 Manchester Lit. Phil. Mem., 1871, 5, 104.

2 Jour. Iron and Steel Inst., 1S88, 129-131.
3 Proc. Chem. Soc. (bond.), 1906, 22, 101.

A Jour. Chem. Soc. (bond.), 1905, 87, pt. 2, 1548.
5 Jour. Am. Chem. Soc, 1903, 25, 394.
°The Analyst, 1905, 30, 232.

THE THEORY OF CORROSION

cleaned by prolonged boiling in pure water, tightly inserted.
After boiling for fifteen minutes longer the clamp at the end of
tube D was opened for a moment and the back pressure allowed
to drive any last traces of air from the tube. After tightly clos-
ing the clamp again, the lamps under flasks A and B were removed
while the water in C was still kept boiling. Boiled water imme-
diately sucked back until the whole apparatus was completely
filled, no trace of air being present. At all events, no slightest
trace of rust appeared on the bright metal strips when kept
indefinitely under this boiled water. Pure oxygen from a cylin-

Fig. 6. — Apparatus to show the action on iron of
pure water and oxygen.

der was now washed perfectly free from last traces of carbonic
acid by passing the gas through a train of wash bottles containing
caustic potash,- barium hydroxide, and lime water. On allowing
this carefully purified oxygen to enter at D and bubble through
the system of flasks, rust appeared on the bright metal surfaces
in five minutes or less and in one hour had become deep and heavy.
The action, just as in Dunstan's experiments, did not take place
evenly all over the surface, but in patches, which had the appear-
ance of a more or less regular pattern following the physical
structure of the metal. This experiment has been frequently
repeated, with every possible precaution to avoid the entrance
of even the smallest trace of carbonic acid. On numerous occa-
sions a few drops of phenolphthalein indicator was added to the
boiling water in the three flasks, and invariably a pink color

38 CORROSION AND PRESERVATION OF IRON AND STEEL

developed, proceeding from the metallic surfaces. This effect
will be discussed at length further on and is mentioned at this
place as contributory evidence that carbonic acid is not neces-
sarily present, as Moody believes, before any reaction between
iron, water, and oxygen can take place.

If pure, dry carbonic acid gas, freed from oxygen by passing
through several wash bottles containing pyrogallic acid dissolved
in sodium bicarbonate solution, was substituted for the pure
oxygen gas and allowed to enter through tube D, no perceptible
action took place on a bright piece of steel after several hours,
although there can be no doubt that iron passed to a slight extent
into solution as ferrous carbonate. Finally, if pure oxygen was
allowed to enter at the same time and mingle with the carbonic
acid, corrosion began in a short time. There was, however, a
difference in the appearance of the rust that was formed with
and without the interaction of carbonic acid. In the presence
of carbonic acid the characteristic blue-green gradually changing
to the red color peculiar to rust was observed. This appearance
invariably accompanies the early stage of attack when iron is
rusting in the presence of carbonic acid. In the experiments in
which pure oxygen alone was permitted to enter the apparatus
the blue-green initial stage of oxidation was never observed, the
red ferric hydroxide making its appearance from the first, as it
usually does, in normal cases of atmospheric rusting of bright
iron surfaces. 1

It may be doubted whether it is possible to boil out all car-
bonic acid from the water contained in the apparatus shown in
Fig. 6. Granting that this is the case in regard to last traces,
it is easily shown that the hydrogen ions which would be supplied
by a minute quantity of carbonic acid are of no more importance
than the hydrogen ions supplied by the normal dissociation of
pure water, and that the assumption that carbonic acid must be
present is quite unnecessary. Whitney 2 shows this very clearly
in the following paragraph:

1 Statements frequently appear in discussions of the corrosion problem
in regard to the importance of analyses of various samples of iron rust. As
a matter of fact little can be learned from such analyses. The ferroso-ferric
hydroxides, carbonates, etc., that are formed are of indefinite composition and
their chemical constitution throws little if any light on the mechanism of
the corrosion reactions.

2 Jour. Am. Chem. Soc, 1903, 25, 397.

THE THEORY OF CORROSION 39

"Assuming the laws of Henry and Dalton to apply to the
solubility of carbonic acid gas in water, also that the solubility
of the pure gas under ordinary pressure is one volume for one
volume of water (which is correct at 15° C), and, finally, that the
normal content of carbonic acid in the atmosphere is 2 parts
in 10,000 by volume, we should expect water in equilibrium with
air containing this concentration of carbonic acid to contain
0.0002 volume carbon dioxide per volume of water. This corre-
sponds to a concentration of the carbonic acid equal to 0.00001
mol per liter, or 0.00002 normal. From the dissociation constant
3040 X 10- 10 determined by Walker, 1 it follows that the first
hydrogen of the acid is 16 per cent, dissociated at this concentra-
tion. From this it follows that 10,000,000 liters of water contain-
ing carbonic acid in equilibrium with ordinary air at 15° contains
16 grams of hydrogen ions, or only 16 times as many as perfectly
pure water contains. At the boiling temperature the carbon
dioxide dissolved would probably yield a concentration of hydro-
gen ions even less than in pure water, for not only is the solubility
of the gas greatly diminished, but the dissociation of water is
greatly increased by rise of temperature. Moreover, the dis-
tilling water would rapidly reduce the concentration of any car-
bonic acid capable of dissolving in water at 100° C."

The carbonic acid theory does not furnish a complete explana-
tion of the phenomenon of corrosion, but it does express partial
truth inasmuch as hydrogen ions must be present before the attack
on the surface of iron can be made. The fact as brought out in
the previous paragraph that even pure water provides a sufficient
number of hydrogen ions to start the action shows that the role
of carbonic acid is only contributory and not the sole cause.

The Peroxide Theory of Corrosion. — The peroxide theory of
corrosion has not been found in accordance with observed facts,
but it will be briefly stated here for the sake of completeness.
This theory is based on the well-known scheme of oxidation
processes advanced by Traube. 2 Thus the chemical reactions
concerned in the formation of iron rust should be written:

Fe + 2 + H 2 = FeO + H 2 2

2FeO + H 2 2 = Fe 2 2 (OH) 2 = Fe 2 3 , H 2

1 Zts. Phys. Chem., 1900, 32, 137.
2 Ber. d. chem. Ges., 18, 1881.

40 CORROSION AND PRESERVATION OF IRON AND STEEL

The excess of hydrogen peroxide immediately reacts with the
iron, forming a further quantity of rust:

Fe + H 2 2 = FeO + H. 2
2FeO + H 2 2 = Fe 2 2 (OH) 2 = Fe 2 3; H 2

One of the arguments which was used by its supporters to
strengthen the theory was the seemingly extraordinary but now
well-known fact that iron cannot rust in solutions of certain
strong oxidizing agents, such as chromic acid and its salts. As
chromic acid is known to destroy hydrogen peroxide, the explana-
tion on first thought seems reasonable. The theory also seemed
to derive some confirmation from the fact that delicate tests for
hydrogen peroxide have beeen obtained during the slow oxida-
tion of zinc and some other metals. On the other hand, in the
case of iron these same delicate tests obstinately refuse to reveal
even its transitory presence during the ordinary process of rust-
ing. The theory has been criticised by Divers, 1 Moody, 2 and
Cribb, 3 — the first named having pointed out that it is not ten-
able to argue that, because such substances as chromic acid and
alkalis gradually destroy hydrogen peroxide, they must prevent
its formation. For instance, ferrous sulphate is oxidized by free
chlorine, but it does not prevent manganese dioxide and hydro-
chloric acid from reacting when brought together in its presence.
Moreover, if the formation of hydrogen peroxide was a necessary
stage in the rusting of iron, and this is inhibited by certain sub-
stances which destroy hydrogen peroxide, why is not the inhibi-
tion extended to strong reducing agents generally? The theory
is an interesting and suggestive one, but in the author's opinion
is not supported by the facts.

The Electrolytic Theory. — From the standpoint of the modern
theory of solutions, all reactions which take place in the wet
way are attended with certain readjustments of the electrical
states of the reacting ions. The electrolytic theory of rusting
assumes that before iron can oxidize in the wet way it must first
pass into solution as a ferrous ion. The subject has been interest-
ingly treated by Whitney, 4 who discussed it from the standpoint

1 Proc. Chem. Soc. (Lond.), 1905, U, 251.

2 Jour. Chem. Soc. (Lond.), 1906, 89, 90, 720.

3 Analyst, 1905, 30, 225.
- hoc. cit., p. 38.

THE THEORY OF CORROSION 41

of Nernst's conception of the source of electro-motive force
between a metal and a solution. When a strip of metallic iron is
placed in a solution of copper sulphate, iron passes into solu-
tion and copper is deposited, this change being of course accom-
panied by a transfer of electrical charge from the ions of copper
to those of iron. Hydrogen acts as a metal and is electrolytic ally
classed with copper in relation to iron. If, therefore, we immerse
a strip of iron in a solution containing hydrogen ions, an exactly
similar reaction will take place, iron will go into solution, and
hydrogen will pass from the electrically charged or ionic to the
atomic or gaseous condition. In such a system the solution of
the iron, and, therefore, its subsequent oxidation, must be accom-
panied by a "precipitation" or setting free of hydrogen. It is
very well known that solutions of ferrous salts as well as freshly
precipitated ferrous hydroxide are rapidly oxidized by the free
oxygen of the air to the ferric conditions, so that if the electrolytic
theory can account for the original solution of the iron the explana-
tion of rusting becomes an exceedingly simple one.

Pure Water a Solvent of Iron. — As iron has been shown by
Whitney, Dunstan, and one of the authors, to rust in the presence
of pure water and oxygen alone, the electrolytic theory as a funda-
mental cause of the wet oxidation of iron must stand or fall on
the determination of one crucial question, viz.: Does iron pass
into solution, even to the slightest extent, in pure water? If
iron does dissolve, the electrolytic theory is so far satisfactory;
if it does not dissolve, we must conclude that the oxygen finds
some way of directly attacking the metal.

Almost every one will admit that in the case of impure iron,
with its unhomogeneous physical and chemical constitution, elec-
trolysis will supervene, but it must be remembered that we are
now concerned with the underlying cause of the wet oxidation or
hydroxylation of iron, regardless of its state of chemical purity.

According to the dissociation theory, even the purest water
contains free hydrogen ions to the extent of about 1 gram in
10,000,000 liters. If iron dissolves in the purest water it should
be by interchange with hydrogen, and as Whitney 1 has pointed
out, pure water is to this extent an acid. In order to get experi-
mental evidence on this crucial point, Whitney describes the
following experiment:

1 hoc. Hi., p. 38.

42 CORROSION AND PRESERVATION OF IRON AND STEEL

"A clean bottle was steamed out for a time to remove soluble
alkali from the glass and was then filled with pure distilled water,
which was kept boiling by passing steam through it for fifteen
minutes. While still boiling, a bright piece of iron was placed
in the bottle. A stopper (in some cases rubber and in others
cork) carrying a tube open in a capillary several inches above the
stopper was inserted into the bottle and firmly fastened in place,
the water being kept boiling. Finally, the glass capillary was
heated hot by means of a blowpipe and sealed by squeezing the
walls together. The bottle was then allowed to cool to a tempera-
ture of about 80° C, and the neck of the bottle was finally covered
with paraffin to prevent leaking. It was thought that in this way
the oxygen, carbonic acid, and other gases in the water were
completely removed. Bottles containing iron and sealed in this
manner have stood without any visible change for weeks. In
some cases a little air was subsequently admitted to bottles
which had stood in this way with the iron apparently unaffected,
and within a few minutes the water became cloudy and assumed
a yellow color. Ordinary rust rapidly deposited upon the glass
and in spots upon the metal. In fifteen or twenty minutes the
production of rust throughout the bottle was perfectly evident.
It seemed plain from the rapidity of formation of oxide and its
precipitation on the glass that the iron had dissolved in the water
before the addition of the air, and that the latter simply per-
mitted the formation of the insoluble oxide."

In order to obtain more light on the subject one of the writers
devised the following experiment, which is sufficiently simple to
be repeated by any one without encountering any difficulties
whatever. The apparatus used is shown in Fig. 7.

The two clean Jena glass flasks A and B are three-quarters
filled with pure freshly distilled water. Two drops of an alcoholic
solution of phenolphthalein indicator (1 gram in 100 c.c. pure
alcohol) are added to the water in each of the flasks. The beaker
C is more capacious than the flasks A and B. The flasks D and E
are used in each experiment as blanks to check the results obtained.
After connecting up as shown, the water in each vessel is simul-
taneously boiled very vigorously until about one-quarter is
boiled off. The rubber stopper in A is then lifted, and clean,
polished strips of iron quickly slipped in. The stopper is again
tightly inserted and the boiling continued for about fifteen minutes.

THE THEORY OF CORROSION

The lamps under A and E are then extinguished, while the water
in B, C, and D continues to boil. As soon as flasks A and E
have sucked back boiling water so that they are completely filled,
the lamps under flasks B and D are also extinguished. When B
is quite full, flasks A and B are quickly cooled by surrounding
them with cold water. The valve at F is then closed. By this
means the bright specimens are immersed in water practically
free from air, oxygen, or carbonic acid, and may be kept under
observation for any desired length of time. This experiment
has been repeated a great number of times with different samples
of iron and steel, and no rusting has ever been observed unless
air was allowed to enter.

Fig. 7. — Apparatus to determine the extent of the solubility of iron

in pure water.

It has been shown that the electrolytic theory of the wet
oxidation of iron is based on the premise that iron must first
go into solution, an equivalent amount of hydrogen being set
free. The resulting ferrous hydroxide in solution betrays its
presence by producing a pink coloration with the phenolphthalein
indicator. In every experiment made the pink color was seen,
although in some cases the color developed slowly and only after
the lapse of a number of hours. That the color was not due to
the action of the water on the Jena glass was shown by the fact
that no color appeared on the blank side of the experiments.

Since it was thought that some doubt might be felt whether
even the small amount of phenolphthalein present could attack
the iron, the experiment was repeated with iron and boiled water

44 CORROSION AND PRESERVATION OF IRON AND STEEL

alone, but the results invariably showed that a small amount of
iron had dissolved. In view of the ease with which these experi-
ments can be confirmed, it would seem needless to yield more
space to this phase of the discussion. It appears to the writer
to be demonstrated that Whitney was right in his assertion that
iron goes into solution up to a certain maximum concentration
in pure water, without the aid of oxygen, carbonic acid, or other
reacting substances.

Rusting of Iron Primarily due to Attack by Hydrogen Ions. —
This point established, it becomes apparent that the rusting of
iron is primarily due, not to attack by oxygen } but by hydrogen ions.
Absolute confirmation of this view will be given later on.

In order that rust should be formed iron must go into solution
and hydrogen must be given off in the presence of oxygen or cer-
tain oxidizing agents. This presumes electrolytic action, as every
iron ion that appears at a certain spot demands the disappearance
of a hydrogen ion at another, with a consequent formation of
gaseous hydrogen. The gaseous hydrogen is rarely visible in the
process of rusting, owing to the rather high solubility and great
diffusive power of this element. Substances which increase the
concentration of hydrogen ions, such as acids and acid salts,
stimulate corrosion, while substances which increase the con-
centration of hydroxyl ions inhibit it. Chromic acid and its salts
inhibit corrosion by producing a polarizing or dampening effect
which prevents the solution of iron and the separation of hydro-
gen. This will be more fully discussed later on.

Brief Explanation of Corrosion of Iron, from Standpoint of
Electrolytic Theory. — From the standpoint of the electrolytic
theory, the explanation of the corrosion of iron is not complicated,
and so far has been found in accordance with all the facts. Briefly
stated, the explanation is as follows: Iron has a certain solution
tension, even when the iron is chemically pure and the solvent
pure water. The solution tension is modified by impurities or
additional substances contained in the metal and in the solvent.
The effect of the slightest segregation in the metal, or even unequal
stresses and strains in the surface, will throw the surface out of
equilibrium, and the solution tension will be greater at some points
than at others. The points or nodes of maximum solution
pressure will be electro-positive to those of mimimum pressure,
and a current will flow, provided the surface points are in contact,

THE THEORY OF CORROSION 45

through a conducting film. If the film is water, or is in any way
moist, the higher its conductivity the faster iron will pass into
solution in the electro-positive areas, and the faster corrosion
proceeds. Positive hydrogen ions migrate to the negative areas,
negative hydroxyls to the positives. As explained in a previous
chapter, by a hydrogen ion is meant a dissociated hydrogen atom

carrying its equivalent static electrical charge, which may be

+
represented by the symbol H. The hydroxyl ion is written OH.

Water, which may be expressed by the symbol HOH, is made up

+ -

of the dissociation products H and OH. An acid like hydro-

+
chloric acid (HC1) dissociates into H and CI. An acid is always
highly dissociated in solution, while water itself is only slightly
dissociated. This explains why the presence of an acid increases
the concentration of the hydrogen ions. Ionization always takes
place in every solution of an inorganic compound, and even the

purest water is slightly dissociated into its constituent ions H

and OH. The more ionized a solution is, the higher its electrical
conductivity, and the more rapid the damage to the underlying
iron.

If the concentration of the hydrogen ions is sufficiently high,
which, as has been shown, is only the same as saying if the solu-
tion is sufficiently acid, the hydrogen ions will exchange their
electrostatic charges with the iron atoms sweeping into solution,
and gaseous hydrogen is seen escaping from the system. This
takes place whenever iron is dissolved in an acid. If, however,
as is usual in ordinary rusting, the acidity is not high enough to
produce this result, the hydrogen ions will polarize to a great
extent around the positive nodes without accomplishing a complete
exchange, and the so-called electrical double layer of Helmholtz
will be formed. 1 This polarization effect resists and slows down
the action. Nevertheless, although it cannot be seen, some
exchange takes place and iron slowly pushes through, as is shown
by the development of the blue nodes in the ferroxyl test. For
every exchange of static charge between iron and hydrogen at the
positive node, a corresponding negative hydroxyl ion appears at
the negative node which is shown in pink with the ferroxyl indica-

1 Wied. Ann., 1879, 7, 337.

46 CORROSION AND PRESERVATION OF IRON AND STEEL

tor. In other words, as fast as the iron sweeps into solution the
concentration of ferrous hydroxide grows, but the ferrous reaction
appears in one place and the hydroxyl in another. It is now that
the oxygen of the atmosphere dissolved in the solution takes up
its work, the ferrous ions are oxidized to the insoluble ferric
condition, which results in the precipitation of rust, and the
action of hydrolysis proceeds. The formation of the insoluble
ferroso-ferric carbonates and hydroxides, changing to the red
ferric hydroxide known as rust, is familiar to every one.

W. H. Walker puts into the following words a very similar
explanation of the action which takes place: 1

"Every metal when placed in water, or under such conditions
that a film of water may condense upon it, tends to dissolve in
the water, or, in other words, to pass from its atomic or metallic
condition into its ionic condition. This escaping tendency of the
metals varies from that shown by sodium or potassium, which is
so great as to cause instant and rapid decomposition of the metal
and water, to gold or platinum where such tendency to dissolve
is zero. Between these two extremes we find the other common
metals, including thereunder the element hydrogen, which may be
considered as a metal. As the atom of metal passes into the
water, it assumes a positive charge of electricity, leaving the
metallic mass from which it separated charged negatively; this
property or escaping tendency of the metal is termed its solution
pressure. It is obvious, however, that this action can continue
for only a short time; owing to the fact that the mass of metal
and solution are of opposite polarity, the electrolytic tension
becomes so great that no more atoms can escape to the ionic
state, and the solvent action ceases. This condition was first
described by Helmholtz, and called by him an electrolytic double
layer. If now there be in the water ions of another metal which
has a smaller solution pressure than the one under consideration,
the action as above described will be reversed and the ion with
the less solution pressure will pass back to the metallic state,
plating out on the first metal and giving up its charge of electricity.
At this point the first metal will be charged positively, and the
solution in the immediate vicinity negative, and there will tend
to be set up a second electrolytic double layer opposite in polarity
to the first. The result is, a current of electricity flows from the
1 Jour. Iron and Steel Inst., 1, 70 (1909).

THE THEORY OF CORROSION 47

metal to the solution at the point where the metal passes into
solution, through the solution to the metal at the point where the
ions of the second metal are plating out, and back through the
first metal to the starting-point again. The electrolytic double
layers are thus destroyed, an electric current passes, and the sol-
vent action of the water on the first metal continues.

"This phenomenon and its relation to the corrosion of iron
are clearly exemplified in the well-known Daniel or gravity cell.
In the case of pure iron in water a perfectly analogous con-
dition is found to exist. Water itself is dissociated to a small
but perfectly definite extent into its ions, hydrogen (H) and
hydroxyl (OH). When a strip of pure iron comes into contact
with water, it sends into the water iron atoms in the form of
positively charged ions. Hydrogen as a metal has a much smaller
solution pressure than iron, and hence an equivalent number of
hydrogen ions plate out on the iron strip (leaving the free hy-
droxyl ions with their negative charges to balance the iron ions
with their positive charges), and an electric current flows from
the iron by means of the iron ions to the solution, and by means
of the hydrogens from the solution back to the iron again, thus
completing the circuit. But here comes an important break in
the analogy of the film of copper in the Daniel cell. Deposited
copper is a good conductor of the current, and offers no resistance
to its flow from the solution' to the iron on which it is attached.
The reverse is true of the deposited hydrogen; here we have a
high insulator — a film of gas which offers a great resistance to
the flow of the current. Hence, although in the case of the iron
strip in water all the conditions for continuous solution are present,
owing to the resistance offered by the deposited hydrogen film
(called polarization) the action must cease.

" Just as in the case of iron in a copper sulphate solution, the
rapidity of the action depended upon the number of copper ions
present in the solution, so here the solution of the iron, in the
first instance, depends upon the number of the hydrogen ions
present. This number of hydrogen ions, or the concentration of
these ions, is increased by the addition of any acid. So weak an
acid as carbonic increases the number, but to a relatively small
amount; while a strong acid, like hydrochloric or sulphuric, adds
to the number to such an extent that the solvent action becomes
violent, and the deposited hydrogen comes off as a stream of gas.

48 CORROSION AND PRESERVATION OF IRON AND STEEL

" Since the presence of the polarizing film of hydrogen arrests
the further solution of the iron, it is obvious that in order for
the reaction to proceed this hydrogen must be removed. The
destruction of the hydrogen film in ordinary corrosion is accom-
plished by the oxygen of the atmosphere, which is dissolved in
the water. The action here taking place is a simple union of the
hydrogen on the iron and the oxygen dissolved from the air,
with the re-formation of water. It follows from this that any
substance which dissolves or reacts with hydrogen should acceler-
ate corrosion. This is found to be in fact the case.' 5

Development and Use of Ferroxyl Reagent. — It is a matter of
common observation that iron usually corrodes rapidly at cer-
tain weak points, the effect produced being known as pitting.
That this effect can best be explained by the electrolytic theory
there can be no doubt, and it was in the effort to obtain an actual
demonstration that the now well-known "ferroxyl" reagent was
evolved. 1 Early in his investigations one of the authors observed
that whenever a specimen of iron or steel is immersed in water
or a dilute neutral solution of an electrolyte to which a sufficient
quantity of phenolphthalein indicator has been added, a pink
color is developed. If the solution is allowed to stand perfectly
quiet, it will be noticed that the pink color is confined to certain
spots or nodes on the surface. The pink color of the indicator is a
proof of the presence of hydroxyl ion's, and thus indicates the nega-
tive poles. It may be added that some specimens of steel exhibit
this phenomenon much more quickly and distinctly than others.

Since phenolphthalein shows only the nodes where solution
of iron and subsequent oxidation does not take place, Walker 2
suggested the addition of a trace of potassium ferricyanide to the
reacting solution, in order to furnish an indicator for the ferrous
ions whose appearance marks the positive poles. If iron goes into
solution, ferrous ions must appear, which, with ferricyanide,
form the well-known TurnbulPs blue compound. Walker further
showed that if the reagent is stiffened with gelatin or agar-agar,
diffusion is prevented and the effects produced are preserved.
For this combined reagent, which indicates at one and the same
time the appearance of hydroxyl and ferrous ions at opposite

1 Announced in a paper before the American Soc. Testing Materials,
June 21, 1907.

2 Jour. Am. Chem. Soc., 29, 1257, October, 1907.

THE THEORY OF CORROSION 49

poles, one of the authors has suggested for the sake of brevity
the name "ferroxyl." If the reagent has been properly prepared,
the color effects are strong and beautiful. In the course of a few
days the maximum degree of beauty in the colors is obtained, after
which a gradual deterioration sets in, although the effects can be
preserved for a long time by flooding the jelly with alcohol.

In the pink zones, as would naturally be expected, the iron
remains quite bright as long as the pink color persists. In the
blue zones the iron passes into solution and continually oxidizes,
with a resulting formation of rust. Even the purest iron develops
the nodes in the ferroxyl indicator, but impure and badly segre-
gated metal develops the colors with greater rapidity and with
bolder outlines. This result would of course be expected, as in
pure iron the formation of poles would be conditioned by a
much more delicate equilibrium than in impure iron, where varia-
tions in concentration of the dissolved impurities would stimulate
the electrolytic effects. Even so-called chemically pure iron
contains small quantities of dissolved gases, and it is not improb-
able that even slight variations in the physical homogeneity of
pure iron will occasion the electrolytic effects which are made
visible by this delicate reagent.

Ferroxyl Indicator shows that Solution Tension of Iron Varies
at Different Points on the Surface. — It should be remembered
that the electrolytic action indicated by the ferroxyl tests is con-
ditioned by a very delicate equilibrium. In order to decompose
water by the electric current a very much higher difference of
potential is needed than ever exists between points on the surface
of a piece of corroding iron or steel, however badly segregated it
may be. The entire value of the ferroxyl test consists in the fact
that it indicates in a very delicate way that the solution tension
of iron is higher at certain points on the surface than it is at
others. Or, in other words, that certain surface points are, be
it ever so slightly, electro-positive to others.

The very delicacy of the test shows that it must be carried
out with precise care. If a sewing needle is taken and rubbed
with the finger at one end and then placed in the reagent and
left there, it will be noticed that where the fingers had come in
contact with the steel, the slight trace of sodium chloride from
the fingers will cause corrosion to commence and continue. 1
1 See Stead, Jour. Iron and Steel Inst., 1, 97 (1909).

50 CORROSION AND PRESERVATION OF IRON AND STEEL

Bresch 1 in a short review of the new theory of corrosion has
made use of a graphic method to show the action which takes
place when a piece of iron is immersed in the ferroxyl reagent.
This diagram, which is self-explanatory, is shown in Fig. 8.

Ferroxyl Reagent

Fig. 8. — Diagrammatic explanation of action in ferroxyl indicator.

Preparation of Ferroxyl Mounts. — The ferroxyl mounts are
prepared as follows: A lj per cent, solution of agar-agar is first
made by dissolving a weighed quantity of powdered agar in the
requisite amount of water. This solution is boiled for one hour,
fresh water being added to replace that lost by evaporation. It
is then filtered while hot and 2 c.c. of standard phenolphthalein
indicator added to every 100 c.c. solution, after which it is brought
to a perfectly neutral condition by titration with a tenth-normal
solution of potassium hydroxide (KOH) or hydrochloric acid
(HC1), as the case may be. The addition of 7 c.c. of a 1 per cent,
solution of potassium ferricyanide to every 100 c.c. of solution is
then made, and the ferroxyl reagent, while still hot, is ready to use.
Enough of the reagent should be poured into a Petri dish to just
cover the bottom, and the dish floated in cold water until the agar
has jellied. A clean sample of iron is then placed on this bed of
jelly and covered with the hot solution. After the final addition
of agar the dish should not be moved until thoroughly cool.
While the colors sometimes show up immediately, they usually
require from twelve to twenty-four hours to attain "their most
perfect development. The mounts may be preserved for many
months by keeping the surface of the agar covered with alcohol.
1 La Nature, Nov. 13, 1909, p. 373.

THE THEORY OF CORROSION

Ferroxyl Indicator Rational Proof of Electrolytic Theory. —
These effects which are produced in the ferroxyl indicator con-
stitute a visible demonstration of electrolytic action taking
place on the surface of iron and causing rapid corrosion at the
positive nodes.

The ferroxyl test in the hands of a number of investigators
has brought out with clearness a considerable body of evidence
to show that the electrolytic theory is in accord with observed
facts when iron is undergoing corrosion. Some of the best re-
sults obtained are shown in Figs. 9, 10, 11, 12, and 13.

Fig. 9. — Showing some results obtained in the ferroxyl test. (Cushman.)

Varied Rust Formation in Different Samples. — It has been
noted by a number of investigators that different samples of iron
and steel do not rust in the same way when subjected to the
action of water and air. While some samples show localized

52 CORROSION AND PRESERVATION OF IRON AND STEEL

electrolytic action, as indicated by deep pitting, others become
covered with a more or less homogeneous coating of hydroxide,
which shows little or no tendency to localize in spots or nodes.
The question naturally arises: In what respect do these two

Fig. 10. — Showing a number of iron and steel objects immersed in the
ferroxyl jelly. (Cushman.)

methods of rust formation differ? A close inspection of Fig. 13
is suggestive if not conclusive of the answer in this respect. The
photographic reproduction exhibits an effect which is frequently
observed in the ferroxyl tests. When the colors first developed,
two dark blue nodes formed at the opposite ends of the test piece,

THE THEORY OF CORROSION

with a large pink area in the center, where for a time the metal
remained quite bright. Very shortly, however, the poles changed,
and the pink central area disappeared and gave way to a large
blue node which enveloped three-quarters of the test piece, with

Fig. 11. — Steel wire nails in ferroxyl jelly. (Cushman.)

a small opposed pinkish spot. Again and again a reversal and
change of poles took place, and at least five such changes are
clearly shown in Fig. 13. As a result of this action the metal
strip was rapidly covered over its entire surface with the same
superficial, loosely adherent coating of hydroxide, which is

54 CORROSION AND PRESERVATION OF IRON AND STEEL

obtained in many cases when certain samples of iron and steel
are allowed to rust under a layer of water. It is presumable that
as the surface of the metal is eaten into by the solution of the iron
at the positive poles, a new condition of equilibrium occurs,

Fig. 12. — Steel nails in ferroxyl reagent. (Walker.)

resulting in changes and even reversals of the positive and nega-
tive nodes. This would indicate that in the case of metals which
suffer from local action or pitting the segregation conditions are
of a different nature from those which exist in the case of metals
which rust more evenly. A rough analogy may be drawn by

THE THEORY OF CORROSION

imagining an imperfect mixture of black and white sand, the
respective grains of which may lie in streaks, spots, and layers,
or may tend to arrange themselves in some more or less uniform
relation to each other. The best demonstration that the rusting

Fig. 13. — Strip of steel in ferroxyl
reagent, showing frequent shifting of
poles. (Cushman.)

and corrosion of iron and steel in all its forms is essentially an
electrolytic phenomenon is afforded by the fact that it has not
as yet been possible to find a specimen of such purity that no
trace of positive and negative nodes will be formed in the ferroxyl
indicator.

CHAPTER IV

APPLICATION OF ELECTROLYTIC THEORY

Two Distinct Effects in Rust Formation. — We may now apply
the electrolytic theory to the actual results obtained in the ordi-
nary rusting of iron. If a section of rolled metal, such as sheet
or plate, is immersed in water, according to the electrolytic theory,
rusting must take place with the establishment of positive and
negative spots or areas. At the positive points iron will pass
into solution and be rapidly oxidized to the loose colloidal form
of ferric hydroxide which is characteristic of rust formed under
these conditions. It is a well-known fact that colloidal ferric
hydroxide will move or migrate to the negative pole if subjected
to electrolysis. 1 We may therefore consider the possibility of
two separate effects that may be produced, viz., when a positive
center is surrounded by a negative area, and vice versa. These
two conditions may be graphically represented by the two circles
A and B shown in Fig. 14.

Fig. 14. — Diagram illustrating the electrolytic action
on the surfaces of iron and steel.

Now, as rusting proceeds we should expect in the case of A
that the ferric hydroxide would be piled up in a crater formation,
while the metal is eaten out at the center. In the case of B
the effect would be reversed, and while the metal would be attacked
in the surrounding area the hydroxide would be piled up in a

1 See page 33.
56

APPLICATION OF ELECTROLYTIC THEORY

cone at the center. That this is precisely what is taking place
whenever a sheet of metal rusts under water a low-power micro-
scope very clearly shows. In Figs. 15, 16, 17, and 18 the writer
has succeeded in showing the existence of both the craters and
cones as they formed on the surface of a piece of wrought-iron
boiler plate. In Fig. 15 a typical crater surrounding the point
of pitting is shown, while in Fig. 16 an excellent example of the

Fig. 15. — Formation of crater with pitting effect in
center. (Enlarged 45 diameters.)

cone appears. Both are photomicrographs magnified about 45
diameters. The source of light was on the right in each case,
and the shadows indicate the crater and cone formation, which
is so clearly discernible under the microscope. Figs. 17 and 18
are from photographs of the rusted metal, showing the craters
and cones as they appeared with very low magnification.

Fig. 19 shows the surfaces of strips of Bessemer steel (s),
puddled wrought iron (z), and charcoal iron (c), prepared in the
following manner: The respective samples were turned off in a
lathe to a bright, smooth finish; they were then immersed under
a thin layer of the ferroxyl reagent and allowed to stand quietly

58 CORROSION AND PRESERVATION OF IRON AND STEEL

for several days. At the end of this time the surfaces were wiped
clean. The electrolytic effects, which had been active on all
three metals, are very well illustrated. The light portions show
the negative areas, where little or no rusting took place, while the
dark spots and areas show the special points of attack, with
the pitting effects. The etching is not, of course, deep in the

Fig. 16. — Formation of cone with pitting effect in sur-
rounding area. (Enlarged 45 diameters.)

case of any of the three samples, and should not be understood
as showing the relative rate of corrosion of the different types of
metal. The specimens simply serve as a demonstration that the
rusting in each case has been accompanied by electrolysis.

In Fig. 20 is shown a photograph of the actual pitting of a
boiler tube, which failed, after eighteen months' service, in a
water-tube type of marine boiler. The conclusion that pitting was
due to electrolysis seems justified by comparing this photograph
with those shown in Figs. 17 and 18.

Practical Application of Electrolytic Theory. — The evidence
advanced in the preceding pages appears to the writers to confirm
the conclusion that the whole subject of the corrosion of iron is

APPLICATION OF ELECTROLYTIC THEORY

an electro-chemical one, which can be readily explained under
the modern theory of solutions. It is an undeniable fact that
some irons and steels suffer corrosion very much more rapidly
than others, and the underlying causes for these differences con-
stitute one of the important problems of modern metallurgy.

Fig. 17. — Photographic representation of rust-spots
formed on the surface of iron.

Although the discussions brought forward up to this point
are mainly theoretical in their nature, it is quite apparent that
they also have an indirect practical bearing. Before advance
can be made in overcoming the difficulties in the way of manu-
facturing iron which shall have the maximum resistance to
corrosion, as well as the preservation of the metal under the
conditions of service, the underlying causes must be thoroughly

60 CORROSION AND PRESERVATION OF IRON AND STEEL

understood. If we accept the electro-chemical explanation of the
corrosion of iron, there can be no doubt that conditions which
inhibit electrolytic effects also inhibit corrosion, and vice versa.
The purer the iron in respect to certain other metals which differ
electro-chemically from iron, and the more carefully lack of
homogeneity and bad segregation are guarded against, the less

Fig. 18. — Photographic representation of rust-spots
formed on the surface of iron.

likely are the electrolytic effects to become serious. These points
constitute the essential problems which confront the manufac-
turer who desires to make a product which shall have a high
resistance to corrosion. The user and consumer, however, are
interested in the protection of the various types of merchantable
iron and steel which are available under market conditions at

APPLICATION OF ELECTROLYTIC THEORY 61

Fig. 19. — Strips of Bessemer steel (s), puddled wrought iron (z), and char-
coal iron (c) after immersion in ferroxyl indicator.

Fig. 20. — Actual pitting of tube in water-tube type of marine boiler.

62 CORROSION AND PRESERVATION OF IRON AND STEEL

the present time. In short, protective coatings and palliative
methods of treatment are in greater demand to-day than ever
before. From the standpoint of the electrolytic theory many
suggestions for experiment under the conditions of service present
themselves.

Walker has used the ferroxyl indicator to illustrate the appli-
cation of the electrolytic theory to practical problems. This
author states : x

"When a piece of iron is placed in ordinary water, exposed
to the air, it will dissolve or rust. If now there be placed in the
water with this piece of iron a piece of platinum, the solvent or
corroding action of the water will not be changed. The oxygen
is present in the solution as before, and the iron ions as they
separate from the metallic iron are being oxidized and precipi-
tated as rust. If now the platinum and iron be electrically con-
nected, a marked increase in the rate of the solution or corrosion
of the iron is noticed. No chemical condition has been changed;
the difference lies in the fact that there is now an electrical contact
between the iron and the platinum, and the platinum furnishes
a surface on which the hydrogen can deposit, and on which, by
virtue of its catalysing action, the hydrogen will be rapidly
oxidized by the dissolved oxygen, and thus removed from the
sphere of action. This is shown in Fig. 21. A piece of platinum
wire is seen in the center of the plate. A nail is also seen corrod-
ing at the ends and setting free hydrogen in the center. A second
nail is connected with a platinum wire. The red of the phenol-
phthalein is seen around this wire, while the nail is corroding
very much faster than the unconnected nail.

"Another example is found in the action of zinc in water
containing an electrolyte. If a strip of zinc and a piece of iron
be placed in water containing a very little salt, the iron will cor-
rode rapidly, while the zinc will be but slightly attacked. The
usual explanation for this phenomenon is that the zinc protects
itself with an adherent film of zinc oxide or hydroxide, while the
iron produces a non-adherent voluminous hydroxide, which does
not protect. If now instead of being separated the two strips
of metal be placed so as to touch each other, the iron no longer
corrodes, but the zinc very rapidly passes into solution. This
action is shown in Fig. 22. A piece of zinc is seen at the center,
1 Jour. Iron and Steel Inst., 1909, I.

APPLICATION OF ELECTROLYTIC THEORY

corroding slightly from the ends, and showing the red due to
the separating hydrogen ions at the center. Zinc ferricyanide
is white, and hence the same reagent demonstrates the separation
of zinc ions just as it does the iron ions in the plates already con-
sidered, in which the positive poles are shown in blue. The
nail which has been connected to another piece of zinc is not
corroding, but is protected by the separating hydrogen. The
zinc thus connected, however, is seen to be rapidly dissolving."

Fig. 21. — A piece of platinum wire is seen in the center of
the plate. Shows the action of the electro-negative
metal, platinum, stimulating the corrosion of iron.
(Walker.)

This point which is so beautifully illustrated here has a
most interesting bearing on the galvanizing problem. It will
at once be seen that the zinc coating which is applied to steel in
the galvanizing processes has a double function to fulfil. In
the first place it is designed to keep the surface of the steel from
coming into contact with water and the atmosphere, and secondly
it enacts the heroic role of self-sacrifice to the protection of the
iron whenever water and oxygen finally succeed in breaking

64 CORROSION AND PRESERVATION OF IRON AND STEEL

through. The practical applications of this theoretical treatment
of the subject must await detailed presentation in a succeeding
chapter.

Walker's next demonstration, as illustrated in Fig. 23, shows
the opposite tendency at work. In this case a piece of mill-scale
is shown in the center, and in addition to this the effect produced

Fig. 22, — A piece of zinc is shown in the center. Shows
the protective action of the electro-positive metal, zinc, on
the corrosion of iron. (Walker.)

by connecting a fragment of mill-scale to a piece of steel. The
mill-scale, which consists of the magnetic oxide of iron, is electro-
negative to the iron, and is therefore acting a part directly oppo-
site to that of the zinc in the preceding illustration. That is to say,
instead of retarding the corrosion of the iron at its own expense,
it is stimulating the attack made upon it by the electrolyte.

These as well as many other observations and experiments,
based upon the electrolytic theory, have a direct bearing upon
many of the practical problems met with in the manufacture
and use of iron and steel. These problems, as they affect the
resistance to corrosion, will be discussed in the following order:

APPLICATION OF ELECTROLYTIC THEORY 65

(1) Effect of purity and chemical constitution of the metal.

(2) Effect of heat treatment, annealing, tempering, condition
of surface, etc.

(3) Effect of state of purity of the rusting medium, electro-
lytes, natural water, the atmosphere.

Fig. 23. — A piece of mill scale is shown in the center. Shows
the stimulated corrosion of iron affected by contact with the
electro-negative magnetic oxide of iron. (Walker.)

Modern Steels seem more Corrodible than the Old. — A mass of
evidence appears to show that the old slowly fabricated irons of
thirty years ago are much more resistant to corrosion than modern
steels. Sang, in summing up from the opinions" of a number of
authorities, discusses this phase of the subject as follows: 1

"From a theoretical standpoint, steel, being negative to iron,
should be the least corrodible of the two. As a general thing,
results of tests between iron and steel have, in the past, resulted
in favor of the iron; in most cases, the experimenters were un-
doubtedly looking for the defeat of the new material, steel, and
their state of mind helped them to find it. There are, however,

iProc. Eng. Soc. West. Pa., XXIV, 10, p. 514.

66 CORROSION AND PRESERVATION OF IRON AND STEEL

a large and ever-increasing number of contrary observations
recorded, especially where the tests have been carried out on a
large commercial scale, and with qualities of recent manufacture.
The opinion one is led to form from a careful examination of
recorded observations is in agreement with that of Ewing Mathe-
son, 1 namely, that properly protected steel and iron rust to about
the same extent, the steel doing so more uniformly; this is, of
course, subject to the variations of structure already referred to,
and those of chemical composition, especially as regards metallic
impurities, which will be considered later.

"A most important paper was presented before the Institute
of Civil Engineers in 1881 by David Phillips, 2 'On the compara-
tive endurance of iron and mild steel when exposed to corrosive
influences 7 ; excellent tables are given, and the general conclu-
sions favor iron. A distinction must here be made between the
cast and wrought metal; cast iron will not rust as readily as
wrought iron, unless the skin is removed, in which case it will
rust faster.

"It must be borne in mind, as a limitation to all results ad-
duced, that while the initial rusting may be greater with either
material, iron or steel, the rates of progression may be different
and may bring about a complete reversal in the final result; in
material which rusted faster at first may outlive the other. This
is especially apt to be the case with forged, rolled, and drawn
metals. Future tests should, therefore, either be carried out to
destruction, as advocated by Howe, or else to the point at which
failure of the material in service would result from loss of useful
area." Again in another paragraph Sang 3 very truly says:

"Carelessness of manufacture, which tends to heterogeneous-
ness, is an invitation to corrosion, and in itself goes far to explain
why modern steel, which is tortured into shape at such a high
speed that the molecules are not permitted to readjust them-
selves, is said to be more corrodible than the metals produced a
generation ago; in those days iron and steel were produced in
small quantities, without the addition of other metals, and were
rolled slowly and allowed to cool naturally. The internal strains
due to mechanical treatment are not to be confounded with the

1 Proc. Inst. C. E., Vol. 69 (1882), p. 1.
2 Proc. Inst. C. E., Vol. 65 (1881), p. 73.
3 Proc. Eng. Soc. West. Pa., XXIV, 10, p. 511.

APPLICATION OF ELECTROLYTIC THEORY 67

unevennesses in the distribution of the impurities due to segrega-
tion in cooling; these mechanically induced strains are really
equivalent to straining the metal beyond the elastic limit, which,
as will be seen later, makes it more corrodible. Moreover, the
tonnage-craze, from which the quality of product in so many
industries is to-day suffering, is causing to be placed on the market
a great mass of material, only a small proportion of which is
properly inspected, which is not in proper condition to do its
work — rails and axles which fail in service and steel skeletons
for high buidlings which may carry in them the germs of destruc-
tion and death."

Chemical Purity an Aid in Resisting Corrosion. — While wish-
ing to avoid taking sides in the more or less heated controversy
that has gone on for many years between the advocates of differ-
ent types of metal, the authors are of the opinion that, other
things being equal, the approach to chemical purity, whether the
metal is classed as iron or steel, should lead to added resistance
to corrosion. Indeed, it is impossible to accept the electrolytic
theory without reaching this conclusion, for it follows naturally
that the presence of impurities invites segregation in the structure
of the metal with inevitable surface differences of potential and
stimulated electrolytic action. All authorities who accept the
electrolytic theory agree on this point. Walker says on this
subject:

"Since before the iron can form rust it must first pass into
solution, and in so doing cause an electric current to flow from
the iron at that point to the iron at some other point, any cir-
cumstance which will aid the flow of this current will accelerate
the solution of the iron. In other words, any differences in poten-
tial which may exist upon the surface of the iron will in itself
cause a flow of electricity which will result in a solvent action on
the iron. Such differences of potential inevitably result from a
segregation or uneven distribution of any impurities which the
iron or steel may contain. /Hence we should expect that the
speed of corrosion would increase in accordance with the per-
centage of impurities present, and it should decrease in accord-
ance with the care bestowed upon the iron or steel during its
manufacture to prevent a segregation of these impurities. The
simplest way, of course, to insure an absence of segregation is to
eliminate altogether those materials not needed in the iron. It

68 CORROSION AND PRESERVATION OF IRON AND STEEL

has been found that a steel made under such conditions that the
total impurities are reduced to not over five hundredths of 1 per
cent. (0.05 per cent.) resists corrosion to an extent equal to the
iron of our forefathers. The steel companies have been slow to
accept this general proposition, but it is gratifying to know that
material of this purity may now be obtained on the open market."

Difficulties Introduced by Improper Heat Treatment, etc. — To
sum up this phase of the subject we may say that, other things
being equal, the electrolytic theory points to the fact that the
purer and least segregated metals should be most resistant to
corrosion. At the same time it should not be forgotten that it
is always possible to leave the frying-pan for the fire, and that in
the effort to arrive at chemical purity in the processes of manu-
facture the metal may have come to harm by burning, gas occlu-
sion, or faulty heat treatment. It is also true that in the effort
to reach chemical purity the physical characteristics may be
changed, to the damage of the metal produced. Unfortunately,
analysts do not as a rule consider the question of occluded gas,
and yet this factor undoubtedly has much to do with resistance
to corrosion. Sang * has collected a mass of information from
the literature of the subject bearing on this point. This author
says, " Occluded gases must next claim attention. Graham found
that iron cooled in hydrogen absorbed 46 per cent, of its volume.
John Perry in 1872 2 detected the presence of hydrogen in steeL
Ledebur found 0.0017 per cent, of hydrogen in a soft open-hearth
steel. These observations are of interest because, hydrogen
being negative to iron, it will, as already stated, promote its solu-
tion and corrosion. The electrolytic activity of hydrogen was
pointed out by Roberts- Austen. 3

" According to Lenz, 45 per cent, of the absorbed gases in
iron may be hydrogen, the balance being carbonic dioxide, car-
bonic oxide, and nitrogen in about equal proportions. According
to F. C. G. Muller, 4 about 67.8 to 90.3 per cent, of the gas in steel
is purely hydrogen.

"It is a well-known fact that iron or steel containing occluded
hydrogen, due to pickling in acids, is hardened to a considerable

i Proc. Eng. Soc. West. Pa., XXIV, 10, p. 504.

2 Jour. Iron and Steel Inst., Yr. 1872, p. 240.

3 Fifth Report, Alloys Research Comm., Inst. Mech. Engrs., 1889.
4 Deuts. Chem. Gesell., Vol. XII (1878), p. 11.

APPLICATION OF ELECTROLYTIC THEORY 69

extent, and is readily oxidized while in that condition; thorough
washing and neutralizing of the acid will not correct the hardness
nor the readiness to oxidize. Gas occlusion by this method may,
normally, reach 12 times the volume of the iron, proving that
most of it must be alloyed or in a liquid or solid state. The
greater proportion of this absorbed gas is hydrogen, and it must,
necessarily, be as impure as that which rises in the pickling vats,
containing, therefore, hydrogen sulphide, arsenide, etc.

"On the other hand, electrolytically produced iron, which is
quite difficult to corrode, is hardened to a considerable extent
by the absorption of hydrogen during its deposition. The hard-
ness of electrolytic iron is 5.5 as against 4.5 for ordinary iron.
According to Cailletet/ electrolytic iron will hold as much as
250 times its own volume of hydrogen, and the alloy containing
0.028 per cent, (by weight) of hydrogen will scratch glass. This
absorbed hydrogen must be relatively pure, and while this may
preclude electro-chemical activity among the gases themselves,
it can hardly have much bearing on the difference of behavior
between it and pickled iron, when exposed to corroding agencies.

"The hydrogen contained in pickled iron can be almost
entirely baked out of it at a low temperature ; not so with the
hydrogen absorbed electrolytically. This tends to show that in
the pickled iron the gas is not so permanently or stably combined
— if combined at all — as in electrolytic iron. Furthermore,
the great volume of the gas taken in by the electrolytic iron
shows that a very large percentage must exist in solution as an
alloy with the iron. The coexistence of three states of matter
has been supported by Graham, Wiedermann, and Spring. While
there may be just as much free dissociated hydrogen contained
in the pores of both classes of iron, and the tendency to rust from
that cause may be the same, yet the larger amount of hydrogen-
iron alloy in the electrolytic iron may resist corrosion much better
than iron alone. The quality of resistivity to corrosion is inti-
mately connected with the rise in electrical conductivity which
is brought about by the chemical union of hydrogen with metals.
Hot iron when quenched in water absorbs hydrogen, and Richards
and Behr 2 have found that the electrode potential was raised by
0.15 volt, the nature of the gas being apparently the same as that

1 C. R., Vol. LXXX (1875), p. 319.

2 T. W. Richards & G. E. Behr Jr., Zeits. Phys. Chem., March 5, 1907.

70 CORROSION AND PRESERVATION OF IRON AND STEEL

which is absorbed in the presence of nascent hydrogen, and there-
fore by electrolysis. The hydrogen taken up by finely powdered
iron reduced at a low temperature was not found to affect the
e.m.f . ; we may infer that the physical conditions attending
the production of this iron were insufficiently powerful to cause the
alloying on which the change of e.m.f. seems to depend. Dr.
Steinmetz finds that electrolytic iron has a very high hysteresis
loss, but attributes it to occluded nitrogen. 1

"From an examination of all these facts, it would appear
that the increase of potential due to the alloyed hydrogen in
electrolytic iron overcomes the effect, as an electro-negative
catalyzer or otherwise, of hydrogen in a free ionic state only. In
all classes of iron the hydrogen exists in both conditions, free
and combined, just as carbon does in pig iron, but the proportion
of hydrogen-iron alloy in electrolytic iron is very much greater
than in the other metals. Hydrogen, like carbon, when present
in a free state will by contact action promote corrosion; like car-
bon also, when chemically combined with the iron it will resist
corrosion, but if the alloy is unevenly distributed the pure iron
in contact with the alloy will be attacked.

"According to Roberts-Austen, silicon, manganese, and alu-
minium prevent the escape of hydrogen from iron; Ledebur
claims, 2 however, that brittleness after pickling, due to hydrogen, is
greater if the combined carbon is high, while silicon has the reverse
effect; he is in accord, therefore, with Troost and Hautefeuille, 3 who
claim that silicon diminishes absorption. These seemingly oppo-
site statements may be reconciled by assuming that while silicon
may reduce the absorption of hydrogen, it will also retard its
subsequent removal, just as non-conductors which absorb heat
with great difficulty will, on that very account, retain it the easier.
Manganese is said 4 to greatly increase the absorption of the gas
while diminishing that of carbonic oxide, which is, in any case,
very slight. Manganiferous pig iron retains more gas than does
ordinary pig.

1 For u, discussion of the important influence of nitrogen on the physical
characteristics of steel, see Giesen, The Special Steels 'in Theory and Practice,
Carnegie Scholarship Memoirs, Iron and Steel Institute, I (1909). See also
Stromeyer, Jour. Iron and Steel Inst., I (1909).

2 Mitt. Kon. Tech. Versuchsanstalten, Ber. Yr. 1890. Suppl., I, 1907.

3 An. Ch. & Ph., 5e S., Vol. VII, p. 1155.

4 An. Ch. & Ph., 5e S., Vol. VIII, p. 1155.

APPLICATION OF ELECTROLYTIC THEORY 71

"Pressure applied during the solidification of metals — as,
for instance, in the Whitworth process — prevents the escape of
the gases. They can be driven out by heating, preferably in
vacuo, or locally by machining or drilling; the combination is,
therefore, not a very close one. To drive the gases out of pig
iron, a temperature of 800° C. is sufficient. Malleable iron con-
tains more carbonic oxide than hydrogen, and it is retained with
greater energy. Steel is said to absorb somewhat less than cast
iron, and wrought iron less than cast iron; these differences are,
in great measure, no doubt, functions of the porosity.

" Occluded gases, and especially hydrogen, must not be lost
sight of when dealing with the problem of corrosion. Hydrogen
is the lightest, and, therefore, kinetically the most active of ele-
ments; it is, in a way, a sort of universal catalyzing 'daemon/
an extravagant statement to the ear, perhaps, but with some
merit of suggestiveness; all chemical reactions take place in the
presence of hydrogen, and it is the only element of which this
is true. Hydrogen, which seems to form the main ejection from
the sun, and may be regarded as closest to the primordial element
from which, according to recent well-grounded theories, all other
elements may proceed, is unique in many of its properties; it
seems to stand apart from the other elements in many ways.
These differences are, in many cases, attributable to the great
activity of its molecules in proportion to their mass, hence, for
instance, the distinct character of its curve representing the
value of pv under different pressures.

"The diffusion through a finely porous material, which gives
rise to dissociation, is similar to, if not identical with, osmosis;
in osmosis the porous membrane causes dissociation resulting in
chemical effects which are the basis of important reactions, and,
among others, of organic growth and life.

"Hydrogen will pass through platinum and red-hot iron.
(Ste. Claire-Deville) and its ready dissociation, which was demon-
strated in Winklemann's important study of its diffusion through
palladium/ suggests a belief in its breakdown, under conditions
of common occurrence, into free and active atoms, ready to take
the first opportunity offered of entering into a combination. The
condition of most common occurrence is, as we have seen, the
contact of dissimilar substances. The occlusion of free hydrogen
1 An. Phys. Chem. Wied., Vol. 6, p. 104, and Vol. 8, p. 388.

72 CORROSION AND PRESERVATION OF IRON AND STEEL

in coal-dust, wheat-dust, zinc-dust, and other dusts, will go far
to account for their detonation by spontaneous oxidation. These
dusts act in the same way as does spongy platinum on certain
gases which it ignites by simple contact. A porous material like
iron or steel should have a similar effect, but its action would be
slow and progressive instead of sudden; instead of spontaneous
oxidation we get slow oxidation, rusting.

"We thus have additional reason to Relieve that free disso-
ciated hydrogen ions, generated by the electro-chemical action on
moisture of iron in contact with its impurities or other substances
exterior to itself, induces by catalytic excitation, or an electrical
effect of its contact with the iron, the solution of that iron as
free ferrous ions which unite with free oxygen to form rust."

It will be seen from the above citations that the occurrence
of gases in steel is a matter of great importance which has not
received the attention from manufacturers that it deserves.
That the resistance to corrosion is influenced by occluded gases
is highly probable. The authors are glad to take this occasion
to recommend to manufacturers who are aiming at the produc-
tion of pure irons, that this subject be given more careful study
in the future. Baker 1 has recently published a valuable paper
on gases occluded in steel, with an account of methods for esti-
mating them.

The Carbon Constituents of Steel. — It is not the authors' inten-
tion to include in this work instruction in the metallography of
steel. For full information on this subject the special works
should be consulted. In view, however, of the possible influence
of the carbon compounds in iron on the electrolytic effects which
are observed, a brief reference to the general substance is here
included. The microscopic examination of almost any steel when
properly mounted and prepared will show that the structure is
not by any means homogeneous, but that it contains recogniza-
ble constituents. Specimens of pure iron, when properly polished
and etched, exhibit a peculiar coarse-grained structure under the
microscope to which the name ferrite has been given. As the
percentage of carbon rises in steel various peculiar and character-
istic structures make their appearance, corresponding to certain
definite compounds, solid solutions, and eutectics. It does not
follow that all these structures are different substances, as they

1 Iron and Steel Inst., Carnegie Scholarship Memoirs, Vol. I, 1909.

APPLICATION OF ELECTROLYTIC THEORY 73

may be allotropic modifications or phases of one and the same
substance. Carbon in steel does not exist in the free state, as it
does in cast iron, but it combines with iron forming the carbide of
iron, Fe 3 C. This carbide appears as dark areas of peculiar and
characteristic appearance to which the name cementite has been
given, because it occurs abundantly in steel, treated by the cemen-
tation process. When examined under a high magnification car-
bon steels show patches that exhibit a pearly luster, owing to a
finely lamellar structure. This constituent has been called pearl-
ite, and is now known to consist of a banded structure of cement-
ite and ferrite, in definite proportions, not being a compound, but
simply an intimate mixture. There are a number of other con-
stituents which correspond to different quantities and conditions
of the combined carbon in hardened steel. To these such names
as austenite, martensite, troostite, osmondite, and sorbite have
been given in honor of eminent metallographists.

Austenite is produced by quenching high carbon steel in ice
water from a temperature above 1050° C. It is not often met
with, and is probably of little importance in any discussion on
corrosive tendency.

Martensite is defined as the constituent that confers hardness
on steel. It is characterized by a finely interlaced structure,
and is found only in hardened steels.

In regard to troostite and sorbite little is known; they make
their appearance when steel is subjected to definite quenching
and tempering processes. Osmondite is a name which has re-
cently been proposed to define a characteristic transition point
that occurs in the tempering of quenched steels. 1

According to Campbell, 2 "Pearlite is an 'eutectic alloy,' a
term which may possibly not be familiar to all readers. An
eutectic alloy is formed by the simultaneous crystallization of
different metals in a liquid mixture, as for example, a mixture of
copper and silver. These metals form an alloy in the propor-
tions of 72 per cent, silver and 28 per cent, copper at a tempera-
ture of 770° C. (1418° P.), and if a melted mixture of these two
metals contain any different proportion than this, and if it be
allowed to cool, the element in excess of this proportion crystal-
lizes out, the crystals remaining uniformly distributed through-

1 Heyn and Bauer, Jour. Iron and Steel Inst., 1, 112 (1909).

2 Manufacture and Properties of Iron and Steel, Campbell, p. 298.

74 CORROSION AND PRESERVATION OF IRON AND STEEL

out the molten mass. When the critical point of 770° C. is
reached, the alloy of 72 silver and 28 copper becomes solid, and
entrains the innumerable crystals of the excess element which
have separated from the mother liquid. A little consideration
will show that under the microscope the element solidifying first
and the eutectic alloy will occupy areas exactly proportional to
the original constitution.

"In steel at high temperatures the same conditions exist as
in the mass of silver and copper just described, save that the
elements are in what is called 'solid solution/ martensite at the
lowest critical point going through a transition into ferrite and
cementite. The element in excess separates by itself, and when
the proper relation has been established the ferrite and cementite
crystallize together in most intimate mixture to form pearlite.
As stated previously, the excess of cementite or ferrite begins to
form by itself at the upper critical point, a small amount being
found in steel quenched just below this, and at the second point
this amount is increased, but this excess is always small except
in the case of low carbon steel.

"The foregoing argument may be summarized as stated by
Sauveur :

"(1) All unhardened steels are composed of pearlite alone,
or of pearlite associated with ferrite or cementite.

"(2) Without taking into consideration austenite and troos-
tite, hardened steel is composed of martensite alone, or of mar-
tensite associated with ferrite or cementite.

" (3) Ferrite and cementite cannot exist together in the same
piece of steel.

" (4) The presence of the lamellar variety of pearlite is almost
certain proof that the steel has been annealed.

"Following the proposition that ferrite is iron free from carbon,
and that cementite is a compound represented by the formula, Fe 3 C,
it is evident that in very low steels, say ranging from .02 to .10 car-
bon, the structure will be almost entirely ferrite, and that in steel
of 2.00 per cent, carbon there will be an excess of cementite. There
will therefore be one point of carbon content at which the compo-
nent ferrite and cementite will both be satisfied, which is to say
that the original proportion will be that of the eutectic alloy. This
occurs in a pure steel containing about .80 per cent, of carbon, the
micro-structure of this grade showing no ferrite or cementite."

APPLICATION OF ELECTROLYTIC THEORY 75

Influence of Iron-Carbon Constituents on Corrosion. — The influ-
ence of the quantity and condition of the contained carbon of a
steel on the tendency to corrosion opens up a most interesting
field of inquiry, in which unfortunately the information is meager,
and to a considerable degree contradictory. A vast amount of
theorizing has been done on this point alone, and the literature
of the subject abounds with statements in regard to the influence
of carbon which are certainly based on insufficient evidence.
Howe believes in a mechanical protection afforded by carbon
as rusting proceeds. This authority has stated: 1

"As steel is gradually corroded away, more and more of its
surface should come to be composed of cementite, and this fact
should tend to retard the corrosion of steel, because cementite
should protect the underlying free iron or ferrite."

And again: "The cementite is in such extremely minute
microscopic plates that the eating away of a very small quantity
of the iron from above them ought to bring very nearly the full
proportion of this cementite to the surface."

Sang claims that, 2 "Carbon, insomuch as it will allow harden-
ing, will act as a protection, provided it is combined with the
iron and uniformly distributed; high-carbon steel is less corrodible
than mild steel or iron."

There can be little doubt that maximum density in steel
will contribute toward maximum resistance to corrosion provided
segregation has not taken place, but in the authors' opinion it has
not yet been proved that the majority of high-carbon steels in
general use are less corrodible than mild steel or iron. This sub-
ject can, however, best be discussed under another heading, on
the effect of heat treatment and tempering.

It is quite certain that more information is needed on the
influence of carbon and the iron-carbon constituents on the cor-
rosion problem. Munroe 3 observed a number of years ago that
in a case of corrosion of a cold-chisel with a soft shank and hard-
ened cutting edge, the soft part was eaten away while the hard-
ened edge had been protected. This observation has been many
times confirmed until we may feel sure that in some cases soft
iron (ferrite) is electro-positive to hardened steel (cementite) so

1 Trans. Am. Soc. Testing Materials, 1906.

2 Proc. Eng. Soc. West. Pa., XXIV, 10, p. 528.

3 Jour. Franklin Inst. 1883, 302.

76 CORROSION AND PRESERVATION OF IRON AND STEEL

that in contact in a corroding medium the latter will be protected
at the expense of the former type. Sauveur states, as quoted
above, that ferrite and cementite cannot exist together in the
same piece of steel. This would seem to indicate that differences
of potential in the same piece of steel would not be occasioned
from this cause. Sauveur, of course, does not refer to the inti-
mate mixture of ferrite and cementite known as pearlite, but in
this case the structure is too definite and intimately mixed to
lead to local electrolytic couples.

Possible Effect of Assembling Metals of Varying Carbon Types.
— The association of steel and iron of different carbon content
in structural work, is a question of the highest importance which
has not received as much attention from engineers as it deserves.
In this connection we may recur to the case already cited of the
Diamond Shoal Light-vessel, in which in eleven years 8400 four-
inch iron bolts were destroyed and had to be renewed. 1 In this
and similar cases the more electro-positive metal is rapidly
destroyed owing to the stimulated electrolysis, which is induced
by the coupling of different types of metal.

Sang treats this subject as follows: 2 "Some tests were made
in 1882 by J. Farquharson 3 on six plates of iron and six of steel;
these were immersed for six months in Portsmouth Harbor, six
of each separately, the other six as connected couples; in this
way the comparative corrosion of the iron and steel was obtained,
and also the increase of corrosion due to galvanic action between
steel and iron. The following table gives the losses observed
in ounces and grains:

(a) Steel) in contact (0-427

Iron [ (7-417

^ S teel l separate (3-340

Iron ) l (3-327

(<) Steel) incontact (0-297

Iron f (7-770

(d) EE?}-*"* [tS

w Ep™ {JS5

1 See illustration (Fig. 4) and description p. 11.
a Proc. Eng. Soc. West. Pa., XXIV, 10, p. 520.
3 Trans. Inst. Nav. Arch., vol. 3 (1882), p. 143.

APPLICATION OF ELECTROLYTIC THEORY 77

(/) ?n ■*— it

£-157
-570

"These results, which were confirmed by Mr. W. Denny,
from his experience in the case of the steamship Ravenna, are
interesting to analyze. They show that in two cases only did
the steel corrode to a greater extent than the iron, but the differ-
ence is so slight that for all practical purposes it can be said that
the steel and iron of the experiments (ship-plates) were equally
affected. They also confirm the theory that the combination
of steel and iron, which is quite frequent in practice, is detri-
mental to the iron, but protects the steel which is the negative
partner. They also throw light on previous observations, and
lead to the conclusion that good homogeneous iron and steel
are about equally corrodible."

Mr. J. P. Snow, Chief Engineer of the Boston and Maine Rail-
road, has called attention to a very significant case of corrosion in
connection with the destruction of some railroad signal bridges
erected in 1894, and removed and scrapped in 1902. These struc-
tures were built at the time that steel was fast displacing puddled
iron as bridge material. The result was that the bridges were
built from stock material which was partly steel and partly
wrought iron. The particular point of interest in this case lies
in the fact that while some of the members of the bridge struc-
tures rusted to the point of destruction in eight years, others
were in practically as good condition as on the day they were
erected. This is clearly shown in the illustration, page 8, which
is from a photograph of these bridge members. The specimen
shown in the middle has suffered only very slight superficial
rusting, while those shown on either side have gone to the point
of destruction. Tests carried on by Snow and examinations made
by one of us appeared to indicate that the badly rusted parts
were steel and the unrusted portion wrought iron. On first
thought this observation would appear to indicate that wrought
iron was far superior to steel as material for such structures. In
the light of the electrolytic theory and of the other evidence that
has been given, it appears highly probable that the steel in this
case was electro-positive to the iron, which resulted in the pro-
tection of one metal at the expense of the other. 1 From this

1 It should be noted that the equilibrium which decides which metal will
be positive to the other is a delicate one and is determined by a number of
factors, which may or may not be known.

78 CORROSION AND PRESERVATION OF IRON AND STEEL

point of view the steel is no more to be condemned for having
failed than the iron for having been the destructive agent. With
the results of recent investigations to guide them, it is not prob-
able that engineers will permit the assembling of different types
of metal in one and the same structure, without first considering
the probable effect on its life.

Effect on Corrosion of Other Constituents in Steel. — We have
now to consider the influence on corrosion of the other usual
impurities found in steel, such as manganese, phosphorus, sul-
phur, and silicon. From the standpoint of the electrolytic theory
there are several reasons why the presence of manganese in steel
should invite corrosion. Manganese decreases the electrical con-
ductivity of iron, and as the percentage of manganese, starting
from zero, rises, the electrical resistance increases up to a certain
specific maximum. It will be seen that if the presence of man-
ganese in iron raises the electrical resistance, any variation in
the distribution of the manganese means that there will not be
a constant electrical conductivity throughout its mass, or on
any given surface. There is abundant evidence to show that
manganese associates itself to a considerable extent with sulphur
when both these impurities are present in steel. 1 That man-
ganese sulphide shows a difference of electrical potential against
iron is also well known. These theoretical reasons which indicate
that manganese should stimulate the corrosion of steel, provided
it is not perfectly homogeneously distributed in the iron, appear
to be justified by a number of reliable observers among whom
may be mentioned Dudley, 2 Drown, 3 Abel/ and Reynolds. 5
Huntley 6 deals with the corrosion of boiler steel, and gives
details of a case of pitting in a boiler. Each pit was found to
be the center of a blister, and the blister contained a slightly
acid solution of ferrous sulphate, while the boiler water was
alkaline with caustic soda. The solid matter in the envelope of
the blister, which consisted of a mixture of iron oxides, acted as
a semi-permeable membrane, keeping apart the ferrous sulphate
within and the caustic soda outside the blister. The theory
advanced is that the particles of manganese sulphide, segregated

1 Fay and Howard, Trans. Am. Soc. Testing Mat., 1908, 8, 74.

2 Trans. Am. Inst. Mining Eng'rs, 1905. Disc. Roe's paper.

3 Ibid. * Proc. Inst. C. E., 1881.

b Ibid. 6 Jour. Soc. Chem. Ind., 28, pp. 339-340.

APPLICATION OF ELECTROLYTIC THEORY 79

in the steel, were oxidized by the oxygen dissolved in the boiler
water to sulphuric acid and an oxide of manganese, the acid then
acting locally on the surface of the boiler plate in the vicinity
of the particles of manganese sulphide. The pitting of the boiler
plate was prevented by adding sodium arsenite to the boiler
water, the reagent taking up the dissolved oxygen.

It should be noted that the above discussion refers to man-
ganese when it occurs as an impurity in steel in amounts not
exceeding about 1 per cent. Very high manganese steels are
said to be unusually resistant to corrosion. It must be remem-
bered, however, that in the latter case we are not dealing with
steel at all, but with one or more special alloys in which stable
eutectics are probably formed. These special manganese alloys
would then fall in line with many others which are very well
known to be highly resistant to corrosion. Among these resistant
alloy steels that have been noted we may mention nickel, 1
chromium, vanadium, tungsten, and silicon. More systematic
investigation of this subject is needed.

Phosphorus has been claimed to render iron more resist-
ant to corrosion, but actual information in the literature of
the subject is too meager to found an opinion on. Sang 2 says:
" Phosphorus and silicon both appear to retard corrosion, and
this effect may, as in the case of carbon, have some connec-
tion with their hardening qualities, or cold-shortening power.
If, however, they are present in patches, like the oft-occurring
phosphide eutectics, the softer parts, through contact action with
the parts rich in phosphorus and silicon, will be destroyed all
the more rapidly. Some authors have claimed that these two
elements increase corrosion, but there is no evidence to support
the contention apart from the case of uneven distribution which
will make any of the impurities rust promoters to a greater or
lesser extent. The fact that common iron does not rust as rap-
idly as the better grades has been attributed by some to the greater
percentage of phosphorus in the former."

Sulphur is usually present in steels in very small quantities,
and probably affects the rate of corrosion mainly by increasing
segregation, as in the case of manganese sulphide already cited.

Silicon in small quantities like manganese may increase

1 Unger, Proc. Eng. Soc. of West. Pa., 24, 10, 549.

2 Ibid., page 259.

80 CORROSION AND PRESERVATION OF IRON AND STEEL

the tendency to corrosion. In larger quantities from 10 to
20 per cent., the solubility in acids as well as the tendency to
rust is checked. Jouve * has called attention to the fact that
iron containing 20 per cent, of silicon is not attacked by acids,
and therefore such material theoretically should not be subject
to rust. One of the authors has confirmed this deduction by
experiment, and found that iron containing more than 10 per
cent, of silicon is almost incorrodible. Unfortunately, such a
metal is not easily workable and has peculiar properties. Since
silicon is much like carbon, chemically speaking, it would seem
as if it might be worked into the surface of steel by modifications
of some of the processes used for case-hardening with carbon.

Various Factors ivhich Modify Effect of Impurities. — It will
be readily understood that even if there were much more data
available than we have it would be quite impossible to accurately
state the influence that any given impurity will have upon the
corrosiveness of iron. Not only will much depend upon the
quantity, distribution, and condition of the impurity, but the
balanced effect produced by other impurities is an ever present
factor in a heterogeneous equilibrium. The discussion and
citations given in preceding paragraphs have been included only
inasmuch as they have a bearing on the electrolytic theory of cor-
rosion. But even as the interior structure of a metal influences its
tendency to corrode, its use in connection with other types of
metal in construction work becomes a matter of even greater
importance. This has been pointed out in regard to iron of vary-
ing carbon contents.

Problem of Combining Structural Metals of Different Chemical
Constitution. — If two metallic elements are joined in a structure
the more electro-positive one will be corroded rapidly, while the
more electro-negative will be protected. It therefore follows
that even if a complete knowledge of the chemical constitution
and interior structure of a metal was in our hands, unexpected
results might still be met with under service conditions. A
case in point is shown by the work of Preuss. 2

In order to study the influence of electrolysis, a number of
nickel steel rivets and iron rivets were driven through several
plates of mild steel, and immersed for two months in a brine

1 Engineer, 190S, 106, 397.

2 Iron and Steel Inst., Carnegie Scholarship Memoirs, I, 1909, p. 81.

APPLICATION OF ELECTROLYTIC THEORY 81

corresponding to sea-water. The loss in weight sustained by the
nickel steel rivets, as compared with the loss sustained by the
rivets of iron, was in the proportion of 11 to 6. As the nickel
steel was essentially more uncorrodible than the iron, we see in
this a reversal of resistibility which only a complete understanding
of the principles of electrolysis can explain. Sang cites the work
of Mallet which is interesting in this connection. 1

"The effects of electrolytic action are clearly demonstrated
by the results secured by Mallet in a series of experiments which
he undertook in order to ascertain the 'amount of corrosion in
equal times in clear sea-water of a unit surface of wrought-iron
plate, exposed in electro-chemical contact with an equal surface
of the following metals electro-negative to it, as compared with
the corrosion of the same surface of the same iron exposed alone
for the same length of time; 7

Relative Corrosion

Iron plate alone 8.63 per cent.

In contact with: Brass (Cu 2 + Zn) . . . 29.64

Copper 42.79

Lead 47.90

Gun-metal (Bronze). 56.39

Tin 74.71

"In connection with the above table, the valuable fact is
mentioned that the brass alloys of composition, Cu 8 + Zn 17 to
Cu 8 + Zn 18 , are without galvanic action on iron in sea-water.
This explains the incorrodibility of the alloy of iron, copper,
zinc (and sometimes tin), which is known as Delta metal, and
which, tested in conjunction with wrought iron and steel, showed
remarkable resistance under test, as follows:

Wrought Iron Steel Delta Metal

Loss 45.9 45.45 1.2 per cent.

"The first copper-zinc alloy for the special purpose of resist-
ing the action of sea-water was patented in 1832 by Muntz.
Muntz metal is used for bolts, valves, etc., and for sheathing
ships; its composition is 2 parts zinc to 3 parts copper. Tobin
bronze is similar to Delta metal, but contains tin and lead."

Walker has used the ferroxyl test to give some very pretty
illustrations of the protection of the electro-negative element

1 Proc. Eng. Soc. West. Pa., XXIV, 10, p. 521.

82 CORROSION AND PRESERVATION OF IRON AND STEEL

in an electrolytic couple. This is shown in Fig 21, where iron
is suffering from stimulated corrosion owing to its being in
connection with the more electro-negative metal platinum, while
in Fig. 22, the reverse action is taking place, the iron being
protected by contact with the more electro-positive metal zinc.
In Fig. 23 the accelerating effect of the electro-negative mill-
scale or magnetic oxide is shown.

Conclusions Based on Measurements of Potential Differ-
ences. — In view of the fact that differences of potential are
known to exist not only between metals of different types, but
also on the surface of more or less homogeneous pieces of
iron, it is natural that a number of experimenters should have
attempted to measure them. Unfortunately, these attempts
have led to results of little or no value. As a matter of fact the
equilibrium is so delicate that the slightest disturbance has an
influence on the electric potential, which is always subject to
change owing to more or less obscure causes over which little or
no control can be exercised. 1 Some experimenters have made
elaborate series of potential measurements which have led them
to conclusions, without apparently understanding that the meas-
urements were in reality nothing but haphazard results, which
should not be interpreted as leading to definite conclusions. In
many cases such results have been largely effected by the poten-
tials of the contact points between the metals under examination
and the take-off wires of the measuring instruments. Walker's 2
ingenious method of adapting the potentiometer to the measure-
ments of potential differences on the surface of iron was published
some years ago, but this authority has since concluded that even
these carefully made measurements cannot be used either to
support or refute the electrolytic theory of corrosion. In the
hands of nearly all investigators the efforts to establish definite
polarity between samples of steel of different chemical analysis
have led to confusing and unintelligible results. In one investi-
gation which came under the notice of one of the authors, the
effort was being made to measure the difference of potential

1 Compare: Punga, Mitteilungen aus dem Material Prufiingsamt zu Gros-
Lichterfelde West, 1908; See also Prenss, Iron and Steel Inst., Carnegie
Memoirs 1909, I, 82.

2 Walker, Cedarholm and Bent. Jour. Am. Chem. Soc, XXIX, 9, 1263
(1907).

APPLICATION OF ELECTROLYTIC THEORY 83

between two steel wires. One wire contained .50 carbon and .81
manganese, the other .50 carbon and .40 manganese. These
were cleaned and polished and after being connected with an
astatic galvanometer were plunged part way into a very dilute
acid electrolyte. The galvanometer needle deflected 15° to the
right, indicating that the low manganese was electro-positive to
the high manganese wire. In another test of the same samples,
made after the wires had slightly corroded, the needle deflected
60° to the left, indicating that now the low manganese was strongly
electro-negative to the higher manganese wire. In experiments
made by one of the authors in which steel electrodes have been
connected with a battery current of very low e.m.f. and amper-
age and plunged into ferroxyl indicator, blue spots have been
seen to form underneath the pink envelope on the negative
electrode. Such an appearance serves as an ocular demonstra-
tion of the delicacy of the equilibrium since it is thus shown that
a negative member in an electrolytic couple can develop a super-
imposed polarity of its own. The same effect can be seen by
close inspection of Fig. 11, page 53, although in this case the
pink zones are not being controlled or demarked by an external
current. In calling attention to these anomalous effects, the
authors merely wish to suggest that the results obtained from
delicate measurements of polarity should be used with caution
in drawing conclusions.

Effect of Heat Treatment, Tempering, Stresses, Strains, etc. —
The effect of heat treatment probably has as much to do with
the resistance tendency to corrosion as any other factor to the
problem. Under the special caption of heat treatment we may
include all effects produced by rolling, whether hot or cold, as-
well as surface stresses and strains. This subject has been
-investigated by a number of authorities. Andrews 1 has published
a paper on tests of steel under tensile, torsional, and flexional
strains, and found that the unstrained parts were electro-positive
to strained parts. Hambuechen, 2 following more or less closely
the method of Andrews, arrived at a diametrically opposite result,
and concluded that the strained metal was electro-positive, and
hence would corrode more rapidly. Both of these investigations
were unsatisfactory, and have been criticized by Walker and

iProc. Inst. C. E. (1894), 355.

2 Bull. Univ. Wisconsin, Engineering Series 8 (1900).

84 CORROSION AND PRESERVATION OF IRON AND STEEL

Dill. 1 Richards and Behr 2 included some interesting experiments
upon the effect of strain upon the potential of iron. No definite
conclusion was reached, but it was clearly indicated that if poten-
tial differences exist, the equilibrium that governs them is an
extremely delicate one. Walker and Dill 3 were led to believe that
although the potential differences were small, the strained metal
had a slightly less tendency to corrode than the same metal
unstrained. Burgess 4 has concluded that test pieces that have
been deformed by straining beyond the elastic limit show a meas-
urable difference of potential. That the potential equilibrium is a
delicate one is again indicated by the fact that Burgess's strained
parts corroded faster than the unstrained. This observation is
confirmed by a great number of ferroxyl tests in which the
deformed or strained portion of a specimen is usually electro-posi-
tive and shows up in blue. It is for this reason that the heads
and points of wire nails are usually electro-positive to the middle
portion in ferroxyl tests. This is not invariably the case, how-
ever, as may be seen in Fig. 10. The exception to the general
rule only serves to show that strains set up in the cold-drawing
of the wire from which the nails are fabricated has reversed
the polarity.

It appears to the authors that these apparently contradictory
results arrived at by different experimenters may be best explained
by the extreme delicacy of the equilibrium which governs the
polarity. That iron has a definite solution pressure has already
been shown. If for any reason, however obscure, the solution
tension is to any extent greater at one point on a given surface
than at another, that point will be to some extent electro-positive
to the other. Whether these points retain the polarity or whether
' it is actually reversed depends upon the nature of the given case,
in which determinative causes may be at work which it is impos-
sible to predicate.

Sang 5 makes the following observations in reference to this
subject: "Some years ago Witkowski found that in a strained

1 Proc. Am. Soc. Testing Materials, VII, 230.

2 Publications Carnegie Institution, Washington .(1906).

3 hoc. cit., p. 237.

4 Trans. Am. Elec. Chem. Soc, 1908.-

6 Proc. Eng. Soc. West. Pa., XXIV, 10, p. 513.
Trans. Roy. Soc. Edin., vol. XXX (1881), p. 413.

APPLICATION OF ELECTROLYTIC THEORY 85

metal there is an increase of electrical resistance in the direction
of the strain. All these observations go to prove the claim that
mechanical treatment, by setting up uneven strains in different
parts of finished pieces, will create variations of potential which
will promote rusting. Whatever the composition of the different
inner parts of the metal may be, and apart from any action
which may be due to difference of composition, if there is a differ-
ence of molecular aggregation, it will promote the rusting of one
or other of those parts. Action, power, everything knowable,
depends on difference of potential, and any chemical or physical
difference between two portions of matter in contact must give
rise to a difference of potential and a flow of electricity.

"If straining a metal below its elastic limit by exteriorly
applied mechanical means will make it electro-negative to the
same metal unstrained, the strains set up by chilling or hardening
should have a like effect; the metal should resist corrosion to a
greater extent and promote the corrosion of more positive metals
in contact with it. This is found to be the case. Eighty years
ago Daniel observed that a certain steel was dissolved by hydro-
chloric acid five times as rapidly when unhardened as when hard-
ened; this is an indication of what we may expect with the agents
of corrosion. What becomes of the energy of a coiled watch
spring when it is dissolved in acid? This is supposed to be one of
the many unsolved mysteries of science. The energy of the coiled
watch spring is indicated by a slight shift of its potential towards
the negative end of the electro-chemical scale, resulting in an
increase of e.m.f . ; when the spring is put in acid, the energy is
expended in retarding the action of the acid, and is equivalent to
a drop of temperature which would restrain chemical action. The
energy of the spring, as increased e.m.f., counteracts the energy
of the acid, is expended and disappears as work of a negative
character."

Relative Stability and Solubility of Strained and Annealed
Metals. — A most interesting paper by Heyn and Bauer 1 on the
influence of heat treatment on the solubility of steel in sulphuric
acid throws much light on the effect of such treatment on the
resistance character and stability of steels. This paper should
be carefully studied by all who are interested in the manufacture
of rust-resistant steel, for it very clearly shows the relative sta-
1 Jour. Iron and Steel Inst., 1, May, 1909.

86 CORROSION AND PRESERVATION OF IRON AND STEEL

bility of strained and annealed metals. In view of the importance
of this work and its direct bearing on the subject under discussion,
the following citations from these authors are worthy of record.

"In the experiments a tool steel with the following com-
position was first used:

Per cent.

Carbon 0.95

Silicon 0.35

Manganese 0.17

Phosphorus .- 0.012

Sulphur - 0.024

It was in the form of a square forged bar of a cross-section
25 X 25 millimeters. From this bar sections of a thickness of 6
millimeters were cut and subjected to the following treatment:

(a) Heated to 900° C, and quenched in water at 14° to 18° C, untreated

(b)

(c)

(d)

(e)

(/)

(<?)

(h) In the original forged state.

(i) Heated to 900° C. and allowed to cool slowly.

reheated at 100° C.

at 200° C.
at 275° to 305° C.
at 405° to 415° C.

at 500° C.
at 600° to 640° C.

"The high quenching temperature of 900° C. was chosen for
the reason that with a small test-piece pure martensite is obtained
by quenching in cold water and observing the necessary precau-
tions. At lower temperatures those constituents indicating the
transition to troostite occur in addition to martensite, and mark
the observation. After the treatment as described above, the
sections of an area of 25 X 25 millimeters were polished and
immersed with the polished surface uppermost in 1 per cent, sul-
phuric acid. The weight before immersion, and the loss in weight
at the end of twenty-four, forty-eight, and seventy-two hours
respectively, were determined. In Fig. 24 the tempering tem-
peratures are taken as abscissae and the respective losses in weight
as ordinates. 1 The letters attached to the ordinates refer to
the kind of treatment given in the foregoing table. For the sake

1 For more detailed information see E. Heyn and O. Bauer, "On the
Structure of Quenched and Tempered Tool-steel," etc. (Communications
of Royal Inst, for Testing Materials at Gross-Lichterfelde, 1906, p. 29.)

APPLICATION OF ELECTROLYTIC THEORY

of clearness the separate values are omitted, and only the average
values of several experiments are given.

"The shape of the curve in Fig. 24 reveals some startling
features. According to the opinions hitherto prevailing, and still
held by some, the transition of the martensite of hardened steel
into the pearlite of annealed steel is continuous throughout the
intermediate states of tempering. This would lead one to sup-

7
6

Troo

itites

C
O

E
O

Sorbi

:es

Pearlite

** S
OC s

C3
LU

^ 3

i/i
i/>
O

_i
2

■

/ i

1 #

\ J

- i0 ^U

s *»^^

jr \jr

^■^^

200

300

400 500 600

TEMPERING HEATS

700

Fig. 24. — Tool steel in various states of heat-treatment. Solubility in
one per cent, sulphuric acid. (Heyn and Bauer.)

pose that the curve of solubility in dilute sulphuric acid, instead
of showing any sharply denned peaks, would form a line indi-
cating the gradual transition from the solubility of martensite
to that of pearlite. The curve, however, runs up to a sharply
defined maximum at a temperature of 400° C. The simplest
and most obvious explanation of this phenomenon is as follows:
The transition of the martensite from the unstable phase below
700° C. into the stable phase of pearlite, due to tempering, does
not proceed directly, but indirectly through an intermediate

88 CORROSION AND PRESERVATION OF IRON AND STEEL

metastable form, to which the name osmondite has been given
by the authors in honor of the celebrated investigator Osmond.
Osmondite is the most soluble of all the forms intermediate
between martensite and pearlite. As the tempering temperature
gradually increases, the martensite first gradually changes into
osmondite, until at 400° C. the whole mass consists of this alone.
In order to prevent confusion in nomenclature, the term troostite
has been hitherto retained by the authors for the designation of
the intermediate states between martensite and osmondite, and
for the stages intermediate between osmondite and pearlite they
have continued to use the term sorbite. If the tempering tem-
perature is raised above 400° .C. there again occurs a gradual
transition from the readily soluble osmondite to the less easily
soluble pearlite. Such transitions from a preliminary condition
to a stable final condition, through one or more less stable inter-
mediate conditions, are of not infrequent occurrence in physical
and chemical processes. As an instance may be mentioned the
combustion of hydrogen, which, according to Nernst, 1 is not
burnt direct to water, but proceeds indirectly through the medium
of the less stable peroxide.

"Another remarkable point in Fig. 24 is the lowest value for
the solubility at a tempering temperature of 100° C, which differs
little from that of untempered steel. The variation lies, however,
within the margin of error for the process, so that whether still
another metastable intermediate form occurs is not yet proved."

Similar work carried out on a sample of pure mild steel is
recorded by Heyn and Bauer as follows:

" The Influence of Quenching and Reheating Soft Mild Steel
upon its Solubility.

"For the experiments a very low carbon mild steel was
used, of the following composition:

Per Cent.

Carbon 0.07

Silicon 0.06

Manganese 0.10

Phosphorus 0.01

Sulphur 0.019

Copper 0.015

The material was in the form of a rolled square bar 25 X 25 milli-
1 Zeitschrift fur Elktrochemie, 1905, Vol. II, p. 713.

APPLICATION OF ELECTROLYTIC THEORY

meters in cross-section. From this bar were cut transverse sec-
tions 23 X 23 X 6 millimeters, which were polished bright on
an emery wheel. Near one corner a hole was bored, by which
the small plates were suspended on glass hooks in the dilute
sulphuric acid.

"The test-pieces were quenched from a temperature of 1000°
to 1030° C, since mild low carbon steel first becomes homogeneous

0,5

0.4

0.3

uiO.2
2

^0.1

t^0. 2

0.1

• Series 1. of Experiments
1 1

o Series 2 of Experiments

o *

^V o

At End of

72 Hours

^Ti

LiJ

U / I

SOLUBILITY IN \% SULPHURIC ACID

la «

► *£-"'*

* \.r.

1 o — i

• S

At End of 48 Hours

° ^^^ '■

"-*^l - J C

'i ,-^" <

• rt

ir>

• . - . ,._

1 o

— — ■S"""'"

o r^X^o

i 8~

o •

^ ^ *"^ s

At End of 24

Hours

rH r 1

200

300 400 500 600

f = TEMPER]NG HEATS

900

Fig. 25. — Mild steel, carbon .07 per cent., quenched at 1000 degrees to
1030 degrees centigrade in water at 18 degrees centigrade; tempered for
two hours at if) degrees centigrade. (Heyn and Bauer.)

from 900° upwards. The duration of reheating was two hours,
and it was effected at the temperatures t given in Fig. 25. As a
rule, at the end of the two hours the specimens at t° were quenched
in water for the purpose of suddenly interrupting the tempering
effect. Only the specimens reheated at 700° were first cooled
slowly to 690°, that is, below the critical temperature 700°, and

90 CORROSION AND PRESERVATION OF IRON AND STEEL

then quenched, otherwise a new quenching effect would have
been added. The samples reheated at 900° were cooled quite
slowly to the temperature of the room.

"The loss in weight in 1 per cent, sulphuric acid was deter-
mined in two series of experiments (I and II) , which were carried
out independently of one another. The Roman numerals against
the numbers of the test-pieces in the table indicate to which
series the particular experiment belongs.

"The test-pieces were suspended by means of glass hooks on
glass rods, the hooks being passed through the holes. They
were entirely immersed in the acid without being in contact
with each other. The quantity of attacking liquid was 2500
cubic centimeters for every ten to eleven test-pieces. The fol-
lowing are the numbers of the test-pieces which were placed
together in the respective vessels:

Series I. — Test-pieces

Vessel A Nos. 1, 3, 5, 7, 9, 11, 13, 15, 17, 52

Vessel B Nos. 2, 4, 6, 8, 10, 12, 13, 16, 18, 53

Series II. — Test-pieces

Vessel A Nos. 19, 22, 27, 30, 38, 41, 44, 54, 57, 60, 62

Vessel B Nos. 20, 23, 28, 34, 35, 39, 42, 50, 55, 58, 63

The two series of experiments I and II were undertaken at differ-
ent times. In each series the test-pieces in both vessels A and B
were exposed to the action of the acid at the same time and in
the same room. The results of the two series are, however, not
immediately comparable, because the temperatures were not
observed; nevertheless they have been amalgamated for the sake
of clearness in Fig. 25. The losses in weight, in grams, at the
end of twenty-four, forty-eight, and seventy-two hours' immer-
sion in the acid, have been entered in the diagram. In the curve
the points relating to Series I are marked thus, #, and those
relating to Series II, thus O. The tempering heats are taken as
abscissae, and the losses in weight as ordinates.

"A maximum degree of solubility is plainly noticeable between
the reheating temperatures 300° and 400°, which again corre-
spond to osmondite. Thus this intermediate constituent, even
with so low a carbon content as 0.07 per cent., still shows its
special characteristic of greatest susceptibility to attack by the
acid. "On comparing Fig. 25 with Fig. 24, it will be observed

APPLICATION OF ELECTROLYTIC THEORY

that in the case of the hard steel the solubility of the test-piece,
with a degree of temper = (martensite), is lower than that
of the slowly cooled steel with degree of temper = 1 (pearlite).
With the mild steel, on the other hand, the solubility of the
untempered test-piece is greater than that of the fully annealed
one. This is no matter for surprise, since it is not possible with
mild low carbon steel, however suddenly quite small samples
may be quenched, to obtain martensite — that is, the degree
of temper = 0."

Influence of Wire-drawing and Subsequent Annealing. — The
work of Heyn and Bauer on the influence of wire-drawing and
subsequent annealing is particularly suggestive. For the details
of this work the original should be referred to. Some of the
conclusions reached are as follows:

"1. By cold-drawing the wire, the solubility of the mild steel
is increased. It rises rapidly until the ratio of elongation N = 4
is reached, but drawing beyond that point produces compara-
tively little further increase of the solubility.

700

800

90Gc°

Fig. 26. — Influence of heating- on the solubility of cold-drawn, hard wire.
(Heyn and Bauer.)

"2. In the case of the harder wire, the curves of relative
solubility for the different periods of attack lie nearer to each
other than those of the softer wire, as may be seen from a
comparison of Figs. 26 and 27. This means that the relative

92 CORROSION AND PRESERVATION OF IRON AND STEEL

loss of weight of the softer material is less during the first
period of attack than that of the harder material, but that
as the period of attack is prolonged, the loss of weight increases
more rapidly in the softer than in the harder steel wire.

Fig. 27. — Influence of heating on the solubility of cold-drawn, soft wire.
(Heyn and Bauer.)

"3. After annealing the wires drawn from the same mild
steel material, the degree of solubility is immediately no longer
affected by the cold-drawing of the wire, however far this may
have been carried. Wire-drawing, therefore, produces no change
in the material which cannot be removed by annealing.

"4. The heating of the cold-drawn wire involves a decrease
in the degree of solubility (Figs. 26 and 27). The decrease in
solubility is plainly noticeable even after heating to 100°."

Cold-working Increases Solubility of Mild Steel. — The solu-
bility of mild steel is shown to be increased by cold-working,
irrespective of whether the cold-working has been effected by
stretching or compression. Part of this effect is shown in Fig. 28.
The investigators considered the solubility of iron from the electro-
chemical point of view, and are led to assume that iron exists in
some allotropic form induced by cold-drawing. They say:

"It has to be further investigated whether iron, with reference
to the increase of its solubility by cold-working, occupies an
exceptional position with respect to all other metals, or whether
there are other metals which behave in the same manner. If
the former is the case, then the increase in solubility of cold-
worked iron can only be attributed to the presence of a more
readily soluble allotropic form of the iron, which is produced
by cold-working.

APPLICATION OF ELECTROLYTIC THEORY

"Since the degree of cold-working does not require to be the same
in all parts of the cold-drawn wire, one might then expect to find
at different parts of it different quantities of the more soluble
modification of the iron. This would explain why the surface
of the drawn wire, on being subjected to the solubility test,
becomes roughened, while the same wire after annealing preserves
its smooth surface under the acid attack. The difference in the
solubility might be accentuated if the galvanic effect differed of the
parts subjected to varying degrees of cold-working. Investigations
upon the electric potentials of cold-drawn and annealed wire in

t 500

2^400

5$ 3

> w

;200

3£

uj O

„96

*72

Hours Att

acK

^^^>24

•m

2 4 6

STRETCHING RATIO n =

Fig 28. — Influence of cold-drawing on the solubility
of mild steel wire. (Heyn and Bauer.)

contact with acid and solutions containing Fe ions were not
undertaken by the authors. On the other hand, the differences
of potential in distilled water were measured, and were found to
be considerable, as appears from Fig. 29.

It cannot, however, be concluded from these figures that the
potentials peculiar to the cold-drawn and annealed wire neces-
sarily differ from each other, since what is measured in water
is not the actual potential of the iron, but the potential of an
oxygen electrode formed on the iron, which certainly does differ
considerably as between cold-drawn wire and annealed wire.
The oxygen electrode formed on the cold-drawn wire appears
more electro-positive than that on the annealed wire, which

94 CORROSION AND PRESERVATION OF IRON AND STEEL

indicates that during the experiments the drawn wire consumed
more oxygen, that is, it rusted sooner than the annealed wire. 1

*SURED

VT END OF

96
HOURS

Fig. 29. — Electric potential drop between a cold-drawn hard wire
and the same wire after annealing. Measured in distilled waters
at 17 to 18 degrees C. The cold-drawn wire is poorer in quality
than the annealed. (Heyn and Bauer.)

"Pairs of wires were prepared as follows: Apiece of wire 100
millimeters long in the original cold-drawn state marked u was

Effi

Fig. 30. — Pairs of wires connected together for
rusting test. (Heyn and Bauer.)

attached as shown in Fig. 30, by means of platinum wire to
another piece of wire marked g, annealed for half an hour at 900° C.

1 E. Heyn and O. Bauer. "Uber den Angriff des Eisens, etc." "Mit-
teilungen aus dem Koniglichen Materialprufungsamtes." Gross-Lichterfelde,
Nos. 1 and 2, 1908.

APPLICATION OF ELECTROLYTIC THEORY 95

These pairs were supported at their free ends on glass rods, and
completely immersed in distilled water of about the warmth of
the room. At the end of twenty-two and forty-nine days the
losses in weight of the single wires due to rusting were ascertained,
and in all cases it was found that the cold-drawn annealed wire
was more strongly attacked than the annealed wire, the propor-
tion being from 103: 100 to 108: 100. Whether this proportion
remains unaltered or is reversed under a long exposure is not yet
determined, but such a possibility is by no means precluded.

"100-millimeter lengths of the 'same kind of steel wires were
then attached together in pairs in the same manner in order to see
whether the attack on iron by dilute sulphuric acid (1 per cent.)
was affected when cold-drawn and annealed wires were in metal-
lic contact with each other. It appears that the strength of
the sulphuric acid attack on the cold-drawn wire is increased by
1.19 to 3.3 times. Therefore if the degree of cold-working differs
in different places in the same wire, the wire will be attacked
unevenly, owing to the contact of the unequally worked parts,
and moreover the attack on the most heavily worked will proceed
quickest. This is the explanation of the furrow and rough-
nesses on the surface of the drawn wires after an acid attack,
and the smooth nature of the surfaces of the wire annealed before
attack.

"On the basis of the foregoing considerations on elastic and
plastic change of form, the state of a cold-drawn wire might be
compared with that of a structure produced in the following
manner. Imagine a pressure cylinder filled with a thick fluid
mass, for instance, pitch, and that mixed with this is a great
number of small cylindrical helical springs. Pressure is then
applied to the mixture, and it is forced out of the cylinder through
a round hole. The pitch undergoes a purely plastic change of
form, but the helical springs suffer an elastic change of form.
As soon as the pressure from outside the mold ceases the springs
have the tendency to return to their state of equilibrium, that
is, to their unstrained state, but they are partly prevented by
the mass of pitch surrounding them. Consequently they remain
partially under stress, and the amount of this stress may vary in
different parts of the mass. If one can imagine a similar condi-
tion with respect to cold-drawn wire, it is intelligible why the
amount of the elastic stresses varies so greatly at different parts."

96 CORROSION AND PRESERVATION OF IRON AND STEEL

The Acid Test as a Measure of Corrosion Resistance. — This
interesting discussion opens up the important question of the
use of an acid test as a measure of corrosion resistance. This
test has been much used by some manufacturers who believe it
to be reliable as an accelerated method for measuring the sta-
bility and hence the corrosion resistance of their products. The
applicability of the acid test is absolutely denied by some author-
ities/ and is treated with caution by others. 2 Committee U of
the American Society for Testing Materials has reported that the
acid test should not be used to decide the relative resistance to
corrosion of different types of iron and steel, but since it seemed
necessary that investigators should use the test in the same way,
the following tentative suggestions for carrying it out were made :

Samples to be tested are first stamped with suitable marks
for identification, and are then accurately machined to a standard
size, 2 inches in length, I inch in width, and -±q inch in thickness.

The samples should be cut longitudinally in the direction in
which the metal was rolled. A hole J inch in diameter is drilled
in each sample about 1 inch from one of the ends.

Samples are polished first with No. 00 emery paper, and
finally with flour emery. The polishing should be done as much
as possible, so that the polish marks run at right angles to the
direction in which the metal was rolled. After polishing, the
test pieces should not be handled with the fingers or allowed to
come in contact with dirt or grease of any kind.

The test-pieces should then be weighed carefully to four deci-
mal places on a chemical balance, and are then strung on a glass
rod with a double reverse right-angle bend at each end, which
will allow the pieces to be suspended in a relatively large beaker
or other suitable dish, so that their upper edges will be submerged
to i inch below the surface of the liquid. The glass rod should
be only slightly less than i inch in diameter, and the distance
between any two adjacent test-pieces when suspended in the
acid should not be less than i inch.

The acid should be exactly 20 per cent. H 2 S0 4 (1.144 specific
gravity at 15° C).

The test-pieces are immersed in this acid for one hour, an
approximately sustained temperature of 15° C. being maintained

1 Saniter, Jour. Iron and Steel Inst., LXXIX (1909), p. 98.

2 Speller, Proc.Eng. Soc. West. Pa., 24, 10, 541, Jan. 1909. Sang., Ibid.

APPLICATION OF ELECTROLYTIC THEORY 97

by any suitable cooling device, such as a large outside container
of cold water.

At the end of exactly one hour's immersion the test-pieces
are quickly removed from the acid, well rinsed with running
water, wiped as dry as possible and kept for one hour in a desic-
cator over sulphuric acid before weighing.

The results should be recorded as actual and not as percentage
loss.

Note 1. — It has been shown that the corrosion of the sample is
not directly proportional to the area exposed. Hence it is essential
in making comparative tests that a standard size be adopted.

Note 2. — Should the samples for investigation be less than
tV inch in thickness, the other dimensions should be adhered to
as closely as possible.

Note 3. — The physical condition of the surface of the sample
is found to materially affect the rapidity of its solution. It is
therefore desirable to finish all samples in the same way.

Note 4. — The strength of sulphuric acid chosen is that
concentration which contains approximately the maximum num-
ber of hydrogen ions, and which experiments have shown to be
the most suitable.

Note 5. — It is necessary that the samples be suspended in
a reasonably large body of acid in order that the ferrous sulphate
formed by the reaction may sink to the bottom or be otherwise dissi-
pated through the solution, so that the concentration of the acid in
the immediate vicinity of the samples be not materially changed.

Note 6. — The greatest care should be taken to employ only
chemically pure sulphuric acid. It has been shown that very
minute traces of arsenic, for example, seriously retard the rapidity
with which iron is dissolved.

The Relation between the Acid Test and Resistance to Corrosion.
— The authors of this work are not willing to state that the sta-
bility of a steel as measured by an acid test bears no relation to
its long-time resistance to corrosion. Nevertheless the test has
not shown itself reliable when used to distinguish between the
rust-resistant qualities of samples of different types of iron and
steel. The authors are in complete agreement with the following
opinion of Heyn and Bauer: 1

"It is quite apparent that the chemical composition of iron
1 Jour. Iron and Steel Inst., 1, 1909, p. 185.

98 CORROSION AND PRESERVATION OF IRON AND STEEL

must considerably affect the solubility in dilute sulphuric acid,
and that the influence of the composition may equal or exceed
in magnitude the influence described in the previous sections.
It must therefore again be specially emphasized that the com-
parative tests of which particulars have been given only apply
as regards their results, subject to the elimination of the influence
of the chemical composition as an unknown variable. The whole
of the solubility tests lead to relative results only, and the abso-
lute degree of solubility has no special value. Only the relative
values of the solubilities of the same material in various states
of treatment are to be regarded as useful in throwing light upon
the present state of the metal. But it is unsafe to draw from the
behavior of one metal conclusions as to the behavior of another,
even though their chemical compositions may be very similar, since
it is always possible that such substances as are not determined in
the course of analysis exercise an effect upon the solubility."

To the above may be added the now well-known fact that
very slight amounts of certain impurities in the acid used produce
wide variations in the quantity of iron dissolved in unit time.
Very small quantities of arsenic contained as an impurity in the
acid will check the solubility of iron to an extraordinary degree, 1
and this is probably true of many other sometimes unknown and
unsuspected impurities. This fact is readily accounted for by
the electro-chemical theory of corrosion. As has been previously
explained, corrosion is a surface action and solution in acid
depends upon the rapid depolarization and disengagement of
hydrogen from the surface. This depolarization is easily affected
one way or the other by impurities in the metal, and also in the elec-
trolyte which provides the attack. It is not to be expected that
the rapid attack furnished by a dilute acid will be comparable
to the slow rusting of iron under the conditions of service. In
the former case depolarization and solution go on rapidly, whereas
in the latter case the difference of potential between positive
and negative points and nodes on the surface is stubbornly main-
tained for long periods of time. Unfortunately the same con-
siderations must apply to all the acceleration tests that have
been proposed and used by engineers and metallurgists, such as
aerated brines and natural waters through which air and car-
bonic acid are bubbled.

1 Burgess, Electro-Chem. and Met. Ind., IV, 384.

APPLICATION OF ELECTROLYTIC THEORY 99

Probably No Reliable Acceleration Test Possible. — Owing to
the nature of corrosion it is probably true that no perfectly
reliable acceleration test for corrosion resistance can be devised.
Corrosion, in the natural process of rust formation, that is to
say, in very slightly acid media, is a question of comparatively
slow growth under special conditions, and any effort to hasten
the action changes all the conditions of equilibrium, producing
an entirely different order of phenomena.

Nevertheless, evidence has been brought out in the preceding
paragraph to show that stresses and strains as the result of cold
rolling or imperfect annealing of steel will affect not only the
degree of solubility in acid, but also the tendency to maintain
differences of surface potential affecting the rapidity of corrosion.
In view of this it is probable that when its limitations are under-
stood and its results properly interpreted, the acid test will be
useful in the hands of competent investigators.

Efficient Annealing Essential to a Rust-Resistant Metal. — A
careful reading of the preceding paragraphs should show that in
order to produce the most rust-resistant steel, it is not sufficient to
work towards purity of the metal and homogeneous structure,
but also efficient annealing as the last stage in the preparation
of the product is absolutely essential. Recent work would appear
to show that the annealing temperature necessary to remove the
stressed condition is not high, provided sufficient time is allowed
for equilibrium to be reached..

Effect of the Rusting Medium, Electrolytes, Natural Waters, etc.
— In regard to two points all investigators are agreed, and as
these furnish some common ground it is interesting to record
them. Iron cannot rust in air or oxygen unless water is present,
and on the other hand it cannot rust in water unless oxygen is
present. The products of the combustion of coal, consisting, as
they do, largely of carbonic and sulphurous acids, aggravate the
corrosion problem for railway engineers, and in the neighborhood
of large cities. It follows from what has been written in the
previous paragraphs that any acid in the atmosphere or the water
by which iron is environed will increase corrosion. In very damp
climates the conditions are aggravated, and on the Isthmus of
Panama they are said to be particularly bad. Natural waters
do not all behave in the same way, some making a much more
rapid attack on iron and steel than others. Sea-water, as is well

100 CORROSION AND PRESERVATION OF IRON AND STEEL

known, stimulates corrosion, as do also all acid mine waters.
These results are easily explained by the electrolytic theory of
corrosion. Provided that the dissolved ions do not actually
inhibit the action, all dissolved salts which increase the electrical
conductivity of the medium result in stimulated galvanic action
and supply hydrogen ions. In some cases, however, it has been
noted that brackish waters from rivers in which sewage flows is
less active in promoting the pitting of boiler tubes than much
purer waters. This fact merely serves to emphasize the conclu-
sion that the speed of the action is functioned by the electroytic
solution pressure of the iron, and that in some cases the effect
of dissolved impurities may be inhibitive instead of stimulative.

Highly Alkaline Solutions Prohibit Corrosion. — Iron does not
rust in alkaline solutions which contain an excess of hydroxyl ions,
that is to say, if the alkalinity is high enough, but in dilute alkaline
solutions rusting goes on in a dangerous fashion, the tendency to pit
being accentuated. This action is incompatible with the theory
that free acid is necessary to induce rusting, but is explained by
the electro-chemical theory. Hydrogen and hydroxyl ions can exist
in a solution at the same time to a considerable extent only when
they are separated or held apart by potential differences, and there-
fore in the rusting of iron the concentration of hydroxyl ions, or,
in other words, the alkalinity, must reach a certain value before
rusting is entirely prohibited. This is because, owing to the very
slight dissociation of water, an excess of hydroxyl ions would be
incompatible with the presence of free hydrogen ions. 1

Natural Waters May Contain Inhibitive Impurities. — In addir
tion to hydroxyl ions there are many other impurities that may
exist in natural waters to inhibit corrosion. It has already been
stated that iron cannot rust in water unless oxygen is present,
for although the first cause of corrosion is due to the electrolytic
exchange between iron and the hydrogen ions, the presence of
oxygen is absolutely necessary for the action to proceed. Walker
develops the point as follows : 2

1 After this paragraph was written the authors' attention was called to
the fact that iron has a considerable solution tension in strong boiling alkaline
solutions, but in such a case the equilibrium is reversed and the metal acts
the part of a negative radical in conjunction with oxygen. This point is at
present being investigated. Sec also Zincite, Chem. Zeit., 12, 355.

-Trans. Am. Electro. -Chem. Soc, XIV, 175 (1908).

APPLICATION OF ELECTROLYTIC THEORY 101

"Since oxygen is necessary to insure the continuous removal
of the hydrogen film, it is obvious that if no oxygen be allowed
to reach the iron through the water, no corrosion can take place.
This fact teaches us much regarding the corrosion of boiler shells
and tubes. Pitting may be entirely avoided if the air be removed
from the free water before its introduction into the boiler. This
may best be done by the employment of an open feed-water heater,
or, what is better still, a feed-water heater connected to the dry
vacuum pumps of the condenser. If such treatment is not pos-
sible, the air may be removed from the feed water by drawing
the water through a closed box containing scrap iron; the oxygen
in the water is used up in corroding the scrap iron instead of the
boiler tubes. Or, the oxygen in the water may be absorbed by
feeding into the boiler with the water a very small quantity of
an alkaline solution of a tannin material. Such a solution of
alkaline tannate will break up under the pressure and tempera-
ture of the boiler, with the formation of a pyrogallate of the
alkali, and this rapidly absorbs the oxygen. Soda ash, or rather
alkali, is of course useful, but not because of its effect upon the
oxygen content, but because, as has already been explained,
corrosion is inhibited by thus decreasing the hydrogen ion con-
centration of the water."

Dissolved Oxygen a Stimulator. — A theory must inevitably
be judged according to whether or not it furnishes a satisfactory
explanation of observed facts. A subject so complex as the
corrosion of iron furnishes many examples of recorded facts that
appear on first thought to completely contradict each other in
the light of the electrolytic theory. Such discrepancies, however,
usually find a rational and simple explanation. The particular
point under discussion furnishes a good example of just such a
case. On first thought it would be assumed that the purer the
water and the lower the electrical conductivity, the less will be
the tendency to rapid corrosion. If it were possible to deal only
with absolutely pure waters, this conclusion would undoubtedly
be justified. Natural waters are, however, only more or less
pure, so that the conditions necessary for corrosion to take place
are always present, as far as this factor alone is concerned. When
we further consider, however, that dissolved oxygen is a necessary
and important factor in the corrosion problem, we get new light
on the subject. Every sanitary engineer knows that the measure

102 CORROSION AND PRESERVATION OF IRON AND STEEL

of dissolved oxygen is a strong indication of the purity of a water
supply. In sewage-laden natural waters, dissolved oxygen does
not exist to any great extent, as the oxygen is fixed in oxidizing
the organic matter present. The fact that certain impure waters
have been found to be excellent boiler waters, thus finds a simple
explanation under the electrolytic theory. It would not, of course,
be safe to conclude that contaminated waters are less corrosive
than pure waters, for this is far from being the truth. If two
forces are present at the same time, acting in opposite directions,
and if these forces are unknown and variable, it is impossible to
tell what the resultant will be except by experiment under all
the possible conditions. In the case in point, the impurity of
the water should tend to increase the electrolytic action, but the
absence of dissolved oxygen counteracts this tendency.

Relative Corrosion in Fresh Water and in Sewage. — In con-
nection with what has been said above, Howe and Lodge's 1
exposure tests on wrought and cast iron plates are most interest-
ing. These tests were carried out for the purpose of studying
the efficiency of various protective coatings, but incidentally
they serve to show the relative amount of corrosion which took
place in fresh water and sewage.

Loss of Weight of Wrought and Cast Iron with Different Pro-
tective Coatings and under Different Conditions, in Pounds
per Square Foot of Surface per Annum.

Wrought-Iron Sheets (No. 32 Gauge, Black.)

Protective Coatings

Exposed to the Weather

Inland

Immersed in —

Average

Canada

New York
State

Fresh Water

Sewage

Bower-BarfTed

Tinned

Nickel-plated

Galvanized

.0
gain, .002.0

.0
gain, .000.4
.001.0
.001.3
.000.2

gain, .000.3
.0C0.1
.000.5

.006.7
.019.4
.050.4
.045.9
.083.9
.137.0
.179.0

.003.6
.007.1
.003.1
.080.5
.117.0
.169.0
.182.0

.002.5

.006.2
.013.5
042

Barffed

Black, i.e. unprotected
Copper-plated

.003.1
.022.6
.005.0

.051.2
.082.5
.091.6

Average

.000.02

.005.1

.074.6

.080.3

.040

^owe, Metallurgy of Steel, 2d Ed. I, 372.

APPLICATION OF ELECTROLYTIC THEORY

103

Cast-Iron Plates.

Protective Coatings

Exposed to the Weather
Inland

Immersed in —

Average

Canada

New York
State

Fresh Water

Sewage

Bower-Barffed

"& paraffined
Galvanized

gain, .004.0
.000.6
.0

gain, .003.1
.001.9
.0

gain, .003.1
.002.5
.005.0
.012.0

gain, .005.5
.000.2
.049.1
.065.5
.131.7
.150.8
.148.3

.001.4
.008.4
.061.0
.061.0
.083.3
.119.2
.272.4

gain, .002.8
.002.8
.027.5

Tinned

.041.1

Nickel-plated

Copper-plated

Black, i.e. unprotected

gain, .003.4
" .004.0
" .006.3

.053.5

.067.8
.106.6

Average

gain, .002.9

.002.1

.007.2

.086.7

.041

A single sheet of No. 23-gaged refined wrought iron was cut
into plates 6 in. by 12 in., and others 6 in. by 6 in. Of the 6 in.
by 12 in. pieces some were exposed without treatment of any
kind, the scale left on; others were tinned; still others were gal-
vanized by the Rhode Island Tool Company. Of the 6 in. by
6 in. plates some were Bower-Barffed by the Yale and Towne
Company, others were Barffed by the Pratt and Cady Company,
still others were nickel-plated and copper-plated, in each case
after pickling. The cast-iron pieces were skin-bearing plates,
4 in. by 3.5 in. by 0.187 in., presented by Professor George W.
Maynard. These were subsequently given the coatings indicated,
their original skin being retained in all cases.

One set of the pieces thus prepared was exposed on the roof
of a dwelling house in the Eastern Townships of the Province of
Quebec, Canada, by Mr. E. C. Hale, of Sherbrook, Canada; a
second was similarly exposed in a village in Rensselaer County,
New York State; a third was immersed in the Chestnut Hill
(Boston) Reservoir, by Mr. Desmond FitzGerald, of Boston; a
fourth was immersed in the Boston main sewer, near the pumping
station, by Mr. H. H. Carter, of Boston.

In each of the conditions of exposure the wrought-iron pieces
were in one open wooden crate, the cast-iron ones in a second,
the corners of the pieces (and in the case of the 6 in. by 12 in.
wrought-iron pieces a small space in the middle of the long sides)
alone being in contact with the crate; and care was taken that
the specimens should not touch each other or any other metallic

104 CORROSION AND PRESERVATION OF IRON AND STEEL

substance. Though exposed nearly a year, including autumn,
winter, and spring, at the end of the experiments the gummed
labels still adhered to twelve out of the twenty-six specimens
exposed in Canada and in New York.

The cast iron lost about as much as the wrought iron; the
loss was about the same in fresh water and in sewage. Comparing
the different conditions of exposure, immersion, of course, greatly
accelerated rusting. Thus, in ten out of the fourteen sets of
cases the pieces immersed in fresh water lost at least twenty
times as much as those exposed to the weather in New York.
The loss in sewage is slightly greater than in fresh water on the
unprotected specimens, but in seven out of the fourteen cases
the loss in fresh water equals or exceeds that in sewage, a result
that was found difficult to explain before the electrolytic theory
of corrosion was developed.

Non-Corrosion of Deeply Immersed or Buried Iron. — If it be
admitted as a fact that both hydrogen ions and free oxygen are
necessary to promote the corrosion of iron under the electrolytic
theory, many curious instances which have been recorded of resist-
ance to corrosion of iron which has been deeply immersed in water
or buried deep in the earth are explained. If the zone of dissolved
and active oxygen has been passed the zone of corrosion has like-
wise been passed, and the problem resolves itself into one of deter-
mining the limits of penetration of active oxygen. These conclu-
sions should be considered in all cases where, as on the Isthmus of
Panama, deep piling in engineering construction is contemplated.

Wemlinger 1 has collected some interesting data on this sub-
ject which we include here. "It has been stated that steel sheet-
piling has often been used as a part of permanent structures, and
the question naturally arises: What is the life of the steel sheet-
piling embedded in the ground? In order to be able to fully
answer this question, it is necessary to consider the various con-
ditions that affect the durability of iron or steel in this case, such
as the nature of the soil and the composition of the metal itself."

"In the first place, it must be remembered that iron or steel
will not corrode in air unless moisture is present, nor will it cor-
rode in water unless air is present. 2 Deep down in the ground the

1 Engineering-Contracting, XXXI, 20, 407.

2 Compare Cushman, Corrosion of Iron, Bull. 30, Office Public Roads, U. S.
Dept. Agr.

APPLICATION OF ELECTROLYTIC THEORY 105

amount of air dissolved in the water is almost nil, and it is only
near the surface that the ground-water contains free air and
carbonic dioxide carried down from the upper layers.

11 There is every logical reason to believe that steel embedded
in the ground or left in contact with standing water will be pre-
served indefinitely. However, it is appreciated that, no matter
how much such evidence may be in favor of this conclusion,
practical experience will have greater influence. It is fortunate
that we can point to actual cases that seem to confirm the logical
reasoning. There is in the possession of one of the New York Gas
Companies quite a number of different articles of iron and steel
that have been unearthed from time to time while excavating for
laying gas pipes. Among these are arms and implements that
were found in lower New York, and proved to have been origi-
nally in the possession of British soldiers. Inasmuch as the British
troops sailed away from New York on November 25, 1783, it
follows that the articles are now at least 125 years old. While
pretty well rusted, these things are by no means entirely destroyed
as would be expected. As a matter of fact, it is believed that
most of the rusting occurred before the implements were com-
pletely embedded, and it is obvious that for some time these
articles must have been lying very near the surface, if not alto-
gether exposed.

"I have also seen a piece of gas pipe 2 in. in diameter that
was taken up after it has been in the ground for twenty years.
This is in such excellent condition that it is impossible to detect
any effect of corrosion. The original mill-scale is intact, and, in
fact, the appearance of the pipe is such that, except for the thin
coating of earthy matter, it has all the appearance of a pipe
fresh from the mill. It is the most remarkable evidence of the
durability of iron or steel embedded in the ground that I have
ever seen.

"At a meeting of the Engineers' Society of Western Pennsyl-
vania, held recently, during which the question of corrosion was
fully discussed, Mr. R. B. Woodworth made the following remarks:

"'It may be interesting to state that I have in my possession
a piece of iron that was put into the Mississippi River in the
year 1863, and that is to-day in as good condition as when it was
put in. It was painted with some kind of red paint which looks
very much like red oxide of iron, and the paint is just as good

106 CORROSION AND PRESERVATION OF IRON AND STEEL

to-day apparently as it was the day it was put in. The con-
ditions there must have been extremely favorable to its preser-
vation. The water of the Mississippi River at that point is
probably pure and the iron was buried in the Mississippi River
silt, sand or gravel, so as to be protected from exposure to
atmospheric conditions, and as a consequence there was nothing
to make corrosion.

"'I have also a theory that so far as our own rivers are con-
cerned, although they do carry a great deal of free acid, yet when
the steel is placed sufficiently below the river-bed as to get the
advantage of the filtering action of the sand and gravel, we do
not have to do with an acid-laden condition, but rather with the
condition of practically pure water free from contact with air,
and I think we have every assurance to believe that the life of
steel under such circumstances will be indefinite. I was told
by the engineers who have charge of the construction of the Black
Rock Lock at Buffalo, that the waters of the lake are not con-
sidered in any way dangerous, and that steel of ordinary com-
mercial quality in such installations may be expected to last
for any reasonable length of time.'

"Further evidence of the durability of iron and steel embedded
in the ground is found in the Proceedings of the American Society
of Civil Engineers, Nos. 6 and 7, for the year 1907. At the Annual
Convention held at Mexico City, July 10, 1907, among various
subjects for discussion was the following: Will iron or steel, used
in foundations, independently or in combination with other
materials, last indefinitely when in direct or indirect contact
with water?

"The most interesting contribution to this discussion was,
undoubtedly, that submitted by Mr. John F. O'Rourke, the well-
known foundation engineer and contractor. Mr. O'Rourke stated
that:

"'Iron and steel used in foundations, apart from conditions
where electrolysis may occur, last indefinitely when in direct or
indirect contact with water, provided the water remains un-
changed. The reason for this is obvious. Water attacks iron
or steel, on account of the oxygen it contains, and, if this is a
proportionately small quantity, the amount of oxygen contained
in wet concrete or ground is negligible, and, having once been
exhausted, the metal remains unharmed and protected.

APPLICATION OF ELECTROLYTIC THEORY 107

"'The writer has seen many cases where immersion in stand-
ing water has been a matter of years, and in every case the effect
upon the metal has been no greater than if it had stood for the
same length of time in linseed oil. In one case bolts on the inside
of cast-iron cylinders, filled with concrete, were exposed to the
salt water in the Harlem River for more than thirty years, and
when removed were found to be without rust. In another case,
a pipe was immersed for ten years in an artesian well, the water
in which had not been pumped for ten or fifteen years, and no
corrosion of this inside pipe had taken place, the scale was still as
fresh as when the pipe was new, and the tool-marks of the pipe-
coupling apparatus were still perfectly fresh.

" ' Similar results came under the writer's observation in
reference to the condition of rods and nails found in wood founda-
tions where the surrounding material was impervious to air, and
in one case which came under his observation, at the time of the
removal of the old elevated railroad columns in Greenwich Street,
New York City, prior to making way for the new structure in
1878, the bottom part of these columns and the bolts in the
masonry were found intact, the corrosion gradually increasing
until near the surface, where the material was almost entirely
destroyed by rust. This experience with both wood and iron,
where the renewal of the oxygen in the surrounding water was
prevented, has been uniformly that of finding the material per-
fectly preserved, so that, in the writer's practice, he does not
hesitate to advise the use of either material under conditions
where a fresh supply of oxygen is excluded.' "

"Busy" Iron does not Rust. — We have now to consider the
well-known fact that in rapidly moving waters corrosion does
not take place so rapidly and dangerously as when the conditions
are comparatively quiescent provided dissolved oxygen is present
in both cases. This has been a matter of record by pipe-line
engineers, who have noted that their dangerous corrosion in
buried pipes always takes place from the outside in, and not from
the inside, out 1 . It follows that if the electrolytic explanation of
corrosion is correct,- undisturbed and comparatively quiescent
conditions are most favorable to the action. Differences of
potential establish themselves, as the ferroxyl indicator shows
at specific points and nodes. If the conditions are continually
1 Gaines, Engineering News; also Thomson, Com. Eng. XLII, 67 (1908).

108 CORROSION AND PRESERVATION OF IRON AND STEEL

changing on the surface, it is presumable that the points. of
potential difference will not be maintained, but, on the contrary,
will never remain long enough in one spot to produce deep
rusting. It may safely be said that dangerous corrosion is
localized electrolytic action. In a pipe-line as in many another
case of structural steel, if the corrosion effect was distributed,
little damage would accrue, but as a chain is no stronger than
its weakest link, so a pipe-line, a boiler tube, or a bridge member
can fail and endanger human life only because it has given way
at some specific point. One of the authors of this work observed
a number of years ago that busy iron does not rust, and explained
this observation by the fact that frequent and recurrent vibra-
tion was sufficient to break up specific points of potential
difference on the surface. Sang's observations on this point are
interesting in this connection: 1

"The case of steel rails is an interesting one, showing, as it
does, the effect of vibration on rusting. A rail which has been
in service, but has been laid to one side, will rust all over, but
especially at the ends where the vibration of the fish-plates has
removed the mill-scale, and on the smooth top of the head. On
the other hand, a quite remarkable fact, which has been univer-
sally confirmed and can be easily observed by any one, is that a
rail while in service will not rust nearly as rapidly as one which
is" lying out of service. The rusting takes place in proportion to
the service, and lines over which fast trains pass frequently,
causing much vibration, will practically not rust at all, whereas
the rails of turnouts or sidings, which undergo less service, and
that of a slow nature, will rust to a certain extent. One observer
(J. M. Heppel) has reported the case of some rails at Madras,
India, which lost three pounds to the yard lying in the yard exposed
to the sea air, while the rails in service near by were not percep-
tibly affected. The top of a rail is compressed and smoothed down
in service by the grinding of wheel tires, for there is always a
certain amount of slip, especially during acceleration and retarda-
tion. Galvanic action between the smooth head of the rail and
the rest of it has been suggested to explain- this immunity from
rust, but it is not at all likely that the foot would owe its pro-
tection to the thin stratum of denser metal so far removed from
it. If that dense skin on the top of the rail were not crushed
1 Proc. Eng. Soc. West. Pa., XXIV, 10, p. 523.

APPLICATION OF ELECTROLYTIC THEORY 109

beyond its elastic limit, it would, on the contrary, tend to acceler-
ate the corrosion of the steel in contact with it. The real reason
for this difference of behavior seems to lie in the observed fact
that oxidation is apparently arrested, or at least greatly retarded,
by vibration. 1 Explanation seems to stop at this point, but a
simple theory can be built on the assumption that the vibration
causes a shedding of the rust as soon as it is formed on the spots
that are not protected by mill-scale, and there is, therefore, no
acceleration of the action due to the accumulation of spongy
and electro-negative rust. The average speed of corrosion of a
vibrating body would be that of the formation of a first film of
rust. Most of the actual rust on rails is probably due to the
rapid evaporation of rain on the surface. In the case of rails in
service, the first film of rust would be confined to bare spots and
cracks in the mill-scale, and the vibration would prevent its
working its way under the mill-scale as would happen if the rail
were at rest. The top of the rail being denser might be expected
to resist corrosion better when the rail is out of use; such is not
the case, however. The surface has not only been subjected to
hammering and crushing, but also to abrasion and rolling, and it
has become short and crackeled and sometimes exfoliated; once
laid aside, the smooth top of an old rail rusts very rapidly."

Application of the Electrolytic Theory to Special Corrosion
Problems. — There are many applications of the electrolytic
theory to special corrosion problems that might be properly
included in this chapter. These will, however, be considered
separately in the second portion of the book which is specially
devoted to the preservation of iron and steel. In addition to
this, much of the material given in this chapter might properly
be presented in the next chapter, in which the specific bearing
of the electrolytic theory on the stimulation and inhibition of
corrosion are dealt with. In treating a complex problem of this
kind, the arrangement of the material is not a simple matter,
but it is hoped that enough has been said to make what follows
quite clear to the reader.

1 Edwin Clark: Proc. Soc. C. E., Yr. 1868, p. 554.

CHAPTER V

THE INHIBITION AND STIMULATION OF
CORROSION

Use of the Word "Inhibition." — The word "inhibition" in
relation to the corrosion problem was first used by one of the
authors, 1 to mean the restriction or repression of corrosion, and
not its complete prohibition. As Sang 2 points out: "The pro-
tective effect is, sooner or later, overcome, and clearly indicates
that inhibition furnishes something to the iron, be it substance
or physical state, which under the attacks of corrosive agencies
is slowly expended until destroyed or brought below the safe
limit of protection." Inhibitory treatments have the effect of
rendering iron passive, but the passive condition is not the stable
condition, so that there must be a continual tendency at work
which causes the return to the normal surface condition. The
passive state of the metals has already been described in a previous
chapter and it will now be necessary to discuss the subject from
the standpoint of the protection of iron.

Hydroxyl Ions as Inhibitors. — All substances in solution which
contain hydrogen ions, such as acids, stimulate the corrosion of
iron. This is also true of salts of strong acids and weak bases,
which, though perfectly stable in a dry condition, hydrolyze in
solution to an acid reaction; or which, though neutral in fresh
solutions, undergo slow decomposition under the action of light,
with the formation of acid salts or free acid. With certain excep-
tions, salts which are perfectly neutral in solution do not prevent
oxidation but appear to aid it by increasing the electrolytic action.
All substances which develop hydroxyl ions in solution, such as
the alkalis or salts of strong bases with weak acids, to a certain
extent inhibit, and, if the concentration is high enough, abso-
lutely prohibit, the rusting of iron. 3

Bushman Bull., Office Public Roads, U. S. Dept. Agr. (1907).

2 hoc cit., p. 536.

3 Under exceptional conditions this statement may require modification.
In this connection see note, p. 100.

110

INHIBITION AND STIMULATION OF CORROSION 111

Under the electrolytic theory the explanation of the pro-
tection afforded by hydroxyl ions is a simple one. Owing to
the small dissociation of water, hydrogen ions cannot exist in a
solution in which the hydroxyl ions are in excess. As hydrogen
ions cannot exist or be locally formed in sufficiently strong alka-
line solutions, no attack is made upon the iron, which remains
permanently unaltered. If, however, the concentration of the
hydroxyl ions is not sufficiently great, electrolysis can go on
with an apparent stimulation of the pitting effects similar to
that produced by perfectly neutral electrolytes, such as sodium
chloride.

Bichromate Solutions as Inhibitors. — As has already been
noted, solutions of chromic acid and potassium bichromate
inhibit the rusting of iron. In order to determine the concentra-
tion necessary to produce complete protection, a number of
polished strips of two different samples of steel were immersed
in bichromate solutions of increasing concentration, contained
in tubes which were left quite open to the air. There were twelve
tubes in each series, ranging by regular dilutions from tenth-
normal down to ten-thousandth normal. At the end of two
months the last four tubes showed graded rusting with accumu-
lation of ferric hydroxide. No rusting had occurred in any of
the solutions above tube No. 8, which contained six-hundred-and-
fortieth normal bichromate, a strength corresponding to about
8 parts of the salt in 100,000 parts of water, or about 2 pounds
to 3,000 gallons. Since solutions of bichromate do not hydrolize
with an alkaline reaction, but on the contrary are usually slightly
acid, some other explanation must be found for this remarkable
phenomenon. On first thought it would seem a paradox that a
strong oxidizing agent should have the effect of preventing the
oxidation of iron, and yet this is precisely the case. If, however,
the initial cause of rusting is the hydrogen ion, it is possible to
believe that under certain conditions oxygen would prove the
most effective of all inhibitors. 1

Passivity of Chromated Iron, — One of the authors has ob-
served that if a rod or strip of bright iron or steel is- immersed
for a few hours in a strong (5 to 10 per cent.) solution of potas-
sium bichromate, and is then removed and thoroughly washed,
that a certain change has been produced on the surface of the
1 See explanation of passivity, p. 28.

112 CORROSION AND PRESERVATION OF IRON AND STEEL

metal. The surface may be thoroughly washed and wiped
with a clean cloth without disturbing this new surface condition.
No visible change has been effected, for the polished surfaces
examined under the microscope appear to be untouched. If, how-
ever, the polished strips are immersed in water it will be found
that rusting is inhibited. An ordinary untreated polished speci-
men of steel will show rusting in a few minutes when immersed
in the ordinary distilled water of the laboratory. Chromated
specimens will stand immersion for varying lengths of time before
rust appears, but the induced passivity gradually disappears.

The passivity which iron has acquired can be much more
strikingly shown, however, than by the rusting effect produced
by air and water. If a piece of polished steel is dipped into a
1 per cent, solution of copper sulphate, a 10-second immersion
is sufficient to plate it with a distinctly visible coating of copper
which cannot be wiped off. 1 A similar polished strip of steel
which has been soaked over night in a concentrated solution of
bichromate and subsequently well washed and wiped will stand
from six to ten 10-second immersions in 1 per cent, copper sul-
phate before a permanent coating of copper is deposited. Even
a momentary plunging of the metal into the bichromate will
induce a certain passivity, but the maximum effect appears to
require a more prolonged contact with the solution.

Bichromate Solutions do not Attack Iron Free from Manganese.
— Moody 2 asserts that potassium bichromate prevents the for-
mation of rust, owing to the fact that it slowly dissolves iron and
its hydroxides. He observed that the addition of ammonia to
solutions of chromic acid and its salts which had been allowed to
act on iron produced precipitates of hydroxide. This point has
been carefully investigated by the authors, with the following
results: Iron which is free from manganese is not attacked by
solutions of bichromate, even if boiled for days in a flask fitted
with a return condenser. Manganese is, however, readily soluble
in bichromate solutions, and therefore iron rich in manganese
yields a sufficient amount to the solvent tiction to produce
a small amount of brownish manganese hydroxide when the
bichromate solution is poured off, made slightly ammoniacal, and

1 This experiment has failed in hands of certain experimenters who have
not been careful to use the copper sulphate solution as dilute as directed.
' Proc. Chem'. Soc. (Lond.), 1906,- 22, 15.

INHIBITION AND STIMULATION OF CORROSION 113

allowed to stand. If metallic manganese is boiled in bichromate
solutions it dissolves readily, and subsequent addition of ammonia
produces an abundant precipitate of brown manganese hydroxide.

Experiment with Bichromate to Induce Passivity. — In order to
show beyond doubt that a passive condition is induced by
immersing iron in a strong solution of bichromate the following
experiment was made: Two polished steel electrodes were pre-
pared and chromated by immersion for a number of hours in a
strong solution of potassium bichromate. The prepared elec-
trodes were then thrust tightly through a rubber stopper which
closed the Jena flask A, which was then filled with pure freshly
boiled distilled water in the manner shown in Fig. 7. The
electrodes were then attached to the poles of a primary battery
at about two volts potential. At the end of half an hour although
the potential was not. sufficient to disengage bubbles of gas and no
visible change had occurred, the electrode which was connected
to the zinc pole of the battery had lost its passivity, the other
retaining it.

Inhibitive Power of Pigments Containing Certain Oxidizing
Agents. — Wood 1 in 1895 commented on the power of paints and
pigments containing certain oxidizing agents, notably potassium
bichromate and lead chromate, to form on iron and steel surfaces a
thin coating of oxide which so effectually protects the metallic
surfaces from corrosion that after the removal of the paint the
metal still resists atmospheric effects for a long time, as well as
the stronger effect of immersion in sea-water or acidulated waters
and sulphurous and other vapors. This action, Wood adds, is
very obscure and not thoroughly understood; but the fact remains,
and extended experiments in this field only demonstrate its pres-
ence and usefulness.

Solution System with Two Contending Forces, Inhibitive and
Stimulative. — It has been already pointed out in a previous
chapter that in many cases a stimulative and an inhibitive tend-
ency may be at work at one and the same time. This assertion
is well brought out by the following experiment in which an
inhibitor and stimulator are literally "pitted' 7 against one an-
other. Samples of bright steel wire were immersed in 100 cubic
centimeters of a very dilute one-thousandth normal solution of
potassium bichromate in a series of shallow dishes. The wire-
1 Am. Soc. Mech. Ens- Trans., 1895, 16, 671.

114 CORROSION AND PRESERVATION OF IRON AND STEEL

test pieces were suspended in the solution so that they did not
come in direct contact with the glass surfaces of the dishes. This
precaution should never be omitted in experiments of this kind,
as owing to the absorption of air by glass, rusting is always stimu-
lated at the point of contact between glass and iron. The first
dish was left as a blank, the second received one drop equal to
sV cubic centimeter of a dilute tenth-normal copper sulphate
solution. The third dish received two drops of the solution,
and so on, each dish getting an increased amount of copper sul-
phate until twenty-five dishes had been prepared.

Now it is apparent that we have in this system two contending
forces at work. Iron has a higher solution tension than copper,
and therefore tends to pass into solution, the copper tending to
plate out on the iron. Chromate ions, on the other hand, put
the surface of iron in a condition in which it cannot pass into
solution. In the solution system iron-chromate-copper we have
an equilibrium to be decided between two contending forces
acting in opposite directions. It was interesting and instructive
to note the results of this struggle, which was known to be going
on, although the actual conflict could not be watched. In the
first dish, in which no copper was present, no corrosion took place;
in the second, also, no action was visible; in the third, however,
minute specks of iron rust appeared. These were larger and
more frequent in the immediately succeeding dishes, the test-
pieces showing rust tuberculation with the well-known pitting
effect. As the middle of the series of dishes was approached,
both iron rust and precipitated copper began to appear side by
side on the surface of the iron, and from thereon in the series
more and more copper separated, while less and less rust formed,
until in the end dishes copper and iron were changing places
evenly over the surface without apparent hindrance. These
experiments, and others of a similar nature, were repeated many
times with the same results, and there seems to be no escape from
at least the following two conclusions to which they obviously
lead:

(1) If the surface of iron is subject to the action of two con-
tending influences, one tending to stimulate corrosion and the
other to inhibit it, the result will be a breaking down of the defen-
sive action of the inhibitor at the weakest points, thus localizing
the action and leading to pitting effects.

INHIBITION AND STIMULATION OF CORROSION 115

(2) While the concentration of an inhibitor may be strong
enough to prevent the electrolytic exchange between atom and
ion, it must be still stronger to prevent entirely the solution of
iron and the subsequent oxidation which leads to the formation
of rust-spots.

Practical Bearing of Alkaline Solutions on Corrosion. — In
the opinion of the authors these observations have a direct bear-
ing on, if they do not actually explain, many of the practical
problems of corrosion. It is well known that if sufficiently alka-
line, solutions inhibit and in some cases actually prohibit corro-
sion. If the concentration of the hydroxyl ions is not sufficiently
high, however, local electrolysis goes on in the slightly alkaline
medium and the tendency to pit is actually stimulated, as was
noted by Cribb 1 in an extended investigation. Pennock 2 has
shown that concentrated ammonia solutions not only do not rust
clean iron but prevent its rusting in the presence of corrosive
agents, and yet dilute ammoniacal liquors appear to stimulate
corrosion. It is a curious fact, however, that iron is to some
extent soluble in strong boiling solution of ammonia, although
there is no apparent action in -the cold. This phenomenon is
now being investigated.

One of the authors has suggested that pipe lines trenched
with lime or limed soil should be to some extent protected,
but it is still a question whether or not the tendency to pit
would not be stimulated. Knudson 3 has called attention to
an interesting case of the stimulated corrosion of the bottom
of oil tanks. In the process of purifying oils it is customary
to treat them with sulphuric acid and after washing to neutralize
the residual acid with a slight excess of caustic soda. The
slightly alkaline liquors containing sodic sulphate which remain
and settle to the bottom naturally stimulate corrosion to a very
high degree. The natural cure for this difficulty which suggests
itself would be in increasing the alkalinity of the oil lees until
the prohibition point was reached. It might also be possible
to use an excess of potassium bichromate for this purpose
provided it was found that this oxidizing agent did not react
with the oil.

1 Analyst, 30, 225, also Engineering, 81, 32.

' l Jour. Am. Chem. Soc, 24, 377.

3 Trans. Am. Electro-chem. Soc, XIV, 189 (1908).

116 CORROSION AND PRESERVATION OF IRON AND STEEL

Stimulative Action of Certain Paint Films Acting as Depolar-
izers. — Walker 1 has shown that the effort to protect the interior
of tinned cans by the application of a lacquer made from a high-
grade copal linseed oil varnish only led to stimulated corrosion
in the form of pit holes. This unexpected result Walker was
enabled to explain satisfactorily by means of the electrolytic
theory. It appears that the varnish acted as a depolarizer by
absorption of hydrogen, and therefore at every scratch or imper-
fection in the lacquer a stimulated pitting effect was operative.
Walker, at the suggestion of one of the authors, has also applied
this same reasoning to the general subject of protective paint
coatings, with interesting results. Walker's method of investi-
gation and conclusions are quoted in the following paragraphs:

"To show that the corrosion of the iron occasioned by the
lacquer film is accompanied by the flow of an electric current
through the external circuit, the following experiment was tried:
A U-tube with a KC1 solution and an agar plug, carefully boiled
out, the whole under an atmosphere of hydrogen, contained two
iron electrodes, the one bare and the other covered with a linoxy-
lin film. An electrometer was inserted into the external circuit
and the current flow measured as shown in the table below. It
was found undesirable to use a silver voltameter for the exces-
sively small currents employed and resort was had to a modifica-
tion of Ostwald's Bromide Voltameter. This voltameter was
used for all measurements made, after the first, which was done
with the silver nitrate instrument.

"Deposit obtained from current due to depolarization by two
different linoxylin films:

Time in
Hours

Deposit of
Mgs. Bromine

Deposit Mgs.
per Hour

Film No. 1

Film No. 2

48
76

185

0.6
1.0

1.8
4.1

S.7

0.012
0.013

0.044
0.046
0.047

"These figures prove beyond doubt that the corrosion of the
iron on the unlacquered side is due to the flow of an electric
current through the external circuit from the lacquered surface.
1 Jour. Ind. and Eng. Chem., I, II, 754.

INHIBITION AND STIMULATION OF CORROSION 117

"These experimental facts seem to allow of but one explana-
tion, namely, that the films in question are porous in their nature,
allowing the electrolyte to penetrate them to the surface of the
metal beneath, that some of these films, due to their unsaturated
state, are capable of absorbing nascent hydrogen and in this way
acting as depolarizers, and that aside from this, all the porous
films allow the penetration through them to the surface beneath
of any depolarizer that may exist in the solution, thus rendering
the coated surface cathodic and concentrating the solvent action
at the exposed part of the metal.

"The reason for the failure of the fruit cans mentioned at the
beginning of this article is now easily understood. While the
lacquer film applied to the uncut sheet was probably a perfect
one, still, in the process of making the can, this film was ruptured
at many points. Thus the die stamping out the head of the can
broke the film at the place where the tin was bent; the mandrel,
on which the body of the can was formed, scratched the sides in
long parallel lines upon the removal of the can, and the burning
of the joints destroyed the lacquer in their immediate neighbor-
hood. When the fruit was introduced into the can, the depolar-
izing action of the lacquer itself, coupled with that of the small
amount of air left in the can in packing, threw the protected
areas into the cathodic state, concentrating the solution of the
metal at the exposed points, dissolving in this way both tin and
iron, and maintaining this corrosion until both the air was con-
sumed and the unsaturated state of the lacquer was completely
relieved. Before this point was reached, however, the corrosion
had gone far enough to seriously damage the fruit and even in
many cases to puncture the can. This action is entirely independ-
ent of any possible imperfections in the tin plate. The remedy
would be to find a lacquer impervious to the solution, or if that
prove impracticable to at least furnish one which will not act
as a hydrogen depolarizer. A non-porous lacquer it has as yet
proven impossible to find; a saturated one can be made by suffi-
ciently baking any ordinary varnish, but there still remains work to
be done to develop this into a satisfactory solution of the problem.
"It is self-evident, however, that the importance of these
phenomena is by no means limited to the problem of lacquering
fruit cans. The majority of protective coatings for iron contain
linseed oil or some one of the various substances found by our

118 CORROSION AND PRESERVATION OF IRON AND STEEL

experiments to be either unsaturated, or porous, or both; and so
soon as a piece of metal painted with these substances comes in
contact with water after the abrasion of the paint film at any
point, all the conditions for corrosion as above outlined are ful-
filled, and we may be sure that corrosion at the exposed point
will be accelerated by the presence of such films in its neighbor-
hood. It is true that the porosity of these films is reduced to a
minimum by the use of the best obtainable loading materials,
such for instance as certain pigments of ordinary linseed paints,
the bituminous bodies of asphalt or coal tar paints, etc., and that
these paints offer in consequence a much greater electrical resist-
ance to the flow of the current than otherwise. One must not
forget, however, that the insertion of such a resistance to the
current flow can only reduce the rate of the reaction, and in no
way influence its tendency or driving force. The exposure of
such films to the air for a long period of time finally entirely satu-
rates them, but this again does not affect their porosity and con-
sequently does not preclude the possibility of acceleration of
corrosion at the exposed point due to the depolarizing action of
air through the film. These facts make clear the reason for the
rapid deterioration and pitting of the iron or steel surface at
points laid bare by the breaking down of many paint films, and
show why it is so exceedingly important that a metal surface
should be clean and bright before the application of a paint.

"If a paint or lacquer film be intact, despite the fact of its
porosity, corrosion does not seem to take place at once beneath
its surface. Thus an iron can, carefully painted with a high-
grade varnish and filled with cold 5 per cent, sulphuric acid,
showed a test for iron with ferricyanide first after twenty-four
hours. If heated on the water bath, however, a test could be
obtained in slightly over one hour, but even then the action was
not severe. It may perhaps be that the electrolyte does not
find sufficient continuous surface beneath the film to allow of a
ready separation into cathodic and anodic areas. At any rate,
corrosion does not readily take place if the film be intact, but the
surface below the film does easily become cathodic if an exposed
area in the neighborhood can act as an anode. This cathodic
liberation of hydrogen loosens the film from the metal in some
way not yet clearly understood and likewise softens it. The
resistance of a film diminishes quite rapidly in this way, and it

INHIBITION AND STIMULATION OF CORROSION 119

soon becomes weak and rotten and easily removable from the iron.
It is easy to see how, especially with rough usage, exposed points
in a painted steel surface rapidly grow in size.

"A few of the common commercial paints and paint-making
materials were examined, using the following apparatus: A glass
U-tube of one inch tubing with six inch arms and eight inches in
length over all, contains 200 c.c. of normal KC1. The electrodes
are of commercial soft iron wire, 0.044' diameter, carefully
cleaned with emery. The bare electrode is 25 cm. long and
the lacquered one 100 cm. The electrode is coated by dipping
in the paint to be examined, the excess removed by rapid twirling
and then dried. The water about the cathode is kept saturated
with oxygen by bubbling air through it. The depolarization
current is measured by the use of the bromide voltameter
already mentioned. The bromide deposited is proportional to
the time, up to a point when the film gives way. This point of
disintegration is different for different films and two quantities
can in general be measured, (1) the initial rate of depolarization
and (2) the time of rupture.

Paint film Corrosion in mgs. Bromide per hour.

''Durable metal coating" 0.70

"Copal linseed oil lacquer" 0.76

"Cosmos" (coal tar) 0.34

Graphite 1.87

Carbon black 2.4

Lampblack 2.3

. Zinc chromate 0.11 x

Barytes 1.2

Zinc oxide 0.078 2

Graphite (baked) 0.078

White lead 0.10 3

Linoxylin 0.31

Paraffine 0.00

"This second factor is somewhat difficult to obtain because
our voltameter measures not current but total amount of electric-
ity passed, and it is frequently, if not usually, impossible to tell
at exactly what point the increase in current began. In the above
table is the average of the data obtained. It is evident that
the electricity measured by the bromine deposited is the resultant

1 Broke down in 36 hours, but not badly.

2 Still perfect after 71 hours.

3 Broke down in 30 hours.

120 CORROSION AND PRESERVATION OF IRON AND STEEL

of a number of factors acting at the same time. A comprehensive
study of the relationship between such data and the value of
various paints as a protective coating for iron is now in progress."

It is interesting to note at this place, that this work furnishes
a striking corroboration of the stimulative and inhibitive effects
of certain pigments which had already been indicated by one of
the authors, in a series of preliminary tests carried out on the
inhibitive power of certain pigments. This work will be presented
in detail in a later chapter.

Wounds on Steel Surface Stimulate Rusting. — A considerable
body of evidence has been brought forward from time to time to
show that in addition to the segregation of impurities in steel,
the presence of scratches, sand pitholes, and, in fact, all indenta-
tions or wounds on the surface of steel, will stimulate rusting by
becoming centers of corrosion. 1 Such marks or indentations are
almost invariably electro-positive to surrounding areas, and the
depolarization which results in the rapid disengagement of hydro-
gen at these spots leads to stimulated pitting. This effect can
be very prettily shown by means of the ferroxyl indicator. In
the illustration shown in Fig. 31 a freshly polished steel plate has
had the word £i Rust ,J carved upon its surface with a cutting tool.
On immersion in the ferroxyl indicator the general surface has
come out in red while the carved letters appeared in blue. This
cause of stimulated pitting is probably very generally active on
all surfaces of iron and steel which from the nature of their ser-
vice cannot be treated with any form of protective coating.
Boiler tubes furnish the best example of stimulated corrosion
effects from this cause. The remedy should lie in truing up of all
active contact surfaces even to the point of polishing if necessary.
There are times when a sound boiler tube is of vastly more
importance than a polished gun barrel, and when the added cost
of preparing the boiler tubes would be insignificant in comparison
with the danger of blow-outs from pitting. The above fact
brought out by experimental investigation is also attested by the
results of practical observation and experience. It has long been
known in the Bureau of Steam Engineering of the U. S. Navy
Department 2 that indentations such as sandpits, or injuries on
the water surfaces of boilers, always become centers of corrosion

Bushman, Proc. Am. Soc. Testing Materials, VIII, 605 (1908).
2 Proc. Am. Soc. Testing Materials, VIII, 244 (1908).

INHIBITION AND STIMULATION OF CORROSION 121

and pitting, and inspection has been as thorough as possible to
guard against this danger.

Mill-scale a Stimulator. — It has already been shown in a pre-
vious chapter that mill-scale consisting of the black magnetic
oxide of iron is electro-negative to the metal, and this acts inevi-
tably as a common cause of stimulation of corrosion, whenever the
coating of mill-scale is discontinuous on the surface. This point
is now so well and generally understood that pickling and sand-
blasting are often resorted to in order to obtain clean surfaces of
metal, before protective coatings are applied.

Fig. 31. — Showing the effect produced in the ferroxyl indicator by
cutting the surface of a polished steel plate. (Cushman.)

Various Inhibitive Expedients which have been Tried. — Almost
all efforts to inhibit the corrosion of iron depend either upon
producing and maintaining a passive condition of the surface or
on maintaining a sufficiently alkaline medium in contact with the
surface of the metal. Chromic acid and its more or less soluble
salt furnish the principal chemical means of producing passivity.
It has long been known that, if iron is made the anode in an elec-
trolytic circuit, it will not rust, provided the quantity of current

122 CORROSION AND PRESERVATION OF IRON AND STEEL

passing is sufficient to protect it. Various expedients based on
this fact have been suggested or tried in order to inhibit corro-
sion. Among these should be mentioned the insertion of zincs
in boilers, and the use of zinc-corroding plates in connection with
steel structures. Zinc, as has been shown, is electro-positive to
iron, and when the two metals are in contact and wet with a
corroding medium zinc will pass into solution and the iron will
be protected. The sphere of influence of the zinc does not, how-
ver, extend far from the point of contact between the two metals

i< i [

HP^'

' -jffE*rv^

Wm$?y : -

^h^*

'0§i

*& .... .

ill

■'.■:. . ■ . ■

Fig. 32. — Illustrating the protection on the surface of the steel in the
immediate neighborhood of a button of zinc soldered into the steel. The
sphere of influence of the zinc is shown to be proscribed. (Cushman.)

so that iron can begin to rust in places long before all the zinc
is destroyed. This can easily be shown by soldering a button
of zinc into the machined surface of a plate of steel and immersing
the whole either in water or in the ferroxyl indicator. This is
very well shown in Fig. 32. In general it may be said that even
if the sphere of influence of the zinc was wider, the idea of burning
up zinc in order to protect steel does not commend itself.

Passivity Maintained by Plunging Electrodes into Neutral or
Alkaline Solution. — If two iron electrodes are plunged into a

INHIBITION AND STIMULATION OF CORROSION 123

neutral or alkaline solution and an electric current of sufficient
density be passed through the circuit the pole where the oxygen
disengages becomes passive, and while it is maintained in this
condition rusting is inhibited or actually prohibited. Just as
in the case of passivity induced by solutions of chromic acid and
its salts, this passive condition is maintained to some extent after
the removal of the cathode plate from the circuit. This action
has been studied by a number of experimenters including the
authors. Toch 1 has made a very recent contribution to the sub-
ject, the details of which are not yet available.

Rusting Caused by Escaped Currents. — It is very generally
known that extraneous escaped currents from high-potential
light and power circuits do great damage to structural iron and
water mains, especially in the larger cities. There is a vast
literature on this subject, much of which is cited in Appendix B.
No effort will be made here to abstract the available information,
as this phase of the subject constitutes a special problem on which
a separate volume might be written. In a general wa} r it may be
stated that the best preventive measures against this cause of
rust stimulation are more judicious bonding of the rails carrying
return currents and the use of effective feeders on the electric
circuits, the maintenance of a negative condition of the struc-
tural material or pipe lines, or their more thorough isolation and
insulation. For more detailed information on this phase of the
subject the original literature should be referred to.

Factors Causing Stimulation or Inhibition . — The effort has
been made in this brief chapter to set forth the principal factors
which enter into the stimulation or inhibition of corrosion. Some
of these factors, such as contact action in the air and in water,
had already been treated in previous chapters, but may be con-
veniently summarized here in the following list :

Factors which Stimulate Corrosion

(1) Impure and segregated metal. ' Unhomogeneous or burnt
metal which may contain blow-holes.

(2) Cold rolled or improperly annealed metal which may
maintain an uneven, stressed, or strained condition.

(3) Contact action, in which different types of iron and steel
are used in one and the same structure.

1 Proc. Am. Electro-chem. Soc. (1909).

124 CORROSION AND PRESERVATION OF IRON AND STEEL

(4) The presence of hydrogen ions from any source whatsoever
that may be brought in contact with the surface, in water or
other electrolytes in the presence of oxygen.

(5) The concentration of active oxygen that is present in the
wetting medium.

(6) The presence of electrolytes generally in the wetting
medium. Even hydroxyl ions and other inhibitors may appear
as stimulators if in insufficient concentration.

(7) The stimulating or depolarizing effects of certain coatings
applied to iron and steel with the object of protecting the metal.

(8) The effect of indentations, scratches or other injuries
which become centers of corrosion.

(9) The effect of extraneous or stray currents escaped from
high-potential circuits.

Factors which Inhibit Corrosion

(1) In most cases the reverse of the conditions which stimu-
late corrosion.

(2) Contact with certain substances in solution, such as
chromic acid and its soluble salts which produce a passive con-
dition.

(3) Alkaline solutions of all kinds in which the concentration
of hydroxyl ions is sufficiently high. But this protection may be
overcome and the equilibrium conditions reversed in very strong
boiling alkaline solutions.

(4) Contact with more electro-positive metals.

(5) Electrolysis under certain specific conditions.

CHAPTER VI

THE TECHNICAL PROTECTION OF IRON AND

STEEL

The Different Phases of the Protection Problem. — No purely
scientific discussion of the corrosion problem will be considered
complete unless it also describes the more recent efforts to attain
efficient protection of the numerous forms of iron and steel which
enter into common use. This phase of the problem may be
conveniently considered under three main heads: (1) Protective
coatings of other metals and alloys. (2) Magnetic oxide sur-
faces. (3) Paint coatings, including linseed oil paints, varnishes,
lacquers, bitumens, and cements.

The protection of iron is presumably a problem which in the
ease of its solution varies directly with the rust-resisting character
of the metal to be preserved. Progress in the manufacture of
more rust-resistant metal must keep pace with an added efficiency
in the methods of protection. That there is much still to be
accomplished along both these lines is attested, not only by the
complaints of innumerable consumers, but also by the records of
the proceedings of engineering and other learned bodies and by the
technical and scientific press.

Metallurgists are exhibiting a growing tendency to look upon
the structure of iron and steel as an aggregate in the same sense
that geologists consider crystalline rock structures to have re-
sulted from the more or less gradual cooling of molten magmas.
From this point of view, careful study and control of the heat
treatments which steel receives in the course of its manufacture
is of the utmost importance if constancy of chemical constitution
and physical character is to be developed. Upon just such
factors as these resistance to corrosion will be found dependent.
The mill man's purposes are often expedited by turning a stream
of cold water upon red-hot piles of freshly shaped forms. How
many mill men, however, have ever given consideration to the
structural changes produced in the metal by such sudden quench-

125

126 CORROSION AND PRESERVATION OF IRON AND STEEL

ing? Even subsequent annealing unless properly done will not
restore the original homogeneous and quiescent conditions in
the structure of the steel as the classic researches of Heyn and
Bauer previously quoted very clearly show. The authors would
earnestly recommend to manufacturers who desire to produce
rust-resistant iron, that every step in the heat treatment which
the metal receives from the furnace to the finished shape be
studied with reference to its possible effect upon the final product.

One way to attain homogeneity in the metal is to eliminate
all the impurities and thus produce a practically pure iron. This
has to a very considerable extent been accomplished on a com-
mercial scale within the last few years, and metal of this type is
now on the market. Every metallurgist knows that steel which
is free from manganese is difficult to roll, and that special care
is necessary in order to shape it successfully. Even if the
influence of manganese as one of the many factors which control
resistance to corrosion is denied, it is none the less probable that
a metal which from its very nature must receive careful and even
heat treatment, will be more resistant than one which will permit
itself to be carelessly and rapidly pushed through the mill.

Until the last few years, the methods of protection have been
carried out mainly on an empirical basis. When iron had been
coated with any substance which was believed to furnish a fairly
good and durable waterproof surface it was considered that
nothing more could be done. Of course, serious efforts were made
to improve the physical character of the coatings as to their
water or acid resisting qualities, but little or no attention was
paid to the possible electrolytic effects produced, for indeed the
fundamental causes of corrosion were not understood. For this
reason many of the so-called protective coatings were actually
stimulating the very kind of corrosion that they were designed to
prevent.

Protection with Zinc. — In any discussions of the use of other
metals for the protection of iron, zinc must take the front rank.
Zinc is the most electro-positive metal which can be practically
used for coating iron and is from this point of view better suited
for the purpose of providing an inhibitive coating than any other.
On the other hand the solution tension of zinc is high 1 and its
power of protecting iron is accomplished mainly at the expense
1 Davies, Jour. Soc. Chem. Ind., 18, 102.

TECHNICAL PROTECTION OF IRON AND STEEL 127

of its own destruction. For this reason zinc must be considered
at the very best as an inhibitor rather than as a prohibitor of
corrosion, and the day must inevitably arrive when even the best
of zinc coatings will fail. There is reason to believe, however,
that the last word on the protection by zinc has not been spoken
and that the methods of galvanizing may be so improved as to
lengthen the life of the coatings by many years. It is probable
that more than half of all the zinc produced is used for the pur-
pose of protecting iron and steel. Burgess 1 has well said that
the efficiency of the zinc coating varies greatly with the purity
of the metal, its thickness, continuity, and methods of application,
though the degree to which these various factors affect the effi-
.ciency seems to have received little attention, if one may judge by
published data. As a matter of fact the electrolytic theory of
corrosion has opened up a number of questions which bear directly
upon the efficiency of zinc coatings, and the answers to which,
when they are available, will constitute new and valuable data
on the subject. These points will be brought out in the following
discussion.

Methods of Applying Zinc Coatings, — There are only three
processes in general use for coating iron with zinc," although the
methods of application under the three processes may vary to a
considerable extent in practice. These methods are well known
as the hot dip, the "cold" or electrolytic, and the vapor or Sherard-
izing processes. It is not necessary here to give a detailed account
of the technology of these several methods, as it is assumed that
the reader is either already familiar with the main and well-
known details of the technology of zinc coating or can easily
become so by referring to any standard work on the metallurgy
of iron and steel. The principal effort that will be made here
will be to apply the most recent theories and results of investi-
gation to the problem of securing greater longevity for the pro-
tected iron. An examination of the literature of the subject as
given in Appendix B will reveal the fact that many claims and
counter claims have been put forward and vigorously defended
in the effort to show that one or other of the processes is the more
efficient or superior. Thus one author 2 states: "The zinc applied
by the hot method usually contains lead, tin and iron. It has

1 Electro-chem. and Met. Ind., Ill, 1, 17 (1905).

2 Cowper-Coles, quoted by Burgess, loc. cit.

128 CORROSION AND PRESERVATION OF IRON AND STEEL

been found that iron above 3 per cent, makes the zinc too brittle
to bend. Lead up to 1 per cent, is harmless, but above 1.5 per
cent, will not dissolve and the excess collects (segregates) and
forms weak spots. On the other hand zinc applied by the elec-
tric process is very pure and it is found to resist the corroding
action of a solution of copper sulphate very much better than hot
galvanized iron."

This statement indeed brings out some of the disadvantages
of the hot-dip process, although by no means all of them; on the
other hand, it does not refer to those which argue against the
cold process. As a matter of fact it is probably true of the three
methods of applying zinc that each will find its special field of
usefulness, and that each will be found to have certain advan-
tages and disadvantages. Further than this, it is quite certain
that all the processes are at the present time more or less imper-
fect and are therefore subject to improvement as time goes on.
These processes will now be criticised from the standpoint of the
electrolytic theory as it applies to the corrosion and preservation
of iron and steel.

The Hot-dip Method of Galvanizing. — In the hot-dip method
the iron to be galvanized is dipped or drawn through a molten
bath of spelter, which necessarily becomes more and more alloyed
with iron up to the point where a definite eutectic alloy of iron
and zinc crystallizes out, and being heavier than the molten
magma falls to the bottom of the bath. This separated alloy
is technically known as dross and is occasionally raked out of the
bath and thus partially removed. The layer of partly oxidized
zinc and impurities which forms on top of the bath is scum, and
this also is removed from time to time. It will be seen that even
if the original spelter was pure zinc, it would not remain so in the
bath, as it would soon become contaminated by the iron which is
immersed in it as well as by other impurities which might acci-
dentally find their way in. As a matter of fact, however, the
original spelter contains a certain amount of impurity in the form
of lead, cadmium, iron, and other materials, and is therefore really
an alloy to begin with. Burgess 1 truly remarks on this subject:
"Such an alloy will dissolve very readily in an acid solution,
owing to local action, which is set up between the different metals.
This results in a more rapid corrosion than where a single metal

1 hoc. tit.

TECHNICAL PROTECTION OF IRON AND STEEL 129

is immersed in a corroding solution." In other words, all that
has been said in previous chapters in regard to the accelerated
corrosion of impure and segregated iron applies equally to zinc.
Not only is the solution tension of zinc alloyed with iron and other
impurities greater than that of pure zinc, but even its electro-
chemical relation to iron is said to change so that it may appear
electro-negative instead of electro-positive. These facts will
always operate against the efficiency of zinc coatings applied
by the hot-dip process, but nevertheless, with proper care, im-
provement in the results obtained should still be attainable.
It is common practice in the hot-dip process to pass the iron or
steel to be coated through a hydrochloric acid fluxing bath just
before it enters the spelter bath. In many cases zinc chloride or
ammonium chloride, "sal ammoniac," is used in conjunction
with hydrochloric acid. Certainly no more corrosive mixture
than this could be applied to the surface of iron, and it is equally
certain that small portions of these acid products become included
in the zinc coatings. One of the authors has dissolved a number
of samples of hot-dip galvanized coatings in pure dilute nitric
acid and never failed to get slight white precipitates with silver
nitrate, proving the presence of chlorine ions. In some of the
foreign practice much more attention is given to the control of
the acidity of the fluxing bath than is usual in this country, and
this fact undoubtedly is contributory to higher rust-resistant
qualities. For this reason the authors have suggested the sub-
stitution of alkaline fluxing and soldering solutions. Adopting
the suggestion, experiments have been carried out in a prominent
wire mill, on the use of a flux made by dissolving zinc oxide in a
fairly strong solution of caustic soda. No trouble was experi-
enced in getting heavy coatings of zinc on the wire with this flux,
but occasional black spots appeared which interfered with the
general efficiency of the coating and which have not as yet been
explained. It is therefore not now certain whether a practical
substitute for the acid flux can be found.

All galvanized wire manufacturers have been to some extent
troubled by the appearance on their finished products of a whitish
efflorescence which often appears in little spots on the surface
and which is technically spoken of as "mold." This appearance
is undoubtedly due to zinc oxide which, formed as the result of
electrolysis, marks a preliminary step to the final failure of the

130 CORROSION AND PRESERVATION OF IRON AND STEEL

coating. The fact that chlorides included between the iron and the
zinc will stimulate such an effect even if it is not the sole cause
cannot be contested. The system Fe — FeCb — ZnCl 2 — Zn is
a galvanic couple which, however well it may justify the name of
"galvanized" iron, would certainly tend to hasten corrosion. If
a practical galvanizing pan could be devised in which the zinc
was melted and maintained in a molten condition out of contact
with the air or in an atmosphere of inert gas, better results would
be obtained. The excessive amount of scum which forms on top
of the spelter bath dirties the coating and prevents the forma-
tion of an even and homogeneous layer of zinc.

The Weight of the Zinc Coating. — In addition to the purity,
density, and homogeneity of the zinc coating, insufficient thickness
is also a contributory cause of rapid corrosion, especially in the
case of galvanized wire. Thin coatings of zinc are to some extent
porous, a condition which leads naturally to electrolytic destruc-
tion of the zinc. By increasing the thickness of the coating the
pores are gradually filled up so that less opportunity exists for
centers of corrosion to form. The method which is in most
general use for determining the thickness of a zinc coating is the
well-known copper sulphate test. The test depends upon the
fact that if zinc is dipped into a strong solution of a copper salt
it goes into solution and copper comes out. As soon as the zinc
is removed from the iron the copper begins to plate out slowly
on the iron and can be easily seen. The difference of solution
pressure is much greater as between zinc and copper than as be-
tween iron and copper. The consequence of this is that copper
replaces the zinc so rapidly that it makes a crumbly deposit
easily wiped off. The copper iron exchange on the other hand
takes place so slowly that the copper plates on the surface of the
iron so that it cannot be wiped. The appearance of copper
plate is therefore an indication that the zinc is gone.

The Preece or Copper Sulphate Dip Test, — This test which
is known as the Preece Test is only an approximate method for
determining the thickness of a zinc coating and cannot be inter-
preted as a measure of the life of the coating or its resistance to
corrosion under the conditions of service. A difficulty in carrying
out the test has been found because there is no sharp point in the
reaction which indicates the exact point when all the zinc is gone.

This point is denned by the Chief Engineer of the Western

TECHNICAL PROTECTION OF IRON AND STEEL 131

Electric Company as follows: "The standard color of a bright
metallic copper deposit on a sample will be denned as the one
obtained by taking a piece of the same class of sample and im-
mersing in .hydrochloric acid until the action ceases, after which
the sample is immediately washed and wiped dry, and immersed
in the standard copper sulphate solution for not over a minute,
then removed, washed, and wiped dry. The copper deposit
thus obtained will be the bright metallic copper deposit referred
to in the specification. These standards should be made every
test day. The specific gravity of the standard solution will be
determined by a standard 6-inch hydrometer with only one
mark on the scale which indicates 1.186 specific gravity at 65° F.
In reading these hydrometers the following must be observed:
(a) Clean and dry hydrometer before placing in standard solu-
tion, (b) Depress the hydrometer so that the solution wets the
tube at least \ in. above the 1.186 mark, (c) Sight just under
level of solution to see if the 1.186 mark is on the solution level
outside of the capillary."

The American Steel and Wire Company have adopted the
following instructions for carrying out the copper sulphate test,
which has been found useful in maintaining a standard weight of
zinc coating on wire :

Standard Solution. — The standard solution of copper sul-
phate shall consist of commercial copper sulphate crystals dis-
solved in cold water about in the proportion of 36 parts by weight
of crystals to 100 parts by weight of water. The solution shall
be neutralized by addition of an excess of chemically pure cupric
oxide (CuO). The presence of an excess of cupric oxide will be
shown by a sediment of this reagent at the bottom of the contain-
ing vessel. The neutralized solution should be filtered before using.
This solution shall have a specific gravity of 1.186 at 65° F.

Small differences in the density of the solution may be cor-
rected by adding water or copper sulphate crystals according as
the solution is too heavy or too light, but if the solution is not
nearly correct or has become dirty or impaired for any reason it
must be thrown out and a fresh supply obtained.

Cleaning of Samples. — Samples must be cleansed thoroughly
of oil and dirt by dipping in benzine or gasoline, then thoroughly
rinsing in clean water and wiping dry with clean white cotton
waste before making the test.

132 CORROSION AND PRESERVATION OF IRON AND STEEL

Apparatus. — The apparatus required for the test is as follows:
1 Fahrenheit thermometer with large scale to read up to
at least 80° F.

1 Specific gravity hydrometer.

1 Hydrometer cylinder 3 in. by 15 in.

2 Jars for washing samples.

4 Glass test jars 2 in. in diameter and 5 in. in height.

1 Copper or galvanized sheet steel box equipped with run-
ning water and waste pipe connection and with a rack for holding
thermometer and a tray over part of the box through which test
jars may be immersed in the water to maintain constant tempera-
ture. White cotton waste free from grease. Benzine or gasoline.

The place for testing should be clean and with good natural
light. A cupboard should be provided for the apparatus in which
it can be placed when not in use and there should be both hot and
cold running water connections. For tests in summer time ice
may sometimes be required, but this should seldom be necessary
as the cold water supply is nearly always lower than the testing
temperature.

Test. — Fill one of the glass test jars with standard solution
to a mark one inch from the top. The temperature of the solu-
tion must not be lower than 65° nor higher than 70° F. at any
time during the test, and the water used for washing samples
must be the same.

Not more than seven wires shall be simultaneously immersed
in one jar, and they must be grouped together but must be well
separated so as to permit the action of the solution to be uniform
on all immersed portions of the samples. The ends outside of
the solution must not be grouped together. The ends inside jar
are likely to touch, but this is not objected to, as indications on
the lower inch of the samples are disregarded, as stated below.
After each complete test of seven wires or less the solution must
be thrown away and fresh solution taken for the next set.

The specified dip must be made on each sample and the periods
of time accurately observed. After each dip the samples must be
immediately rinsed in water having a temperature within the speci-
fied limits of the solution temperature, thoroughly cleaned with
soft cotton waste (not with a brush) and wiped dry. Samples
must be dry at the time of immersion. For an extended series of
tests the wash water should be frequently renewed, for obvious

TECHNICAL PROTECTION OF IRON AND STEEL 133

reasons, and in order to maintain the proper temperature of the
wash water the jars should be placed in the same tray as the test
jars.

Specified Dips. — The standard method of making a test is as
follows, using a "two-minute" immersion test for the purpose of
illustration :

The cleansed, washed and dried samples are immersed in fresh
standard solution, within the temperature limits of 65° to 70° F.,
for exactly one minute. They are then removed, rinsed in water,
of the proper temperature, and wiped with cotton waste so as to
remove the dark deposit which forms, and until they are dry.
They are then immersed again in the same solution for exactly
one minute, removed, rinsed and wiped as above. Samples so
treated should show no trace of metallic copper on the steel more
than one inch from the end, although they may be black, indicat-
ing nearly complete removal of zinc. If they do show copper on
the steel they will be considered to have failed on the test and will
be so marked in the record book. Occasionally copper deposits
on the zinc, without removing the latter, and can be scratched
off without destroying the zinc coating. Such cases, after the
second immersion, are not counted as failures.

The engineering department of the American Telephone and
Telegraph Company have adopted the following test specifica-
tions for galvanized iron and steel wire:

These specifications give in detail the test to be applied to
galvanized material. All specimens shall be capable of with-
standing these tests.

(a) Coating. — The galvanizing shall consist of a continuous
coating of pure zinc of uniform thickness, and so applied that it
adheres firmly to the surface of the iron and steel. The finished
product shall be smooth.

(b) Cleaning. — The samples shall be cleaned before testing,
first with carbona, benzine or turpentine, and cotton waste (not
with a brush), and then thoroughly rinsed in clean water and wiped
dry with clean cotton waste.

The samples shall be clean and dry before each immersion in
the solution.

(c) Solution. — The standard solution of copper sulphate shall
consist of commercial copper sulphate crystals dissolved in cold
water, about in the proportion of 36 parts, by weight, of crystals

134 CORROSION AND PRESERVATION OF IRON AND STEEL

to 100 parts, by weight, of water. The solution shall be neutral-
ized by the addition of an excess of chemically pure cupric oxide
(CuO). The presence of an excess of cupric oxide will be shown
by the sediment of this reagent at the bottom of the containing
vessel.

The neutralized solution shall be filtered before using by
passing through filter paper. The filtered solution shall have a
specific gravity of 1.186 at 65° F. (reading the scale at the level
of the solution) at the beginning of each test. In case the fil-
tered solution is high in specific gravity, clean water shall be added
to reduce the specific gravity to 1.186 at 65° F. In case the
filtered solution is low in specific gravity, filtered solution of a
higher specific gravity shall be added to make the specific gravity
1.186 at 65° F.

As soon as the stronger solution is taken from the vessel
containing the unfiltered neutralized stock solution, additional
crystals and water must be added to the stock solution. An
excess of cupric oxide shall always be kept in the unfiltered stock
solution.

(d) Quantity of Solution. — Wire samples shall be tested in a
glass jar of at least two (2) inches inside diameter. The jar with-
out the wire samples shall be filled with standard solution to a
depth of at least four (4) inches. Hardware samples shall be
tested in a glass or earthenware jar containing at least one-half (i)
pint of standard solution for each hardware sample.

Solution shall not be used for more than one series of four
immersions.

(e) Samples. — Not more than seven wires shall be simultane-
ously immersed, and not more than one sample of galvanized
material other than wire shall be immersed, in the specified quan-
tity of solution.

The samples shall not be grouped or twisted together, but
shall be well separated so as to permit the action of the solution
to be uniform upon all immersed portions of the samples.

(/) Test. — Clean and dry samples shall be immersed in the
required quantity of standard solution in accordance with the
following cycle of immersions.

The temperature of the solution shall be maintained between
62 and 68° F. at all times during the following test:

First : Immerse for one minute, wash and wipe dry.

TECHNICAL PROTECTION OF IRON AND STEEL 135

Second : Immerse for one minute, wash and wipe dry.

Third : Immerse for one minute, wash and wipe dry.

Fourth : Immerse for one minute, wash and wipe dry.

After each immersion the samples shall be immediately washed
in clean water having a temperature between 62 and 68° F. and
wiped dry with cotton waste.

In the case of No. 14 galvanized iron or steel wire, the time
of the fourth immersion shall be reduced to one-half minute.

(g) Rejection. — If after the test described in section "F"
there should be a bright metallic copper deposit upon the samples,
the lot represented by the samples shall be rejected.

Copper deposits on zinc or within one inch of the cut end shall
not be considered causes for rejection.

In case of a failure of only one wire in a group of seven wires
immersed together, or if there is a reasonable doubt as to the
copper deposit, two check tests shall be made on these seven
wires and the lot reported in accordance with the majority of
the sets of tests.

Note. — The equipment necessary for the test herein out-
lined is as follows :

Commercial copper sulphate crystals.

Chemically pure cupric oxide (CuO).

Running water.

Warm water or ice as per needs.

Carbona, benzine, or turpentine.

Glass jars at least 2 inches inside diameter by at least 4£
inches high.

Glass or earthenware jars for hardware samples.

Vessel for washing samples.

Tray for holding jars of stock solution.

Jars, bottles, and porcelain basket for stock solution.

Cotton waste.

Hydrometer cylinder 3 inches in diameter by 15 inches high.

Thermometer with large Fahrenheit scale correct at 62 and 68°.

Hydrometer correct at 1.186 at 65° F.

Filter paper.

Objections to the Standard Copper Sulphate Test — In the
authors' opinion there are grave objections to the copper sul-
phate test as carried out under either of the above specifications.
The attack on the zinc coating which is made by the copper

136 CORROSION AND PRESERVATION OF IRON AND STEEL

sulphate is most active when the wire is first immersed in the solu-
tion and slows down to almost zero toward the end of the immer-
sion. This slowing effect is of course due to the growing accumu-
lation of deposited copper on the surface of the zinc. After the
wire is withdrawn and wiped at the, end of one minute the action
starts off rapidly again on the section immersion only to again slow
down as before. The consequence of this action is that it is quite
usual to find a sample that will stand a three-minute immersion
without failure, whereas an additional one second plunge after
wiping will suffice to completely plate the surface with adherent
copper. It follows from this that the test as carried out is not giving
the copper sulphate an opportunity to exert its maximum effect
throughout the period of the test. In order to obviate this difficulty
one of the authors carries out the test under the following specifica-
tions which have been found to yield fairly concordant results: 1
The clean wire sample is immersed with a quick motion in the
copper solution, and at the end of exactly five seconds quickly
withdrawn and at once dipped into pure water and wiped. The
immersion should be timed with a stop-watch if possible. As
long as no copper is visible plated out on the iron, the five-second
immersions are repeated until the copper deposit which begins
to form wipes off with difficulty. At this point a series of rapid
two-second immersions are made, with alternate wipings, until a
bright metallic copper surface, which cannot be wiped off the wire,
appears. The total time of immersion in minutes and fractions
of a minute is taken as an indication of the percentage of zinc
carried by the wire. Thus, if 150 seconds' immersion were neces-
sary to reach the end point, the weight of the galvanizing would
be called 2.5 per cent. It must be remembered that although the
results of this test are approximate only, their value depends
entirely upon careful accuracy in following out the directions.

It will readily be seen that the authors' method of making the
test will give much lower results than those obtained under the com-
mercial specifications given above. On the other hand, the results
will be more accurate and more comparable. One great objection,
however, to the more accurate method is that so much more time
is consumed in carrying it out. In a wire mill in which samples
are being tested from every bundle of wire made, this is a serious
consideration.

1 Cushman in appendix to Farmers' Bull., 239, U. S. Dept. Agr.

TECHNICAL PROTECTION OF IRON AND STEEL 137

Walker has very recently published some interesting and
valuable data on the testing of galvanized coatings. He criticises
the copper sulphate test in the following words :

Walker's 1 Contribution to the Testing of Galvanized Coatings. —
"When the copper sulphate reaction is used in testing galvanized
iron it is assumed that the speed of solution of the zinc is a direct
function of the time of immersion, and by the number of one-
minute immersions necessary to dissolve the zinc from the iron
and to deposit a film of bright copper thereon, the thickness of
the zinc can be estimated. It will be seen that this method meas-
ures the thickness of the coating only at its thinnest point, and
that the assumption is made that no bright copper will be formed
until the iron base is reached. To appreciate the limitations of
this test, it is but necessary to note the structure of the three
classes of zinc-coated iron."

"Apolished oblique section of ordinary hot-galvanized iron etched
with iodine reveals three distinct layers. First, a coating of zinc,
which varies in thickness in accordance with the temperature of the
zinc and the amount of squeezing or wiping applied to the article
before the zinc solidifies; second, a distinct layer of a zinc-iron alloy,
termed alloy B, which varies in thickness with the temperature of
the zinc bath and length of time the iron was subjected to it; third,
the iron base itself. Between the zinc and alloy B there is generally
a second alloy, alloy A, which is a thin, discontinuous layer, richer
in zinc than alloy B, and between the iron and alloy B is a third
alloy, relatively unimportant, which is very rich in iron, alloy C.

"The structure of wet-galvanized iron is comparatively simple,
the layer being practically pure zinc."

"Sherardized iron, on the other hand, presents a relatively
complex structure. The metallic zinc penetrates the iron, form-
ing deep layers of the alloys B and C, and in place of alloy A
there occur$ a number of compounds, as yet unidentified. Upon
the surface there is generally a layer of relatively pure zinc,
although frequently the process is carried to the point where only
a deep layer of alloys is formed. When examined under the micro-
scope this alloy is seen to be covered with deep cracks or fissures
as though the alloy in forming had contracted. It is thus apparent
that in testing galvanized iron made by these different processes
we are dealing with three very different materials."
1 Electro-chem. and Met. Ind., VII. 10. 440.

138 CORROSION AND PRESERVATION OF IRON AND STEEL

Walker's Desiderata for a New Test for Determining Durability
or Resistance to Corrosion. — " It is the present purpose to treat
of zinc-coated iron with regard only to its durability or resistance
to corrosion, and not to its tensile strength. From this stand-
point a test should, if possible, indicate the following:

(1) The uniformity and thickness of the zinc coating;

(2) the continuity of the coating with reference to pin holes;

(3) the purity of the zinc, and

(4) the toughness and ductility of the coating.

The corrosion of pure zinc in water is very slow. If, however,
the zinc is in electrical contact with any material upon which the
hydrogen can be liberated, the corrosion of the zinc is relatively
rapid. The iron-zinc alloy B, and the iron itself, are both sur-
faces on which this depolarization action can take place, and,
hence, so long as neither the iron nor the alloy is exposed, other
things being equal, the zinc coating will not corrode. The impor-
tance, therefore, of maintaining a uniform coating of metallic zinc
upon the iron can be appreciated.

To determine whether the ratio of the time of immersion
required to produce a bright copper surface was in reality a
measure of the thickness of the zinc coating, samples of galvanized
sheets were obtained and tested. The weight of zinc per square
inch area was determined by dissolving the zinc from the sheets
without attacking the iron-zinc alloy. This can readily be
accomplished by heating the galvanized sample, together with a
piece of metallic iron, in boiling caustic soda until the generation
of the hydrogen ceases. As might be expected, the ratio between
the time of immersion required to show the presence of bright
adherent copper and the amount of zinc present per unit area
depends upon the uniformity with which the coating of zinc is
spread upon the iron base. In the case of 14 pieces of galvanized
sheet iron, for example, the average ratio found was 23, while
the greatest variation in either direction was but 20. This test,
therefore, so far as indicating the uniformity of the coating and
measuring the thickness of the zinc are concerned, is fairly satis-
factory; but alloy B is always indicated as iron, and hence, when
the test is applied to sherardized articles, very erroneous and mis-
leading results may be obtained.

The presence of the channels or pin holes caused by the free
iron surface coming in contact with the zinc and causing it to pass

TECHNICAL PROTECTION OF IRON AND STEEL 139

into solution cannot be detected by the copper-sulphate method,
as these pin holes down through the zinc to the iron fill up with
black, spongy copper and cover up the bright copper spot at the
bottom.

The purity of the zinc cannot be determined, as the bright
copper particles precipitated by this method are so small that they
are lost in the mass of black, spongy copper.

Finally, the fourth factor, the toughness or the ductility of
the coating, cannot be determined, due to the same tendency of
the spongy copper to cover up the bright copper deposits, which
indicate cracks extending down through the zinc coating to the
alloy or iron underneath."

Walker's Test for Detecting Pin Holes and Cracks, — "The rela-
tion between the presence of pin holes in the zinc surface and
cracks due to a brittle coating and the durability of the structure
as a whole has not heretofore been studied, largely because no
method of detecting these imperfections has been available. The
following phenomenon serves as a basis for such a test. If a piece
of zinc be placed in a strong solution of caustic soda heated to about
100° C. no action is noticeable. If now the zinc be touched with
a piece of iron, hydrogen is liberated in great volume from the
iron. Small cracks in the zinc coating may be easily detected
in the same way."

"Ordinary hot-galvanized ware is generally very free from
imperfections of this kind, while wet-galvanized ware, on the
other hand, is frequently very porous, generating hydrogen at
numberless small points over its entire surface. Theoretically,
the best electroplated surface should be that carrying the greatest
weight of zinc per unit area, deposited at the slowest rate. That
is, the lower the current density at the cathode in plating, the less
porous will be the deposited metal, and the thicker this dense
deposit is the better protection will it be. A great many tests
on electroplated zinc-coated iron were made, the samples varying
in both these particulars. In every case, those deposited most
rapidly were the most porous, and the results showed that the
time of plating was more important than the weight of zinc per
unit area, although there was a minimum below which it was not
safe to go."

Walker and Campbell's Test for Determining the Combined
Effect of Imperfections.— "As the rapidity of hydrogen generat : on

140 CORROSION AND PRESERVATION OF IRON AND STEEL

from zinc is a function of the presence of an alloy of zinc and iron,
hence the more impure the zinc coating, the more rapid will be
its solution in acid. Accordingly, the impurity of the zinc, the
presence of pin holes and cracks, and the thinness of coating, are
all factors which act in an accumulative way to increase the rate
at which hydrogen is generated when the zinc surface is exposed
to acid. This method has been developed by Mr. Charles L.
Campbell in his thesis for the B. S. degree at the Massachusetts
Institute of Technology, and found to give very concordant
results. The apparatus employed may take any convenient
form, it being but necessary to expose a known area of the zinc
coating to a standard acid solution under uniform conditions and
to measure the hydrogen generated per minute. In almost every
case a sharp maximum is reached, giving the resultant of all the
different factors which make for rapid dissolution of the zinc.
Thus, in a series of wet-galvanized sheets, those electroplated for
15 minutes reached a maximum of 7 cu. cm. in 3 minutes; those
plated for 30 minutes a maximum of 14 cu. cm. in 10 minutes,
while those plated for 45 minutes showed a rather poor maximum
of 6 cu. cm. in 30 minutes. In many samples of sherardized pro-
duct there were two maxima on the curve showing the amount of
hydrogen generated per minute. This indicates that there is first
an action between the outside zinc coating and the iron-zinc alloy,
and later a well-defined action between the alloy and the iron."

"Service tests with galvanized iron are of necessity very slow
at best, and the relation of the above phenomenon to the real
durability of the material can be determined only after a number
of years. Something can, however, be learned by the study of
ware which has already been exposed for sufficient time to
determine its durability. Most of the material available for the
purpose was in the form of fence wire ; from a study of these the
following conclusions are drawn":

First, Thickness of Coating, — "In every instance where a very
durable fence was found, the coating of zinc proved to be relatively
very thick. On the other hand, the wire fences which showed
marked corrosion in from one to two years proved to have almost
no zinc on the iron, the zinc color being due to a layer of alloy
alone. It is doubtless true that the purity of the iron used in
the wire itself plays an important part in determining the ulti-
mate durability of a fence, yet in the opinion of the writer the

TECHNICAL PROTECTION OF IRON AND STEEL 141

thickness of the zinc coating calls for more immediate attention.
The modern method of close mechanical wiping the wire as it
emerges from the zinc bath produces too often a wire covered not
with zinc but only a thin layer of electro-negative iron-zinc alloy,
which affords but little protection to the iron. To produce a
wire with a liberal coating of zinc would, of course, cost more,
both on account of the extra zinc and on account of a somewhat
smaller production per machine."

Second, Purity of Zinc. — "It is probably a necessary conse-
quence of hot galvanizing that the zinc becomes somewhat con-
taminated with dissolved iron. It is apparent that this should
be kept at a minimum in order that the coating should be of maxi-
mum durability."

Third, Flexibility of the Zinc Coating. — " The important objec-
tion to placing a thick coating of zinc upon wire used for fencing
is that such a coating cracks off when running through the machine.
This objection applies only to those wires which are subjected to
very sharp bends or turns. It is possible that by passing the zinc-
coated wire through dies or under grooved rolls that the crystal-
line condition of the zinc could be destroyed and the flexibility
and ductility materially increased."

Walker sums up his results as follows:

1. "The copper sulphate immersion test is of value in deter-
mining the uniformity of coating and relatively the thickness of
coating in hot and in wet-galvanizing products; but it is worth-
less in the case of sherardized products, and gives no idea of other
important factors involved in the durability of the structure.

2. By immersing a galvanized product in hot, strong caustic
soda, the presence of any unprotected iron may easily be detected,
however small such area may be.

3. The weight of zinc per unit area apart from the zinc-iron
alloy may be analytically determined by dissolving the zinc from
the plate through treatment with hot caustic soda while in con-
tact with metallic iron.

4. Theoretically, the rate at which hydrogen is evolved when
the galvanized product is immersed in dilute acid should indicate
its relative durability; inferior products should reach a maximum
in a few minutes compared with a much longer time for better
products.

5. In the very important matter of fencing wire, while the

142 CORROSION AND PRESERVATION OF IRON AND STEEL

purity of the iron used is of consequence, a more immediate
necessity is a heavier coating of zinc on the wire.

6. The flexibility of a zinc-coated wire may possibly be
increased by mechanically working the wire in rolls or dies, to
destroy crystallization in the zinc coating."

The authors desire to record here that the conclusions reached
by Walker are in complete accord with their own opinions, as will
appear in the following paragraphs which were written before
Professor Walker's publication was available.

Circumstances Affecting the Weight of the Zinc Coating and the
Corrosion of Fence Wire. — Galvanized sheet metal from different
sources show generally less variation than wire in the weight of
zinc carried. The reason for this is perfectly obvious: the sheets-
leave the spelter bath through light rollers and are set aside to
cool. The rate of cooling affects the size of the " spangled'''
crystalline structure of the zinc coating, but generally no effort
is made to thin or smooth the zinc surface. Wire, on the other
hand, is either smoothed or actually wiped as it leaves the spelter
bath. This fact has an important bearing on the durability of
wire galvanized by the hot-dip process. Telegraph, telephone,
and in fact all straight-line wires which are not designed to be
fabricated into any form of woven wire, are usually smoothed by
drawing through a loose bed of damp charcoal or sand. Such
wires can easily, and should always, be made to stand four-
minute immersion in the copper sulphate test, this is sometimes.
held to mean that the wire carries about 4 per cent, of zinc. This
represents about the practical limit of the quantity of zinc that iron
in any form can be made to carry in a smooth coating. If the coating
is unsmoothed the zinc will to some extent collect in lumps and
nodules which add to the cost without increasing the efficiency
of the coating. The American market demands a form of wire
fencing which for various reasons is not used to any extent in
Europe. This is the woven wire or fabricated fence, so many
types of which are in common use in this country. In the usuaL
practice, wire for these fences is galvanized before the weaving
or welding process of fabrication is begun, and if the coating of
zinc is thick, it is liable to crack in the bends formed by the weav-
ing machines. If a welding process is used too thick a coating
of zinc interferes with the formation of strong bonds at the joints.
For this reason it is customary to wipe the galvanized wire for

TECHNICAL PROTECTION OF IRON AND STEEL 143

fabricated fences as it issues from the spelter bath with asbestos
wipers. The result of this treatment is that this type of wire
generally will not stand more than two-minute immersions in the
copper sulphate test, and trie authors have tested many. samples
that would not stand even this length of time. Undoubtedly
this is one of the principal contributory causes of the rapid corro-
sion of woven-wire fences.

It is well known to every one familiar with the process of gal-
vanizing wire that it is not possible to get as heavy a coating of
zinc on light-gage wires of small diameter as on heavier wires.
It would therefore be unreasonable, even if unwiped wire could be
used in fabricated fencing, to expect a number 14 gage wire to
stand the copper sulphate test as long as a number 9 or 7 gage
wire. The only cure for the fence problem will consist in the
public demand for fences fabricated of heavy gage wire carrying
the maximum weight of zinc consistent with the practical possi-
bility of its manufacture. In the authors' opinion no wire of
lighter gage than number 9 should ever be used in a fence if
durability is an important consideration. From an economical
point of view, it is cheaper to cut down the number of wires in a
fence rather than to lighten their gage and weight. The authors
would further suggest that the slight crackling of the zinc coating
at the bends of wire or other flexible material should no longer
be considered as objectionable, as it has been in the past. In the
effort to reduce this tendency to crack the galvanizers have been
tempted to wipe closer and closer, until in some of the modern
practice almost no zinc is left at all. Cracking at sharp bends as a
matter of fact is a good test of honest galvanizing and shows that
at least a fairly heavy coating of zinc has been applied. Zinc is
electro-positive to iron and when the two metals are in contact
and wet with a corroding solution, zinc will be destroyed and iron
will be protected. The extension of the sphere of influence of the
protecting zinc will depend upon the severity of the attack. The
more corrosive the action, the larger the protected area, but the
more rapid the destruction of the zinc. This can easily be shown
by soldering a button of zinc to a steel plate and immersing the
whole in water or various corroding solutions. The appearance
of such an experiment is shown in Fig. 32. From this point of
view a heavily zinc-coated wire, even with slight cracks at the
bends, should be preferred to a very light-coated wire which

144 CORROSION AND PRESERVATION OF IRON AND STEEL

exhibits a smooth appearance to the eye. The test wire fences
which are about to be described show that after fourteen months'
exposure to the Pittsburg atmosphere the cut-off ends of the
galvanized wires are all bright and free from any appearance of
rust, although, of course, no zinc covers the cross-section of the
cut-off ends. Since this is the case, slight cracking of the super-
ficial coating of galvanized material should not be made the
"bugaboo" of the honest galvanizer. It will often be noted that
the cut ends of galvanized iron will rust in a warehouse if under
cover, and will remain bright for long periods when exposed to
more corrosive influences in the open. This apparently contra-
dictory observation is accounted for by what has been said above.
If the attack is very slight, such as is produced by condensed
moisture from the atmosphere, the protective electrolytic action
of the zinc does not come into play, leaving the iron to slowly
coat itself with a superficial layer of hydroxide. Under the more
severe outdoor conditions, however, the zinc continues to protect
the iron in its neighborhood as long as it lasts.

Summary of possible improvements in the hot-dip process:

(1) Careful annealing of the wire and avoidance of stresses
and strains.

(2) Use of purer spelter and more careful attention to the
drossing and scumming of the spelter bath.

(3) Control of the acidity of the fluxing or soldering bath, with
the possibility of substitution of an alkaline or inert fluid.

(4) Development of methods for obtaining heavier coatings
of zinc, with the possibility of devising a practical method for
galvanizing wire fences after they are fabricated.

The Electrolytic or Cold Method of Galvanizing. — Since it is
possible by means of the electric current to deposit a coating of
zinc on iron which is practically free from other metallic impuri-
ties, it would seem that this method has some advantages over
the hot-dip method. The fact that it has found and is maintain-
ing a place in large industrial processes is the best indication
that for certain purposes the method is successful. Like the
hot-dip process, however, the cold electrolytic process has its
weak points and is open to criticism. Burgess, 1 after an extended
investigation, formed a favorable opinion of electro-galvanized
material, but speaks of some of the disadvantages of the process
1 Electro-chem. and Met. Ind., Ill, 1, 22.

TECHNICAL PROTECTION OF IRON AND STEEL 145

as follows: "It is seen that to obtain the best quality of coating
the disadvantages of the electrical method are brought out,
the principal one of which is the great length of time required to
apply the deposit. This can be remedied only by the use of high
current densities, and with solutions at present available it seems
impracticable to run the density beyond 18 or 20 amperes per
square foot on account of the liberation of hydrogen and unde-
sirable physical quality of the deposit produced. In the tests
made, 14.4 amperes per square foot was taken as best represent-
ing practical conditions where a cold, unagitated solution is used.
With this current density the metal may be deposited with a
current efficiency of nearly 100 per cent.

" A serious limitation upon the electrolytic process as at present
applied is that after the coating has reached a certain thickness
there is a tendency for it to become rough and crystalline, and
this tendency makes it extremely difficult to apply a coating
having more zinc than 30 grams per square foot."

In another place the same investigator says: "A great diffi-
culty lies in securing an even distribution of current over the object
being plated, as demonstrated on corrugated sheet iron and other
irregularly shaped articles. This causes, also, greater thickness
and roughness around the edges.

"The tests of durability show that the best results are attained
with solutions operated as nearly neutral as possible, and that the
addition of free acid, which seems desirable for keeping a clear
solution, not only decreases the current efficiency, but also materi-
ally reduces the efficiency of the zinc as a protection.

"There seems, therefore, abundant opportunity for improve-
ment in zinc-plating solutions, even though present methods are
giving good results industrially. Such improvements may come
through the discovery of new compositions for the bath which will
enable higher current densities to be used, thicker deposits to
be obtained, and greater stability and uniformity of operation
acquired. Circulation has been suggested as an improvement,
and it is to be expected that the advantages which efficient agita-
tion gives in copper depositions may also be realized with zinc-
plating vats.

"The adherence of zinc coatings to iron is frequently ascribed
to an alloying which is supposed to take place. The tests for
adherence indicate that no such alloying occurs, either with hot or

146 CORROSION AND PRESERVATION OF IRON AND STEEL

electrolytic galvanizing, since the tenacity with which the coatings
adhere is far less than would be expected if alloying existed."

Although it is true that the electro-galvanizing process pro-
vides a method of obtaining a closely adherent coating of zinc
which is not likely to be alloyed with other metals, evidence col-
lected by the authors tends to show that material protected by
this method is no more resistant to corrosion than that coated
by hot dipping. Mowry 1 has claimed as the result of experi-
ment that electrolytic coatings are inferior to those made by the
hot-dip method. The difficulty seems to be, as has already been
pointed out, that certain impurities from the electrolytic bath
are included in the deposited zinc and become accelerators
of corrosion. The Trunkhahn 2 system of electro-galvanizing
attempts to surmount this difficulty by avoiding an acid bath,
and by keeping the electrolyte saturated with nascent hydrogen
by a special method which it is claimed prevents oxidation of the
coating, renders it more dense, and prevents the occlusion of cor-
rosion-stimulating impurities from the electrolyte. The authors
have not been able up to the present time to determine to what
extent the claims made for this system of electro-galvanizing are
justified as far as the durability of the coatings are concerned.

The Vapor Deposition Process. — The Cowper-Coles 3 method
of vapor galvanizing which is known as Sherardizing is a develop-
ment of recent years. There is a considerable literature on the
subject as shown in Appendix B which should be consulted for
full information. 4 "Vapor galvanizing is distinctive from all
other forms of galvanizing inasmuch as the vapor of zinc is
employed for coating metal surfaces instead of dipping them into
molten zinc or into an aqueous solution through which a current
of electricity is passed. Two forms of vapor galvanizing are
practised, namely the 'zinc dust process' and the 'molten zinc
vapor process.' In the first process the zinc dust and the iron
articles to be coated are placed in an air-tight iron drum, which is
heated in a gas-fired furnace to obtain a temperature within the
drum of about 600° F. The length of time at which the drum is
kept at that temperature depends on the thickness of the coat

1 Mowry, Iron Age, 77, 352.

2 See Dingler's Polytech. Jour., 320, 47, 1 (1905).

3 J. Soc. Chem. Ind., 28, 399-403.
" Chem. Abs., 3, 23, 2792.

TECHNICAL PROTECTION OF IRON AND STEEL 147

required. The drums are emptied over an iron grid, which retains
the zinc-coated article and allows the zinc dust to collect into a
receptacle in order to use it over again for charging the drums.
The coating is distributed evenly and is alloyed with the iron
surface. According to claims made, the alloy is non-corrosive and
does not rust even when the zinc surface has been removed by
abrasion. The advantages of vapor galvanizing are said to be
that it is cheaper than the ordinary hot galvanizing, forms a bet-
ter protection, and at the same time enables screw threads and
machine work to be coated with an even distribution of zinc, so
that the parts fit together properly. In the 'molten zinc process'
the materials to be coated are placed in a cage or hollow drum
which is slowly rotated inside an outer cylinder, composed of
wrought iron, to ensure an even coating. The metallic zinc is
heated by means of gas, or any other suitable means, to a tempera-
ture sufficiently high to volatilize the zinc at the same time that
hydrogen or any other gaseous reducing agent is forced into the
apparatus. A modification of the vapor process is applied to the
inlaying and decorating of metallic . surfaces which surpasses all
other methods in artistic effects. Furthermore it has the advan-
tage that a number of metals may be blended together, which
hitherto has been impossible, and alloys of many tints and colors
can be obtained in one operation."

Sherardized material has not been available in this country
for a sufficient time for complete tests of durability to be made.
The vapor deposition process avoids some of the above-mentioned
difficulties of the other two methods, but it is by no means cer-
tain that it does not introduce other undesirable factors. Although
zinc is electro-positive to iron, certain alloys of zinc and iron are
said to be electro-negative and have in addition to this a higher
solution pressure than pure zinc. If these alloys are formed in
this process the main purpose of using zinc as a protective agent
is defeated. On the other hand, the dry process can be used for
galvanizing articles that have fine interstices, such as the threads
of pipe fittings, etc., so that while it may not prove to be efficient
for wire and some other special forms of iron, it will probably have
a distinct field of usefulness. 1 The results of durability tests of

1 Since the above paragraph was written Sang has described his Electro-
Cementizing process for wire. This is a modified form of the vapor deposition
method. See Proc. Elec. Chem. Soc, XVI, 257.

148 CORROSION AND PRESERVATION OF IRON AND STEEL

sherardized material will be watched with interest. One curious fact
that has been noted in connection with exposure tests on sherardized
material is the peculiar black sooty color that is assumed by the
exposed surfaces. After once acquiring the black surface, sherar-
dized articles appear to be very resistant to corrosion.

Tests of Galvanized Wire. 1 — At the request of one of the writers
a number of interesting tests have recently been undertaken for
the purpose of ascertaining the relative resistance to corrosion
of steel wires of varying chemical composition, and also to show
the relative value of protective coatings.

With these ends in view twelve samples of galvanized steel
wire were manufactured by the American Steel and Wire Com-
pany. They have been put up in the form of fence and exposed
to the action of the weather on the grounds of Carnegie
Technical School at Pittsburg (Fig. 33). These twelve samples
fall into three groups, which will be described in the order in which
they were made.

The first sample was a wire of the following composition:

Per Cent.

Carbon 0.66

Manganese 84

Sulphur 028

Phosphorus 016

This was made by the basic open-hearth method. The ingots
were cast July 23, 1908, and were reheated and rolled into billets
the same day. On the following day they were again heated to
redness and rolled into A-inch rods. One billet was taken at
random from the lot for the purpose of this test. The rod after
cooling was cleaned by immersion in hot dilute sulphuric acid,
then rinsed with water and dipped into milk of lime. After drying,
the rod was drawn cold into 9-gage wire. It was then galvanized
by passing it through a furnace in which it was heated to dull
redness, and then into a bath of hydrochloric acid containing zinc
chloride, and lastly into a bath of molten zinc. This wire was not
wiped; that is to say, it was allowed to retain as much zinc as

1 The expenses of these tests were incurred by the American Steel and
Wire Co., and the fences were mounted on the grounds of the Carnegie Tech-
nical School at Shenley Park, Pittsburg. The tests have been placed under
the general supervision of Committee U of the American Society for Testing
Materials. A description of them is included here as the intention is to keep
them in place for a number of years until final failure occurs.

TECHNICAL PROTECTION OF IRON AND STEEL 149

would adhere to it. This sample was designated C-l, and it is
purposed to test it against any of the low-carbon, low-manganese
wires to be described later, and which may for this purpose be
regarded as C-2.

The second group consists of six samples of "American" style,
8-strand fabricated fencing, 45 inches high. The object of this
group is to determine the effect, if any, of segregation of the

150 CORROSION AND PRESERVATION OF IRON AND STEEL

impurities in the iron. Therefore the six samples were taken
from the same heat. The steel was made at the same mill and
on the same day as sample C-l, but by the Bessemer process,
and was cast in six ingots. As these ingots were rolled, two billets
were taken from the top of the first ingot, two billets from the
middle of the first ingot, and two from the bottom of the first
ingot. These three pairs of billets were marked, respectively,
A-l, A-2, and A-3. Similarly, from the last ingot of the heat
two billets were taken from the top, two from the middle, and two
from the bottom. These three pairs of billets were marked,
respectively, B-l, B-2, and B-3. These were all rolled hot into
rods, and after cooling and cleaning, in the manner described
above, were drawn into wire. The six samples were each divided
into three portions. About half of each sample was drawn into
11-gage wire, a quarter of each sample into 9-gage wire, and the
remaining quarter into 12-gage wire. This was because these
three sizes of wire are all used in making the standard "Ameri-
can" fence, the 9-gage for the top and bottom horizontal strands,
the 11-gage for the intermediate strands, and the 12-gage for the
vertical connecting wires. Immediately after being drawn the
wire was galvanized in the same manner as described for sample
C-l, except that on emerging from the molten zinc it was wiped
by a mechanical device which left a thin, smooth coating of zinc
on the steel. A separate piece of fencing about 300 feet in length
was then woven from each of the samples. These six pieces of
fencing then represented the top, middle, and bottom of the first
and last ingots of the heat, respectively. The samples were not
analyzed separately, but an analysis of the heat as a whole showed
percentages as follows:

Per Cent

Carbon ;■> 0.09

Manganese 55

Sulphur 045

Phosphorus 092

The third group consists of five samples of basic open-hearth
steel showing increasing amounts of manganese from 0.07 per
cent, to 0.37 per cent. These were all made in one furnace,
between August 18 and August 22, 1908. The process of manu-
facture throughout was identical with that employed on the
sample designated C-l, except that the wire was wiped during

TECHNICAL PROTECTION OF IRON AND STEEL

151

the process of galvanizing. These five samples were ail drawn
to 11-gage and are designated by their "heat numbers," which
axe given here together with the analyses.

Analysis of Five Samples Forming the Third Group.

Heat Numbers

Carbon

Manganese

Sulphur

Phosphorus

Per Cent.

8118

0.05

0.12

0.014

8119

.04

.07

.014

8120

.05

.16

.010

8121

.05

.24

.018

8122

.06

.37

.013

.016

All the wire described above was shipped to Pittsburg, Penn-
sylvania, and between September 23 and September 26, 1908, was
put up in the grounds of the Carnegie Technical Schools. The
site assigned for this purpose is a level strip of ground at the
bottom of a deep and narrow natural depression to the north of
the quadrangle of school buildings. Because of this location none
of the wire will be exposed to the action of wind, but all will
show the effects of fog and dampness, which are very prevalent
in this hollow.

The fence was put up in an extremely substantial manner.
The posts are 2 by 8 inches, dressed oak; the corner posts are 8
by 8 inches, sunk 3 feet in the ground. The lower ends were
painted with a mixture of benzine and asphalt, and the portion
above ground with ordinary lead paint. The post holes were
filled in with concrete.

There are four parallel lines of posts, as shown in Fig. 33,
3 feet apart and 275 feet long. These lines run almost due east
and west. The one farthest north is called row No. 1, the next
row No. 2, the next row No. 3, and the line farthest south No. 4.
The arrangement of the wire on these posts is as follows :

On the north side of row No. 1 is placed sample A-l.

On the south side of row No. 1 is placed sample A-2.

On the north side of row No. 2 is placed sample A-3.

On the south side of row No. 2 is placed sample B- 1.

On the north side of row No. 3 is placed sample B-2.

On the south side of row No. 3 is placed sample B-3.

152 CORROSION AND PRESERVATION OF IRON AND STEEL

On the north side of row No. 4 at the top are 8 strands of 8119.

On the north side of row No. 4 at the middle are 8 strands
of C-l.

On the north side of row No. 4 at the bottom are 8 strands of
8118.

On the south side of row No. 4 at the top are 8 strands of 8122.

On the south side of row No. 4 at the middle are 8 strands of
8121.

On the south side of row No. 4 at the bottom are 8 strands
of 8120.

In the first three rows the posts are 15 feet apart, in the fourth
row 30 feet apart. The wire was all stretched as tightly as possible
to prevent sagging, and none of it is less than 1 foot above the
ground.

In addition to the test fences described above, a number of
panels have been mounted on either side of the original rows
in order to test different methods of providing protective coat-
ings for ordinary steel wire fences. One panel of American style
woven-wire fence is galvanized by the cold or electroplating
process; another panel is made of double galvanized wire carrying
about the same weight of zinc as is usually specified for telegraph
wire. This galvanizing has been done by a special adaptation
and variation of the ordinary hot-dip process, and has not been
subjected to the acid bath before galvanizing. A third panel is
coated with zinc by the Sherardizing process. A number of
panels, in part galvanized by the ordinary hot-dip process, and
in part ungalvanized, have been protected by painting with vari-
ous protective paint coatings. The formulae of these stimulative
and inhibitive paints will be published in a special report. All
of these extra panels have been mounted in exactly the same
manner and side by side with the samples already described in
the foregoing, and the tests should in the course of time yield
results of great value. A key to the arrangement of all the panels
is shown in Fig. 34.

Preservation with Tin. — Roofing, cans, and tinned utensils
of all kinds depend for their preservation on a metal coating
which, though not electro-positive to iron, has such a low solution
tension that it is known to be practically incorrodible. The
principal difficulty encountered in the use of this metal is that it
seems impossible to apply it economically to the surface of steel

S1VIU3J.VH 0NI1S31

aod Aiaioos nvoiubwv

30N3J 3HIW 'A1VO aavaNVJ-S-a-coefr

SISOd 3IIHM

sisod NMoaa

154 CORROSION AND PRESERVATION OF IRON AND STEEL

so that pinholes shall not exist in the coating. Walker has shown
this by an ingenious modification of the ferroxyl test. 1 A slightly
acid solution of gelatin containing a small amount of potassium
ferricyanide is flooded, while hot, upon the surface of the tin
plate to be tested. The gelatin becomes hard upon cooling, and
in a very short time every pinhole on the tin surface is marked
by a spot of Turnbuli's blue. The results obtained by Walker
are illustrated in Figs. 35 and 36. They are made even more

A • . '%

,-.**• ' ■ *

' '*

f '

■ ? "« * \ .~ »»-i ■-.,* -■„ . ..

"-""'* . * . • -

„ * y * '>;' I

* . ' ' *•-.',-

i '*'* -- 1 -' '

;V ; ; - ■-. ! '*-'\ > \. / * ' '

. *' *t ■-'.■:'

• * ' \v/$

V '■ . '.' / ' ■

\ * ; " " -

• 1

, v . J ' N 1

" ■ >"■" . *

*■'.,■ t "**■"

t '', " »r ;;i.--' J, . "'^

Fig. 35. — Showing Walker's method of determining pin holes in tin

plate.

interesting by comparison with the actual corrosion of a tin plate
under service conditions shown in Fig. 37. After the number
and position of the pinholes is known, it becomes possible to study
ways and means of eradicating them. Hot rolling and ageing of
the plates under pressure are among the solutions that have been
suggested in the hope that the pinholes will disappear.

The passage of the Food and Drugs Act of 1906 has had a
curious and interesting bearing on the tin-plate industry in its
application to canned goods. Artificial coloring matter is debarred
by this law, and it was found that the natural color of preserved
food-stuff is reduced and destroyed by the action of the tin. In

1 Jour. Iron and Steel Inst., 1, 79 (1909).

TECHNICAL PROTECTION OF IRON AND STEEL 155

the effort to meet this difficulty the can manufacturers sought to
use less tin, and also to lacquer the tin plate so as to prevent its
contact with the fruit and vegetable juices. The formation of
pinholes has gone on rapidly, however, so that it is not an unusual
occurrence to find tinned preserves discharging their juices in
tiny streams through small holes resulting from rapid corrosion.
Walker's more recent researches on this subject have already
been described in a previous chapter.

. '

. •

•

' ".

Fig. 36. — Showing Walker's method of determining pinholes in tin

plate.

Preservation with Copper, Lead, and other Metals and Alloys. —
It is possible to make copper-coated steel, but this makes an
efficient and rust-resistant material only when the coating is
thorough and homogeneous. Methods are being developed to
extend the use of copper for this purpose, and it is now applied
to wire as well as to sheet and plate metal. If the cost is not
a prohibitive factor, there would seem to be no reason why steel
protected in this way should not come into more general use.
It follows, however, from the electrolytic theory of corrosion that
if openings in the copper coating occur so that water can pene-
trate to the point of contact, the corrosion of the iron will go on
very rapidly. Tassin, 1 speaking for the new metallurgical product
1 Jour. Industrial Chem., 1, 9, 670.

156 CORROSION AND PRESERVATION OF IRON AND STEEL

known as copper-clad steel, makes the following claims: " Copper
and steel in the presence of moisture form a galvanic couple, and
the corrosion of the steel proceeds with great rapidity. The resist-
ance to corrosion of copper-clad so far as the copper coating is
concerned is, of course, the same as that of copper. But, in

* . """4

l "

%,.

■ "

% v 1

**'

• i: .-.

•lli 1

■■;■■■ '■■ ■ ■."■/' ■' '

mm s

* . %

■7

• *^

Fig. 37. — Charcoal iron tinned plate, showing the corrosion lines.

view of the marked galvanic action set up between copper and
steel, it would be assumed that corrosion would quickly occur at
the exposed ends, and be a constant factor so long as an elec-
trolyte was present. This supposition is not so, at least in
the presence of fresh or salt water. Test samples placed in
water, through which a current of air has been allowed to bubble

TECHNICAL PROTECTION OF IRON AND STEEL 157

continuously for three months, demonstrated that after a certain
period of time, somewhere between 15 and 40 days, corrosion
practically ceases, as shown by taking the loss in weight of the
tests at varying periods. It is believed that this stopping of
corrosion is a result of the following conditions :

"After a certain amount of rust has been formed, it appears
that a thin film of copper mixed with some copper oxide is plated
out or deposited between the iron oxide and the unattacked steel,
and that this film will act as a preservative coat as long as it
remains intact. If broken, further oxidation sets up and the
process simply repeats itself. While it is true that the corrosion
on the end of a wire is not a factor in its life, yet the corrosion
of the end of a relatively large diameter may become very serious.
If the above observations hold true on larger sizes (so far it has
not been tried on sizes above f in.) it will have quite a bearing
on material suitable for marine work. Tests along this line are
now being carried out, and the evidence to date points to a con-
firmation of the observations made on the smaller sizes."

Lead has to some extent been used for covering iron, and for
certain purposes this material is useful. It is doubtful, however,
whether the application of lead for this purpose will find a very
wide field of usefulness. Terne plate covered with an alloy of
lead and tin is, of course, extensively used. There have been
many complaints recorded in regard to the rapid -corrosion of
this material, to which much that has been said in previous
paragraphs of this book about other forms of protection will
apply.

Many other alloys have been proposed and experimented with
for the protection of iron. In more recent work the effort is
made to change the electro-negative character of lead or tin by
producing electro-positive alloys of these metals that will have
the proper physical characteristics to warrant their use as pro-
tective coatings. Among these may be mentioned an alloy of
lead and antimony, and of tin with antimony and aluminum.
None of these special alloys have as yet come into very general
use.

Processing after Manufacture. — A number of metallurgical
processes, which depend upon the production of a fully oxidized
surface on iron and steel, have been in use for a number of years
and are too well-known to require description. Among these are

158 CORROSION AND PRESERVATION OF IRON AND STEEL

the Bower-Barf and Wells methods of producing a surface of
magnetic oxide. If perfectly coherent and homogeneous such a
surface is, of course, unrustible. The protection of the metal
depending as it does upon a coating that is strongly electro-
negative to iron, it follows- that any manipulation or accident
that injures the continuity of the covering surface must neces-
sarily lead to accelerated and deep corrosion. This fact together
with certain other practical difficulties has confined the use of
these processes to special cases where protection from corrosion
is a matter of prime importance. It is claimed for the process
that it is specially valuable for builders' hardware, domestic
articles, wrought-iron grilling, railings, and the better class of
scroll and fancy work. It has also been used for pipe of various
kinds, and can be used on general structural material of a rigid
nature. It could not, of course, be used with wire or flexible forms,
owing to the brittle nature of the magnetic oxide.

The well-known Russia iron, which is in common use for stove-
pipes, locomotive casings, etc., depends upon a similar oxide
coating. When taken care of, iron treated in this way will last
for many years without failure from corrosion.

Sang 1 in an article on the inoxidation processes for protecting
iron and steel gives the following information: In the Barf
process the articles to be coated are heated to 1000° and steam
superheated to 538° is injected into the coating chamber. In
the Bower process, producer gas rich in carbon monoxide is said
to reduce the red oxide obtained by air treatment or by subject-
ing the work to acid fumes. In the Bower-Barf process the articles
to be coated are heated in a closed retort to 871°, after which
superheated steam is injected which forms a coating of the black
magnetic oxide. The operations are repeated alternately until
a sufficient depth of oxidation is obtained. The cost of treatment
is expensive and ranges from $5 to $20 per ton. It is claimed
that enamel will hold tenaciously to iron so treated and that paint
can be applied without trouble. The coatings are said to resist
acid fumes better than galvanized iron; they are not affected by
solutions of copper sulphate. This method has been used to
some extent for coating pipes in order to make them withstand
corrosion. A modification of the Bower-Barf is the Wells process,
which consists of introducing the steam and carbon monoxide at
1 Electro-chemical Metallurgical Industry, 7, 351-3.

TECHNICAL PROTECTION OF IRON AND STEEL 159

the same time. In Gesner's process, a compound of hydrogen,
iron, and carbon is said to be deposited on the iron. Steam at
low pressure is introduced through a red-hot pipe into the coat-
ing retort, which is kept at a temperature of 538° to 649°, while
some hydrocarbon gas is allowed to enter slowly. By this means
the higher oxide of iron is reduced and the surface carbonated.
The charge for so treating small articles is said to range from 4
to 7 cents per pound, and for large pieces about 1 cent per pound.
The Dewees Wood's process is similar to Gesner's process. Ruf-
fington employs potassium nitrate in his method for oxidizing
the surface of the iron. Claudius uses a solution of a manganese
compound for the production of durable black coatings or patina.
Hydraesfer's process is essentially similar to Gesner's, but requires
only one furnace operation.

A number of electric processes have been proposed, both for
oxidizing the surface of steel and also for producing a passive
condition of the surface. One of the authors called attention
to the possibility of the use of electric currents in producing a
passive condition some years ago. Meritens' method uses a bath
of distilled water at a temperature of 70° to 80° and applies a
weak current. The nascent oxygen which is given off in the
presence of the hydrogen occluded in the metal produces a film
of black oxide of iron. It is, however, impossible to produce a
deep coating unless a sufficient amount of hydrogen is present
in the iron, and on this account it is generally necessary to run
the article first as cathode in order to enable it to absorb hydro-
gen. Lately Toch has been investigating the use of electric
currents to produce a more or less lasting passive condition of
the surface. This work has not yet been published in detail, but
it is understood to consist in making the iron to be treated an
electrode in an alkaline solution using a current of large amperage
at low potential. Krassa x works in a hot caustic soda solution
and makes the iron the anode. Strong currents make the anode
passive at once, while weak ones cover it with a black coating
of Fe 3 4 . To get active iron again the specimens must be boiled
for a long time. Further work along these lines will be watched
with interest.

Several other methods have been proposed from time to time
for processing after manufacture in order to make the surface
1 Zeit, Electro Chem., 15, 490.

160 CORROSION AND PRESERVATION OF IRON AND STEEL

resistant to corrosion. The Coslet process is said to consist in
immersing the iron in a hot, phosphorized solution containing
an iron compound. The surface, it is claimed, is converted into
a ferroso-ferric phosphate, which is to some extent resistant to
corrosion. Jouve * has called attention to the fact that iron
containing 10 per cent, of silicon is not attacked by acids, and,
therefore, such material should theoretically be unrustible. The
authors have confirmed this deduction by experiment, and found
that iron containing 10 per cent, of silicon is almost incorrodible.
Unfortunately, such a metal is not workable, and has peculiar
properties. Since silicon is much like carbon, chemically speak-
ing, it would seem as if it might be worked into the surface of
steel by modifications of some of the processes used for case-
hardening with carbon.

Unprotected Steel. — A large proportion of the iron and steel
which is in use cannot, from the very nature of the service to
which it is put, be protected from corrosion. To this class belong
rails, heavy chains, implements, boiler tubes, etc. The only hope
of meeting this particular phase of the problem consists in the
improvement of metallurgical processes, to the end that perfectly
homogeneous metal, as free as possible from segregation, may be
manufactured. It is well known that some of the alloyed steels,
such as nickel steel, are very highly resistant to corrosion. The
same thing is probably true of the chromium, vanadium, and chro-
mium-vanadium steels. Unfortunately, such materials are too
costly to be used on a large scale, although there is always the possi-
bility that some alio} 7- or combination will be discovered which will
be at the same time comparatively inexpensive and incorrodible.

The Corrosion of Boilers. — One of the most important phases
of the technical preservation problem is the necesssity for pre-
venting the pitting of boilers and boiler tubes. In a bulletin
published some years ago, one of the authors 2 stated that the
pitting of boiler tubes was probably due to electrolysis, and it
was shown that Wood had reached the same conclusion in 1894.
Wood 3 said: "That there is a continual electrical action of a most
complex character present in all boilers under steam can scarcely

1 Engineer, 1908, Vol. CVI, p. 397.

2 The Corrosion of Iron. Bui. 30, Office of Public Roads, U. S. Dept.
Agr. (1897).

3 Am. Soc. Mech. Eng. Trans. (1894), 15, 998.

TECHNICAL PROTECTION OF IRON AND STEEL 161

be doubted, and the same action, but less apparent, is possibly-
present in all constructions of iron when the different members
formed of iron and steel of various compositions, made by different
processes after various torturing methods of manufacture to bring
them to the desired shape, are assembled and put into duty under
strains and conditions foreign to their nature. It would be strange
indeed did not some electrical energy manifest itself and call for
some palliative if not protective means of arresting decay."

Recently Burgess 1 without reference to previous authorities
has reached the same conclusion. Burgess describes an investi-
gation which shows the importance of electrolytic action in boiler
corrosion, and states that if electrolytic action were entirely
absent or could be entirely neutralized, there would be practi-
cally no internal corrosion. Iron may itself set up electrolytic
couples even if it is not segregated owing to physical strains
occasioned by previous heat treatment. Burgess suggests as a
possible remedy the use of an external iron bar maintained as an
anode by a current from a dynamo, the negative pole of which is
connected to the boiler. This suggestion had been made many
times before 2 in connection with boiler tubes and also for the
protection of pipe lines; it has not, however, as far as the authors
are aware, been very generally adopted.

The expedient of using metallic zinc in boilers to overcome
local electrolytic effects in the iron by producing a still greater
electrolytic effect at the almost exclusive expense of the more
positive zinc is well known and has been in use for a long time.
Great difficulty has, however, been experienced in maintaining
good metallic contacts between sufficiently large surfaces of the
two metals under the conditions which maintain in a boiler;
moreover, as has been previously shown, the protective action
of zinc takes place only in a restricted area.

The most practical method for protecting boiler tubes depends
upon the exclusion of oxygen from the feed water. It has been
shown that oxygen plays an important contributing role in the
electrolytic corrosion of iron. In order to remove this action,
the use of pyrogallol, amidol, tannic acid, and other reducing
agents has been proposed, but far more practical ways than this
are known. Feed-water heaters in which all the dissolved gases

2 Jour. Western Soc. Eng., 14, 375.

2 See Cushman in Journal Iron and Steel Inst,, 1 (1909), 37,

162 CORROSION AND PRESERVATION OF IRON AND STEEL

may be removed are not difficult to design, and if these were
made a necessary adjunct of all boilers there would be fewer cases
of boiler-tube corrosion to cope with.

Special Cases Met with in the Preservation of Iron. — In the
opinion of the authors the most interesting and significant dis-
cussion of recent date on a special corrosion problem is to be
found in the Transactions of the American Institute of Mining
Engineers for the years 1907 and 1908. The special significance
of this discussion lies in the fact that it exhibits the more or less
divergent opinions of a number of engineers and experts of wide
experience. We find in it the views of engineers who are con-
sidering their problems in the light of the electrolytic explanation
of corrosion, as well as the opinions of those who refuse to consider
that any form of electrolysis is a factor.

Theories are manifestly of little value unless they can come to
the aid of those who are charged with the great industrial prob-
lems of the world, and it is in the court of practical experience
that a working theory must be tried and either proved or found
wanting. At the same time it frequently happens that an engi-
neer will at first consider as "new-fangled" or " fanciful" a new
theory only because he has not fully grasped its meaning or
recognized its application to his problem. The authors have
thought it best to include the entire discussion in an appendix 1
as its re-reading will undoubtedly stimulate thought. Indeed, the
problems discussed should be reviewed in the light of the modern
theories and explanations of corrosion.

As has been pointed out, the technical preservation of iron
and steel presents a great number of separate problems, each one
of which requires special consideration and different treatment.
As an instance of this, it may be stated that the cases of wire,
of ships' bottoms and boiler tubes involve special difficulties, in
which the environment and conditions of service are so different
as almost to constitute three separate problems. Nevertheless,
the same underlying principles apply to the general problem of
the corrosion of iron in all its phases. It has been the object
of the authors to make these principles clear in the foregoing
pages. The protection of the great bulk of finished iron and steel
must inevitably remain a paint problem, and to the consideration
of this the remaining portion of this book will be devoted.
1 Appendix A, p. 279.

CHAPTER VII

RELATION OF PIGMENTS TO THE CORROSION OF IRON

The many theories which have attempted to explain the
rusting of iron, during the last century, have stimulated a large
amount of original research on the relation of various pigments
to the corrosion problem. In the course of the investigations
undertaken, the subject of protective coatings for iron and steel
was naturally brought into prominence and received a consider-
able amount of attention. The study of protective coatings
for iron has led many paint manufacturers, as well as scientific
investigators, to make a closer study of the causes of corrosion.
It is evident that the electro-chemical explanation of corrosion
must have a direct bearing on paint problems.

In the course of researches carried on by one of the authors
upon the corrosion of iron, it was found that certain substances
in contact with iron possessed the property of exciting electrical
action and stimulating corrosion, while still other substances
exhibited a tendency to inhibit or prevent corrosion. Bichro-
mates of soda and potash were found to be the most eminent
examples of the so-called "inhibitives," and it was found that
as long as steel or iron remained in contact with these salts, even
in fairly dilute solution, rusting could not take place. This
naturally suggested the preparation and use of slightly soluble
chromates as pigments. It was found that the chromic acid salts
could be precipitated with certain other compounds to produce
chromates applicable as pigments. A series of these compounds
were prepared and tested; but it soon became apparent that
while some afforded protection to steel, others did not. The raw
materials used, the method of preparation, the amount and
character of the contained impurities, together with other factors,
had a marked influence upon the efficiency of these compounds.
It was even found that a series of chrome salts, all of which theo-
retically should have possessed inhibitive value, were in many
cases actually stimulative.

Testing Pigments in Water Suspension. — The results obtained

163

164 CORROSION AND PRESERVATION OF IRON AND STEEL

from these investigations led to a series of experiments, first
suggested by Thompson, in which the effort was made to determine
the relative inhibitive value of a number of typical pigments for
application to iron. A series of fifty of the most important pig-
ments used in the manufacture of paints was therefore procured
and submitted to test. The apparatus in which these pigments
were tested consisted of a series of eight-ounce bottles, into each of
which was placed a measured amount of pigment, afterward adding
the same amount of distilled water to each bottle. Into each bottle
was then placed a strip of steel, numbered and carefully weighed.
The bottles were then connected up in series by the use of rubber
stoppers and bent glass tubing, through which a constant current
of washed air was passed for a period of three weeks. At the
end of this time the train of bottles was disconnected, the steel
plates removed, washed off, carefully brushed with a tooth brush,
to remove any extraneous or adherent pigment particles, and
afterward dried and carefully reweighed. A loss in weight repre-
sented the amount of iron which was removed by corrosion.
From the results obtained it was possible to tentatively divide
the pigments into three groups. 1 The pigments which caused
the most active corrosion were termed "rust stimulators/' those
which showed the least action were termed "rust inhibitors/'
while those which were found to be intermediate in the group
were called "inerts." 2 The results were charted, Fig. 38, show-
ing the position of each pigment tested, and the results proved
so interesting that it was decided by Committees "E" and "U"
of the American Society for Testing Materials to carry on a sup-
plementary series of experiments. A considerable quantity of
pigments similar to those used in the first test were collected in
the open market, and samples of each, from the original package,
were put in cans, carefully marked, stamped, and forwarded to
members of the Committee, who had volunteered to repeat the
work. A series of small plates of steel of the same size, cut from
the same sheet of metal, and all stamped with numbers running
from 1 to 50, were also supplied to each investigator. A descrip-
tion of the apparatus used for the test was also sent to the mem-
bers of the Committee.

1 See Cushman, The Inhibitive Power of Certain Pigments on the Corrosion
of Iron and Steel. Proc. Am. Soc. Testing Materials, VIII, 605 (1908).

2 Some objection was made to this use of the word ( inert" and the joint
Sub-committee of Committees U and E, Am. Soc. Testing Materials,
adopted the word ( 'indeterminate."

O O ou j
? Z Z £

i_j5>mo=Oer oa><J -oof
if- vraNr T«ini-wwinra--itc

ui i y uj < < c

> o 2 > o £ ;

U. -£ 3 ■" < UJ UJ 3; t

(zziiiHiir-O,
m £ > ozoQtno i.

J O Q. C
J (E E C

J X, < 3

in o» «i « ■- o o

'i-iory(otNin<j-tcr

' PI IO "BMrVlnlit

- -* uj t- r __

; < < «j O Z

: (E cc uj uj I I O

j uj uj z -i a. 0- cd

] z z — Q- < < cc

j _ _ X < a or <

JEEoZuou

166 CORROSION AND PRESERVATION OF IRON AND STEEL

The results obtained by the various investigators, working
independently and in different laboratories, were afterwards
compared at a special sub-committee meeting, called for the
purpose. The agreement between the results obtained by the
several investigators corroborated the results of the original work,
and appeared to indicate that the theory of inhibitives should
have some practical application.

Effect of Impurities and Other Factors Governing Nature of
Pigments. — Since this work was done, so much progress has
been made following this line of investigation that it now seems
possible for certain rules to be formulated in regard to the
inhibitive value of various pigments. Not only the nature of the
pigment itself, but also the quantity and kind of impurities
contained in it, decide in which one of the three groups a given
material must be classed. The solubility of the pigment, and the
ease with which it is ionized when brought into contact with
water, are also important considerations. If the pigment is of a
basic nature, and therefore tends to increase the number of
hydroxyl ions in solution, it is sure to be in the inhibitive list,
provided the concentration of the hydroxyl ions is sufficiently
high to furnish protection. If the pigment is acid in nature, or
contains any free acid which is easily hydrolized, it is quite
certain to appear in the stimulative group.

The question naturally arises as to whether a pigment which
has been shown to possess active stimulative properties may be
mixed with a pigment possessing inhibitive values, and in this
manner become inert or non-stimulative. This depends largely
upon the percentage of inhibitive pigment added. In some cases
the addition of a large quantity of inhibitive pigment would be
of little value, as the following statement already published by
one of the authors illustrates: 1 "If the surface of iron is subjected
to the action of two contending influences, one tending to stimu-
late corrosion, and the other to inhibit it, the result will be a break-
ing down of the defensive action of the inhibitor at the weakest
points, thus localizing the action and leading to pitting effects. "

It is interesting to note that the two samples of Prussian
blue which were included in the tests described in a previous
paragraph showed widely divergent results, one proving to be an
inhibitor and the other a stimulator of corrosion. As a matter of
fact, these two samples of Prussian blue were prepared by different
1 Proc. Am. Soc. Testing Materials, 1908, VIII, 606.

RELATION OF PIGMENTS TO THE CORROSION OF IRON 167

methods. In one, acid impurities were sure to have been included.
As has been already pointed out in a previous chapter, it is impossi-
ble to throw down a colloidal precipitate from a solution without
the inclusion of ions and salts from the mother liquors. The
methods of manufacture used in these two Prussian blues served
as an advance indication that one would appear in the stimulative,
and the other in the inhibitive class, and they were so marked.
The fact that the result of experiment brought out this relation
presents additional evidence in support of the general theory.

Fig. 39. — Apparatus used in testing inhibition value of pigments.
(Cushman.)

It is well known that a good working theory should enable
an investigator to predict to some extent in advance what the
results of experiment should show. Thus with the theory of
periodicity of atomic weights in hand, Mendeleeff was enabled
to predict the existence of unknown elements which were subse-
quently discovered. Although it is not claimed that the obser-
vations in regard to the Prussian blues are of the same order
of magnitude as the predictions of Mendeleeff, nevertheless the
results obtained are none the less suggestive. 1

The results of these tests are given below. The apparatus
used in their determination is illustrated in Fig. 39.

1 It is of course not yet shown that the relative durability of paint coat-
ings under service conditions will agree with these results, but the authors
think that the point made is none the less interesting.

168 CORROSION AND PRESERVATION OF IRON AND STEEL

Chart of Findings of Members of Committees "E"and "T7", American
Society for Testing Materials. Loss of Steel in Grams in Tests
Carried Out on Pigments to Ascertain Their Value as Rust
Inhibitors.

Pigment

Gardner

No. 1
20 days

Zinc Chromate

Zinc and Barium Chromate . . .

Zinc and Lead Chromate

Zinc Oxide

Zinc Lead White

Barium Chromate

Ultramarine Blue

Chrome Green (blue tone)
Prussian Blue (Inhibitive)

Lithopone

Willow Charcoal

Litharge

Dutch Process White Lead ....
Quick Process White Lead ....

Calcium Sulphate

Prince's Metallic Brown

Orange Mineral French

Calcium Carbonate (Whiting). .

Sublimed Blue Lead

Lemon Chrome Yellow

Orange Chrome Yellow

Medium Chrome Yellow

Chrome Green

Venetian Red

Bone Black

Asbestine

Keystone Filler

Orange Mineral (American) . . .

Umber

China Clay

Calcium Carbonate Precipitated

Red Lead

Prussian Blue (Neutral)

Indian Red

American Vermilion

Sublimed White Lead

Sienna

Naples Yellow

.0050
.0153
.0094
.1524
.0842
.2333
.0247
.0860
.1438
.0160
.1694
.4325
.2040
.2120
.3966
.3774
.3950
.3828
.3177
.2767
.2826
.4090
.3265
.2682
.3392
.2394
.3560
.4416
.1365
.3493
.3574
.3112
.3584
.3546
.4328
.4176
.2876
.6482

Walker
P.H.
No. 2

7-£ days

.0300

.0468

.0277

.0296

.1712

.0101

.3185

.2269

.2267

.3791

.2795

.1932

.2895

.3352

.2143

.2620

.2724

.3620

.3425

.4067

.4203

.3767

.3670

.4756

.3245

.4025

.4651

.4336

.5961

.4770

.4910

.3555

.4463

.3739

.4147

.5856

.5432

.4800

.0094

.0034

.0153

.1002

.0515

.0429

.0137

.0548

.0448

.1274

.1439

.0309

.1781

.1288

.1759

.1983

.1495

.1384

.1001

.1365

.1700

.1319

.1348

.1955

.0921

.1748

.1366

.1719

.1498

.1248

.1828

.1495

.1218

.2617

.2612

.0982

.2949

.1512

Cush-

man

No. 2

10 days

.0130

.0140

.0085

.0856

.0094

.1865

.1240

.1130

.1792

.1362

.1584

.1150

.1848

.1597

.1408

.1467

.2380

.2365

.1972

.1907

.1763

.1453

.2375

.1413

.2240

.3349

.2065

.3817

.2445

.2625

.1717

.2415

.1905

.1877

.2372

.3085

.2347

Walker
W. H.
No. 1

.0396
.0351
.0620
.0504
.0456
.1932
.0496
.2346
.2671

.2110

.2743
.2274
.2174
.1974
.2526
.1208

.2150
.2288
.3521
.1564
.4401
.3405
.1481
.2315
.2403
.3212
.2616
.5707
.4173
.4334
.3387
.3116
.4462
.3846

Aver'ge

.0194

.0229

.0246

.0682

.0876

.0978

.1186

.1453

.1591

.1754

.1880

.2038

.2122

.2176

.2328

.2352

.2432

.2484

.2492

.2543

.2557

.2645

.2651

.2666

.2674

.2762

.2881

.2970

.3009

.3034

.3111

.3117

.3171

.3228

.3270

.3300

.3761

.3797

RELATION OF PIGMENTS TO THE CORROSION OF IRON 169

Chart of Findings, etc. — Continued

Pigment

39
40
41
42
43
44
45
46
47
4S
49

Prussian Blue (Stimulative)

Mineral Black

Barytes

Natural Graphite

Bright Red Oxide

Acheson Graphite

Ochre

Carbonith White

Carbon Black

Precipitated Blanc Fixe
Lamp Black

Gardner Cush-
man

No. 1
20 days

.5113
.3050
.4454
.4342
.3878
.5262
.4022
.2655
.5003
.5247
.7180

Nos.l&2
10 days

.4559
.8018
.5883
.5437
.7896
.6337
.8408

.6955

.8806

1.3098

Walker
P. H.
No. 2

74 days

.2055
.2017
.2547
.2606
.2920
.3723
.2119

.4069
.3132
.2838

Cush-
man

No. 2
10 days

.2195
.3529
.3841
.3173
.3707
.2789
.4315

.3751
.5085
.7096

Walker
W.H.

No. 1

Aver'ge

.5202
.3353
.5636
.7165
.4429
.5095

.7152

.5716
.5064
.6257

.3825
.3993
.4472
.4545
.4566
.4641
.4716
.4904
.5099
.5467
.7294

Classification of Pigments Based on Results of Tests.

Inhibitors

Indeterminates 1

Stimulators

Zinc Lead Chromate

White Lead (quick process,

Lampblack

Zinc Oxide

Basic Carbonate)

Precipitated Barium

Zinc Chromate

Sublimed White Lead (Basic

Sulphate (Blanc

Zinc and Barium

Sulphate)

Fixe)

Chromate

Sublimed Blue Lead

Ochre

Zinc Lead White

Lithopone

Bright Red Oxide

Prussian Blue (In-

Orange Mineral (American)

Carbon Black

hibitive)

Red Lead

Graphite No. 2

Chrome Green (Blue

Litharge

Barium Sulphate

tone)

Venetian Red

(Barytes)

White Lead (Dutch

Prince's Metallic Brown

Graphite No. 1

process)

Calcium Carbonate (Whiting)

Prussian Blue (Stim-

Ultramarine Blue

Calcium Carbonate (Precipi-

ulative)

Willow Charcoal

tated)
Calcium Sulphate
China Clay
Asbestine

American Vermilion
Medium Chrome Yellow

1 See note, page 164.

Electrical Conductivity of Paint Films. — Certain objections to
the foregoing tests have been made on the ground that they
were made in water solutions in which the pigments were sus-

170 CORROSION AND PRESERVATION OF IRON AND STEEL

pended. It was claimed that pigments which might stimulate
or cause corrosion in the presence of water would have no such
action when enveloped in an oil medium. It was said that the
oil acted as an envelope for the pigment particles, and being
in itself a non-conductor of electricity should prevent electrolytic
action taking place upon iron. While this criticism is worthy
of the most careful consideration, it appeared to the authors
that the objectors had not paid attention to the well-known fact
that linseed oil films have the power of absorbing both water
and carbonic acid. For this reason the oil film cannot be con-
sidered a non-conductor of electricity, and must depend to a
large extent upon the pigmentary substances with which it is
reinforced. In order to get further data on this interesting sub-
ject, the following experiments were carried on by one of the
authors.

A series of glass slides such as are used in microscopic work
were painted with a number of separate pigments ground in
linseed oil. After the paint films were thoroughly dry, small
strips of copper were attached to the opposite ends of the paint
films, and these were in turn attached to the wires of an ordinary
dry cell, a galvanometer being included in the circuit in order to
measure the current passing through. It was found that abso-
lutely no current could pass a perfectly dry paint film, and the
galvanometer needle remained in its original position, at zero.
The glass slides were then removed from the apparatus and im-
mersed in water for four hours, during which time the films were
inevitably penetrated to some extent by water. The action of
rain and weather exposure upon painted metal surfaces in the
open of course produces a similar result. The slides were then
removed from the water and, after being carefully wiped, again
inserted in the apparatus and tests made to determine whether
any passage of current took place. It was found that those
pigments which are naturally good conductors of electricity per-
mitted to a slight extent the passage of current, which was shown
by the deflection of the galvanometer needle. Those pigments,
on the other hand, which are not good conductors of electricity,
permitted so little current to pass that the needle remained prac-
tically stationary. The pigments referred to as being conductors
of electricity, which were used, belong in the carbonaceous group,
and appeared in the stimulative class, as may be seen by reference

RELATION OF PIGMENTS TO THE CORROSION OF IRON 171

to the foregoing table. These "experiments seem to indicate that
when iron is painted with the so-called stimulators of corrosion, if
for any reason water finds its way into the oil film, electrolysis, with,
the resulting current flow, is able to take place, and is sure to
cause active corrosion of the steel surfaces to which such pigments
are applied. On the other hand, it appears to the authors that
it is not unreasonable to conclude that the use of insulating pig-
ments which prevent the flow of current, even when the film of
lynoxyn, in which they are enveloped, is to some extent damaged,
will tend to check the corrosion of painted metal surfaces.

If. the immunity from rust, of painted surfaces, depends mainly
upon the inhibitive properties of the pigments with which the
steel is coated, it is fair to believe that this immunity from attack
will be reinforced by providing that the principal base content
of the pigment shall be a good non-conductor of electricity. In
addition to the nature of the pigment, the character of the vehicle
must also be considered. Inhibitives in liquid form may be in-
corporated with the vehicle, and thus a double protective effect
produced. The use of a paint inhibitive in both pigment and
vehicle antagonizes the forces that cause corrosion, and the
greater the inhibitive value of the pigment and the vehicle, the
greater will be the defensive action working against the factors
which produce corrosion and rapid decay.

Water-shedding and Excluding Properties of Pigments. — A
careful study of the nature of paint films has shown that some
pigments produce and lend peculiar properties to the oil in which
they are enveloped, which give to the resulting films an excluding
or a water-shedding nature. An excluding paint is one that has
the property of excluding and preventing the admittance of
water to the underlying metal, thus protecting the steel from
the water, which is the essential medium in which corrosion can
alone take place. A water-shedding paint is one which, because
of its peculiar physical character, appears to show a certain
inability to become wetted. Plates painted with pigments of
the latter class dry very soon after rains, while paints made of
another type of pigments retain upon the surface drops of rain
which tend to be absorbed by the paint coating and ultimately
work through to the metal. The porosity which is generally
noted in every dried paint film may be prevented by the use of
gums, so that a coating is formed which is a good excluder, but

172 CORROSION AND PRESERVATION OF IRON AND STEEL

such coatings are not always the best water shedders. The size of
particle of the various pigments used, and their adaptability for
filling the pores of the film, is an extremely important considera-
tion in the composition of a paint which will form films, both
excluding and water-shedding. Linseed oil when used by itself,
will be found to possess neither excluding nor moisture-shedding
properties. The film will soon become tacky, and show a peculiar
blistered appearance, which indicates the spots where moisture
has penetrated, and results in the formation of a soft, easily
disintegrated film.

It would appear from the conductivity experiments which
have been described in a previous paragraph that the carbona-
ceous pigments, which sometimes are good excluders, often fail
to serve their purpose in the end, for the unctuous nature which
has given them their water-shedding value sooner or later dis-
appears, and the ultimate break-down of the coating admits
moisture.

That other authorities have accepted this explanation of the
authors is shown in part by the following citation from Toch: 1
"It is universally admitted that the presence of water is necessary
before corrosion can take place, since this is the medium which
contains or supplies the hydrogen ions which are necessary for
interchange with the iron. Thus, in order to prevent rusting,
you must either exclude water entirely, or have some substance
present which will prevent the formation of the hydrogen ions.
The only way in which this problem can be solved is by the appli-
cation of a paint which will either protect the surface by being
absolutely impervious to water, and therefore resistant to the
hydrolizing action, or one which contains in itself either as a
part of the vehicle or pigment some substance which can pro-
duce the passive state."

Moisture Penetration Tests of Paint Films. — The penetration
of moisture into and through paint films is evidently to some
extent dependent upon the peculiar nature of the pigments con-
tained in the paint. Some paint films have the power of pre-
venting the admission of water, while others containing different
pigments seem to easily permit the entrance of moisture. A
series of tests were designed by one of the authors, to determine
the water-excluding power of a number of typical pigments,
1 Trans. Amer. Electro-chem. Soc. XIV, 1908. 210.

RELATION OF PIGMENTS TO THE CORROSION OF IRON 173

when ground in oil and made into films. It has not been found
possible up to the present time to devise a method of making
films of absolutely even gauge. In the test which is about to
be described there may have occurred a very slight difference
in the thickness of the films.

A series of small glass bottles with wide mouths, holding about
two ounces, were half filled with concentrated sulphuric acid,
and paint films were tightly sealed over the mouths of the bottles,
using Canada balsam. The bottles were then carefully labeled,
numbered, accurately weighed on chemical balances, finally
exposed to air saturated with moisture and kept at constant
temperature under a large glass receptacle. The bottles were
removed from the receptacle every week and weighed. The
increase in weight, due to the amount of moisture which had
penetrated the films, and which had been taken up by the sul-
phuric acid, owing to its hygroscopic nature, was thus determined.

In another series of bottles, lumps of calcium chloride were
substituted for the sulphuric acid. The results obtained from
these tests corresponded to those of the former series, and led
to the conclusion that linseed oil films are undoubtedly porous,
and also that certain pigments have the property of reducing
this porosity to a greater extent than others.

As has been pointed out, however, this test cannot be used
as an accurate measure of relative permeability, owing to the
difficulty of making films of definite and constant thickness.
The following table gives the results of the test:

MOISTURE EXPERIMENTS.

Figures Given Express Gain in Weight, e.g., Water Absorbed

7 days

14 days

21 days

28 days

35 days

49 days

Iron Oxides (with 2 per cent. Zinc
Chromate and 2 per cent. Gum)

White Lead, D. D

White Lead and Zinc Oxide

China Clay

Whiting

Zinc Oxide, Barytes and Blanc Fixe

Zinc Lead White

Red Lead

Basic Sulphate' — 'White Lead

0.032
0.040
0.043
0.044
0.044
0.048
0.049
0.049
0.049

0.048
0.078
0.081
0.086
0.079
0.092
0.095
0.092
0.092

0.072
0.111
0.115
0.122
0.114
0.125
0.130
0.130
0.128

0.092
0.162
0.163
0.182
0.167
0.183
0.181
0.187
0.185

0.110
0.187
0.192
0.219
0.197
0.190
0.211
0.215
0.213

0.140
0.264
0.266
0.317

0.277
0.290
0.284
0.295
0.292

174 CORROSION AND PRESERVATION OF IRON AND STEEL
MOISTURE EXPERIMENTS. — (Continued)

7 days

14 days

21 days

28 days

35 days

49 days

Zinc Oxide and Whiting

Zinc Chromate

Barytes and Zinc Oxide

Zinc Oxide

Calcium Sulphate

American Vermilion

White Lead, Barytes and Blanc Fixe

Barytes

Willow Charcoal

Lithopone

Carbon Black

Lead and Zinc Chromate

Chinese Blue (Stimulative)

Venetian Red

Natural Graphite

Medium Chrome Yellow

Bright Red Oxide

Barium and Zinc Chromate .......

Ultramarine

Prussian Blue (Inhibitive)

Raw Linseed Oil

Lampblack

Blanc Fixe

0.060
0.064
0.064
0.065
0.066
0.069
0.074
0.074
0.077
0.083
0.084
0.086
0.092
0.093
0.104
0.106
0.116
0.116
0.119
0.125
0.143
0.199
0.210

0.110
0.121
0.118
0.122
0.140
0.140
0.137
0.138
0.154
0.156
0.168
0.161
0.185
0.190
0.223
0.207
0.240
0.211
0.230
0.246
0.300
0.411
0.472

0.156
0.176
0.169
0.172
0.212
0.202
0.200
0.202
0.236
0.228
0.250
0.226
0.276
0.279
0.350
0.300
0.365
0.298
0.336
0.361
0.449
0.641
0.744

0.221
0.270
0.240
0.244
0.313
0.311
0.294
0.298
0.378
0.332
0.391
0.319
0.405
0.418
0.539
0.429
0.548
0.430
0.484
0.521
0.679
1.033
1.144

0.256
0.298
0.278
0.285
0.377
0.349
0.344
0.336
0.459
0.380
0.448
0.369
0.470
0.508
0.632
0.505
0.662
0.481
0.578
0.619
0.803
1.234
1.414

0.352
0.417
0.386
0.391
0.555
0.501
0.490
0.466
0.694
0.550
0.654
0.497
0.671
0.770
0.951
0.725
0.976
0.660
0.814
0.733
1.201
1.873
1.944

The pigments appear in this table in the diminishing order
of their excluding values.

The action of a high-grade gum, in sealing up the pores and
filling the interstices of a vehicle, is also indicated by this test.
It will be found that iron oxide used alone appears about half-
way down the list, while iron oxide, containing 2 per cent, of
gum in addition to the linseed oil, was found to be an excellent
excluder. The apparatus used is shown in Fig. 40.

Industrial Application of the Theory of Inhibitors. — That the
manufacturer is rapidly coming to the realization of the value
of the new theory of rust inhibition, as outlined in the preceding
chapters, and is adapting it to commercial practice, is shown by
the change that has been made in the formulas of several pro-
tective paints now on the market. Heckel 1 has said: "Out of

Railway Age Gazette.

RELATION OF PIGMENTS TO THE CORROSION OF IRON 175

this mass of tentatively accepted facts has developed a provisional
theory along which the more advanced manufacturers are now
engaged in working out a new mode of procedure in the painting
of steel.

"The theory is that rust-stimulating pigments should never
be placed in contact with the steel surface, but that an inhibitive
priming coat should always intervene. This inhibitive coating
may be suitably compounded of the chromes, zinc oxide, white

Fig. 40.— Apparatus for Testing Permeability of PaintFilms.
(Gardner.)

lead, red lead, willow charcoal, etc., among the inhibitors, or of
any of the neutral or indeterminate pigments reinforced with a
small proportion of the stronger inhibitors, such as zinc chrome,
zinc oxide, zinc and lead chrome, etc.

"Over this priming coat the air — and moisture — excluding
coats can then be safely applied; these coats being designed for
protection only, with regard to inhibitive qualities/!.

Commenting on the action of gums in rendering paint films
less porous, Perry states: 1 "Turning, therefore, to the conserva-

1 Coatings for Conservation of Structural Materials. Bulletin No 14
Scien. Soc, Paint Mfrs. As/n. of U.S.

176 CORROSION AND PRESERVATION OF IRON AND STEEL

tion of structural iron and steel, and to its rust inhibition through
particular coatings, we have the problem of choosing the proper
materials for manufacturing a paint which will both exclude the
agencies of rusting, and which, when moisture and gases do
penetrate the coating, will inhibit the iron from rusting; and we
also have the problem of giving to the chemist, engineer, and
architect some simple method of determining whether any given
paint is, in at least rough measure, harmful, safe, or beneficial.

"In the case of structural steel we have a condition of the
surface to be protected, essentially different from the conditions
existing with lumber. Lumber, no matter how well kiln dried,
invariably has some moisture in its structure. It has been abso-
lutely demonstrated, and all technical men acquainted with the
subject are agreed, that linseed oil, when dried to a coat of linoxyn,
has voids or pores in it. The paint manufacturer uses a com-
bination of pigments which largely fill these voids, but the most
carefully prepared coat of paint for lumber, when the vehicle is
linseed oil, without the addition of varnish, is always somewhat
porous, and the moisture has some opportunity to pass out with-
out blistering, etc.

"In the case of structural steel we have no moisture in the
structural material to contend with, 1 as we have with lumber,
and therefore we can safely add varnish gums in solution to the
linseed oil, thus producing a more nearly water-proof coat to
the agencies of decay. The action of the gum, diffused throughout
the linseed oil by solution, is to help to fill up or fuse together
the pores or voids, or, in other words, to render the linoxyn film
more resistant to the entrance of moisture through the paint
coating to the structural material. This is the conclusion reached
from the standpoint of technical research, and practical experi-
ence has also demonstrated that those coatings which have been
proved the best protection for steel have invariably contained a
vehicle in which some gum or varnish material has been present."
In the above connection, Toch's 2 views on the subject of the

1 With this statement the authors are not in complete agreement, as it
has been contended that more or less moisture is always absorbed on the sur-
face of iron and furnishes one of the many difficulties met with in the attempt
to provide a perfect coating on iron. This point has been discussed in
another chapter.

2 M. Toch. The Chemistry and Technology of Mixed Paints, p. 87. D.
Van Nostrand Co., New York, 1907.

RELATION OF PIGMENTS TO THE CORROSION OF IRON 177

physical characteristics of linseed oil films, and the part they
play in corrosion, should be added: "There are questions in
regard to the physical and chemical characteristics of linseed oil
on which there has been considerable discussion and naturally
a difference of opinion. The first is whether linseed oil dries in
a porous film, and the second is whether linseed oil while drying
goes through a breathing process and absorbs oxygen, and gives
off carbonic acid and water." With reference to the porosity of
the dry film of linseed oil, the same authority has stated: 1 "In
a paper before the American Chemical Society on March 20,
1903, I gave it as my opinion that a dried film of linseed oil is
not porous, excepting for the air bubbles which may be bedded
in it, but that any dried film of linseed oil subjected to moisture
forms with it a semi-solid solution, and the moisture is carried
through the oil onto the surface of the metal. We then have
two materials which, beyond a doubt, have sufficient inherent
defects to produce oxidation under the proper conditions, and
granted that the percentage of carbon dioxide in the air of the
tunnel is not beyond the normal, the fact that carbon dioxide,
together with moisture, would cause this progressive oxidation is
sufficient warrant for the discontinuance of paints that are not
moisture and gas proof. Lewkowitsch demonstrated in his Can-
ton lectures that the fats and fatty oils hydrolized with water
alone, and linseed oil is hydrolized to a remarkable degree in eight
hours when subjected to steam. It can, therefore, be inferred
that water would act on linseed oil without the presence of an
alkali, and calcium hydroxide added to water simply hastens the
hydrolysis, by acting as a catalyzer. This, then, bears out my
previous assertion that a film of linseed oil (linoxyn) and water
combine to form a semi-solid solution similar in every respect to
soap, and inasmuch as we have lime, lead, iron, and similar bases
present in many paints, it is almost beyond question that these
materials aid in the saponification of oil and water."

In summing up the discussion of the possible porosity of
paint films as presented above, the authors can only reiterate
that the results of their experiments appear to them to furnish
actual proof of the permeability of paint films by moisture.
Whether or not the result is due to porosity must depend upon

1 Journal of the Society of Chemical Industry (May 31, 1905), "New Paint
Conditions Existing in the New York Subway."

178 CORROSION AND PRESERVATION OF IRON AND STEEL

just what is meant by that term. The formation of a solid or
semi-solid solution must be dependent on the existence of molec-
ular interstitial spaces. Whether spaces of these orders of mag-
nitude should be called pores is quite beside the question as the
final result is the same in either case.

Thompson's Work on Solubility of Paint Films. — That the
solubility in water of various paint coatings must also be taken
into consideration, as well as the inhibitive and excluding proper-
ties, seems to be indicated by the work of Thompson. 1 In a series
of tests of oil films immersed in water, Thompson found that
calcium sulphate was dissolved in great quantity from films
containing this pigment. Thompson's method of conducting the
solubility tests, as described by him, is as follows: "The paint
is coated on antiseptic gauze in the following manner: Strips of
gauze 2 in. wide and about 15 in. long are dipped in the paint for
about 13 in., then drawn through a wringer consisting of two
test tubes (J by 6 in.) drawn together by elastic bands. Gauze
having these dimensions weighs about 1.25 grams. The weight
of each piece of gauze is previously obtained, and after it has
been dipped in the paint it is placed on a clock glass. The
clock glass and its contents are then weighed, from which, by
deducting the weight of the clock glass and any adhering paint,
we obtain the amount of paint which has been placed on the
gauze. This gauze is then hung up to dry for about ten days
with a tared clock glass under it, so that paint drippings may
be caught and weighed and deducted from the amount of paint
taken. After ten days the gauze and paint are weighed, and if
desired the percentage gain in weight may be calculated. The
painted gauze is then placed in a 150 c.c. beaker and covered
with about 130 c.c of distilled water. After 24 hours this water
is poured off and evaporated and another 130 c.c. of water is
placed on the gauze. The water poured off is placed in a tarred
vessel, so that after evaporation to dryness the amount dissolved
can be determined by weighing. This dissolved matter can then
be analyzed for pigment constituents or treated with nitric acid,
and carefully ignited and weighed. The treatment with water
can be repeated as often as may be desirable, and, if it is thought
best, the extract from several treatments can be combined and

1 G. W. Thompson. Certain Solubility Tests on Protective Coatings.
Froc. Amer. Soc. Testing Materials, 1908, Vol. VIII, p. 601.

RELATION OF PIGMENTS TO THE CORROSION OF IRON 179

treated as one." The results obtained by this test have brought
out some interesting points. It has shown, for instance, that in
paints compounded of different pigments, some of these may be
dissolved out by the action of water. As a specific instance of
this, the case of Venetian red may be cited, which contains a
large percentage of calcium sulphate. Such a pigment would
prove dangerous to use on iron or steel, on account of the solubility
of the calcium sulphate.

The tables previously presented in this chapter, in which is
given the list of inhibitive pigments, and also the list of excluding
pigments, should prove of great value to the manufacturer in
aiding him to select the best materials for the composition of
paints for the protection of iron, and it is hoped that the tests
described will be given full consideration in conjunction with the
more practical field tests which are now being conducted.

The architect and the engineer are constantly clamoring for
protection for the enormous structural work under their super-
vision. More attention than ever is being paid to recent investi-
gations on the subject of paints, and better materials for painting
purposes are being insisted on. It is fair to believe that those
in charge of engineering work will demand in the future vital
improvements in steel protecting compounds, based on modern
knowledge of the problem. As has been previously pointed out,
the day of empiricism in the selection of protective coatings has
passed, and the subject must now be considered from the stand-
point of scientific investigation based upon a satisfactory working
theory. The treatment of steel surfaces by the metallurgist,
either by physical or chemical means, may soon be developed to
such an extent that metal will be made almost non-corrodible,
but until such a day comes there will be an immense demand for
protective coatings that will protect, and too much attention
and study cannot be given to the subject.

CHAPTER VIII

RECENT FIELD TESTS ON PROTECTIVE COATINGS
FOR IRON AND STEEL

Steel Test Panels at Atlantic City. — The results obtained by
recent investigators on the inhibitive value of pigments, as de-
scribed in the previous chapter, suggested the erection of a series
of steel panels on which could be tested the same range of pig-
ments under the practical conditions of service in the open. The
Paint Manufacturers Association of the United States offered
to erect the panels, and place them under the supervision and
inspection of Committee "U" of the American Society for Testing
Materials. It was decided to erect the panels near Atlantic City,
N. J., and the work was commenced during the fall of 1908, at
this place. In order to broaden the test, and furnish data in regard
to the durability of the various paint films when placed on differ-
ent types of metal, three kinds of steel were used on the panels.

The following table shows the analyses of the three grades of
metal tested:

Bessemer

Open Hearth

.08

.16

.08

.02

.05

.024

.35

.44

Extra Mild
Open Hearth

Carbon
Phosphorus
Sulphur . . .
Manganese .

.03

.005

.024

trace only

Pickling and Preparation of Plates. — The three types of metal
selected for the test were rolled to billets, the middle of which
were selected, and worked up into plates 24 in. wide, 36 in. high,
and i in. in diameter — approximately 11 gauge. A number of
plates of each of the metals selected, in all 450, were pickled in
10 per cent, sulphuric acid, kept at 180 to 200° F., in order to
remove the mill-scale. The plates were then washed in water,
and later in a 10 per cent, solution of caustic soda. Finally the
plates were again washed in water and wiped dry. They were

1 This metal is called "Ingot Iron' ; by the manufacturers.
180

RECENT FIELD TESTS ON PROTECTIVE COATINGS 181

then packed in boxes containing dry lime, in order to prevent
superficial corrosion. By this method the plates were secured
in perfect condition, the surfaces being smooth and free from
scale. Upon these pickled plates, paints were applied with a
definite spreading rate of 900 sq. ft. per gallon. The unpickled
plates, coated with mill-scale, were painted with the same paints,
but without adopting any special spreading rate, thus following
more closely the ordinary method of painting structural steel.
A few extra plates of special Bessemer steel and Swedish charcoal
iron were also included in the test, some of which were painted,
while others were exposed without any protective coating. Plates
of the three types of metal already mentioned were also exposed
unpainted, both in the black and pickled condition.

Fence Erection and Preparation for Work. — The fences which
were erected for the holding of the plates were constructed of
yellow pine, the posts being set deeply in the ground and properly
braced. The framework of the fence was open, with a ledge
upon the lateral girders, upon which the plates might rest, and
to which the plates were secured by the use of steel buttons.
After the framework had been erected, painted, and made ready
for the placement of the panels, a small shed was built upon the
ground, and the materials for the field test placed therein. The
steel plates were unpacked from the boxes in which they were
shipped, brushed off, and stacked up ready for painting. Small
benches were erected, and the accessories of the work, such as
cans, brushes, pots, balances, etc., were placed in position.

Method Followed in Painting Plates. — A frame resting upon
the work bench served to hold the plates in a lateral position
while being painted, room being allowed beneath the plate for
the operator to place his hands in order to lift the plates from
the under surface, after the painting had been finished.

A pickled plate having been placed upon the framework,
everything was in readiness for the work. The specific gravity
and weight per gallon of the paint to be applied was determined,
and the amount, in grams, to be applied to each individual panel
was calculated according to the following formula:
Spreading rate sq. ft. in plate grams paint in gal.

900 sq. ft. : 6 : : 4500 : x

The reciprocal of x being the number of grams of paint to be
applied to the panel.

182 CORROSION AND PRESERVATION OF IRON AND STEEL

An enamel cup was then filled with the paint and a brush
well stirred within. The cup, paint, and brush were placed upon
the balances and accurately weighed in grams. After most of
the paint had been applied to the panel, cross-brushing of the
panel was continued until the pot with brush and paint exactly
counterbalanced the deducted weight. The painted panel was
then set in a rack, in a horizontal position, to dry.

A period of eight days elapsed between the drying of each
coat. The greatest care was taken in the painting of the edges
of the plates, and the racks for containing the plates after they
were painted were so constructed that the paint would not be
abraded while sliding the plates back and forth. The working
properties of each paint, and the appearance of the surface of
each plate after painting were carefully noted and included in
the report. No reductions were made to any of the paints applied
except in three cases, where the viscosity was so great that it was
necessary to add a small amount of pure spirits of turpentine.
The amount of paint was proportionately increased in such cases,
so that the evaporation of the turpentine would leave upon the
plate the amount of paint originally intended.

The appearance of the completed series of test panels is shown
in Fig. 41.

Vehicles Used and Reasons for Avoidance of Japan Driers. —
The pigments used were selected with the view to securing as
nearly as possible purity and strength, and as already noted were
out of the same lots used in making the preliminary laboratory
tests on inhibitives. They were ground in a vehicle composed
of two parts of raw linseed oil and one part of pure boiled oil.
Paint is generally caused to dry rapidly by the use of japans or
driers. These materials contain a large amount of metallic
oxides which might have some effect in either exciting or retard-
ing corrosion. To prevent the introduction of such a factor,
these materials were not used in the test. The boiled oil, with
its small precentage of metallic oxides, was sufficient, however,
to cause the paints to dry in a short time after they were spread.

The table on page 184 gives the number of pounds of pigment
to a gallon of oil, used in the manufacture of the pigment paints
for this test. The formula for determining the amount is as fol-
lows:

Sp. Gr. of pigment X 3 = lbs. pigment to gal. oil

RECENT FIELD TESTS ON PROTECTIVE COATINGS 183

184 CORROSION AND PRESERVATION OF IRON AND STEEL

The excessive amount of oil required to grind pigments Nos.
19 and 21 made it necessary to introduce a percentage of barytes
(barytes requiring very little oil), in order to form paints which
would contain relative amounts of oil.

Pig-
ment

No,

Name

Sp. Gr. of
Pigment

Weight of

Pigment to

Gal. Oil

Weight of

Paint
per Gal.

24
27
2S
29
30
31
32
33
34
36
39
40
41
44
45
48
49
51

Dutch Process White Lead

Quick " " "

Zinc Oxide

Sublimed White Lead

Blue Lead

Lithopone

Zinc Lead White

American Orange Mineral

Red Lead

Bright Red Oxide

Venetian Red

Prince's Metallic Brown

Natural Graphite

Acheson Graphite

( Lampblack

( Barytes

Willow Charcoal

J Gas Carbon Black

\ Natural Barytes

French Yellow Ochre

Natural Barytes

Precipitated Barytes (Blanc Fixe) . .

Calcium Carbonate (Whiting)

Calcium Carbonate Precipitated

Calcium Sulphate (Gypsum)

China Clay (Kaolin)

Asbestine (Silicate of Magnesium) . .
American Vermilion (Chrome Scarlet)

Medium Chrome Yellow

Zinc Chromate

Zinc and Barium Chromate

Chrome Green (Blue tone)

Prussian Blue (Stimulative)

Prussian Blue (Inhibitive)

Ultramarine Blue

Zinc and Lead Chromate

Magnetic Black Oxide

6.83

6.78

5.56

6.45

6.39

4.26

4.42

8.97

8.70

5.26

3.1

3.17

2.60

2.21

1.82

1.49
1.85

2.94
4.46
4.23
5.48
2.56
2.33
2.67
2.75
6.83
5.88
3.57
3,45
4.44
1.96
1.93
2.40
4.76

Lbs.

20.49

20.34

16.68

19.17

12.78

13.26

26.91

26.10

15.78

9.30

9.51

7.80

6.63

1.82

8.92

4.47

1.39

10.03

8.82

13.38

12.69

8.22

7.68

6.99

8.01

8.25

20.49

17.64

10.71

10.35

13.32

5.88

5.79

7.20

14.28

Us.
20.4
20.56
17.68
19.65
20.1
15

16.37
24.74
24.4
17.07
12.6
12.49
11.4
10.2

13.32

8.99

13.94

12.16
15.24
15.32
11.41

11.24
10.41
11.19
11.49
21.36
18.8
13.07
13.16
16.16
9.95
9.76
10.74
15.99
15.99

RECENT FIELD TESTS ON PROTECTIVE COATINGS 185

Note. — The oil used in grinding all these pigments and
used throughout the test was composed of two parts raw linseed
oil and one part boiled linseed oil. The specific gravity of this
mixture was .937.

Data of application of each pigment was accurately made
during the tests. There is given herewith the data on four of
the pigments applied.

No. 2, Quick Process White Lead:

Sp. Gr. of Pigment 6.78

Lbs. to Gallon Oil 20.34

Sp. Gr. of Paint as Received 2.47

Wt. of Paint per Gallon 20.56

Grams to Panel 62

Condition of Paint Good

Working Properties Works easy

Drying 24 hours all coats

1 Coat Oct. 26 T 60 B 29.94 W fair

2 Coat Nov. 3 T 54 B 30.23 W clear

3 Coat, Nov. 7 T 52 B 29.66 W cloudy

No. 9, Orange Mineral (American) :

Sp. Gr. of Pigment 8.97

Lbs. to Gallon Oil 26.91

Sp. Gr. of Paint as Received. . . 2.97

Wt. of Paint per Gallon 24.74

Grams to Panel 74.7

Condition of Paint Good

Working Properties Smooth — no brush marks.

Drying Good

1 Coat Oct. 28 T 58 B 30.01 W cloudy

2 Coat Nov. 4 T 65 B 29.61 W cloudy

3 Coat Nov. 9 T 58 B 29.91 W clear

No. 14, Venetian Red:

Sp. Gr. of Pigment 3.1

Pigment to Gal. Oil 9.30

Sp. Gr. of Paints Received .... 1.52

Wt. of Gallon Paint 12.6

Grams to Panel 38

Condition Good Smooth Paint

Working Properties Smooth — no brush marks.

Drying Good on all coats

1 Coat Oct. 29 T 56 B 29.82 W cloudy

2 Coat Nov. 4 T 65 B 29.61 W cloudy

3 Coat Nov. 9 T 65 B 29.91 W clear

186 CORROSION AND PRESERVATION OF IRON AND STEEL

No. 19, Lampblack:

Sp. Gr. of Pigment 1.82

Lbs. to Gallon Oil: Lampblack ... 1.82

Barytes 8.92

Sp. Gr. of Paint as Received 1.60

Wt. per Gallon Paint 13.32

Grams to Panel 40.2

Condition Good

Working Properties Works fair

Drying Slow. Tacky after 3 days.

1 Coat Oct. 29 T 56 B 29.82 W cloudy

2 Coat Nov. 5 T 54 B 29.92 W clear

3 Coat Nov. 10 T 60 B 30.02 W clear

Note. — T stands for temperature (F)
B " " barometer

W " " weather

Testing Effect of Various Prime Coats. — Some of the special
tests made, included a series of plates prime-coated with different
inhibitive pigments, and these tests were designed to determine
which pigments offer the best results for such work. These
plates were all second-coated with the same paint. It is the
opinion of the authors that any good excluding paint may be
used whether it is inhibitive in action or not, provided the
contact coat is inhibitive. If, however, both coats can be
designed so as to have the maximum possible value from both
these points of view, the best results would, of course, accrue.
The only way such data can be obtained is by careful observation
of the results of exposure tests.

Combination Formulas Tested. — By selecting a series of pig-
ments which in the water tests showed inhibitive tendencies, and
properly combining these pigments into a paint, it was thought
possible that a more or less inhibitive paint would be produced.
If this proved to be the case, it would follow that the selection
and introduction into a paint of the stimulative pigments would
inevitably produce a paint unfit for use on iron or steel. Formulae
of both types were therefore used in these tests, and below the
percentage composition of two of these is shown.

RECENT FIELD TESTS ON PROTECTIVE COATINGS 187
Inhibitive White.

Pigment Formula Paint Formula

35 per cent Zinc Oxide 20.90 per cent.

45 per cent Special White 26.87 per cent.

5 per cent CaC0 3 2.98 per cent.

15 per cent Silex 8.95 per cent.

Japan 1.56 per cent.

Raw Oil 38.74 per cent.

Stimulative Black.

Pigment Formula Paint Formula

40 per cent Lampblack 8.18 per cent.

40 per cent Natural Graphite . . 8.18 per cent.

20 per cent Barytes 4.09 per cent.

Japan 8.33 per cent.

Raw Oil 71.22 per cent.

Drawing Conclusions from Results of Field Tests. — At the
present writing, the Atlantic City test plates have been in place
fourteen months, and the authors do not consider that suffi-
cient time has elapsed to justify definite conclusions or sweeping
deductions. In a number of cases, however, the panels have actu-
ally failed, and the corrosion effects are apparent to all observers.
The authors fully recognize and desire to record here that in
spite of the fact that they were the principal responsible persons
in charge of the design and erection of these tests, nevertheless,
in view of the fact that no official report of results obtained has
yet been made to the American Society for Testing Materials,
it would not be proper to enter upon a discussion of the relative
durability of the different types of formulae, even if definite results
had begun to show.

It is, however, fair to record the results of scientific examina-
tions of the paint surfaces, which have been made from time to
time, leaving the conclusions which will naturally follow from
these observations to be drawn later. The object of the authors
in presenting a detailed description of the design and construc-
tion of these fences has been solely to act as a guide for other
experimenters who may desire at some future time either to
duplicate or develop this form of field testing. It is by no means
easy to carry out work of this kind which involves a very large
amount of thought and labor. The experience gained by one
set of experimenters will, of course, prove valuable to others
who are pursuing the same line of research.

188 CORROSION AND PRESERVATION OF IRON AND STEEL

The Inspection of Painted Surfaces. — In making an inspection
of painted surfaces to determine the durability of various coatings,
the chalking, checking, color, maintenance, elasticity of film,
and other properties should be noted.

By the term chalking is meant that condition of the film
which allows removal of the pigment when rubbed with the hands
or with cloth pads. Destruction of the oil, through the saponi-
fication caused by alkaline pigments, is often the primary cause
of chalking.

By checking is meant that condition of the painted surface
that shows checks, cracks, alligatoring (an appearance resembling
alligator skin), etc. Checking may be seen very clearly by using
a small but powerful magnifying glass. Surface, matt, coarse
and deep are qualifying terms used to define the character of the
checking which takes place. Paints which are not elastic and
which become brittle are subject to checking, as well as some
paints which chalk readily.

Elasticity of film is noted by raising a portion of the film with
the knife blade and peeling it off. Only very elastic paint will
stand this test. Brittle coatings crumble up at once when this
attempt is made. This test cannot, of course, be applied to the
Atlantic City panels, but is in general use by car painters.

Hardness of coating can be gauged to a certain extent by the
appearance of the paint film and its resistance to abrasion. A
small hard object is sometimes used to strike against the painted
surface to determine its hardness. The hiding power and color
maintenance of paints are also important features to consider
in making an inspection.

Continuity of coating and absence of disintegration, as well
as other conditions of the paint film, is best determined by examina-
tion with a microscope. This can be done by using the photo-
microscope designed by one of the authors. This apparatus has
proved of great value in field inspection work. Permanent
records of the wearing of a paint from time to time can be
obtained by using this instrument, and the progressive decay
shown, determined with accuracy.

The photomicroscope consists of a modified form of a Gordon
photomicrographic tube such as is used for plate exposures in
laboratory work. This tube contains a projection lens properly
focused, and also an exposure shutter fixed on a lift pin. On the

RECENT FIELD TESTS ON PROTECTIVE COATINGS 189

rear end of the tube is fitted a disk of metal, into which is placed
a block of wood having a central annular opening the size of the
tube. On the back of the wood is firmly set and screwed into
position a film-pack holder, such as is used for the ordinary photo-
graphic camera. The arm and body of a microscope, containing a
draw tube fitted with objective and eye-piece, is mounted in a hori-
zontal position on a solid iron base, bored and threaded to receive
a screw from the top of a heavy tripod. The objective of the micro-
scope is placed close to the painted surface, and by raising or lower-

Fig. 42. — Photomicroscope for examining and photographing painted
surfaces. (Gardner.)

ing the tripod the microscope may be focused on any spot, by
the regulation of the coarse adjustment. Minute abrasions or
cracking of any nature can easily be detected. When all is ready,
the tube camera is placed directly over the eye-piece of the micro-
scope, and exposure is made by lifting the shutter cap for twenty
or thirty seconds, according to the light conditions. No artificial
illumination is necessary. The apparatus is shown in Fig. 42.

190 CORROSION AND PRESERVATION OF IRON AND STEEL

In using the photomicroscope on the Atlantic City panels,
one of the first effects shown was the appearance of minute crys-
tals of sodium chloride deposited on the films from the saline
atmosphere due to the close proximity of the ocean.

A very interesting point brought out by these examinations
is the fact that some of the paint films appeared to have the
power of holding the salt crystals to a much greater degree than
others, some, indeed, being almost free from salt. A sufficient
amount of work has not yet been done along this line to enable
definite conclusions to be drawn, but some of the results obtained
are shown in order to stimulate further inquiry. In Figs. 43 a,
b, and c are shown photomicrographs of three different paint coat-
ings, to illustrate this point.

The tendency of certain paint coatings to produce a slight
wrinkling of the surface when softened by moisture is well known.
This action, which is probably preliminary to a breaking down
of the paint film, is shown in Fig. 44.

As has been pointed out in a previous paragraph, certain
pigments rapidly destroy the oil films and produce the phenomenon
known as chalking. The appearance of chalked paint films is
very characteristic under the photomicroscope. The effect is
very well shown by Figs. 45 and 46.

The cracking of paint films is a most important point in rela-
tion to the general subject of the protection of iron and steel by
paint coatings. It requires no evidence to prove that the open-
ing up of cracks on the surface will allow water, and the other
corroding influences, to make a ready entrance to the surface of
the metal, and act as centers of corrosion.

The photomicrographic method of studying paint films will
undoubtedly lead to more definite information on this point
than has been heretofore available. A study of the illustrations,
as shown in Figs. 47, 48, and 49, indicates that cracks in different
paint coatings are of a different nature and appearance.

It is well known that although zinc oxide has many properties
which make it an extremely valuable pigmentary substance, its
tendency to crack constitutes its greatest fault. The cracking
of a zinc oxide film is beautifully illustrated in Fig. 47. Another
peculiar form of cracking is shown in Fig. 49 on a gypsum coating.

Other types of cracking which are being closely followed by
corrosion along the line of the cracks are shown in Figs. 50, 51, 52.

RECENT FIELD TESTS ON PROTECTIVE COATINGS 193

The recent work of Walker, calling attention to the probability
that cracks in paint films are responsible for a stimulated corro-
sion effect owing to the depolarizing action of the open paint
film, due to the absorption of nascent hydrogen, is particularly
interesting when considered in relation to the effects shown in
these illustrations.

On Figs. 53 and 54 are shown two typical examples of check-
ing. This action, just as in the case of cracking, must be con-
sidered a preliminary step in the breaking down of a paint coat,
leading to corrosion. Fig. 55 illustrates the effect produced by
uneven brush work, leading to a special characteristic form of
checking which is frequently observed as a preliminary step in
the breaking down of paint coatings.

Figs. 56, 57, 58 exhibit in a very interesting manner the
breaking through of corrosion nuclei to the surface of the paint
film. It must be apparent that although a paint film may be
opaque, corrosion goes on underneath for a long time before
the trouble becomes evident on the surface. A time comes,
however, when little blisters make their appearance on the film.
This appearance is a preliminary step to the breaking out of the
tubercles of rust which are so familiar to all observers.

The rusting of iron has frequently been compared to a malady
or disease, and certainly these photomicrographs appear to show
a striking analogy with the effects produced by disease in living
tissue.

The authors have stated in a previous paragraph that it is
perfectly fair for every observer to call attention to specific cases
of success or failure exhibited by the Atlantic City test panels,
provided no sweeping conclusions are reached or deductions
drawn. While bearing this in mind, it serves our purpose to
present herewith some illustrations representing interesting con-
trasts which the tests have already served to show. The authors
do not hesitate to appear as defenders of the electrolytic theory
of corrosion, and therefore of the value in general of inhibitive
pigments. Fig. 59 illustrates the condition of a plate painted
with zinc chromate, a distinctly inhibitive pigment.

Fig. 60 shows the present condition of two plates painted
respectively with pigments, one of which is a distinct stimulator
of corrosion and the other of which saponifies and destroys the
protective film. An inspection of this illustration shows that

RECENT FIELD TESTS ON PROTECTIVE COATINGS 195

the plate painted with the gypsum is dark in color. This dark-
ness is due entirely to the corrosion which is plainly seen to be
taking place underneath the more or less transparent film. That
this same action is going on under many of the opaque films can
hardly be doubted, although it is not at the present time apparent
to the eye.

Fig. 59. — Perfect condition of plates painted with zinc chromate.

It is well known that calcium carbonate as a pigmentary
substance is used in two different forms — the natural, ground
chalk (or whiting), and the precipitated calcium carbonate. In
view of the fact that calcium carbonate is a useful ingredient in
small percentage in many paint formulse, it is interesting to

196 CORROSION AND PRESERVATION OF IRON AND STEEL

record the condition of the two plates which have served as a
test of the relative value of these two forms. An inspection of
Fig. 61, taken in conjunction with Fig. 62, in the same plate,
appears to indicate that the whiting or natural form of calcium
carbonate should be selected, if it must be used, for all protective
paints for iron.

Fig. 60. — On left, scaling and corrosion on plate
painted with precipitated calcium carbonate. On
right, corrosion showing through gypsum paint film.

Other Field Tests and their Value. — J. Cruickshank Smith, 1
in commenting on paints for the preservation of steel, makes an
appeal for more field tests such as those already described, in
the following words: "Although opinions vary as to what is the-
oretically the best pigment, or, more correctly, combination of
pigments, for the manufacture of protective coatings, the general
consensus of technical opinion seems to be gradually coming to
be that, with certain more or less obvious reservations, the com-

1 A Plea for International Investigation concerning Protective Coatings
for Iron and Steel. Paper submitted to 5th Congress, Internat. Assn. for
Testing Materials, Copenhagen, 1909.

RECENT FIELD TESTS ON PROTECTIVE COATINGS 197

position of the paint is less important than its physical state,
and that one pigment or blend of pigments is more efficient than
another as the basis for a protective paint mainly according to
the degree in which it tends to produce certain definite physical
and mechanical effects in the ultimate protective film. The
problem that remains is the selection of that blend of pigments,

Fig. 61. — Whiting (calcium carbonate) on steel.

suspended in a suitably adjusted vehicle (the paint as a whole
possessing the necessary physical properties), which will, under
known conditions as to exposure and climatic conditions, yield
a satisfactory protective film. Only exposure tests can solve
this problem.

"It is now generally accepted that such factors as fine state
of division and uniformity in size of the pigmentary particles,
viscosity of the paint, elasticity and strength of the paint film,

198 CORROSION AND PRESERVATION OF IRON AND STEEL

and impermeability of the paint film to moisture and gases, are
of paramount importance in determining whether or not a par-
ticular paint will satisfactorily effect the purpose of a protective
coating. In this connection, I would direct attention to the
admirable scientific work recently accomplished, and still being
conducted by the Scientific Section of the Paint Manufacturers'

Fig. 62. — Precipitated calcium carbonate on steel.

Association of the United States, and also by the various commit-
tees of the American Society for Testing Materials. The official
publications of these two bodies teem with technical matter,
which no one who claims to be up-to-date in his knowledge of
this subject can afford to be ignorant of."

Another practical service test of paint is being carried on by
Committee "E" of the American Society for Testing Materials
on a Pennsylvania Railroad bridge at Havre de Grace, Mary-

RECENT FIELD TESTS ON PROTECTIVE COATINGS 199

land. 1 In this test a number of protective paints, selected from
those largely advertised in the American market, were used.
The following outline of the tests will be of interest to the
engineer and bridge builder, who have the protection of similar
structures under their supervision.

Tests on Preservative Coatings Carried on under the Super-
vision of Committe "E" American Society for Testing Materials.

"1st. — Actual Service Tests under Normal Conditions to
be Applied to Full-Size Structures and Normal Panel Tests.

"Under this head the Committee started a long-time service
test in June, 1906, on part of the new double-track deck bridge
of the Pennsylvania Railroad over the Susquehanna River, at
Havre de Grace, Maryland. The bridge is about 4000 feet long,
consisting of 16 spans. The trusses are 30 feet deep, and the
panels 29 feet long. The paint was supplied by 19 manufacturers,
and applied to 2 continuous spans of the bridge directly over the
water, containing 19 panels. Each panel represented one paint
test.

"In addition to the test of the paint on the bridge proper, panel
tests were also made on very carefully prepared steel plates, 2
feet by 3 feet, the paint being applied by a most skilful workman,
under the most perfect conditions possible. These panels were
made in triplicate at 3 spreading rates, 600, 900, and 1200 square
feet per gallon of paint. The panels are exposed on the lower
chord of the bridge.

"Conditions of surface, application, weather, etc., were care-
fully noted and recorded. Complete analyses of the paints used
in the test were made. The paints used in this tests were, in
most cases, the representative proprietary coatings prepared for
use on iron and steel structures.

"The pigments included red lead, graphite, carbon, oxide of
iron, white lead, zinc oxide, sulphate of lead, together with clay,
silica, silicate and barytes. The oil was in the majority of cases
a pure linseed oil, though in one case rosin or a resinate drier
was detected.

"The volatile matter was found to be turpentine or naphtha,
or a mixture of the two.

1 Resume* of Work done by the American Society for Testing Materials
on Preservative Coatings for Iron and Steel, presented at the Annual Meet-
ing of the International Society, September, 1909.

200 CORROSION AND PRESERVATION OF IRON AND STEEL

"In the set was one heavy asphaltic varnish thinned with
naphtha.

"The result of the inspection made in June, 1908, is as follows:

"As was anticipated, no marked differences are noted in the
majority of cases. As was clearly indicated at the inspection a
year ago, the only example of an asphaltum varnish thinned
with a petroleum solvent has failed to a marked degree after
eighteen months' exposure. A carbon paint containing rosin in
the oil has developed minute fissure cracks all over the surface.
While it is by no means conclusive that this failure is due to the
rosin, still it is worthy of note, and will be the subject of further
investigation.

"An example of difference of expansion of red lead under-
coat and a carbon final-coat is shown in panel 14. The carbon
coat has cracked badly on the bridge proper in cobweb-like
cracks, showing the red of the under-coat. On another panel
the same final-coat, however, applied both as an under-coat and
as a final-coat, has not failed in this manner.

"It is also interesting to note that one of the pure red lead
pigments in straight raw linseed oil shows unmistakable evidence
of alligatoring.

"High gloss and tenacious film are shown by Panel 1. In
this case the pigment consists of oxide of iron, red lead, and a
carbon black, mixed with a varnish containing some gum and
thinned with turpentine.

"The other panels, as stated in the report of the sub-com-
mittee, show in most cases a marked loss of gloss with minor
differences in hardness, tenacity, toughness, and elasticity of
film. These differences are, however, too slight to warrant an
opinion at this time.

"There is evidence of rusting to a slight degree on all sec-
tions of the bridge, due to mechanical injury. These spots have
been accurately noted and will be carefully Avatched at subse-
quent inspections. In general, however, the paints are affording
good protection, and it will require longer exposure to differen-
tiate in the majority of cases.

"An inspection made March 26, 1909, confirms the report of
last June, with no marked changes of note to report. As stated
in June, one of the red leads showed unmistakable evidence of
alligatoring. At the inspection just made, the other sample of

RECENT FIELD TESTS ON PROTECTIVE COATINGS 201

red lead shows the same condition though to a less marked degree.
In both cases the alligatoring is most marked at the spreading
rate of 600 square feet to the gallon, but the cracking is apparently
superficial. Good protection is afforded, however, by both
paints.

"In general, most of the paints are affording good protection.

"It is- impossible to give details of the tests in this paper,
but reference is made to reports of Committee ' E ' in Proceedings
of the American Society for 1906, 1907, and 1908."

Considering the fact that nearly all these paint coatings are
reported to be in fairly good condition after about four years'
exposure, it would appear that the conditions of service at this
place are by no means severe. All the panels and sections are
located far down on the bridge structure where they cannot come
into direct contact with sulphurous gases from the locomotives.
It is more than probable that in some other location more definite
results of the tests would have been obtained in the time elapsed.

Tests of Paint Coatings Designed to Resist Sea Air. — The
attack of sea-water and a saline atmosphere upon paint coatings
is very severe. The soluble salts which are constituents of sea
water have the power of penetrating the hardest paint coatings
known, and acting to a certain extent the part of paint removers.
In the opinion of the authors, it is extremely important that the
paint designed to protect metal exposed to sea-water and sea-
air should contain inhibitive pigments. If such paints contain
pigments which are stimulative in nature, it is logical to assume
that in the presence of such easily ionized salts as the sea-water
contains, energetic galvanic action might proceed, with subse-
quent rapid pitting and corrosion. Moreover, the necessity of
an extremely good excluding vehicle for paints, designed for the
protection of marine structures, has been demonstrated by many
tests.

Along the various resorts on the New Jersey coast, thousands
of examples of the rapid decay of paint coatings applied to struc-
tural metal may be seen. On some of the old steel piers which
extend out to the ocean, ordinary coatings which have been
renewed almost yearly are in many cases destroyed after a few
months' exposure. The girders under the board walks, and the
railings above them, offer striking examples of the fugitive nature
of the ordinary paint coating when exposed to the action of salt

202 CORROSION AND PRESERVATION OF IRON AND STEEL

spray and salt air. Steel must be used in some places where
the use of concrete is not applicable, or where the present state
of development of the art of erecting concrete structural material
has not reached the necessary perfection. A combination of
steel and concrete would probably prove the solution of many
of the engineering problems at present met with, but the demand
for an efficient paint to withstand salt exposure is as great as
ever, and the tests which are under way may provide a solution
of the difficulty. The management of Young's Old Pier, at
Atlantic City, working in conjunction with the Carnegie Steel
Company, have recently erected a series of columns or piles on the
beach, under the supervision of one of the authors, in order to
determine the resistance to corrosion, and permanence of a com-
bination of one of the carefully designed types of modern steel
and cement construction. The test consists of a series of five
piers about 18 in. in diameter. The piers are made of steel sheet
piling locked together, filled with concrete, and sunk to two thirds
their depth in the sand, so that at low tide one third of the height
of the piling is subjected to the action of the salt air, while at
high tide the piling is completely covered. Details of construc-
tion are shown in Fig. 63. Previous to assembling and erecting
this piling, the steel parts were all carefully painted with different
paint formulae to determine which was best suited for the purpose.
The rise and fall of the tide, with the constant swirl of the sand
in its grasp, causes an enormous and unusual abrasion effect on
the paint coating, while the alternate exposure to the disintegrat-
ing effect of the salt water and the oxidizing effect of the air offers
one of the most severe tests to which a paint could be put.

On the ocean end of Young's Old Pier, at Atlantic City, the
United States Geological Survey, in cooperation with the authors,
has placed a number of test panels. Among these test panels
many grades of iron and steel, made by different processes and
containing different percentages of impurities, are represented.
Some of the plates are galvanized, some are tinned, and some are
left in their original condition as received from the mill, with the
ordinary coating of mill-scale. Inhibitive paint formulas have
been placed upon a few of the plates, to determine whether a paint
coating would be as serviceable as the tinned or galvanized plates
unpainted. The writers do not feel that they have the delegated
authority to analyze the results which are being obtained in this

RECENT FIELD TESTS ON PROTECTIVE COATINGS 203

-4

12 -35 Ib.U.S. Piling

Concrete

nt Grout

CROSS SECTION OF PILE

High Water

Sand

Experimental Steel Piling Piers.
Jp Young's Old Pier

/ | Atlantic City, N.J.

Pig. 63. — Details of Construction of Test Piling at Atlantic City, N. J.

204 CORROSION AND PRESERVATION OF IRON AND STEEL

series of tests, or to draw conclusions from them. They have been
briefly described here to make the record of contemporary tests

complete. The general appearance of these interesting tests is
shown in Figs. 64 and 65.

Many other tests have been made at Atlantic City, upon
the girders under the board walk, the steel piling under some of

RECENT FIELD TESTS ON PROTECTIVE COATINGS 205

the piers, and upon a series of plates exposed near the piers, and
within a few years an abundance of useful information will be

obtained on. the subject of protective coatings adapted for use
in the neighborhood of salt water.

CHAPTER IX

PAINTS FOR VARIOUS PURPOSES

The Importance of the Special Design of Paint Formulm. — One
single variety of paint can no more fulfil all the purposes for
which protective paints are used than can one kind of paint
serve all the demands for the proper decoration and preser-
vation of wood. The controversy in regard to mixed pigments
as against straight white lead house paints is analogous to that
which contrasts the mixed pigment with the straight red lead
prime coatings for structural metal. In a like manner the gum
varnishes used on the interior woodwork of houses correspond
in a sense to the lacquers so largely used on brass and other metals
for which protection is demanded. It may be said that there is
a paint for every purpose, in the same sense that there is a special-
ist in every field of scientific research and in every profession.

Throughout this book and in the tests that have been outlined,
we have been mainly considering those pigments which, mixed
in definite proportions, will afford the best protection in a general
way for bridges, structural framework for buildings, and all
work on a large scale. We have now to consider the painting
of special forms of metal which, because of certain peculiar
conditions which have to be met, present a special problem.

Galvanized Iron and Modern Methods Used in its Decoration
and Preservation. — The painting of galvanized metal has always
been considered by the master painter one of the difficult problems
that must be met. Very large quantities of galvanized iron are
used for siding and roofing purposes, especially in buildings and
factories designed to be fire-proof. Railings, cornice work,
gutters, drain pipes, fences and other accessories to the construc-
tion of buildings and the maintenance of property, are galvanized
and put into service without paint protection. Efficient protec-
tion is not as a rule provided by the galvanizing processes, if the
object is to preserve the metal for long periods of time, and paint-

206

PAINTS FOR VARIOUS PURPOSES 207

ing must be resorted to in order to erect another barrier against
the attacking forces which produce rapid corrosion and decay.
The peculiar greasy nature of a zinc-coated metal prevents the
proper adhesion of ordinary paints, so that peeling and blistering
result. In order to secure the proper adhesion of paint to gal-
vanized iron, several different methods of treatment of the metal
have been suggested. Acetic acid in the form of common vinegar,
acid salts, ammonia, sal soda, mixtures of copper salts with
muriatic acid and sal ammoniac, are in more or less general use
among painters, the object, of course, being to produce a slightly
roughened surface to which the paint will adhere.

That these preliminary treatments are wrong in principle and
quite unnecessary in practice, when proper paints are selected,
is the firm conviction of the authors. The metals should, as a
matter of fact, be given a priming coat of a sharp-toothed silicious
pigment, ground in varnish. This should be well brushed onto
the metal, so that the sharp angles of the pigment will form a
union with the zinc surface. In short, the varnish serves the
purpose of closely cementing the pigment to the zinc. In addi-
tion to the sharp pigment, a certain proportion of a soft pigment,
such as sublimed or corroded lead, is believed by some experts
to be beneficial. Subsequent coats of paint of any desired for-
mula may be applied provided the first or priming coat is properly
designed. Varnish vehicles have been tried in connection with
soft pigments, such as carbon black and lampblack, with some
success, as far as proper adhesion is concerned, but in the opinion
of the authors silicious pigments properly mixed with some form
of lead pigment are much better and safer to use.

Painting of Tinned Surfaces. — In the course of exposure
tests near the seashore, which have been under the observation
of the authors, in which the test pieces were tinned with five to
forty pound coatings, the great importance of affording efficient
paint protection of tinned surfaces immediately after erection
was clearly shown. Some of the test plates actually showed
evidences of rust at the end of two days' exposure, and at the end
of a month nearly all were corroding rapidly. Tinned surfaces
do not present as great painting difficulties as galvanized surfaces,
and any good inhibitive composition may be selected for their
protection. It is interesting to note that the use of graphite has
been condemned by the Master Sheet Metal Workers' Associa-

208 CORROSION AND PRESERVATION OF IRON AND STEEL

tion as positively dangerous for application to tin roofs. 1 As evi-
dence accumulates, it becomes more and more apparent that this
pigment is a natural stimulator, and when used as a contact
coat has been a contributing cause to the failure of many tin
roofs. The authors have designed an inhibitive formula based
on knowledge gained from research and practical experiments
which has given satisfaction in this class of work. This formula
is composed of 98 per cent, pure iron oxide with 2 per cent, zinc
chromate, ground in properly aged linseed oil, with or without a
certain percentage of varnish. The use of inert natural iron
oxides or ground hematites for this purpose is recommended.
It should be stated, however, that before painting, the tin should
be wiped so that the palm oil used in the process of manufacturing
tin plate will be removed from the surface. The use of turpentine
or benzine is recommended for this purpose.

Bituminous Coatings and their Application for the Protection
of Iron. — From the standpoint of the electrolytic theory of cor-
rosion, it has repeatedly been stated in the foregoing pages that
a natural insulator or non-conductor of electricity should be well
adapted for the protection of iron. In view of the truth of this
general principle it would seem that bitumonius compounds
were peculiarly well adapted for the purpose. Unfortunately,
many materials which are well adapted for the protection of iron
from one point of view are too prone to fail from another. As a
matter of fact, most bitumens will not stand free exposure to
atmospheric conditions, water, and sunlight. It is a significant
fact that in the cases in which bitumens have been successfully
used it has been under conditions in which the material was not,
so to speak, normally exposed to sunlight and the atmosphere.
Asphalts (and by this term is meant the natural bitumens found
in nature) do not withstand the action of water without under-
going change. On the other hand, artificial bituminous sub-
stances such as coal tar, both crude and refined, although they
are sufficiently waterproof, undergo certain changes under the
action of sunlight and the atmosphere, which materially change
their nature. 2 It is for just these reasons that bituminous coat-
ings instead of coming into general use as protective agents have

1 Proc. Second Annual Convention, Nat. Assn. Master Sheet Metalworkers
Indianapolis, August, 1906.
-■ See illustration, p. 214.

PAINTS FOR VARIOUS PURPOSES 209

found only a limited and special field of application. A careful
study of the subject serves to show that the useful application
of bituminous paints and dips is confined to the cases in which
the action of sunlight is excluded. Thus we find the use of bitu-
men confined principally to the painting of pipe lines, smoke
stacks, tunnel work, and piping generally. In short, to all cases
in which either sunlight is excluded as a factor or where the tem-
perature of the structure to be protected is so high that no other
form of protective coating will serve. It is not the intention of
the authors, however, to contend that properly refined coal tar,
asphalts, and bituminous mixtures of various composition have
not been used with more or less success.

The value of such compounds has been increased by admix-
ture with fossil gums, among which kauri gum may be mentioned.
By mixing lime with hot coal tar a valuable mixture has also been
produced for pipe dipping and other purposes, and in the authors'
opinion such a mixture as this is well adapted for many purposes.
The possible development of the use of steel for railroad ties will
demand efficient protection, for which, owing to economical
reasons, the bitumens must be made available. This phase of
the subject will be discussed in a later paragraph.

Baked japan coatings have come into general use in recent
years. These coatings are compounded largely of high-grade
asphaltic gums such as gilsonite and elaterite, sometimes mixed
with a certain percentage of kauri or other fossil gums. The
material to be coated is generally heated and then dipped in the
hot mixture. In some cases the material is baked in ovens until
the coatings become hard.

The well-known Angus Smith method of baking pipes coated
with linseed oil, and afterwards immersing in tanks of heated
coal-tar pitch, or coal-tar pitch admixed with linseed oil, was the
original process which has been followed by many others, notably
the Sabin process. Asphaltum or gilsonite mixed with linseed
oil or mineral oil has also come into some use for this purpose
and with considerable success. For information in regard to the
technology of these processes, the original authorities should be
consulted. 1

Regarding the Use of Coal Tar. — With regard to the use of

1 See Sabin — Technology of Paint and Varnish. First Edition, 1906.
John Wiley & Sons, New York.

210 CORROSION AND PRESERVATION OF IRON AND STEEL

coal tar. Wood, 1 a well-known authority, says: "A coal-gas tar
paint that has given very good results in the coating of gas-holder
tanks and other situations where the metal is exposed to ammonia
and sulphurous acids in solution, and to alternate melting and
drying under a great range of temperature, is made as follows:
Coal-gas tar is well boiled to evaporate the water and light hydro-
carbon elements, and then 20 to 25 per cent, of caustic quicklime
is sifted and well stirred in to neutralize the acid elements in the
tar. This is to be kept hot for a few hours and then an equal
quantity of good Portland or hydraulic cement is sifted and
stirred in thoroughly. The mixture is applied hot to the clean
dry iron, and can be repeated soon as cool or dry if the exposure
conditions are to be very severe. In the latter case, a little more
cement should be added, so that the caustic lime and cement
mixture will contain 50 per cent, of each. The pigments thicken
the coal tar and prevent it from running under sun temperatures
and give a bond to the brush coating of neat Portland cement
that should be applied to the coal-tar coat as soon as either the
first or second coat of the mixture is dry. This coating can be
repeatedly applied with advantage. It is impervious to gases
and water and has no tendency to run at temperatures under
130° to 140° F."

The views of Spennrath 2 are also worthy of citation in regard
to the use of bitumens as protective agents: "Tar paints were
at one time largely used for the prevention of rust, but their
employment has of late fallen into disrepute, the alleged protect-
ive action having been shown to be of very doubtful efficacy.
It is true that a coating of tar is altogether insusceptible to the
chemical influence of the atmosphere and also to acids and alka-
lies; but. at the same time, the tar itself contains substances
capable of causing iron to rust. Crude tar, for instance, always
contains moisture, and, in addition to this, acetic acid is present
in wood tar, the water and acid quickly causing the iron to rust.
Hence it is that sometimes very extensive rusting can be de-
tected, often to a considerable degree, under the tarry coating.
Water-free or acid-free tar would not exert such action; but the
purification of the tar from these admixtures is troublesome,

1 M. P. Wood. Rustless Coatings; Corrosion and Electrolysis of Iron and
Steel. First Edition, 1905. John Wiley & Sons, New York.

2 Iron Corrosion. L. E. Andes. Scott, Greenwood & Son, London, 1900.

PAINTS FOR VARIOUS PURPOSES 211

and increases the cost of this preparation, so that the advantages,
as compared with oil paint, disappear. Finally, tar paints become
softened by the sun's rays to such an extent that the mass runs
and leaves the iron bare. Thus no security against rusting can
ever be obtained by the use of tar paints."

An immense amount of work has been done in recent years
on the chemistry of the bitumens, and more reliable information
is now available in regard to their refinement and treatment than
ever before. As a matter of fact, engineers have been inclined
to speak of tar and asphalt as though these materials were stand-
ard substances in the same sense, for instance, as sugar and com-
mon salt are known as standard substances. We now know,
however, that this is very far from being the case. The character
and quality of these substances, as a matter of fact, varies within
wide limits and is dependent not only upon their source but upon
the methods of preparation and refinement. As an instance
of this, it is only necessary to refer to the widely differing quantity
of free carbon or soot carried by different samples of refined coal
tar. It is not only possible, but highly probable, that a coal tar
containing, say, less than 20 per cent, of free carbon would be
better adapted for a protective coating than one carrying from
40 to 50 per cent, of this constituent. The following paragraph
is included for the information of readers who may be especially
interested in this phase of the subject.

Tar Paints. 1 — Coal tar, if properly refined, makes an excellent
paint or varnish, especially for metal surfaces which are subjected
to the action of corroding gases. Crude coal tars vary greatly
in composition, but always contain ammoniacal water which
makes them unsuitable for use as paints. The presence of water
greatly reduces the adhesive qualities of the tar, and ammonium
salts tend to saponify some of the oily constituents, thus making
them more or less soluble in water. A coating of crude tar is
likely to blister and peel off in places shortly after application.
Tars which have been boiled to remove the ammoniacal water
are somewhat better than crude products, but are never as satis-
factory as properly refined tar paints.

Besides ammoniacal water, coal tars contain at least two other
objectionable constituents from the standpoint of paint manu-

1 In the preparation of this paragraph the authors are indebted for assist-
ance to Prevost Hubbard, who has made a specialty of the study of bitumens.

212 CORROSION AND PRESERVATION OF IRON AND STEEL

facture. These are naphthalene and free carbon. Both are the
products of high temperatures to which the coal gases are sub-
jected during the process of tar formation, and as they are usually
present to a less extent on low temperature than in the high tem-
perature tars, the former are to be preferred in the preparation
of paints. Naphthalene is a crystallizable solid which is held in
solution by the tar oils. It volatilizes slowly at moderate tem-
peratures, although it is found mainly in the higher boiling frac-
tions. If present in large amount, it makes a pasty paint which
after drying commences to crack and disintegrate as the naph-
thalene volatilizes. Excessive quantities of free carbon produce
a lumpy, uneven paint, which is deficient in waterproofing
properties. Free carbon also gives the bitumens a false consist-
ency. If present in moderate amounts, it is perhaps not an alto-
gether undesirable constituent, as it adds to the body of the tar
and serves to some extent as a filler. Twenty per cent, free carbon
would seem to be the maximum allowable limit.

When refining a tar for the production of paint, it may merely
be distilled until sufficient volatiles have been driven off to give
it the desired consistency, or distillation may be carried to the
formation of a moderately hard pitch which is afterwards cut back
with certain of the distillates. The latter type is usually to be
preferred. In the European countries, it is customary to employ
an upright form of still with convex bottom for tar refining, but
in the United States large horizontal cylindrical stills are com-
monly used. The process of distillation may be briefly described
as follows: The still is first charged with crude tar, all openings
except that leading to the condenser closed, and firing commenced.
As the tar becomes heated it begins to froth and foam and if not
carefully fired will boil over into the condenser. This is due to
the presence of water which is converted to steam while the tar
is still viscous. Boiling over may also be caused by an excessive
amount of free carbon which causes the vesicles of gas and vapor
to adhere to the carbon particles and thus swell the volume of
tar. The removal of water in the final stages gives rise to a
peculiar noise made by the steam, which is sometimes called the
rattles. After this noise ceases, distillation usually proceeds
without further trouble, and the distillate is caught in fractions
according to the needs of the distiller. Thus, the following
fractions may be collected:

PAINTS FOR VARIOUS PURPOSES 213

(1) First runnings to 110° C.

(2) Light oils, 110°-170° C.

(3) Carbolic or middle oils, 170°-240° C.

(4) Heavy, dead, or creosote oils, 240°-270° C.

(5) Anthracene oils above 270° C.

In the United States the distillate is usually divided into only
two fractions, the light oils having a gravity of less than 1.00 and
the creosote oils having a gravity greater than 1.00. The second
fraction is collected until the residual pitch is of the desired
consistency. If this pitch is sufficiently fluid it may be employed
cold as a paint. It is preferable, however, to carry the distillation
further and afterwards cut back the harder pitch with dead oils
from which the naphthalene has been removed. Pitches which
contain only the higher boiling oils as solvents for the heavier
bitumens take a long time to dry, and for this reason it is some-
times preferable to use a harder pitch which requires to be heated
before application, but afterwards drys rapidly. A still more
satisfactory paint is one produced by cutting a moderately hard
pitch with about three-fifths its volume of light oils. Such paints
may be applied cold and dry in a comparatively short time.
Rapid drying paints and varnishes may be prepared by substitut-
ing the more volatile tar naphtha for the light oils. They may
be applied in very thin coats if desired by using a large amount
of naphtha. Water-gas tar pitch makes an excellent base for
these varnishes, as it contains but very small amounts of free
carbon and naphthalene and produces a glossy black coating of
uniform texture.

To sum up the discussion of the use of bituminous paints and
compounds as protective coats for iron, the authors wish to
reiterate their opinion that in special cases, and if properly selected,
there are no better materials available. For use, however, under
general exposure conditions where sunlight and free atmosphere
are elements in the problem, bitumens cannot be successfully
used. The well-known effect which has been mentioned in pre-
vious paragraphs as "alligatoring" is almost always produced
on bituminous coatings exposed to the action of sunlight. This
effect, which has been photographed from one of the panels on
the Altantic City tests, is shown in Fig. 66.

An accurate explanation of the phenomenon spoken of as
alligatoring is not easy to give, but it probably depends on a

214 CORROSION AND PRESERVATION OF IRON AND STEEL

superficial hardening of the outer or skin surface of the coating,
underneath which the body of the material still remains to some
extent plastic and elastic. The regular and diurnal expansion
and contraction of the iron, which have sometimes been referred

Fig. 66. — Alligatoring of coal tar over red lead.

to as the breathing of steel, must gradually produce this tendency
to cracking or checking of the superficial layer of the bituminous
coating. The effects produced are exactly similar to the crazing
of vitreous glazes which has produced so much discussion in the
ceramic industry. Exactly the same effects are produced in some

PAINTS FOR VARIOUS PURPOSES 215

cases of linseed oil films in which an excess of unctuous or soft
pigment has been used. It follows that if this checking effect
spoken of as alligatoring continues for any length of time, the under-
lying surface will become exposed to the corrosion influences.

Paper Paints and Paper Preservatives. — A process proposed
by Cross and Bevan has been described by Andes/ by which cellu-
lose paper is dissolved in caustic soda lye, producing a highly pro-
tective paint. Paraffined paper has also been used successfully as
a protecting material. For further information in regard to these
rather unusual methods the reader is referred to the literature
as given in Appendix B.

Painting Metals Subject to Marine Growths. — The dockage
of steel ships which have been in the salt water for six months
or more generally discloses a condition of surface below the water
line that has caused naval constructors much concern. The
thick crust of barnacles and marine growths must be scraped off
if the ship is to make speed with the least coal consumption, but
the tenacity that is shown in the adherence of the marine growth
makes it impossible to avoid abrasion of paint coatings, in the
cleaning process. That such growths are stimulators of corrosion
there seems to be no doubt, and the only remedy seems to be in
the application to the steel hull of the boat of a good, hard under-
coat, with a top coat of paint containing materials which will act
as poisons to the various growths, and prevent their action. Cer-
tain semi-drying compositions, containing pine tar or crude oil,
and having a percentage of some powerful poison, such as strych-
nine, have been used as a vehicle for high-grade oxide of iron and
lead pigments, and seem to have shown excellent results.

Soaps of copper, arsenic, and other poisonous compounds,
made by precipitating a saponified oil with salts of the above
metals, are often dissolved in the vehicle of marine paints and are
undoubtedly of value, especially the soaps of copper. Corrosive
sublimate and oxide of mercury (mercury vermilion) are probably
the most efficacious poisons to use for this purpose.

The first, or under coating, for the ship plates should contain
in the vehicle a good hard drying varnish to act as an excluder
of water, and the outer coats should be those containing the
poisonous pigments. A certain percentage of oil of a semi-dry-
ing nature is sometimes advisable.

1 Der Eisenrost, p. 252.

216 CORROSION AND PRESERVATION OF IRON AND STEEL

Paints resistant to the action of sea water and saline atmos-
pheres have already been discussed in a previous chapter.

Painting Steel Cars. — The equipment of modern railroads
has brought about great changes during the last few years in the
construction of both freight and passenger cars. The old type*
of wooden cars are rapidly being replaced by those of all steel
construction, which are found to be safer and more efficient from
every point of view.

Before painting steel cars, the surface should be, and usually
is, cleaned with the sand blast. The sand blast has been found
more efficient than paint removers for removing old paint from
steel surfaces. Old paint, under the abrasive action of the sand
from the blast, readily wears down, and the continuous action
of the sand upon the steel furnishes a good surface for the appli-
cation of new paint. In order to give proper smoothness, the
surface may be coated with a primer which generally consists
of a silicious pigment, such as mineral black, with the possible
addition of some willow charcoal ground in oil and varnish to a
thick paste. After this has thoroughly dried, a coat of filler is
usually applied, and then a coat of surfacer, these materials being
preferably of inhibitive pigments and of different colors so that
the inspector may know by the color the number of coats that have
been applied. Two color coats are generally applied upon the
surfacer and then several coats of the finishing varnish may be
used. The authors do not wish to convey the impression that the
above general outline of car painting is recognized as standard,
but it is their desire rather to indicate that in the various primers
and fillers that are used for car painting, the replacement of stimu-
lative pigments with inhibitive pigments should be made, and
when this condition prevails, not only on the first and second
coats but throughout on all coats, the best results will probably
be obtained.

The ordinary type of sheet-steel coal carrier sometimes receives
sand blast treatment and then spraying with paint, the small
amount of time allotted the painter in his work requiring the use
of excessive drier in order to get the cars second-coated, ready
for lettering and stenciling, and out of the way to make room for
others. Such methods should be revised. The authors have
yet to see a car painted in the above manner that is not badly .
corroded in all too brief a period, and it is doubtful whether

PAINTS FOR VARIOUS PURPOSES 217

many steel coal cars even survive their maiden trips without
showing rust tuberculation.

The use of pneumatic paint atomizing machines by large
railroads has reduced materially the labor cost per car, and has
probably resulted in a more uniform distribution of the paint,
especially on rough and irregular surfaces. The use of excessive
quantities of thinners should, however, be avoided when paint
machines are used.

The authors are aware of the numerous practical difficulties
that the car painter and railroad engineer has to contend with.
The modern steel coal-carrying freight car provided with lower
dumps furnishes a specific instance of just such a difficulty. Our
winter climatic conditions frequently cause the coal to freeze
in the cars, so that in order to free the dump hoppers the railroad
operatives find it necessary to beat upon the side of the cars with
the nearest available weapon. It is evident that no paint coat-
ing that has ever been designed will stand such heroic treatment.
The only practical solution of this difficulty which presents itself
is some system of inspection and maintenance which will provide
for repainting at frequent intervals in the yards without waiting
until complete failure demands the return of the car to the shops.
The old wise saw that a " stitch in time saves nine" finds a pecu-
liarly appropriate application not only to the conservation of
steel shaped into cars and other containers, but also to the subject
of the conservation of iron in general. As a matter of fact, such
a system of maintenance as is outlined above has to some extent
been tried, and it is not altogether an unusual sight in freight
yards to see corrosion spots on cars being touched up with red
lead. As far as this work has been carried on it constitutes in
the opinion of the authors one step in the right direction. There
is no reason, however, why an inhibitive paint compound should
not be designed and compounded to a definite color scheme, so
that retouching under this plan of maintenance could be carried
on by expert painters in the freight yards. Surely the cost of
maintenance as carried out by such a scheme would not in the
end be greater than the steadily fixed charges laid upon all owners
and operators of structural steel by the never ceasing attack of
corrosive influences.

Paints for Locomotives and Tenders. — In painting locomotives
and tenders, large quantities of mineral black (a special grade of

218 CORROSION AND PRESERVATION OF IRON AND STEEL

bituminous slate), ground in varnish, is generally used for the
primer. After rubbing down, this coat is often followed with one
containing white lead, carbon black, and silica. This coating
is followed, after rubbing, with a body coating of ivory black in
japan, and finally with several coats of special locomotive finish-
ing varnish. The use of mineral black is not bad practice, as
it has no special stimulative properties. In the opinion of the
authors, however, 2 to 5 per cent, of special inhibitive pigment
should be used in admixture with the pigment.

Protection of Iron in Tunnel Work. — The protection of iron
in tunnel work presents a problem of unusual difficulty. Con-
stant seepage causes drippings of an extremely corrosive nature
to come in contact with all exposed surfaces. Very often this
exuded moisture found in tunnels is rich in chloride and other
corrosive electrolytes. These salts act to a certain extent the
part of paint removers. The atmosphere in tunnels is often rich
in carbonic acid, sulphur dioxide, and invariably high in moisture
content. These conditions call for the very best excluding and
inhibitive paints that it is possible to design. It has already
been stated that properly refined coal tar, treated with lime,
is well adapted under certain conditions for use in the dark, and
therefore should be useful in protecting the various forms of iron
and steel used in tunnels. If oil paints are to be used, they
should have a varnish vehicle of the highest type, made with
gums of low saponification value, such as copal. A varnish paint
of this nature has been found to give extremely good results on
the cars running in the tunnels of the Pennsylvania Railroad
under the Hudson River in New York City.

Painting Train Sheds. — The train shed with its gigantic
network of structural metal is constantly filled with the sulphur-
laden gas belching from the numberless engines that operate
therein. This gas condenses with the rising steam upon the
painted metal surface, and often causes most disastrous results.
In the large terminals a force of men are kept constantly at work
repainting the maze of steel. The- engineer is ever on the look-
out for some material that will last for a greater length of time.
Excluding the fumes and gases by the use of a varnish paint
seems to offer the best solution of the above problem. The
varnish for such a paint must be, however, of the highest grade,
and one that will withstand exposure without cracking or checking.

PAINTS FOR VARIOUS PURPOSES 219

The use of high-grade copal gums for this purpose is to be recom-
mended.

Black japan made by fusing and properly treating eight pounds
of kauri or copal gum in two gallons of oil, and mixed with about
twenty gallons of boiled linseed oil containing 1 per cent, of
litharge, forms a base into which may be added five gallons of
high-grade asphaltum. When this composition is applied, it
dries to a high glossy surface and forms an excellent excluder.
Other paints, made of inhibitive pigments and possessing exclud-
ing vehicles, have also proven most satisfactory for the above
class of work.

Paint Protection for Water Tanks. — Water tanks and other
large steel containers present a double problem, as protection
must be provided on both the inner and outer surface. The
painting of the inside of steel water tanks has been in practice
on American railroads for some time past. F. C. Peterson 1 of
the Southern Pacific Railroad has recorded satisfactory results
with the use of a formula composed of 25 lbs. dry red lead, 5 lbs.
litharge, 5 lbs. Venetian red, 1 gallon boiled linseed oil, and 1
gallon of turpentine. Three coats are usually given with this
mixture. The use of natural hematite (iron oxide) in place of
the Venetian red would be better practice. Venetian red gener-
ally contains fifty per cent, or more of gypsum, which is partially
soluble in water and is often the cause of the rapid failure of
paint films. The results obtained with gypsum on the Atlantic
City test panels would certainly not recommend its use in an
inhibitive formula.

Painting Steel Railroad Ties. — The rapid diminution of our
forests, and the increasing cost of wood, may ultimately require
the substitution of steel for railway ties. Steel ties are at present
being experimented with on several railroads, and, while they
have not as yet proved wholly satisfactory, some progress has been
made, and we may feel certain that their final adoption in some
form is sure to come. Steel ties must be protected or corrosion
will soon weaken and destroy them, reducing their durability
below that of the wooden tie. If, however, they are properly
protected, there is no reason why they should not last for many

1 Painting Insides of Steel Water Tanks. Paper read before the Mainte-
nance of Way Master Painters' Association, Niagara Falls meeting Septem-
ber, 1909.

220 CORROSION AND PRESERVATION OF IRON AND STEEL

years. The authors believe that dipping in hot properly refined
coal tar free from acid is the best treatment that has been suggested
for the purpose. A mixture of 80 parts of coal tar, 5 parts of dry
lime, and 15 parts of Portland cement is recommended as a good
formula for a tie dip. The lime which in itself is an inhibitor
neutralizes any free acid in the coal tar, while the cement appears
to lead to a harder surface, which gives more resistance to the
abrasive action of roadbed dust. In addition to this, it might
be advisable to apply a preliminary dip coating designed to be
inhibitive. The first cost of these treatments will be high, but
if steel is to be used for the purpose, some such protection must
be supplied.

The operation of refrigerator cars would constitute a special
menace to steel ties, unless mechanical devices are contrived to
catch and discharge the brines. The Master Car Builders' Asso-
ciation has appointed a committee to study this problem. In a
paper on this subject, A. B. Phelps 1 of the Lake Shore Railroad
says: "To those of you who have the care of bridges, interlocking
plants or other iron or steel apparatus or other structures, I need
not enlarge on the destructiveness of salt brine; to those of you
who are not on trunk lines, and consequently have few, if any,
refrigerator or stock cars passing over your road, I will simply say
that salt is one of the most destructive agencies to steel and iron
of which I know, and when brought in contact with steel bridges,
if neglected, soon forms large, thick scales of corroded metal,
greatly weakening the structure, and requiring the use of heavy
hammers or chisel-pointed steel tools to remove them. Ordinary
steel brushes will have no effect whatever on them.

"This salt brine, of which I speak, is formed by the melting
ice and salt used in cooling the contents of the car, and is allowed
to drip directly from the bottom of the car, almost exactly on,
or only a few inches from the rail, alighting on both the top and
bottom flanges of the bridge, where it proceeds to get in its work
of destruction. To prevent, not to cure, this evil is my theme.
The present general construction of these cars is to provide an
outlet underneath the icebox of about a 2-in. iron pipe, pro-
truding into a metal cup, or box, holding about a quart, which

1 The Destruction of Railroad Property by Refrigerator Car Drippings,
Sixth Annual Convention Maintenance of Way Master Painters' Association
1909.

PAINTS FOR VARIOUS PURPOSES 221

when filled, forms a trap, so that warm air will not enter the ice-
box from beneath. This cup is allowed to fill and spill, -and slop,
and drip its contents the entire length of the road, and do its
hellish work — that is strong language, but no stronger than the
solution of which we are speaking. These conditions should never
exist, and should not be allowed for one single day. Provision
should be made, in building these cars, to retain the brine and
carry it to and empty it at the terminals, or at the icing stations."

Inhibitive Pigments for Railroad Equipment in General. — Red
lead, basic lead chromate, zinc chromate, and bright neutral
oxide of iron are inhibitive pigments which may be used in pro-
ducing reds for semaphore, signal and switch targets on railroads.
Genuine oxide of chromium is a most permanent green pigment,
when this color is desired, for the above use. When this pigment
is reinforced with a silicious pigment, its wearing value is increased
without materially detracting from its color depth. Zinc-chrome
green made from zinc chromate and inhibitive Prussian blue, is
also a very permanent inhibitive color that should meet favor
for signal work.

All these pigments have been experimented with and carefully
observed by both the authors under a wide range of service con-
ditions, and are known to have given satisfactory results!

The Preservation of Steel Mine Timbers. — The question of
the substitution of wood by steel for timbering mines, is an im-
portant one from several points of view. An enormous amount
of valuable timber which might possibly be employed to better
purpose above ground is used in mining operations. On the
other hand the fear of rapid and dangerous corrosion of steel
induced by the acid character of most mine waters has led many
engineers to feel that the substitution of steel for wood might
lead to extensive failure in the course of time. Woodworth has
made an extensive research on this subject. 1 in the progress of
which he has reached the following conclusions:

"The application of economic principles to the use of steel
demands that the life, on which its economical use is predicated,
be lengthened to as long an extent as possible, and the preven-
tion of waste and loss in the steel itself by preservative treatment.

*R. B. Woodworth, M. Am. Soc. C. E., Engineer Carnegie Steel Co.
Paper read before the West Virginia Coal Mining Institute, Huntington,
W. Va., December 7, 1909.

222 CORROSION AND PRESERVATION OF IRON AND STEEL

No iron has ever been or ever could be manufactured that would
not rust in moist air unless it is protected by some sort of covering,
for the reason that pure iron itself has a tremendous affinity for
oxygen and the slightest contact between iron and oxygen in a
more or less unstable condition causes immediately that combina-
tion of iron and oxygen which is known by the name of oxidation
or rust. This tendency to corrode is characteristic of the metal
itself and is independent of the presence of the impurities it may
contain. 1 Iron does not rust when entirely submerged in pure water
or water not containing free air; neither does it rust in a perfectly
dry atmosphere, the beginning of oxidation being due to the pres-
ence of sufficient quantities of free oxygen, due to impurities con-
tained in the air, in the presence of sufficient moisture to act as
water of disassociation, this condition permitting the escape of
hydrogen and the consequent oxidation of the steel or iron.
The great variation which occurs in the rusting of iron and steel
is probably due to unequal distribution of the impurities neces-
sary in their manufacture. Iron does not occur in a pure state
and in the process of the manufacture of wrought iron and
steel, small traces of impurities occur more or less unequally dis-
tributed throughout the mass. It has been claimed that the small
proportion of manganese which steel contains adds greatly to its
oxidizing tendencies. This, however, is not borne out by actual
experience and the removal of manganese does not carry with it
freedom from oxidation. 2 The difficulty lies not in any one cause
or element contained in the steel but in their unequal distribution
throughout the steel which permits differences of potential at
different points. The prevention of this oxidation, however, is
a very simple matter. Such steel as has already been installed
in the mines of the United States has given entire satisfaction
without any serious corrosion and without any protective treat-
ment other than the use of preservative paint. In wet mines
in England steel girders are frequently tarred before being put
in, but the actual loss from corrosion is so small as to be a minor
quantity even when the steel is not painted. In dry locations
there is not much danger of any serious corrosion. In the speaker's
judgment, however, true economy will require the painting of
all steel for underground operations with one shop coat of good

x The authors are not prepared to accept this unqualified statement as
has been shown in preceding discussions.

2 This statement may also require qualification.

PAINTS FOR VARIOUS PURPOSES 223

paint and with at least one field coat. If these are well worked
in, the steel should need no further attention for years. What
these coats should be has been the subject of considerable investi-
gation, attention being had as well to atmospheric conditions as
to the possibility of contact with acid waters, and while laboratory
experiments have not been finally completed, the broad lines of
preservative treatment may now be indicated. Steel should not
be painted with carbon paints in whose manufacture sulphuric
acid has been used and the use of coal tar products is, therefore,,
to be avoided. The natural carbons, such as graphite, and hydro-
carbons, such as asphalt and gilsonite, may be recommended
for second coat work if properly ground and mixed with a good
vehicle. For the first coat pigments should be used of a more or
less inhibitive character, though it has been demonstrated that
some of the best paints are not entirely inhibitive. The oxides
of iron, such as Venetian reds, are usually manufactured by
chemical processes and their use is to be avoided, though a good
natural oxide of iron paint may be used under dry atmospheric
conditions.

" The correct theory which underlies the prevention of corro-
sion in iron or steel is based on the use of a practically inhibitive
pigment to prevent the inception of corrosion in the steel and the
use of a second coat to protect the first from atmospheric and
temperature conditions and to fill up thoroughly any vacancies
or voids which may occur therein. To meet these conditions
the pigment should be practically inert and its particles extremely
small. While red lead is not absolutely inhibitive, it has been
demonstrated in all kinds of exposures to be a first-class pigment
for use in the preservative treatment of steel in almost any situa-
tion and its particles are extremely minute. It is a very heavy
pigment and settles quickly, necessitating the mixing of a new
lot if it has stood over any length of time. The settling may
be retarded by the use of barytes or asbestine, and the use of the
latter is to be recommended by reason of its permitting a firm
hold on the pores of the steel. The oil should be pure, as the
matter of vehicle is of large importance, and the raw oil is better
than the boiled. A mixture of red lead, oil and asbestine, in the
proportions of at least 15 pounds of red lead and 2 pounds of
asbestine to a gallon of oil, with sufficient japan drier to work
well, may be recommended as probably the best under present

224 CORROSION AND PRESERVATION OF IRON AND STEEL

conditions, and the use on this of a first-class graphite paint will
serve thoroughly to protect the shop coat and to fill up any
vacancies or voids therein, the particles of well-ground graphite
being also extremely small. In the painting of steel the surface
should be most carefully cleaned from all scale, rust, dirt, etc.,
and the paint should be applied in dry weather. No painting
should be done in wet or freezing weather. When, for any reason,
it is necessary to repaint, the repainting should be done on clean
surfaces absolutely free from all rust, paint skins, dirt, etc. It is
not sufficient to apply a new coat of paint over an old paint sur-
face under which traces of corrosion appear; the new paint will
cover the old surface and may adhere firmly thereto, but the
corrosion goes on underneath just the same. This small atten-
tion to detail characterizes all thorough work and is necessary
for the entire preservation of steel or iron in any location. It is
doubly necessary by reason of the fact that underground steel may
not be examined and repainted with the readiness possible in above
ground construction. At the same time it may be added that
underground conditions are not nearly so secure on the steel as
above ground conditions, and certainly the paint itself is not
exposed to those alternating conditions of high and low tem-
peratures, dryness and wetness, strong light and darkness, with
which above ground construction has to do and which are especially
accelerative to the deterioration of protective coating. The only
serious condition underground is due to the presence of acid-laden
waters, and the speaker's chemical analyses and laboratory
investigations indicate that the objections due to this cause have
been largely overstated. Experience and theory all indicate that
only the simplest means are necessary for the absolute guarantee
of an extremely long life for steel used in underground conditions."
Painting of Steel Mine Timbers. — In a special communica-
tion to the authors Mr. Woodworth gives an account of extended
laboratory researches on the subject of painting steel mine tim-
bers: "Mine water usually carries with it more or less free acid
and its corrosive action is one of the reasons advanced why steel
should not be used in substitution for the standard wooden con-
struction, and this in spite of the fact that an experience of fifteen
years in the use of square timber sets has proven that material
can be very well protected from this corrosive action by the use
of methods recommended in ordinary outdoor structural work.

PAINTS FOR VARIOUS PURPOSES 225

Conditions in mines vary greatly; in the long, dry passageways
there is little danger of deleterious action; in damper situa-
tions there may be running water carried in the side ditches,
which may come in contact with the feet of the timber sets; and at
shafts the timber may be exposed to the constant action of running
water which may be more or less acid in accordance with the
character of the overlying strata which may or may not contain
beds of coal. To find the best methods of protective treatment
to meet the worst conditions, laboratory tests were instituted.

Samples of mine water were obtained from a number of anthra-
cite and bituminous coal mines which were carefully analyzed in
the laboratory of the Carnegie Steel Company at Homestead Steel
Works for total solids and total free acid in parts per million, as
shown on tables attached. From a comparison with a fuller
analysis made by the Lehigh Valley Coal Company it was decided
that practically the worst condition to be met would be that of
an acid water containing 2320 parts of free sulphuric acid (H 2 S0 4 )
in a million and 128 parts of ferric sulphate Fe 2 (S0 4 )3. To expedite
the action of the acid water on the samples to be tested it was
decided to intensify ten times and to subject samples to a solu-
tion containing 23,200 parts of sulphuric acid and 1280 parts of
ferric sulphate to the million. In this connection it may be stated
that investigations indicate the intensifying action of ferric sul-
phate in solution as compared with a solution containing sulphuric
acid only.. A mixture was prepared with 180 cubic centimeters
of commercial sulphuric acid of 1.84 specific gravity and 79 drams
of commercial ferric sulphate dissolved in one meter of water,
giving 2.616 per cent, free sulphuric acid and .432 per cent, ferric
sulphate, or 2.135 per cent. S0 3 and .432 per cent. FeS0 4 .

One-half in. round steel rods 3J in. long were used for test
purposes, with wide mouthed bottles 4£ in. high and 1? in. in
diameter with necks If in. in diameter. The rods were painted
with the pigments to be tested, inserted in holes bored in the
corks and immersed in the above-mentioned solution, the bottles
being properly labeled with the number of the pigment tested and
the rods being marked by a file with the same number. Exami-
nations were made at intervals of two or three days apart in
the inception, of the test and at longer intervals on such pig-
ments as remained. The rods which showed a deterioration of
pigments under solution action or corrosion underneath the films

226 CORROSION AND PRESERVATION OF IRON AND STEEL

were removed when such deleterious action became pronounced.
Solutions were renewed when the reddish discoloration produced
by solution action became pronounced provided the pigment
showed no apparent signs of disintegration. The bottles were
freely agitated at all such examinations and repeatedly between
examinations as well. In some instances the solution continued
clear and the pigment intact until the completion of the tests,
which extended from December 23, 1908, to June 4, 1909, at
which time the relative values of the pigments had been demon-
strated.

The pigments tested were in part commercially prepared ready
mixed paints, in part paints prepared from the raw materials in
the laboratory, and in part special formulas furnished ready mixed
by the Paint Manufacturers 7 Association of the United States.
The vehicle used in the preparation of pigments in the laboratory
was selected raw linseed oil purchased from the Pittsburg Plate
Glass Company. The material painted was carefully cleaned
from all mill rust, scale, etc., and the pigments applied carefully
by brush. The first coat was allowed to harden thoroughly,
some seven or eight days, or longer if necessary, before the appli-
cation of the second coat, and the second coat was firm and hard
in all cases before immersion in solution.

In one or two cases the pigments failed utterly in two or three
days from the inception of tests. In others the action of the acid
water produced almost immediately a marked discoloration with-
out further change or deleterious action for long periods and with-
out any precipitation of solid matter. One hundred and thirteen
different combinations were tested and after expiration of the
tests thirty-one remained. The purely carbon paints, presum-
ably manufactured from coal tar products, gave no good results;
natural oxide of iron showed up fairly well. The best pigments,
both for adhesion and for minimum deterioration of film and
minimum corrosion underneath the film, were red lead and the
graphites.

On the basis of the results obtained in the execution of these
tests it is recommended that steel mine timbers be painted at the
shops with one coat of red lead and oil, mixed in the proportion
of at least 15 pounds of red lead and 2 pounds of asbestine to
a gallon of pure raw linseed oil, with sufficient japan drier to work
well under the brush, and that a field coat be applied of first-class

PAINTS FOR VARIOUS PURPOSES 227

graphite paint. Material to be thoroughly clean from rust and
mill scale before painting. The theory which underlies this rec-
ommendation is based on the use of a practically inert pigment,
thoroughly tried in all sorts of conditions, of firm adherence to
metallic surfaces and of extremely small particles; and also on
the use of a first-class excluder of likewise fine particles and firm
adhesion thoroughly to protect the shop coat and to fill up any
vacancies or voids therein.''

The tabulated results of Woodworth's laboratory experiments,
which were courteously turned over to the authors for use at their
discretion, were not included here owing to the fact that a large
number of the paints experimented with are proprietary com-
pounds. Woodworth's own summing up of his experiments has,
however, shown that he came to the conclusion that a good grade
of graphite on top of a good grade of red lead primer was best
adapted to the purpose, and provided the most efficient protec-
tion among all the compounds and methods of painting tried.
With this conclusion the authors find themselves in complete
accord, under the conditions which maintain in the experiments.
The attack* of free acid here was very strong, and, as Woodworth
points out, ten times stronger than is found in the average mine
conditions. It is quite evident that under such conditions as
these the primary object must be to select a pigment which will
be highly resistant to acid attack. It is clear that of all pigments
in common use no one is more resistant to the attack of mineral
acids than carbon in its graphitic form. The conclusion of
Woodworth, therefore, that a good reinforced inhibitive prime
coating, topped with a good acid resisting excluder, is the best
combination that could be used under the circumstances is in
accordance with the opinion of the authors. The question,
however, as to whether red lead is the best inhibitive primer which
could be selected for the purpose has not yet been convincingly
shown.

The Painting of Field Fence Wire. — Galvanized zinc coatings
are almost exclusively depended upon for the protection of field
fence wire. This subject has been discussed in detail in a pre-
vious chapter. There can be no doubt of the fact that it is
possible to prolong the life of wire for many years if the trouble
is taken to place upon the zinc surface a top coating of properly
selected paint.

228 CORROSION AND PRESERVATION OF IRON AND STEEL

The painting of fence wire presents certain difficulties that
are not met with in any other class of work. The round surface
of the wire causes paint of the ordinary consistency to run from the
upper part, where a thin film is left, to the under part of the wire,
where it collects in quantity, forming thick, non-drying clots
which absorb moisture. The authors have found that the best
consistency for paint to be used on wire fences should be that
possessed by a highly viscous cream; in fact, a thin paste is often
advisable, if properly brushed out. A good soft hair, flat varnish
brush about i-m. wide has been found the best conveyor for
the paint. After the paint has been applied to the top and side of
a wire fence for a considerable distance, the painter should go
back and run the brush along the bottom of the wire so that the
paint that has collected there will be uniformly distributed.
Turpentine or other volatile diluents should not be used, and if
reduction is found necessary, it should be with linseed oil, either
raw or boiled.

A quart of paint will cover something over 300 ft. of ordinary
eleven-strand fabricated wire field fence. For the painting of the
fencing surrounding a large farm, only a few gallons, therefore,
would be required. It is not necessary to employ expert painters
to apply paint to wire, and when ordinary care is exercised the
farmer or his men can easily paint fencing at their leisure. The
increase in the durability of the fence which painting provides
should induce the farmer and other consumers to undertake the
work even though it involves a certain amount of time and trouble.

A recent inspection of the Steel Wire Test Fences at Pitts-
burg by the authors indicated very clearly that the wire panel
which was painted with an inhibitive formula was outlasting all
other panels on which ordinary paint coatings had been used.
The use of finely ground hematite with a small percentage of
zinc chromate, ground in pure linseed oil with or without the addi-
tion of varnish, should prove efficient for use on fence wire.
Silicious primers, mixed with zinc or red lead, are also being used
for a first coat on galvanized wire. Paints containing stimulative
pigments, in the opinion of the authors, should not be employed
on wire fencing, as they are almost certain to cause rusting and
may even hasten the decay of the underlying zinc. Fig. 67
shows the effect produced on ungalvanized wire in less than one
year when a stimulative pigment has been used.

PAINTS FOR VARIOUS PURPOSES 229

Care should be taken to paint when the wire is perfectly dry.
Moisture will lead to subsequent failure, causing non-adherence,
blistering, and destruction of the film. Painting in very cold
weather should also be avoided, as puckering and contraction of
the paint skin will result.

In nurseries, vineyards, and certain other places, a spray of
copper sulphate is sometimes used as a destroyer of the animal-
cule that live upon the growing plants. When this copper

Fig. 67. — Section of wire painted with stimulative pigment. Enlarged.
Failure resulted in less than twelve months.

sulphate strikes the zinc-coated wire upon which the plants are
supported, the zinc at once goes into solution and the copper is
precipitated in a loose spongy mass upon the wire. Moisture
at once attacks the imperfectly protected iron and serious corrosion
proceeds.

If the wire supports for plants are coated with a good pro-
tective paint, the above difficulties are to a large extent overcome,
and chemical insecticides can be used without producing serious
damage.

The authors have discussed with the manufacturers the pos-
sibility of shop coating galvanized wire fencing with inhibitive
compounds, but the perfectly reasonable objection has been raised
that wire so treated would not withstand the abrasion due to
transportation, handling, and mounting on the fence. For this
reason it would seem that the painting of fencing should be
done in place and after they are mounted on the posts.

Prime Coati?igs for Structural Metal — It is always difficult
for the engineer to decide what paint formula should be used
for the prime coating on structural framework for modern build-
ings. Probably red lead is in most general use for this work,
where it has been found to give, generally speaking, a hard and

230 CORROSION AND PRESERVATION OF IRON AND STEEL

satisfactory base upon which to place the finishing coat. Andes 1
has stated in this connection : " We know that red lead forms a paint
that strongly adheres to and hardens on iron, which it preserves
from rusting; and we also know that the same paint is very
durable when covered over by subsequent coatings' of good paint.
On the other hand, red lead paint, by itself, must, by reason of its
small proportion of oil, soon perish under atmospheric influences,
and is therefore unsuitable for use as a finishing paint. Now
there is another point of great importance in connection with
painting on iron, namely, the teaching of experience that although
a paint rich in oil or varnish may be applied over a paint contain-
ing a high proportion of pigment to varnish, yet the converse
practice must not be pursued; and this is the reason why the idea
of painting with red lead was originally hit upon. This experi-
ence primarily applied only to wood, but it seems also to have
held good for iron. The words ( it seems ' are used advisedly,
since no reliable experience to the contrary appears to have been
gained, or at any rate published.

"In most cases, though not always, iron structures receive
a first coat of iron or red lead oil paint in the works when com-
pleted, i.e. j when the individual girders, stays, etc., have been
riveted and wrought into the most suitable condition for delivery.
Nobody, however, troubles about whether the girders have been
riveted out of doors — where they are exposed to the air and, more
especially, to the deposition of moisture brought about by fluc-
tuations of temperature — or under cover; and still less is any
thought bestowed on the painting of the rivet holes, the stems of
the rivets or the under side of the rivet heads. In many works
the paint is laid on by day laborers who simply daub it on, often
without suitable brushes, and the work is regarded as properly
done if the metal is outwardly covered over with paint. Now the
author readily admits that painting the rivet holes, and the
rivets themselves, is a very tedious and often expensive task, and
that very often there is insufficient space available for the rivets
to dry after painting. Nevertheless, he holds that when it is a
question of properly painting ironwork so as to fulfil the require-
ment that all parts of the metal shall be thoroughly covered, and
that the connections, in particular, shall be so carefully painted
as to prevent the incursion of water, then all these considerations

1 L. E. Andes. Iron Corrosion. Scott, Greenwood & Son, London, 1900

PAINTS FOK VARIOUS PURPOSES 231

of reducing the cost of production to a minimum must be put
into the background."

Zinc chromate, high-grade oxide of iron, and other pigments
have proven equally as efficient as prime coaters, as red lead.
For further information on this subject, see Chapter VIII on the
results which are being obtained on the Steel Test Panels at
Atlantic City, which should be watched with interest by engineers.

As has been pointed out in a previous paragraph, red lead is
not a standard substance and its inhibitive value may be easily
affected by the process of manufacture by which it is prepared.
As is described in a separate paragraph in a succeeding chapter,
red lead is manufactured by two separate and widely different
processes. It is not at all probable that the two products would
show the same inhibitive value, as this value would depend upon
the amount and character of the included impurities.

The Use of Oil as a Shop or Prime Coating. — The subject of
shop coatings for machinery and structural iron has received a
great deal of attention. It has long been known that linseed oil
is an active stimulator of corrosion when applied to iron, when
its surface becomes abraded in the least degree. This opinion
is borne out by recent work done by Walker. Walker's work on
this subject has been referred to in a previous paragraph, but in
view of its direct bearing on the subject under discussion, it will
be well to recapitulate it here. Walker finds linseed oil to be,
under certain conditions, an accelerator of corrosion. When a
steel or iron surface painted with linseed oil becomes abraded at
a particular spot, corrosion proceeds more rapidly in the presence
of the coating of oil than without the coating. This is ascribed
to the fact that the hydrogen, which is always evolved when iron
rusts, is constantly removed by the linseed oil, which (being an
unsaturated hydrocarbon) has the power of absorbing hydrogen
and therefore acts as a depolarizer much in the same way that
mill-scale acts by destroying the so-called " electrolytic double
layer." 1 When, however, linseed oil contains pigments there is
a marked decrease in the power of the oil to remove the hydrogen
and act as a depolarizer.

Perry 2 in commenting on the subject of shop coatings for

1 See Chapters II and III.

2 Bulletin No. 15, Scientific Section, Paint Manufacturers' Association
of U. S.

232 CORROSION AND PRESERVATION OF IRON AND STEEL

metal, states: "The addition of high-grade fossil resins, carefully
compounded with a carefully treated oil, adds greatly to the
power of a paint to resist penetration by gases and moisture,
producing a better excluding paint and at the same time
adding to the appearance. The glossy surface which a paint
made along these lines possesses, renders the paint a better
repellent or resister of moisture. The quality and percentage
of gum used influences to a great extent the wearing properties of
this kind of paint.

"During the transportation of machinery and structural steel
from the factory to the field and the workshop, there is met a
state of condition that causes rapid corrosion. Moisture and
gases attack the metal and assert their destructive action. In
the past these results have been partially overcome by swabbing
the metal with crude oil, in some cases, and, again, by giving the
metal a dip in hot linseed or other drying oils or by applying tar
and cheap paints as shop coats. The crude oil leaves upon the
surface of the metal, even after wiping, a quantity of non-drying
mineral oil which interferes with the drying of the paint coat
which is afterward applied at the time of the assembling of
the metal. It also prevents the paint coat from properly adher-
ing to the steel surface, and this coat of crude non-drying oil,
which still exists between the metal and the paint coat, is a
source of never-ending trouble, causing peeling and shriveling.
This crude oil treatment, therefore, should be avoided when-
ever it is intended that the steel is to be subsequently painted
with oil paints.

"Where linseed oil instead of crude oil is used, a film of the
oil is left upon the metal and rapidly oxidizes to a coat of linoxyn.
This coat will protect the metal for a certain period of time, but
is extremely porous and ultimately admits moisture. If, within
this coating of linseed oil, there had been contained a proportion
of pigment, or if the linseed oil had been developed by gums into
a varnish or lacquer, then the excluding properties of the linseed
oil would have been increased, and, if the formula were inhibitive
in nature, the steel would be better protected from corrosion, and
the application of future coats of paint, after assembling the steel,
would have been practical and facilitated/'

Mulder 1 states his views on the subject of shop coatings as

1 Iron Corrosion. L. E. Andes. Scott, Greenwood & Son, London, 1900.

PAINTS FOR VARIOUS PURPOSES 233

follows: "Since boiled linseed oil dries, on wood, to form a not
very hard layer of varnish, it cannot be anything else but preju-
dicial to the adhesion of the subsequent paint, on iron, when the
paint is separated from the metal solely by a layer of such linseed
varnish, especially since the latter is, according to instructions,
applied to the warm iron, a method which, in turn, cannot pro-
mote adhesion.

"This meaning will become clear if we imagine what would
take place if the iron were coated with copal varnish. It is true
that the residual varnish left by the drying of the linseed oil is
tougher, and in so far better, than copal varnish; still it is a var-
nish all the same. Now it is essential that the iron should be
coated with a very adhesive paint, which will afterwards dry
hard and fast, and that all intermediate layers should be avoided
and dispensed with; consequently the use of linseed oil as a first
coating for iron is to be discouraged."

The authors desire to call attention to one point in connection
with shop coatings, which has very generally been overlooked.
The point referred to is the possible adhesion or absorption of
water on the surface of steel. When kept in a humid atmosphere,
at ordinary temperatures, the surface of steel is never quite dry,
however much it may appear to be so. The consequence is that
these layers of moisture are enclosed between the surface of steel
and the protective coating. That this is a contributing cause
to future trouble, if not failure, can hardly be denied. Mr. J. E.
Stead, F. R. S., 1 a member of the council of the Iron and Steel
Institute of Great Britain, has stated: "That when ships' plates
were made at steel works, identification marks were frequently
made with paints while the plates were hot; and it was remarkable
that ten or twelve years afterwards, when the ships' bottoms
were cleaned they found the steel uncorroded where the marks
had been made. That pointed directly to the best method of
protecting steel and iron. They ought to heat the metal so as to
remove every trace of moisture from the surfaces, and then apply
the paint to the hot metal. There was always a little moisture,
which adhered to the surface of cold steel plates, after they had
once been moistened, even when they appeared dry, and probably
that was responsible in some measure for starting the corrosion
under the paint applied to cold surfaces."

1 Journ. Iron and Steel Institute, I, 1909, p. 98.

234 CORROSION AND PRESERVATION OF IRON AND STEEL

Commenting on this important subject, Wood 1 states: "There
are hundreds of records of the painting of important railway
structures, where the first coat of boiled oil method was used,
and, in the great majority of instances, the utter and rapid failure
of the coating and the extra corrosion of the structure could be
directly assigned to this so-called method of protection. The
weather-resisting power of an oil coating is almost nil as compared
with paint."

"A foundation coat of oil is a direct cause of the blistering
and peeling of the coatings spread over it. It is seldom dried
enough before the other paints are spread over it to ensure a close
adherence to the metal it covers. When the subsequent coats of
paint are spread, the solvents and oils in them soften to some
extent the underlying coat of oil, and a moderate heat from the
sun causes the whole coating to blister or peel. Too much oil
in a paint coating, particularly if the surplus oil is in or near the
foundation coat, whether on a wooden or metallic surface, will
generally cause peeling, regardless of the pigment used in the
coatings.''

That linseed oil when used without pigments is not a desirable
material for the prime coating of metal seems to be the universal
opinion of authorities on the subject of the protection of iron.
Further comment on the subject may be of interest. Brown 2
says: "The most insidious enemy of the iron bridge is rust, and
the primary object of painting it is to protect it from those ele-
ments which cause destruction by rust. Rust is caused by the
combination of the metal with oxygen to form the hydrate oxide
of iron. This oxygen may be obtained from the air, from water,
or from some other substance which acts as a carrier of oxygen
or an oxidizing agent — always, however, in the presence of
moisture. Now, one of the primary things to be considered in
choosing a paint for ironwork is that it shall not contain in its pig-
ment or vehicle any substance which is chemically active in such
way as to convey oxygen to the iron. For if such a chemically
active agent be introduced into the paint, sooner or later it will

1 M. P. Wood. Rustless Coatings. First edition, John Wiley & Sons,
New York, 1905.

2 Edward Hurst Brown. Conditions that Must Be Met in the Ideal Paint
for Railway Bridges. Proc. Third Annual Convention Maintenance of Way
Master Painters' Assn. of U. S. & Canada, New York, Nov. 13 and 14, 1906.

PAINTS FOR VARIOUS PURPOSES 235

promote rather than prevent rust. Of course, so long as the oil,
in an oil plant, remains intact, it envelops the particles of pigment
and keeps them away from the iron, but in time the oil, which
has hardened by absorbing oxygen from the air, begins to dis-
integrate by the action of water coming from rain, hail, snow, or
fog. Moreover, even the freshly applied oil is not absolutely
impenetrable to moisture, as has been shown by numerous experi-
ments, and, however completely the particles of the chemically
active pigment may be covered by an oil film, they will neces-
sarily come in contact with moisture — will decompose the water
and absorb its oxygen, and convey it, together with the hydrogen,
to the surface of the iron to cause rust. 1 For this reason the ideal
paint for a steel or iron bridge should not contain a chemically
active pigment, nor any strongly oxidizing agent in the way of
driers.

We have also seen that linseed oil is permeable to moisture and
to the gases and steam from locomotives. This was first clearly
demonstrated, we believe, by Dudley, and to this fact may be
ascribed the corrosion of the metal under an apparently intact
coating of paint. It is true that in the mixture of oil with pigment
in a very finely divided form, the tendency is for the pigment
particles to more or less fill up the interstices in the oil film and
render it less porous. How completely this is done depends more
or less upon the shape of the minute particles of pigment, and,
as Robert Job demonstrated in a paper read before the Franklin
Institute of Philadelphia, it depends even more largely upon the
fineness of grinding of the pigment particles. The finer the pigment
is ground, the more perfect will be its protective power. This
was shown very clearly by examination of the paint film under
the microscope as well as by actual service test.

Paint Coatings Used at Panama. — At Panama, up to the
beginning of the American engineering regime, the isthmus re-
sembled a vast junk shop, where millions of dollars 7 worth of
metal could be seen in various stages of oxidation and decay.

Little evidence was left to show that paint had been applied
to most of the structural and fabricated metal used by the French

1 This article was written without reference to the electrolytic explana-
tion of corrosion, and is included here to show that independently of the theory
engineers were beginning to recognize that pigments enveloped in oil were
capable of stimulating corrosion.

236 CORROSION AND PRESERVATION OF IRON AND STEEL

engineers. Some of the old locomotives, however, used by the
French were pulled out of the mire and found to be in fairly good
condition. Sections of the paint still adhering were removed by
Speller 1 and, upon his return to the United States, analyzed and
found to consist largely of zinc, lead and small percentages of oxide
of iron used probably as a coloring material. The analysis of
some of these paints appears below, and would almost seem to
indicate that the French possessed years ago, some knowledge of
the inhibitive properties of pigments.

No.

White
Lead

Oxide of
Zinc

Silica

Alumina

Oxide
of Iron

Oil and

Undetermined

matter

White Paint

White Paint (Exca-
vating Machine) . .

White Paint (La
Boca)

Red Paint (French
Locomotive) . . . .

P1A

PIC
P1D

P1B

66.30
58.01
37.86

Red

Lead

70.06

12.10
28.38
13.67

Carb. of
Lead

11.90

11.14

2.27

37.12

3.18

5.21
8.42
6.84

5.10

2.81
1.12

1.66

5.86

2.44
1.80

2.88

3.90

Recent progress in the art of compounding protective coat-
ings makes it possible to design paints superior to the above, and
it is probable that in the future the problem of protecting the
iron work at the canal zone will not be found as difficult as it
has been in the past.

It should be remembered, however, in this connection that
the metal used by the early French engineers and contractors
was superior in quality to much of the metal that is in common
use to-day. The authors must reiterate their opinion that the
problem of protecting iron and steel depends largely upon the
rust-resisting quality of the metal to be protected.

Painting Various Municipal A ccessories, etc. Iron poles for trol-
ley wires are sometimes imbedded in the earth after receiving a coat
of tar. The upper part is then painted. The addition of lime to
the tar used in painting the lower part of the poles is to be advised;
in fact the addition of 20 per cent, of dry lime to the earth used in
tamping around the pole will furnish some protection to the metal.

1 Published by permission of F. A. Speller, National Tube Co.

PAINTS FOR VARIOUS PURPOSES 237

The paint for the upper part of the poles should be inhibitive.
Carbon paints, which have been so largely used for this purpose,
in the authors' opinion are generally stimulative in their action
and should not be used unless on top of a good inhibitive primer.

Lamp posts and letter boxes, which in the past have been
painted with aluminum paint (a mixture of aluminum and zinc
alloy powder suspended in bronzing liquid or collodion lacquer)
have rusted rapidly. The vehicles of such paints are generally
poor excluders and allow the moisture to get through to the
metallic powder spread over the iron surface of the box or post.
The metallic powders, being good conductors of electricity, inevi-
tably lead to stimulated corrosion, for reasons which have been
fully discussed in previous chapters.

Proper protection of lamp posts, boxes, and other articles of this
sort, can best be secured by using a good inhibitive paint formula
ground in oil with or without a percentage of varnish. A good
bronze green, made from zinc oxide, white lead, zinc chromate, and
willow charcoal, makes a most excellent coating for such work, both
from a decorative and inhibitive standpoint. The use of red lead,
provided it is inhibitive in its nature, for fire signal boxes is to be
recommended. Another composition equally as good from an inhib-
itive standpoint may be made from bright red oxide of iron and
basic chromate of lead, the latter possessing excellent inhibitive
values.

Painting Refrigeration Machinery. — The painting of conden-
sers and pipes for carrying brine has presented a problem that has
puzzled the makers of the best paints upon the market. Instances
of the complete failure of several different types of paint tested
on this class of work have come to the attention of the authors
who have recommended the application of a special formula for
this purpose. The formula which has given the best results is
made of asbestine, willow charcoal, red lead and zinc chromate
ground in oil containing forty per cent, of a high-grade kauri gum
varnish. The wonderful results obtained with this inhibitive
paint, which has withstood the formation upon its surface of nearly
two inches of salty ice for over a year, only adds to the weight
of the plea that has been put forward in successive chapters in
this book, namely, that both the pigment and the vehicle should
be given consideration from the standpoint of inhibition, exclud-
ing value and moisture-resisting properties.

238 CORROSION AND PRESERVATION OF IRON AND STEEL

Ornamental Iron Work Protection. — For the protection of
iron grill work, porch railings, etc., etc., the Bower-BarfT and Wells
processes have to some extent been successfully used, but the
great majority of material of this kind must depend upon paint
coatings for its protection. Red lead is very generally used for
prime coating, and if inhibitive in nature serves the purpose well.
In many cases the iron decoration of buildings has been painted
a flat black, using lampblack in oil thinned with a large quantity
of turpentine for flatting. The rapid evaporation of the turpen-
tine leaves a porous surface. The unsightly rusted condition of
much of this sort of work is a matter of common observation.
If flat black is required for the finish, it should be put on over a good
inhibitive primer.

CHAPTER X

THE TESTING AND DESIGN OF PROTECTIVE
PAINTS

The Testing and Design of Paints. — Spreading value should,
of course, receive careful attention in the selection of a protective
paint, but this should not be allowed to be the first consideration.
A paint that has a high spreading value would seem from one
point of view to be the most economical, but if it fails to provide
protection it will prove much more expensive in the end. No
one knows this better than the painter, who is called upon to labor
over rusted surfaces before repainting. Within reasonable limits,
the paint that gives the best protection is the cheapest paint to
use, no matter what its first cost may be. In order to determine
in advance the efficiency of a paint, designed to protect iron from
corrosion, recourse must be had to some form of preliminary
test. In the following paragraphs the authors have endeavored
to present a description of a number of such tests that have been
proposed and used by various authorities, together with some
directions which have a general bearing on protective paint testing.

General Directions in Regard to Testing. — When a series of
commercial paints are to be tested in the field for their relative
merits, it is well to use both the definite and indefinite spreading
rates, making two sets of tests. For instance, if there is sub-
mitted for test six samples of paint, designed for the protection
of iron, two steel plates should be used for each paint. The
plates should be either sand-blasted or pickled in acid, and prop-
erly prepared for the test, as previously outlined. 1 One plate
should be painted with the quantity of paint demanded by the
surface. The operator should, of course, properly brush the
paint out and not leave too much upon the surface, otherwise
wrinkling will ensue. The other plate should be painted with
just the number of grams of paint that is required to produce a
spreading rate of say 900 square feet, if this rate is adopted.
Great care should be taken to have the same operator apply

1 See p. 180.
239

240 CORROSION AND PRESERVATION OF IRON AND STEEL

each paint tested, so that no great variable will be introduced in
the preparation of the plates. The value of such tests is added to
by including plates covered with the ordinary black mill-scale, as
most structural steel is painted in this condition. Positive infor-
mation may be obtained regarding the value of a paint when
the tests are made upon both cleaned and uncleaned plates, as
above outlined, and with definite and indefinite spreading rates.

If the purchase of large amounts of paint is under considera-
tion, samples should be submitted to the chemist for examination.
If no specifications have been laid down for the composition of
the paint, the chemist, if he so desires, may omit the analysis of
the paint and may then proceed with the accelerated laboratory
tests that are best suited to each paint, according to the pur-
poses for which it is to be used. The results of the acceleration
tests may be taken as fairly good evidence of how the paint will
work in practice, but a series of practical exposure tests should
also be started so that definite knowledge can be obtained in
regard to their protective value.

Throughout such tests the greatest care should be taken that
the number of grams of paint applied to each panel is recorded,
and when reporting on the value the number of grams of paint
should be calculated to the spreading rate per gallon, so that the
final report will include this factor.

When testing pigments, if the proportions of pigment to
vehicle, according to specific gravity, as used in the Atlantic City
tests, are not satisfactory to the mind of the engineer making tests
of this nature, the following method may appear more practical.

The pigments should be ground in well-settled raw oil, the
amount of oil being regulated according to the amount necessary
to form a medium paste with the pigment. This paste should
readily break up for thinning, by the addition of more oil, so
that the final product will be in the proper condition for brushing,
without leaving streaked marks upon the plate. For instance,
91 lbs. of white lead requires 9 lbs. of oil to make a grinding paste.
The subsequent addition of 31 lbs. of oil and 2 lbs. liquid drier pre-
pares the composition for the brush. Eighty-four lbs. of zinc grinds
with 16 lbs. of oil to a medium paste, which may be thinned with
60 lbs. of oil and 4 lbs. drier to brushing consistency. In the case
of lampblack, 25 lbs. of the pigment require 75 lbs. of oil to grind
to a paste. The addition of 100 lbs. of oil is necessary to form

TESTING AND DESIGN OF PROTECTIVE PAINTS 241

a paint thin enough to apply to steel. After the oil has been
added for the necessary flowing consistency a small amount of
drier, generally about 12 to 15 lbs., should be added, in propor-
tion to the nature of the pigment, which invariably affects the
drying. For instance, in a paint high in lead and zinc, small
quantities of drier are sufficient, whereas a paint containing
lampblack or carbon black requires a large amount of drier.

In nearly all exposure tests which have been recorded the
pigments have constituted the variable, while the vehicle has
been kept as constant as possible. There can be no doubt of the
fact that more information is needed in regard to different com-
binations between varied vehicles and the principal pigments.

Laboratory Acceleration Tests. — Laboratory tests to deter-
mine the protective power of pigments are of great value pro-
vided they can be depended on to furnish advance information
in regard to the probable action of a given paint compound. A
great deal of work has been done along this line. In 1896, Smith
carried out a series of tests to determine the value of various
pigments, by boiling them in water contained in small iron cups,
and observing the rate of corrosion. Other tests have been
used, of a similar nature.

Small iron cups painted on the inside have been filled with
water, either pure or containing impurities, so as to approximate
the action of the water or gases that the coatings would be
exposed to in service. The liquid is allowed to evaporate, some-
times with the aid of heat, the cups are then again filled and the
operation continued in this way for some time. Observations
are made from day to day, to see whether rusting is taking place,
and a careful record of the first and successive breakdowns is made.

A valuable apparatus for testing the value of paints which
are to be submitted to alternate exposure to liquids and atmos-
pheric gases was designed by Norton. It may be constructed
from a box about 6 feet long, 2 feet wide, and 2 feet deep. In
the middle of each end, on the top of the box, 2-in. steel collars
may be placed to accommodate a 2-in. wood rod. Upon this
rod may be placed a series of iron plates or disks, less than 24
inches in diameter, painted with various paints, and punched in
the middle so that they will slide along the rod. These plates
are kept about 2 inches apart, when making the test. The box
is then half filled with the liquid, which may be salt water, alka-

242 CORROSION AND PRESERVATION OF. IRON AND STEEL

line, or acid water. By mechanically turning the rod, the plates
are kept slowly moving through the liquid in which they are half
submerged. The test may be kept up for several weeks. The
alternate exposure is severe, and some paints last in this test but
a short while.

Another method, which is satisfactory for large samples, is the
acceleration test box, devised by one of the authors, for testing
painted surfaces in an artificial atmosphere, containing various
percentages of carbonic acid gas, ammonia, etc., duplicating
atmospheric conditions in those localities in which the paint is to
be used. This box may be also used for testing various steels,
for their resistance to corrosion. The box is about four feet long,
2 feet wide, and 2 feet deep, and is fitted with glass sides and
top. The top of the box contains a row of girders, from which
the samples are suspended from small hooks. The side of the
box is fitted with a tubular opening, through which carbon dioxide
may be passed. Water is kept in a tray at the bottom of the
box, and thus the atmosphere of the box is kept constantly humid.
The temperature of the box may be adjusted to suit the require-
ments of the test. Ozone may be developed with a Ruhmkorff
coil and sparker. In fact, any desired atmosphere can be intro-
duced into the box. Samples of painted surfaces exposed to
this test have failed in a remarkably short time.

Another acceleration test has been proposed by Loesner, and
is described in Andes' work on Iron Corrosion, as follows: "When
the plates of sheet iron are coated with paint, rust forms — as
is well known — on the surface of the metal after a short time.
Owing, however, to the slowness of the reaction, immersion in
water does not afford a suitable means for determining the dura-
bility of paint, whereas, on the other hand, steam places a very
convenient method at our disposal for this purpose. With this
object, plates of sheet iron are cleaned perfectly bright on one
side, by means of emery paper, the clean surfaces — which must
not on any account be touched with the fingers — being then
coated over with a thin uniform layer of the paint to be tested, and
left to dry at the ordinary temperature for four days. The plates
are next set, painted side downwards, over a boiling water-bath,
so that the surface of the paint is just 2 inches (5 cm.) above
the constant level of the water. At the end of fifteen hours, the*
plates are dried for a short time at 100° C. (not higher), and the

TESTING AND DESIGN OF PROTECTIVE PAINTS 243

layers of paint impregnated with aniline oil applied with a soft
brush. Being thus softened they can be removed, and the metal
is then dried by means of alcohol. For the paint to be classed
as good, the metal must have remained perfectly intact, a condi-
tion readily recognized by the appearance of the scratches formed
on the surface by the emery paper used in the initial cleaning.
Many paints will stand this steam test after the painted iron has
been heated to 100° to 105° C. for about five hours."

A novel method for testing the pigment portion of paints, to
determine whether the pigment is inhibitive or stimulative, was
developed by one of the authors during his researches. The
pigments to be tested are rubbed up with sufficient water to
make thick water-color paints, and are then flowed or brushed
upon the clean blades of steel table knives. After the coatings
are dry the knives are laid on a wet blotter and covered with a
sheet of wet blotting paper. At the end of forty-eight hours the
surfaces are cleaned off with running water and a stiff brush.
This acceleration test has been found to give results which agree
with the oil film tests.

The appearance of the knife blades after being used in this
test is shown in Fig. 68a and 6. 1

Perry developed the foregoing test so that it could be made
applicable to the examination of paints ground in oil. He
recommends that a portion of the paint when it is received
for examination be placed in a test tube and shaken several
times with benzine, in order to extract the vehicle, the extractions
being disregarded and the separated pigment used for the test.
Small centrifugal machines for the accommodation of the test
tubes may be used for the extraction. The pigment is removed
from the test tube and placed upon a sheet of druggists' filter
paper. It is then moistened and with the finger rubbed to a
soft paste. A razor blade, or other small piece of steel highly
polished, is then covered with the paste together with the paper,
and laid upon a porcelain dish within a cigar box. Several differ-
ent pigments may be tested in the same way and included in the
same box. The box is kept constantly humid by lining the sides
and top with druggists* filter paper kept constantly moist with

1 One of the authors has recently devised a new form of acceleration
test which promises to give valuable results but which has not been suf-
ficiently developed to publish.

244 CORROSION AND PRESERVATION OF IRON AND STEEL

fee v

Fig. 68a and b. — Showing the appearance of knife blades used in
Cushman's test.

TESTING AND DESIGN OF PROTECTIVE PAINTS 245

water. At the end of a few hours the box may be opened and
the steel plates removed from their wrappings of paper and pig-
ment. After washing and scrubbing with a tooth brush, in order
to remove any pigment which persists upon the metal, the amount
of corrosion which has taken place may be easily ascertained.
Another acceleration test which gives valuable information has
recently been devised by one of the authors, but as the work
was done in a government laboratory and has not yet been pub-
lished, it cannot properly be included here.

It is natural that there should be considerable difference of
opinion among both paint manufacturers and consumers in
regard to the value of acceleration tests. It is noteworthy that
the same discussion has been going on for years in regard to
Portland cement, and yet all cement is bought and sold subject
to acceleration tests. It is the authors' firm conviction that
the same conditions will eventually be developed in the purchase
of protective paints. Extravagant claims made by manufac-
turers will not always satisfy the consumer, and "best by test"
will inevitably replace "best by claim." There are perhaps a
thousand "best" protective paints in the market, and among
them the consumer is at present at a loss to decide. In no section
of technical industry are standard tests more necessary than in
the one under discussion here.

Design of Protective and Inhibitive Paint Coatings. — In the
manufacture of paints, the various pigments are usually ground
in oil to a paste, and stored away in that form in mixing tanks.
This practice is followed by the larger manufacturers who have
the proper facilities. When an order is received for any special
combination, the proper weights of each paste, according to the
percentage composition of the paint desired, are mixed together
and thinned with the necessary amount of oil and drier. For
instance, if a white paint was desired for iron and steel, and, in the
judgment of the engineer ordering the paint, the following compo-
sition was considered best to use in the locality where the work
was to be done:

Pigment Composition

Basic Carbonate — White Lead 67 per cent.

Zinc Oxide 20 per cent.

Asbestine 3 per cent.

Calcium Carbonate 10 per cent.

246 CORROSION AND PRESERVATION OF IRON AND STEEL

with this working formula the paint-maker could proceed with the
manufacture of the product. If about five gallons of this paint was
desired, the following amount of paste would be used, these amounts
being proportioned to the percentage composition of the mixture:

Basic Carbonate — White Lead in Oil

Zinc Oxide in Oil

Asbestine in Oil

Calcium Carbonate in Oil

Raw Linseed Oil (2§ gals.)

Drier (& gal.)

Turpentine ( T \ gal.)

Weight per gallon, 16 lbs., 5 oz.

Analysis of this paint would show the following composition:

42.5

lbs.

14.2

3.3

6.6

20.6

1.25

88.95 lbs.

Per Cent,
in Pigment

Per Cent,
in Vehicle

Per Cent, in

Total Pigment

and Vehicle

Basic Carbonate — White Lead

Zinc Oxide

Asbestine -

Calcium Carbonate

Raw Oil

Drier

Turpentine

67.01

19.89

3.86

9.24

100.00

94
3
1

100

89
81

42.84

12.71

2.47

5.91

34.02

1.40

.65

100.00

This paint could be tinted to any color desired. For instance,
100 lbs. of the white base could be tinted with 5 lbs. zinc chromate,
in order to get inhibitive yellow paint, or with 3 lbs. of a good
inhibitive black pigment, in order to secure a good gray tint.

For special formulas, however, the pigments specified in the
formula are generally weighed out in their proper proportion
and mixed and ground in linseed oil. If the formula called for
200 gallons of green paint, and the pigment specified was as follows:

Special White 52 per cent-
Silica 22 per cent.

Zinc Yellow , 7 per cent.

Medium Chrome Yellow 8 per cent.

Prussian Blue (Inhibitive) 6 per cent.

Whiting 5 per cent.

TESTING AND DESIGN OF PROTECTIVE PAINTS 247

this could be made by weighing out the following amounts of
pigments in oil, mixing and grinding:

750.0 lbs.

Special White

312.5 "

Silica

112.5 "

Zinc Chromate

125.0 "

Neutral Lead Chromate

87.5 £f

Prussian Blue (Inhibitive)

62.5 "

Whiting

600.0 "

Raw Oil

2050.0 lbs.

This paint could be thinned for application with:

2050 lbs. of above paste
100 £< Drier
850 " Raw Linseed Oil

3000 lbs.

This paint would weigh about 12J lbs. per gallon.

Inhibitive Paint Formulas. — The authors have frequently
been asked to design protective and inhibitive paints for special
purposes.

While it would appear unwise for the authors to attempt
standardization of paint formulae, or to define the percentage
composition of paints intended for iron and steel, it is believed
that to give the formulas of a few paints which have proven of
value in field tests would be a matter of interest to the manufac-
turers of protective coatings, as furnishing a basis upon which
to make further tests. In various parts of this book there has
been given a general outline of the composition of paints of espe-
cial value for special purposes. The following formulas would
be of value for general use on iron work :

Brown-Zinc

Metallic Brown (neutral) 60 per cent.

Zinc Lead 20 per cent.

Zinc Oxide 20 per cent.

Chocolate Color

Metallic Brown (neutral) 90 per cent.

Willow Charcoal 5 p er cent.

Zinc Chromate 5 per cent.

248 CORROSION AND PRESERVATION OF IRON AND STEEL

Red

Bright Red Oxide (free from acid or sulphur) 95 per cent.

Zinc Chromate 5 per cent.

Red

Bright Red Oxide 65 per cent.

China Clay 15 per cent.

Red Lead 15 per cent.

Zinc Chromate 5 per cent.

Black

Willow Charcoal and Bone Black 68 per cent.

Zinc Chromate 2 per cent.

Inert 30 per cent.

Steel-Black

Sublimed Blue Lead 60 per cent.

Willow Charcoal 20 per cent.

Inert 20 per cent.

Green

Zinc Oxide 45 per cent.

Sublimed White Lead 25 per cent.

Inert 15 per cent.

Zinc Chromate 5 per cent.

Inhibitive Prussian Blue 5 per cent.

Medium Chrome Yellow 5 per cent.

Green

Inhibitive Prussian Blue 4 per cent.

Zinc Chromate 3 per cent.

Chrome Yellow 8 per cent.

Asbestine 15 per cent.

Zinc Lead or Zinc Oxide and Corroded 1

White Lead 70 per cent.

White

Zinc Oxide 60 per cent.

Corroded or Sublimed White Lead 30 per cent.

Asbestine 10 per cent.

White

Zinc Lead 20 per cent.

Zinc Oxide 20 per cent.

Corroded or Sublimed White Lead 50 per cent.

Kaolin or Silica 10 per cent.

- If corroded lead is used, it must be in small proportion on
account of its action on the green.

TESTING AND DESIGN OF PROTECTIVE PAINTS 249

Wire Fence Paint. — On the wire fences at Pittsburg several
paints of the inhibitive and stimulative pigment types were
tested. Excellent results were obtained with the use of a paint
of approximately the following composition:

Natural Bright Iron Oxide 65 per cent.

Silica 20 per cent.

Willow Charcoal 5 per cent.

Red Lead 5 per cent.

Zinc Chromate 5 per cent.

100 per cent.

This paint, ground in oil alone, would afford good protection
to wire, but when ground in a vehicle of the following nature,
it would prove of even more value:

Boiled Linseed Oil 23 per cent.

Raw Linseed Oil 45 per cent.

Kauri Varnish 20 per cent.

Turpentine and Drier 12 per cent.

100 per cent.

Another paint showing good results on the Pittsburg wire
fences was of the following composition :

Zinc Chromate 90 per cent.

Inhibitive Prussian Blue 10 per cent.

100 per cent.

This paint, however, is quite expensive on account of the
large percentage of zinc chromate.

The carbon black and graphite paints applied to the wire
fences showed very bad results, corrosion becoming apparent
within six months after their use.

Condenser Paint. — The use of a formula of the following com-
position will be found of value in painting condensers and pipes
subjected to low temperatures:

Pigment

Willow Charcoal or Drop Black 40 per cent.

Red Lead 25 per cent.

Asbestine and China Clay 30 per cent.

Zinc Chromate _5 per cent.

100 per cent.

250 CORROSION AND PRESERVATION OF IRON AND STEEL

Vehicle

Boiled Linseed Oil 20 per cent.

Treated China Wood Oil 20 per cent.

Raw Linseed Oil 50 per cent.

Turpentine and Drier 10 per cent.

100 per cent.
Or the Boiled Linseed Oil and China Wood Oil may be
replaced with Copal Varnish.

Paint for Iron Piping. — A paint that would prove efficient
for use at the seashore for protecting piping exposed to the weather,
could be made of the following composition :

White Lead 55 per cent.

Zinc Oxide 20 per cent.

Silica 10 per cent.

Willow Charcoal 10 per cent.

Zinc Chromate 5 per cent.

100 per cent.

Formula Labeling. — The purchasing agent of the large cor-
poration is often at a loss, when buying paint, to know whether
he is securing that grade for which he has contracted. The
labeling of house paints in various States by certain manufac-
turers was brought about by legislation as well as by a desire for
honesty. The analysis printed upon the label is generally con-
sidered as an evidence of the manufacturer's desire to make a
good paint, otherwise the analysis would not appear. Formula
labeling on paints for iron and steel may come in the near future,
and . such an innovation will probably be welcomed by the
engineer, architect, purchasing agent, and consumer.

CHAPTER XI

THE PROPERTIES OF PIGMENTS

The Requisites of Protective Coatings. — Although this book
deals largely with paints for the protection of steel and iron, the
wide interest of late that has been exhibited by the engineer,
regarding the composition and value of various paints, demands
that a brief description of the physical properties of pigments
and the prime requisites of paints should also be included.

A proper understanding of certain basic laws regarding
materials for the fabrication of protective coatings is absolutely
essential to the factory superintendent or chemist who is called
upon to specify paints, and should also be of interest to the painter,
architect, and engineer who constantly use and depend upon
these products.

Meaning and Cause of Hiding Power. — Structural material
of various kinds differ in the demands made upon the paint used
for their protection and decoration. That the film should
have proper covering and hiding power, or opacity, is essential.
Opacity in a paint is dependent upon the difference in the refract-
ive indices of the pigment and the vehicle with which the pigment
is mixed. The farther apart the indices of refraction of the com-
posite pigment and the vehicle, the greater the opacity. Greater
protection from the powerful rays of the sun is afforded the struc-
tural material upon which opaque paints are applied. Proper
hiding of the surface is necessary from a protective as well as a
decorative standpoint.

Whiteness. — White pigments should not darken when made
into paint. White lead, zinc oxide, and barytes are all examples
of white pigments, which when ground in oil produce white paints.
The barytes in such a formula becomes transparent in its oil
coating, but has no darkening influence on the composition.

Lithopone, when ground in oil and painted out, gives a whiter
surface than any other pigment. For outside work, however, it
is not well suited on account of its liability to turn dark. Zinc

251

252 CORROSION AND PRESERVATION OF IRON AND STEEL

oxide is generally taken as a standard of whiteness, and is used
for comparison with various other pigments when determining
their degree of whiteness.

Stable and Chemically Active Pigments. — Pigments are some-
times classified according to their behavior in oil, as " chemically
active" or "inert." The term "inert pigments" generally applies,
however, to that group of pigments which when ground in oil
show relatively slight opacity but extreme stability toward the
oil. There are some white pigments which are of great value
for their hiding power and other properties, but which, because
of their basic or alkaline nature, show chemical activity towards
the oil in which they are ground, causing saponification, and final
disintegration of the paint film. Such pigments are chemically
active, and are often the cause of chalking, checking, and dis-
coloration.

Spreading Value. — The number of square feet over which a
paint is spread is an important consideration. Some pigments
in oil have a relatively low, while others have excessive spreading
quality. Too high-spreading quality causes the formation of
thin films which do not properly resist the action of the elements.
A number of authorities believe that a combination of pigments,
in proper percentages, of both types, gives a paint of improved
spreading quality, and makes it easier to produce a film of the
proper thickness.

Effect of Gases on Paints. — Some pigments are not affected
by sulphurous gases, while others darken under such exposure,
and are materially changed in color. Zinc oxide and basic sul-
phate-white lead are good examples of base white pigments upon
which sulphur gas has practically no effect, while basic carbonate-
white lead is an example of a pigment easily affected, the black
sulphide of lead being the product formed.

Paint Coat Strengthened and Preventives to Settling. — Mag-
nesium and aluminum silicates (asbestine and talcose) have
been suggested and much used for reinforcing paint coats. These
pigments, the former in needle-shaped particles, and the latter
in flake form, are believed by many authorities to possess
value in overcoming the defects of certain prime white pigments
when used in too large a percentage. It is claimed that they not
only increase the abrasion resistance of the film, but render it
much stronger. They are also largely used by the paint manu-

THE PROPERTIES OF PIGMENTS 253

facturer to-day as anti-settlers, as they have the property of hold-
ing up heavy pigments and preventing the settling and hardening
of paints in which they are used. The holding up or non-settling
property of paints is vitally necessary. Anti-settlers should be
used with moderation, however, as excessive use may be con-
sidered an adulteration.

Excluding Properties and Elasticity. — A paint coating for iron
should possess the property of being an excluder of moisture. It
should be impervious to rains and storms. It should also possess
enough elasticity to conform with the expansion or contraction of
the materials upon which it is used. This elasticity is largely
dependent upon the proper regulation of the percentages of pig-
ments contained in the paint. Linseed-oil films are very elastic,
but through progressive oxidation they soon become deficient
in this property, whereas linseed-oil films, containing various
pigments in the proper proportion, maintain their elasticity and
increase in tensile strength.

That the pigments which go into the formation of a paint should
be of different sizes, is the conclusion of Perry 1 in his elaborate
researches upon the physical properties of paint films. Following
is a brief resume of the work of this authority : — The law of mini-
mum voids which applies to concrete also applies to a paint
coating. The pigments which are made by sublimed processes
are extremely fine, and serve the same purpose as sand in con-
crete, while pigments coarser in nature are comparable to the
broken stone or rock. The use of asbestine has been likened to
the use of reinforcing rods and wire mesh in concrete, and adds
strength to the paint coating. — If Perry's conclusions are
accepted, it is evident that a paint, made by combining in certain
percentages pigments of each type, is far superior to a paint
made of one single pigment.

Working Properties of Paints. — The so-called tooth of pig-
ments is a term applied by the master painter to paint which
works well under the brush, and gives proper penetration without
too much slipping greasiness or excessive flow. The use of pig-
ments, such as silica and silicates and barytes, gives this so-called
tooth.

In concluding this general and brief discussion of the proper-
ties of pigments in their relations to paint films, the authors have
1 The Physical Properties of Paint Films.

254 CORROSION AND PRESERVATION OF IRON AND STEEL

endeavored to present all statements made, from an unbiased point
of view. There are a number of paint authorities who believe that
a paint compounded of a single good pigmentary substance, and
an efficient vehicle provides as adequate protection as it is pos-
sible to obtain. While not agreeing with these authorities, the
writers believe that their views should be given the fullest con-
sideration. In the course of time thorough testing will remove
these varying opinions from the realm of theory to that of fact.

Description of Pigments in General Use for Painting Iron. —
The following descriptions of the properties of the various pig-
ments is given with a view of presenting information in a concise
form, which it is believed will be useful to many persons on whom
the responsibility for protective paint design falls.

Basic Carbonate-white Lead. — Basic carbonate-white lead is
made by several processes:

The Old Dutch Process;

The Carter (or Quick) Process;

The Precipitation Process;

The Electrolytic Process;

The Mild Process, etc.

In the Old Dutch Process, pure lead buckles are placed in
clay pots with dilute acetic acid, the pots being subsequently
stacked up and covered with boards and tan bark. The fermen-
tation of the tan bark raises the temperature and causes the
formation of carbon dioxide, which acts on the acetate of lead
formed within the clay pots, producing a basic carbonate of
lead. After two months or more the action is completed and the
white lead is broken up, ground in water, floated to separate the
"blue" or uncorroded lead, then dried in copper pans. It is
sold dry or ground in oil.

In the Carter Process, the lead is made in two weeks by act-
ing upon finely atomized lead particles with dilute acetic acid and
carbon dioxide gas (the latter generated from burning coke)
within large revolving cylinders.

The Mild Process uses no acid, depending upon the oxidation
effect of air blown upon finely divided particles of lead, agitated
in water. The hydrate of lead thus formed is subsequently
carbonated. This material does not require the washing neces-
sary for lead made by other methods of manufacture, as it
contains no free acetic acid.

THE PROPERTIES OF PIGMENTS 255

By the above processes, a white product having the composi-
tion 2PbC0 2 • PbOH 2 , is formed, generally containing about 85
per cent. PbO, and 15 per cent, carbon dioxide and water. It
is soluble in dilute mineral acids or acetic acid. The acetic acid
solution is generally used for the assay of white lead, and may be
precipitated as lead chromate with standard volumetric bichro-
mate solutions. Lead salts may be precipitated with hydrogen
sulphide, to form the black sulphide of lead, and this reaction is
also used in analytical work. All forms of basic carbonate-white
lead are easily reduced to metallic lead, by the use of the blow-
pipe, in the presence of a reducing substance such as charcoal.

Basic carbonate-white lead has a specific gravity of 6.8, and
grinds in 9 per cent, of oil to a stiff paste. It may then be thinned
to a working consistency with 38 per cent, of oil. As generally
reduced by the master painter, 100 lbs. are mixed with 4 to 6
gallons of oil, with the addition of one quart of turpentine and a
pint of drier, the amount of oil or turpentine varying with the
nature of the surface to be painted and the conditions prevailing.
Basic carbonate-white lead is an extremely opaque pigment and
possesses excellent body. It is somewhat deficient in spreading
rate and is, therefore, generally mixed with zinc oxide or other
pigments of high spreading values, when used for general paint-
ing purposes. It is easily affected and darkened by sulphurous
gases, and should not be used with other pigments containing
sulphur in any form.

It is claimed that on account of its alkaline nature it acts
upon the saponifiable oil in which it is ground, forming lead soaps,
which are the active cause of the chalking of white lead — the
greatest evil attending its use. Solubility in carbonic acid of
the atmosphere and decay in the presence of sodium chloride may
be active causes of the rapid chalking of this pigment at the
seashore. Checking in some climates appears to proceed rapidly
on white lead, in a deep hexagonal form, leaving a series of rough
crests and cracks. This checking is secondary to the chalking
which takes place.

It generally has inhibitive tendencies, and is a valuable con-
stituent of certain paints intended for the protection of iron and
steel.

Landolt 1 says: "White lead, used alone and in a pure state,
1 Iron-Corrosion and Anti-Corrosive Paints. L. E. Andes.

256 CORROSION AND PRESERVATION OF IRON AND STEEL

is not a good paint for iron-work. Apart from the fact that the
pure white of the pigment will speedily become impaired and
dirty, the paint also sets hard in a short time, the elasticity dis-
appears, and cracks are formed. For this reason white lead is
mixed with other substances, in the first place to impart color
to the paint, and furthermore for the purpose of increasing its
power of absorbing oil, this latter, or linseed varnish, being the
principal agent determining the durability of the coating.

"The more oil required by a pigment in order to produce a
distributable paint, the better will it be adapted for use on iron,
especially in the open air, provided it satisfies the other condi-
tions, of covering power, neutrality, and capacity of resistance
to acids."

Zinc Oxide. — Zinc oxide is a very valuable prime white base
pigment, the larger quantity being produced by the roasting and
sublimation of Zincite and Franklinite, the latter being found in
large quantities at Franklin Furnace, New Jersey. The New
Jersey zinc oxide runs about 99 per cent, pure, while that pro-
duced in Wisconsin contains upward of 5 per cent, of lead sulphate.

Another variety used, and which is usually about 99.5 per
cent, pure, is produced by the sublimation and oxidation of
spelter, this variety being the standard for whiteness, and known
as French Process Zinc Oxide.

This pigment has a specific gravity of 5.2, and grinds in about
16 per cent, of oil to a stiff paste, 100 lbs. of which may be thinned
for application with 64 lbs. of oil. It is a very opaque pigment,
and possesses excellent spreading properties, being used in admix-
ture with basic carbonate-white lead for outside paints. Being a
very white fine pigment by nature, it acts well with the lead
pigments, and helps to overcome the chalking tendency which
the latter may possess. Because of its stable nature, and extreme
whiteness, it is an excellent base for delicate colors, bringing
out their purity of tone. Zinc oxide has good drying properties
of itself, but the use of litharge aids its drying and is the safest
kind of drier to use with this pigment. Progressive oxidation
gives zinc oxide a hard surface, and consequently does not resist
the changes in temperature as well as when in combination with
a softer pigment. Zinc oxide is invaluable as a pigment base for
enamels, being possessed of great body and whiteness, and giving
a high, permanent gloss when ground in varnish. The checking

THE PROPERTIES OF PIGMENTS 257

of zinc oxide is very regular and characteristic, being of a distinct
triangular form. Nearly all zinc oxides are fairly good inhibitive

pigments.

Zinc oxide is soluble in most mineral acids to a clear solution.
Solutions of zinc oxide in acetic acid may be titrated with standard
ferrocyanide of potash solutions, this being a quick volumetric
method used in its determination. Solutions of zinc may be
precipitated with sodium carbonate, and the white zinc carbonate
thus formed ignited to the oxide and weighed as such.

Basic Sulphate-White Lead (Sublimed White Lead) . — Basic
sulphate-white lead (sublimed white lead) is made from galena,
an ore containing a very high percentage of sulphide of lead.
This ore is mixed with fuel, placed in furnaces, and volatilized.
The fume is brought into contact with air, and becomes oxidized
to a basic sulphate of lead. The fumes are then drawn through
flues and air-cooled pipes to bags or receivers.

Basic sulphate-white lead is an amorphous pigment, of great
fineness, and almost of the same characteristic whiteness that is
exhibited by basic carbonate-white, lead. It has a specific gravity
of 6.2, and grinds in about 10 per cent, of oil to a stiff paste that
may be thinned for application, to brushing consistency, by the
addition of 38 per cent, of oil to 100 lbs. paste.

This pigment possesses extreme chemical stability, and is
not subject to the blackening caused by sulphurous gases, as is
the case with some pigments. Delicate colors, such as greens,
blues, and yellows, may be mixed with this pigment without any
action on the color. This stability renders it invaluable for
many tinted paints.

The tendency of certain lead pigments to liver or become
thick in cold weather is often due to the presence of sulphites,
but a pure basic sulphate-white lead, free from sulphites, does
not exhibit any great amount of stiffness under the brush during
cold weather.

Commercial basic sulphate-white lead usually runs very uni-
form in its composition, analysis showing it to contain approxi-
mately 70 per cent, of lead sulphate, 20 per cent, of lead oxide,
and 5 per cent, of zinc oxide. The lead sulphate is chemically
combined with the lead oxide, forming a stable and definite
chemical compound. An examination of the pigment under the
microcsope shows absence of crystals, thus proving that the lead

258 CORROSION AND PRESERVATION OF IRON AND STEEL

oxide present is chemically combined and not in its regular crys-
talline form. Lead oxide is insoluble in ammonium acetate, but
when chemically combined with lead sulphate it readily goes
into solution when treated with ammonium acetate. This solu-
bility of basic sulphate of lead is utilized by the analytical chemist
in the detection, separation, and estimation of sublimed white
lead in mixed paints.

Sublimed Blue Lead. — Sublimed blue lead is made by burn-
ing coarsely broken lumps of galena, admixed with bituminous
coal, in a special form of furnace. The fumes which are vola-
tilized from this mixture are very complex in their chemical make-
up, and in color are white, blue, and black. After being drawn
through the cooling pipes by the suction of huge fans, whereby
the fumes are cooled, the pigment is deposited in bags. This
pigment is bluish black in color, and has been highly recommended
for use on iron and steel. Its composition runs approximately
as follows:

Lead Sulphate 50 per cent.

Lead Oxide 35 per cent-
Lead Sulphide 5 per cent.

Lead Sulphite 5 per cent.

Zinc Oxide 2 per cent.

Carbon 3 per cent.

The color of the pigment is largely due to the carbon and the
lead sulphide. Its specific gravity is 6.4, and it grinds in 10
per cent, of oil, to a stiff paste, 100 lbs. of which may be thinned
with about 26 lbs. of oil to working consistency. Some manu-
facturers use it in mixture with iron oxide and other pigments for
the production of paints for metal surfaces. Wood and others
have found it of value for this purpose. It has a tendency to
chalk, but this may be overcome by admixture with other pig-
ments such as zinc oxide and iron oxide. Lane has found it of
especial value admixed with lampblack.

The selection of this, as well as all other pigments, should,
in the authors' opinion, be decided by tests, and as far as possible
by the observation of its protective action under service condi-
tions.

Lithopone. — Lithopone is probably the whitest pigment
known, and is extremely desirable for the manufacture of high-
grade enamels, the amount consumed for this purpose being very

THE PROPERTIES OF PIGMENTS 259

large. It is manufactured by the double decomposition of zinc
sulphate and barium sulphide, two soluble salts which interact
to produce two pigments chemically combined, namely, zinc sul-
phide and barium sulphate. The resulting product is filter-
pressed, and furnaced, after which it is rendered more opaque by
disintegrating the heated pigment in cold water. It is after-
ward thoroughly washed, and again filter-pressed, then dried and
ground. The production of this pigment is carried out in the
factory under certain definite and critical temperatures. Proper
control of the solutions for their purity and strength is constantly
maintained, and the greatest care throughout the process is neces-
sary in order to produce a high-grade product.

Although lithopone is a very stable pigment, it has very
peculiar photogenetic qualities, and when exposed to the action
of the actinic rays of the sun, in the presence of moisture, darken-
ing of the pigment results. The pigment sometimes resumes its
normal color, however, in a very short time after the darkening.
In the absence of moisture, this darkening cannot take place.
It is most excellently suited as a pigment for inside use, but for
outside use it requires a very large percentage of some more stable
pigment, such as zinc or calcium carbonate, in order to prevent
this rapid darkening and to prolong its life. .

Lithopone has a specific gravity of 4.25, and grinds in about
13 per cent, of oil to a -paste that may be subsequently thinned
with 60 lbs. of oil to 100 lbs. paste in order to form a paint of
the proper flowing consistency. When spread, it shows high cov-
ering value and brushes exceedingly well. This pigment is some-
times a fairly good inhibitive pigment, but must be admixed with
other pigments when designed for use on iron or steel, because of
its rapid disintegration when exposed outside. Lead driers should
not be used with lithopone.

Zinc Lead White. — This pigment is made by the reduction
of sulphur-bearing ores of lead and zinc, in especially constructed
furnaces. Volatilization of the lead and zinc, and subsequent
oxidation, takes place, with the formation of a pigment fume
consisting of about equal parts of zinc oxide and lead sulphate.
The fume is cooled by passage through long metal pipes, and
collected in bags. Subsequent treatment for whitening and
desulphurization is then made, and the pigment as placed upon
the market is extremely fine in its particle size.

260 CORROSION AND PRESERVATION OF IRON AND STEEL

This pigment has a specific gravity of 4.4, and grinds in about
12 per cent, of oil. It is a most excellent pigment for use in
tinted paints, but is seldom used alone for white paints, because
of a very slight yellowish tint. It has proved of considerable
value when mixed with varying percentages of white lead and
zinc oxide, and some inert pigments. In its chemical stability
it resembles basic sulphate-white lead, and is very well suited
for the manufacture of paints containing delicate colors, such as
blues and greens. It is generally somewhat inhibitive, and in
this case may be used with safety on iron and steel.

Leaded Zinc. — In manufacturing this pigment, various grades
of lead and zinc ores are mixed, roasted, and furnaced. The
fume therefrom is cooled through pipes, and collected in bags;
the process in some ways resembling those used in the manu-
facture of zinc lead white and sublimed white lead. This pig-
ment has a specific gravity of 5.8 and the analysis shows the
presence of about 25 per cent, of lead sulphate and 75 per cent.
of zinc oxide, with traces of zinc sulphate and sulphur dioxide.
The product is a very stable pigment, resembling zinc oxide and
zinc lead white. This pigment oftimes contains a small percent-
age of zinc sulphate which may affect the life of the paint in which
it is used. Zinc sulphate also has the effect of causing paints to
liver to some extent. Because of the presence of a soluble salt
such as zinc sulphate it must be used with care when intended as
the basis of a paint for iron and steel.

Barium Sulphate (Barytes). — Barium sulphate, or barytes, is
one of the most important pigments used in the manufacture of
paints. It occurs as a mineral in large quantities and very widely
distributed. In its preparation for the market, the mineral is-
ground, washed in acid to free it from iron and to whiten its
color. It is then washed several times and dried. Flotation of
the pigment produces a very fine grade. The pigment is used
in large quantity as a base upon which to precipitate colors, and
also together with other white pigments in the manufacture of
ready-mixed paints. It renders the paint coating more resist-
ant to abrasion, and gives to the paint certain very important
brushing qualities. It is a very stable pigment, not being ma-
terially affected by either acid or alkali and can be used with the
most delicate colors. In oil, it is transparent and must be mixed
with opaque pigments when used in mixed paints. It is a very

THE PROPERTIES OF PIGMENTS 261

heavy pigment, having a specific gravity of 4.4, and grinding in
about 10 per cent, of oil. It is generally used with lighter pig-
ments, such as asbestine, in order to prevent its settling.

Blanc Fixe is an artificial form of barium sulphate, made by
mixing solutions of soluble barium salts with sodium sulphate
or other soluble sulphates, causing precipitation of barium sul-
phate. This material has a somewhat lower gravity (4.2) than
the natural barytes and does not have the same tendency to
settle out when used in paints. This pigment is also used as a
base for colors and also in ready-mixed paints, giving good brush-
ing qualities to the paint. It has more opacity in oil than barytes
and for some purposes is better suited than the natural form of
barium sulphate. Both barytes and blanc fixe are liable to con-
tain traces of acids or acid salts, and they must be carefully
tested, before using in paints to be applied to iron or steel.

Gypsum (Calcium Sulphate). — Gypsum, or calcium sulphate,
is found in nature very widely distributed. In its natural form it
contains about 20 per cent, water of combination. It has come
into wide use in the manufacture of various colors, and is often
found in ready-mixed paints. It is a very stable pigment, and,
although it lacks any hiding power in oil, when ground in water
it is very opaque, and is largely used as a base for distemper
colors. It is slightly soluble in water, however, and when pres-
ent in a paint coating it is liable to leach out. It should never
be used upon iron or steel, because of the corrosion which it is
certain to cause, due to the ease with which it is ionized in the
presence of water. It has a specific gravity of 2.3 and grinds
in about 22 per cent, of oil. This pigment is present to a great
extent in Venetian red.

Magnesium Silicate (Asbestine and Talcose). — This pigment
comes in two forms: as asbestine and as talcose. The former is
very fibrous in nature and is a very stable pigment to use in the
manufacture of paint on account of its inert nature and tendency
to hold up heavier pigments, and prevent settling. It also has
the property of strengthening a paint coat in which it is used.
The talcose variety is very tabular in form, and is also somewhat
largely used in the manufacture of mixed paints. Both varieties
are transparent in oil, and they are very inert toward iron or steel.
They have a gravity of about 2.7 and grind in about 32 per cent,
of oil.

262 CORROSION AND PRESERVATION OF IRON AND STEEL

Whiting (Calcium Carbonate) . ■ — Whiting, or calcium car-
bonate, occurs very widely distributed in nature, as chalk. After
proper purification' and grinding it forms a pigment with a specific
gravity of about 2.8, which grinds in about 20 per cent, of oil.
It is used largely in distemper work and also to a considerable
extent in the manufacture of oil paints, being considered very
valuable in small percentages in neutralizing any free acid con-
tained in the linseed oil. It spreads very well and does not settle
to any marked extent. The artificial, precipitated form of cal-
cium carbonate is much lighter in gravity (2.5), and requires
about 25 per cent, of oil for grinding. This form possesses more
hiding power and, although it has a tendency to chalk, it is very
valuable for certain purposes. The precipitated form should not
be used in inhibitive paints, owing to the fact that it generally
contains occluded impurities which tend to stimulate corrosion.

Aluminum Silicate (China Clay). — Aluminum silicate, or
China clay, is a widely distributed pigment found in granitic
formations. This pigment plays an important part in the
manufacture of paints. It is a fine, amorphous, white powder,
extremely permanent, and when ground in oil shows very little
hiding property. It has a specific gravity of 2.6 and grinds in
28 per cent, of oil. It is a very stable compound, not being
affected by ordinary acids.

Silica. — This is a most valuable white pigment, used in
immense quantities for wood fillers, and in moderate percentages
in mixed paints, both for iron and for wood. It is usually of
great purity, often running over 97 per cent. SiCh, and when
ground in oil becomes perfectly transparent. In combination
with white lead and zinc oxide it produces a most excellent wear-
ing paint, both for seashore and inland use. It has marked tooth
and spreading properties. It is a good extender for iron and steel
paints, and gives the coating a harder surface.

Litharge. — Litharge (PbO), or lead monoxide, in color is a
yellowish red pigment, made by submitting metallic lead for several
hours to oxidation under intense heat, in reverberatory furnaces.
It has a specific gravity of 10 and grinds in about 9 per cent, of
oil. It comes into large usage as a drier in the manufacture
of boiled oil or japans. It is generally very inhibitive and gives
a good, hard drying, elastic film.

Red Lead. — Red lead is a very heavy, brilliant red pigment,

THE PROPERTIES OF PIGMENTS 263

made by submitting litharge to further oxidation in reverberatory
furnaces. It has a specific gravity of 8.7. It varies in color
and strength of tone, according to the degree of oxidation and
physical structure. It is considered as one of the most valuable
pigments known for the protection of steel and iron, and has been
used for this purpose for many years.

Red lead is also produced by heating litharge with sodium
nitrate, in large iron pots, the interaction taking place forming
red lead and sodium nitrite, the latter being a very valuable salt
in the manufacture of para reds. Red lead, when used pure, is
almost always mixed at the time of application, the average for-
mula being 30 lbs. of red lead to the gallon of oil. No drier is
required, because of the rapid drying nature of the red lead itself.
Because of the stiffness of red lead in oil and the difficulty in
spreading a properly proportioned mixture, thinning with turpen-
tine and volatile thinners is sometimes resorted to by workmen,
causing a loss in the protecting value of the mixture. This pig-
ment is more or less affected by sulphurous gases, under the action
of which it turns brown. Admixture with certain inert pigments
is to be recommended, in some cases. The inhibitive nature of
red lead will vary according to the method used in its manufac-
ture and the quantity and kind of impurities which it carries. It
should always be tested, if possible, before its selection as a pro-
tective agent.

Orange Mineral. — Orange mineral is produced by the oxida-
tion in reverberatory furnaces of white lead which is slightly off
color. It is of the same chemical composition as red lead, but
it possesses a different tone. Because of the tendency of oxides
of lead in linseed oil to absorb oxygen, and cause stiffening and
hardening of the mixture, these pigments are seldom used alone
in prepared paint. Admixture with other pigments prevents this
hardening for a long period. This is another pigment the inhibi-
tive power of which is found to vary, owing to indefinite impuri-
ties introduced in its manufacture.

Artificial Iron Oxides. — There are a great number of iron
oxide pigments used for the protection of iron and steel, and they
vary in their specific gravity as well as composition. There can
be obtained iron oxides of over 99 per cent, purity, and these
are generally made by the burning of copperas (ferrous sulphate),
but the resulting materials are apt to contain traces of sulphuric

264 CORROSION AND PRESERVATION OF IRON AND STEEL

acid which have not been thoroughly burned out. Oxides made
by the above process should be carefully tested previous to use
on iron or steel.

Venetian Reds. — The so-called Venetian reds are made by
the calcination of mixtures of copperas and lime, the interaction
taking place forming ferric oxide and calcium sulphate. The
percentage of iron in these Venetian reds varies from 15 to 45
per cent. Because of the soluble nature of the calcium sulphate
in this pigment, when not dead burnt, and the ease with which it
is ionized in the presence of water, it is considered as a dangerous
material to use upon iron or steel. Some Venetian reds are made
by mixing iron oxides with gypsum and calcium carbonate. These
are considered safer pigments to use, being free from acid, espe-
cially when the latter pigment is present.

Metallic Brown. — There are many other oxide of iron ores
which are used in the making of paints, and one very important
ore is called Prince mineral, or Prince metallic brown. This ore is
mined largely in Pennsylvania, and is found as a natural hydrated
iron oxide and also as a carbonate (siderite). It is roasted for
several hours at a cherry red heat, the hydrated oxide or carbonate
being changed to the sesquioxide. It is then ground and made
ready for shipment. This pigment contains a considerable quan-
tity of silica and alumina. It has been used both on wood and
steel with considerable success for many years, and is considered
as one of the standard pigments for protective paints.

Indian Red. — Natural hematite ores and Persian Gulf ores
are pigments of great value, and when of the proper shade are
termed Indian red. The term Indian red is also applied some-
times to Artificial Iron Oxides made by calcining copperas. These
pigments generally run over 95 per cent, oxide of iron, with varying
percentages of silica. The shade and tone of Indian red, although
varying in many samples, is distinctively more pleasing than many
of the brown oxides. The natural iron oxides, made from hematite
ores, some of which are termed bVight red oxide and Indian red, have
a gravity of 3.5 to 5.2 and grind in about 20 to 25 per cent, of oil,
while the Venetian red and Prince's metallic brown have a gravity
of 3.1 and grind in approximately 25 per cent of oil. The natural
iron oxides make valuable body pigments for inhibitive paints.

Ochre, Sienna, and Umber. — There are several other iron
oxide pigments, such as ochre, sienna, and umber, the ochre being

THE PROPERTIES OF PIGMENTS 265

a hydrated ferric oxide admixed with clay. The umber is similar
in composition to the sienna (consisting of iron and aluminum
silicate with varying percentages of manganic oxide which gives
it the brown color), but containing a much higher percentage of
manganic oxide. These pigments vary in gravity from 3 to 3.5
and they grind in from 35 to 50 per cent, of oil, according to their
chemical composition. These pigments are seldom used alone
in paints for iron and steel.

Graphitic Pigments. — There are two forms of graphite: the
natural, and the artificial. The natural has a gravity of about
2.6, and the artificial of about 2.2. Both varieties grind in approxi-
mately the same amount of oil, namely 45 per cent. The natural
product is an allotropic form of carbon found in many localities,
and contains varying percentages of carbon admixed with silica,
and sometimes oxide of iron. The artificial or Acheson graphite
contains about 90 per cent, carbon. Burning of this latter variety
gives an ash consisting of carbide of silicon, with a very small
percentage of silica, iron, and alumina.

Both pigments have been used to a great extent for painting
steel and iron, but inasmuch as they have a very excessive spread-
ing rate, a very thin film is produced, which sometimes suffers
early decay. Graphite paints are therefore generally mixed with
heavier pigments, such as red lead, or sublimed blue lead, pro-
ducing superior paints. Graphites are not considered good
inhibitors by the authors on account of the ease with which they
conduct electric currents and thus excite corrosion. The pig-
ment itself is very inert, and has practically no action upon the
oil in which it is ground, except to retard drying. It is a very
unctuous or greasy pigment and unless reinforced with pigments
that have more tooth, it slides under the brush and the particles
tend to segregate.

Bone Black. — Bone black is a pigment made from ground
bones burned in highly heated iron retorts for several hours. This
pigment sometimes contains traces of organic matter and oil
which retard its drying. In composition, it runs about 85 per
cent, calcium phosphate and 15 per cent, of carbon. Its gravity
is 2.68, and it grinds in 50 per cent, of oil to a -stiff paste. The
authors consider it generally an excellent inhibitor, and it has
been used in certain inhibitive paints where dark colors are de-
manded, to replace carbon black and lampblack.

266 CORROSION AND PRESERVATION OF IRON AND STEEL

Lampblack. — Lampblack is a very pure form of carbon,
often being over 99 per cent. pure. This pigment is made from
the combustion of oils, and is very uniform in its composition.
It is extremely permanent and has wonderful tinting strength.
It has a gravity of 1.82 and grinds in 75 per cent, of oil. It is
an extremely slow drying pigment. It possesses great tinctorial
values and is used in large quantities for the tinting of paints.
In the authors' opinion, it is not safe to use as a contact coat on
iron surfaces, as, like the graphites, it is a good conductor of
electricity and therefore acts the part of a stimulator. As a
top coating it possesses distinct merit, because of its resistivity
to water and its action in preventing early decay of the oil with
which it is used.

Carbon Black. — Carbon black is made from the combustion
of natural gas, and contains approximately 99 per cent, pure
carbon. It has a specific gravity of 1.85 and grinds in 84 per
cent, of oil. It has been largely used in admixture with white
lead for paints to be used upon steel and iron, but recent investi-
gations have proven that it is a stimulative compound and is
dangerous to use as a prime coating material. It may be used
with perfect safety as an excluder, when an inhibitive pigment
is used for the under coating.

Vine Black and Willow Charcoal. — These pigments are made
from the charring of certain grades of wood, and may contain
slight traces of alkali which probably account for the excellent
inhibitive values which they seem to show. They have a specific
gravity of about 1.4 and grind in about 33 per cent, of oil.

Mineral Black Pigments. — Mineral black is a pigment made
by grinding an especially black form of slate. Although this
pigment does not possess very much tinting value, it has proven
of considerable merit as an inert pigment for addition to certain
paints. The so-called Keystone filler is made from a bituminous
schist ore and forms a dark-colored pigment containing over 50
per cent, of silica, the balance being alumina and carbon, with
small quantities of calcium carbonate and iron compounds. This
pigment is used extensively as a filler for steel and iron surfaces,
especially for machinery.

Orange Chrome Yellow. — Orange chrome yellow is a pigment
used largely for tinting purposes and is made from the nitrate
or acetate of lead, chromate of soda, and alkali. It is really a

THE PROPERTIES OF PIGMENTS 267

mixture of the neutral chromate of lead and the basic chromate
of lead. It is liable to contain small quantities of chromate of
soda, basic nitrate or basic acetate of lead, and sulphate of soda.
It has a specific gravity of 6.9 and grinds in 20 per cent, of oil.
It has good inhibitive values when the impurities are not acid in
nature.

Medium Chrome Yellow. — Medium chrome yellow is a pure
neutral chromate of lead made from either the nitrate or acetate
of lead, and chromate of soda. It has a gravity of 5.8 and grinds
in 30 per cent, of oil. It is used in large quantity as a tinting
material, its strength depending largely upon the method of
manufacture. Like all precipitated pigments, the yellow is
likely to carry down small quantities of impurity from the mother
liquor. While theoretically an inhibitor, this pigment has not
given a very good account of itself on test.

American Vermilion. — American Vermilion is made by boil-
ing white lead and chromate of soda, and adding small quantities
of sulphuric acid in order to brighten the shade. A basic chromate
of lead is thus formed. This pigment sometimes contains free
chromates which account for its excellent inhibitive value. It
contains 98 per cent, of lead compounds, and has a specific gravity
of 6.8. It grinds in 16 per cent, of oil. It could probably be
improved as an inhibitor, if the acid treatment for brightening
was omitted.

Lemon Chrome Yellow. — Lemon chrome yellow is manu-
factured from the acetate or nitrate of lead, bichromate of soda,
and sulphuric acid, or a salt of this acid, and is really a mixture
of sulphate of lead and chromate of lead. It sometimes contains
traces of acetate of soda, It may also contain up to 60 per cent,
of lead sulphate. It has a gravity of 6.2 and grinds in 20 per
cent, of oil. It is not a good inhibitive, for the reasons given in
the description of medium chrome yellow.

Barium Chromate. — Barium chromate is a pale yellow pig-
ment in color, made from barium chloride and chromate of
soda. It sometimes contains traces of barium chloride. Like
many other precipitated, very slightly soluble chromates, its
inhibitive value is not high.

Zinc Chromate. — Zinc chromate is a beautiful yellow pigment
manufactured from zinc oxide, sulphuric acid, and bichromate
of potash, or zinc salts and bichromate of potash. It has a fairly

268 CORROSION AND PRESERVATION OF IRON AND STEEL

high solubility and always contains some impurities. It gen-
erally contains free chromates and uncombined zinc oxide. This
pigment has a specific gravity of 3.5 and grinds in 25 per cent,
of oil. Like the rest of the chromate pigments, it is a very slow
drying material, often requiring over a week to set up, unless
considerable drier is added. In spite of the impurities which it
carries, it has shown itself to be the most inhibitive pigment
known and has demonstrated its value in even small percentage
in paints for iron and steel. It dries to a hard adherent film that
tends to protect metal from corrosion. The authors recommend
at least a 2 per cent, addition of zinc chromate to all inhibitive
contact coats for iron and steel surfaces.

Zinc-and-B avium Chromate. — Zinc-and-barium chromate is
made by precipitation of a solution of zinc and barium chlorides,
with chromate of soda. It is a pigment less soluble than zinc
chromate, and has proven very efficient as an inhibitive compound.

Chrome Green (Blue Tone) . — Chrome green (blue tone) is
made from nitrate of lead, bichromate of soda, and oil of vitriol,
precipitated on white lead and Chinese blue. The resultant
product contains chromate of lead, sulphate of lead, Chinese
blue, and white lead. It sometimes contains traces of nitrate
of lead and nitrate of soda. It has a gravity of 4.4 and grinds in
25 per cent, of oil. On account of the contained impurities and
the comparatively low solubility of this chrome pigment, it does
not appear high in the inhibitive class.

Chrome Green (Yellow Tone). — Chrome green (yellow tone)
is made from nitrate of lead, Chinese blue, and bichromate of
soda. It is liable to contain small traces of lead salts. It has
a gravity of 4 and grinds in 22 per cent, of oil.

Chrome Green (Oxide). — Pure oxide of chromium is pro-
duced in the wet process by precipitation of reduced chromium
salts with alkalies, or, by Guignet's process, by fusion of bichro-
mate of potash with borax, subsequently washing the green
powder that is formed, to remove the potassium borate or other
soluble salts. It is a very valuable, permanent green, possess-
ing unequalled permanence. It is therefore used for signal and
semaphore work on railroads. It possesses good body and cover-
ing capacity, and it is inhibitive rather than stimulative. It is,
however, rather high in cost.

Prussian Blue. — Prussian blue is a ferri-ferrocyanide of iron

THE PROPERTIES OF PIGMENTS 269

made from prussiate of potash and copperas, by interaction
of their solutions, the precipitate formed being subsequently
oxidized. This pigment has a gravity of 1.9 and grinds in about
55 per cent, of oil. Like all precipitated pigments, the Prussian
blues should be tested for inhibitive value, before their selection.
Prussian blue has a marked action in preserving the oil with which
it is used, from early decay. It presents a very glossy surface
even after long exposure.

Ultramarine Blue. — Ultramarine blue is a very bright blue
pigment made from silica, china clay, soda ash, and sulphur,
chemically combined by burning in pots and furnaces. The
resulting product is ground and bolted. It has a specific gravity
of 2.4 and grinds in 30 per cent, of oil. It is not considered a
good inhibitive on account of its sulphur content.

Naples yellow, cadmium yellow, verdigris, cobalt, and Bruns-
wick greens and blues, Vandyke brown, and the various alizarin
lakes and coal tar colors are seldom used in paints for iron and
steel. No description therefore is given of these pigments.

CHAPTER XII

THE PROPERTIES OF PAINT VEHICLES

Linseed Oil. — Linseed oil is the Only vegetable oil possessing
the necessary drying power and other properties that is produced
to-day in sufficiently large commercial quantities to meet the
demands of the paint and varnish manufacturer. Attempts have
been made to introduce other vegetable oils such as China wood
oil, soya bean oil, etc., but their use has been limited by their
scarcity or on account of peculiar characteristics which render
them unsuited for general purposes.

Linseed oil is produced by crushing and grinding flax seed in
specially constructed mills, either of the plate or screw type,
the oil being forced out, leaving a residuum of oil cake or flax
meal which is used for cattle food The use of heat in conjunction
with the crushing of the oil secures a better yield, and causes the
oil to assume a golden yellow color; the color of cold pressed oil
is considerably lighter.

Another process utilizes 'the solvent power of naphtha upon
linseed oil. After the crushed seed has been well extracted, the
naphtha is distilled from the mixture, leaving the linseed oil.
The naphtha is recovered by condensation and used over again
for the same purpose.

The flax plant is grown in Russia, India, North America, and
Argentine, the latter country producing the largest quantity and
representing an average output of over 100,000,000 gallons per
year, while the amount of seed produced in the United States
gives a yearly production of over 60,000,000 gallons of oil. Minne-
sota, North and South Dakota, are the states in which most of
the seed produced in this country is grown.

The best method of bleaching linseed oil is by proper ageing
and exposure to sunlight. "Unfortunately this method cannot
always be resorted to. Several methods are in use, such as blow-
ing with steam or air, the acid treatment, the use of artificial
oxidizing agents, etc. Linseed oil that is treated with sulphuric

270

THE PROPERTIES OF PAINT VEHICLES 271

acid must be carefully tested before using in paint intended for
iron and steel. Blown linseed oil must also be guarded against,
as it is rather viscous and is very apt to cause curdling of paint.
Mechanical processes of producing pure oil, such as rapid cooling
after the oil has been subjected to heat, throws out a large per-
centage of foots and other objectionable materials, Oil pro-
duced in this manner is generally of good quality.

Linseed oil is a triglyceride of oleic, linoleic, isolinoleic, and
other organic acids. It is a saponifiable oil, and when treated
with alkali the acid part of the oil forms salts or soaps, liberating
free glycerine.

When spread upon glass in a thin film, raw linseed oil dries
in about four days. This drying takes place by progress ve oxi-
dation, the oil increasing in weight by the addition of oxygen,
without any appreciable decomposition or elimination of its own
constituents. The initial oxidation is slow, but the formation of
peroxides causes the oxidation to proceed very rapidly. The use
of certain non-drying pigments with linseed oil, such as carbon
black or lampblack, retards this drying for several days, while the
use of other pigments, such as zinc oxide, red lead, litharge, and
white lead, increases the drying action of the oil. It is therefore
apparent that in the manufacture of paint, the amount of drier
to use must depend on the composition of the paint.

Chemical Characteristics of Linseed Oil, — Linseed oil pos-
sesses certain characteristics which are utilized by the chemist
to determine its purity. Among the most important of these
characteristics are the saponification value, the iodine number,
the acid value, and the specific gravity. If the oil is suspected
of adulteration with mineral oils, treatment with caustic alkali
will at once yield evidence whether such is the case. Linseed
oil is almost totally saponifiable, and in the presence of alkali
forms a soap, while mineral oils are unsaponifiable and are
unacted upon by the alkali. The saponification value of lin-
seed oil is approximately 192, other saponifiable oils varying to
greater or lesser extent in this value.

Linseed oil always contains traces of free vegetable acid-
uncombined with glycerine, the amount generally averaging not
over one per cent. The presence of a high percentage of free
acid in an oil would indicate the addition of rosin or rosin oil, as
the latter oils both possess a very high percentage of free uncom-

272 CORROSION AND PRESERVATION OF IRON AND STEEL

bined acids. Excessive free acid in linseed oil is often the cause
of the gelatinization or thickening of paints containing lead and
zinc, this action being due to the formation of soaps with the pig-
ments of basic nature, in the paint.

All vegetable oils differ in the amount of iodine which they
are capable of absorbing; this iodine absorption property being
in direct relation to the ability of the oil to absorb oxygen. The
iodine absorption test is very valuable in the determination of
the purity of an oil.

The specific gravity of an oil is also a good criterion of its
purity, and often will readily indicate whether the oil has been
adulterated with lighter or heavier mineral oils. The presence
of petroleum products may be noted by their odor, and this latter
test is also used when linseed oil is suspected of containing fish
oils, the characteristic fishy odor being developed when the oil
is slightly heated.

The Maumene test, or rise in temperature, when oil is treated
with sulphuric acid, has been of value in detecting the presence
of petroleum spirits. The flash test is also in vogue for this
purpose. The refractive index and oxygen absorption test are
also used in the determination of the purity of linseed oil.

When a sample of raw linseed oil is spread upon glass, and the
drying observed, the condition of the film is sometimes indicative
of the presence of mineral or fish oils, cloudiness of film or exces-
sive tackiness indicating their presence or the presence of some
semi-drying oil such as cottonseed oil.

The following characteristics of linseed oil gives the average
results of several analyses of the pure oil:

Specific Gravity at 15.5° C 932 to .935

Acid Number 5 to 7

Saponification Value 187 to 192

Unsaponifiable Matter 8 to 1.5%

*Iodine Number ISO to 190

*(May sometimes be as low as 160.)

During the drying of linseed oil, certain changes in its molecular
structure, due to the absorption of oxygen, takes place, and the film
may become more or less porous, allowing the admission of moist-
ure which would cause the corrosion of underlying iron. Toch 1

: M. Toch. The Chemistry and Technology of Mixed Paints, p. 88.
D. Van Nostrand Co., New York, 1907.

THE PROPERTIES OF PAINT VEHICLES 273

has stated that moisture goes through a linseed oil film by form-
ing a semi-solid solution with the oil, as microscopic examination
of films fails to reveal the existence of pores. While Toch's state-
ment is undoubtedly true, it does not argue that a form of porosity
does not exist. Semi-solid solution demands intra-molecular
spaces that can be occupied by water molecules. Whether we
call these spaces pores or something else, they must exist, even
though the microscope is incapable of resolving them.

The use of linseed oil as a shop coating is not good practice
and has probably been the cause of much damage. Repainting
over such a coating, with good results, is almost an impossibility.
When a linseed oil film on iron is abraded at any point on the
surface, corrosion will proceed rapidly. As Walker 1 has shown,
the hydrogen which is evolved during the initial start of corro-
sion would, under ordinary conditions, form an electrolytic double
layer, to some extent preventing further corrosion. The linseed
oil, being an unsaturated hydrocarbon, acts the part of a depolari-
zer and therefore the action is accelerated. Inhibitive pigments
in the oil, however, retard this action and overcome the stimu-
lating effect of the oil film.

The presence of mucilaginous and albuminous matter and
certain solid fats causes the "foots" or cloudy sedimentation
in linseed oil. A high percentage of foots should be avoided,
because of the high acid number and destructive effect on the
dried film of linseed oil.

Boiled Oil. — Boiled linseed oil is made by adding lead or
manganese salts or oxides to raw linseed oil, heated to 350° F.
Solution takes place, with the production of an oil which has
very rapid drying power. When spread in a thin film on glass,
eight hours is sufficient to set the oil up very firmly without tacki-
ness. This oil is extremely valuable in the manufacture of paints
for iron and steel.

Other methods of making boiled oil produce inferior products;
the so-called "bung hole" oil, in which a liquid drier, or thick
mixtures of boiled oil containing salts and oxides, is added to cold
raw oil in the barrel, is quite common. This method is to be
deplored, as it produces an inferior grade of oil.

Chinese Wood Oil. — Chinese wood oil is obtained from the

1 Wm. H. Walker. The Function of Oxygen in the Corrosion of Metals.
Trans. Amer. Electrochem. Soc, Vol. XIV, 1908.

274 CORROSION AND PRESERVATION OF IRON AND STEEL

nuts and seeds of the Chinese tung tree, and imported in large
quantities for use in the manufacture of special varnishes and
paints. The raw oil is unsatisfactory for use, drying white and
unelastic, but when heated at certain temperatures, less than
250° F., and treated with salts or oxides, then thinned with oil
and benzine, an excellent product is obtained. The treated oil
dries rapidly, with a high gloss and heavy body, producing a hard,
elastic film which seems to be relatively moisture proof. On
account of these properties, it is largely used in the manufacture
of floor paints and enamels,- but has not as yet received any
special application in the manufacture of paints for iron or steel.
Films, however, made with this oil seem to possess excellent
excluding and moisture resisting properties.

Toch l says: "By the use of China wood oil, paints are made
which dry in damp atmospheres. The advantage which the
Chinese and Japanese have had over the Europeans on this
subject has been recognized for a long time. It is now known to
have been due to their knowledge of the proper manipulation
of China wood oil. For the making of marine paints and water-
proof paints China wood oil is indispensable." It is probable
that this material will be more extensively used in the future for
making up protective paints for iron.

Soya Bean Oil. — This oil is produced from the soya bean
which grows in Manchuria, 2 the product being used largely in the
manufacture of soaps. It possesses such close resemblance to
linseed oil that it is difficult for the analyst to determine its pres-
ence. Experiments are being made to utilize it in place of lin-
seed oil, but only a few hundred gallons have been received in
this country up to the present time. Soya has a slightly darker
color than linseed oil. The following chemical characteristics
will show how closely it resembles the latter:

Specific Gravity 924

Saponification Value 192

Iodine Value 122-130

It dries rather slowly, with the production of an elastic film.

1 M. Toch. Chemistry and Technology of Mixed Paints. D. Van
Nostrand Co., New York, 1907.

2 Attention is called to the fact that the Soya bean grows well in many
parts of the United States although it has not been used as a source of paint oil.

THE PROPERTIES OF PAINT VEHICLES 275

Special treatment may render this oil of great value in the manu-
facture of protective paints.

The Use of Driers. — The necessity of having the oil dry
gradually, but still with enough speed to prevent undue tacki-
ness with the resulting dust-catching properties, demands the
use of a drier that will cause oxidation to proceed uniformly and
with the proper rate of speed. The too rapid oxidation of lin-
seed oil, with the formation of linoxyn, is the cause of the non-
elasticity and brittleness of some paints that have been recently
used. This effect is more noticeable in paints containing japan
driers than in those containing oil driers, the latter being the
safest of the two varieties to use. Japan driers are manufactured
by fusing resins with salts and oxides of zinc, lead, manganese,
and other metals, reducing the mass at the proper time with
turpentine or light mineral oils. The drying action of japan
driers is very rapid. The oil driers are made by heating linseed
oil with salts or oxides of metals, at a high temperature, subse-
quently thinning down with more oil and volatile diluent, prefer-
ably turpentine.

Oxides of lead and manganese are the most widely used chemi-
cals for the making of driers. Manganese starts the drying action
and causes rapid surface drying. The lead drier causes the
oxidation to proceed throughout the film and should generally be
in the larger proportion Red lead and litharge are also used,
but red lead causes a brittleness of paint films, while litharge gives
a paint film very elastic properties.

Turpentine. — This remarkable and highly prized paint
diluent is distilled from the sap of the pine tree. The largest
quantity is being produced in the pine forests of Georgia and
South Carolina. In its pure form it is a water white liquid,
easily recognizable by its distinctive odor. When added to paint,
it prevents thickening of the paint and causes the paint to work
very freely under the brush, increasing its spreading properties.
It is largely used for flatting purposes, and it evaporates slowly
but completely, causing more rapid oxidation of the oil in which
it is used. It evaporates on paper without a stain, and this test,
accompanied by tests for its specific gravity, polymerization
value, and flash point, makes its adulteration easily detected.

Sulphuric acid, when mixed with turpentine, results in the
formation of a brownish colored, thick liquid, the turpentine

276 CORROSION AND PRESERVATION OF IRON AND STEEL

being polymerized with the acid to other compounds. The
presence of mineral oils which are unacted upon by sulphuric
acid, is readily shown by this test, and they separate and float
upon the top of the polymerized liquid.

A pure turpentine will show the following characteristics,
within very close limits :

Specific Gravity 86 to .87

Boiling Point 156° C.

95% will distil between 150 and 165.

Polymerization test — less than 5% should be unacted
upon by sulphuric acid.

Spot test — should completely dry on filter paper with-
out stain, and without affecting the water-absorptive
properties of the paper.

Flash point — should not be under 40° C.

Wood turpentine is made by distilling pine debris and logs.
This process produces a liquid having an odor very much more
penetrating than pure gum spirits, and its effect upon workmen
is sometimes serious. Proper refining may remove pyroligneous
and formic acids, pyridine bases, and creosote, which it contains
in small quantities, producing a product that is almost identical
in painting values with the pure gum spirits. It has been found,
however, that the most highly refined wood spirits contains a
slight difference in the relative percentage of pinene, and turpene,
to that contained in pure gum spirits.

Paraffine Spirits. — These are produced to-day by the refin-
ing of petroleum and asphaltum products. In the refining of
petroleum, several light oils, such as benzine and naphtha, are
produced and are largely used by the painter as cheap diluents.
They evaporate very rapidly, but their excessive use should be
condemned. The very low flash point of these products make
them somewhat dangerous. Several higher boiling point oils are
produced in the distillation of petroleum and asphaltum com-
pounds. In their flash point and boiling point they resemble
pure turpentine, and they answer to the spot test satisfactorily.
They possess the same flowing and flatting properties as pure
turpentine, and are being used in considerable quantities as a
substitute for this material.

Benzol. — This product is obtained by the distillation of coal
tar, and is a chemical compound corresponding to the symbol C 6 H 6 .

THE PROPERTIES OF PAINT VEHICLES 277

It is a water-white liquid of excellent penetrative and solvent prop-
erties. It has been found valuable as a diluent for coal tar and
asphaltum paints, and is largely used in the manufacture of paint
and varnish removers.

Lacquers and their Application. — By the term "lacquer" is
meant the pale varnishes so largely used in the protection of
brass, copper, and silver ornamental work. Lacquers have been
used to some extent in the protection of steel and iron, and a
short outline of their manufacture will not be out of place.

There are three distinct varieties of lacquers, known as gum
lacquer, cotton lacquer, and the combination lacquer. Gum
lacquers are produced by dissolving broken gums, such as kauri,
sandarac, and shellac, in cold grain alcohol, fusel oil, acetone, and
other solvents. They may be colored with various coal tar
dyes. They are applied with a brush, and rapidly dry to a very
transparent film with a high gloss. Materials so lacquered are
often baked to insure a harder film.

The most widely used lacquer is cotton lacquer. This lacquer
is made by dissolving soluble nitrated cotton in amyl acetate.
The cotton is manufactured by immersing perfectly clean absorb-
ent cotton in a mixture of nitric and sulphuric acid. The cotton
used may be either the long, staple variety, or the short-fibered,
pulp variety; the former producing the best and most elastic
lacquer. During immersion in the acid, the cotton becomes
nitrated, taking up two molecules of nitric acid. The strength
of the acid, and temperature at which nitration takes place, are
extremely important, as several varieties may be produced, all
of which have different solvent values. The cotton after nitra-
tion is removed from the acid and thoroughly washed and neu-
tralized. It is then dried and treated with solvent. Amyl
acetate, acetone, wood, and grain alcohol are used as solvents,
but only the first-named solvent produces a transparent film.
Articles to be coated with cotton lacquer must be dipped, as
brushing causes streaking.

Combination lacquer is made by combining a gum lacquer
and a cotton lacquer, the solvents of which have been properly
adjusted to prevent settling out or curdling of either constituent.
Combination lacquers are very serviceable and of extreme value
for protecting metal ornaments.

One of the authors has used a chromated lacquer which has

278 CORROSION AND PRESERVATION OF IRON AND STEEL

given good results in protecting iron articles on a small scale.
Gum shellac is dissolved in grain alcohol and the insoluble wax
filtered out. Chromic acid is then introduced into the lacquer
by adding tiny pinches at a time until saturation has taken place.
If the chromic acid is added carelessly or in large increments,
violent oxidation will take place, and the lacquer may take fire.
For this reason it is well to have the vessel containing the solu-
tion surrounded by ice water while the lacquer is being chromated.

Perry has developed a chrome resinate which gives promise
of proving itself a valuable constituent of inhibitive vehicles.
This compound is soluble either in benzol or linseed oil, and when
used in very small percentage may be of value as an additional
fortifier against corrosion.

Varnishes. — Varnishes will not be discussed here, as they
are used almost entirely for interior decorative purposes. It has
been demonstrated, however, that a percentage of high-grade
varnish in a paint renders that paint a better excluder of mois-
ture. If the varnish be of a nature that will stand exposure, its
use is desirable, in some instances.

APPENDIX A

Presenting a Discussion before the American Institute
of Mining Engineers on the Corrosion of Water
Jackets of Copper Blast Furnaces.

The Corrosion of Water Jackets of Copper Blast Furnaces. 1 —
During the two years in which the new reduction-works of the
Copper Queen Consolidated Mining Company have been in oper-
ation at Douglas, Arizona, there has developed a remarkable
condition in regard to the corrosive action of the water used to
cool the jackets of the blast furnaces.

Were it not for the many contradictory features, it might
pass as one of the unavoidable troubles due to the composition
of the water. This water, obtained from wells 600 ft. deep, is
also used in the steam boilers, and its composition, as shown by
the following analysis, does not indicate the presence of any
ingredient which would explain the corrosion:

Grains per U. S.
Gallon.

Silica 0.861

Iron oxide and alumina 0.223

Calcium carbonate 0.261

Calcium sulphate

Magnesium carbonate

Sodium and potassium sulphates 14.850

Sodium and potassium chlorides 9.732

Sodium and potassium carbonate 6.482

32.409

The jackets are made of inner plates 0.5 in. thick, and an
outer plate f in. thick, with f in. stiffeners between the inner and
outer plates. In from ten to twelve months the inner plates have
been reduced by corrosion to a thickness of from i to T V in., while
the outer plate in the same time is reduced by less than T V in.,
and " the stiffeners show very little corrosion. The plates are

1 Transactions of the American Institute of Mining Engineers. By George
B. Lee, Douglas, Arizona. (Toronto Meeting, July, 1907.)

279

280 CORROSION AND PRESERVATION OF IRON AND STEEL

pitted and eaten away in some places more than in others. There
is practically no scale found on the jackets, but when cleaned
considerable iron oxide is found in the bottoms.

With an action so marked, serious trouble would be expected
in the boilers, but, on the contrary, a recent inspection by a
boiler insurance company gave an almost perfect report on the
large boiler plant, which consists of eight 500-h.p. Sterling boilers.
There was no pitting in the tubes. The inspector's attention
was particularly called to the pitting of the jackets. The cast-
iron impellers of rotary pumps that pump to the cooling tower
from the hot-well are pitted in spots quite as deeply as the jackets.
A sheet-iron pipe 0.25 in. thick, that has carried all the hot water
for the jackets of eight furnaces, has never leaked. Some wrought-
iron pipes, handling water at a temperature of from 65° to 80° F.,
have been almost destroyed by pitting, while others in the same
line have not shown a leak.

These notes are offered in the hope that some member of the
Institute may have met this problem before, and can throw light
on this interesting subject.

Discussion

William Kent, Syracuse, N. Y. (communication to the Secre-
tary *) : The analysis of the water shows it to be somewhat
unusual; it is rather high in sodium and potassium sulphates and
chlorides, 24.58 grains per gal., and very low in calcium car-
bonate, 0.261 grain per gal. There are three theories which may
account for the corrosion:

1. Air Bubbles Lodging on the Iron. — It is well known that
even pure water, such as the water of condensation from steam-
heating systems, is an active agent in causing the pitting of the
nipples used for connecting cast-iron radiators and the iron or
steel return water-pipes, and the presence of air in the water is
supposed to be the real cause of corrosion.

2. Electrolytic Action. — The water containing sulphates and
chlorides may act as an electrolyte, and different portions of the
steel plate, having slight variations in chemical composition, may
act as two different metals or electrodes. 2

1 Received July 17, 1907.

2 See Dr. Cushman's paper on Corrosion, read at the 1907 meeting of the
American Society for Testing Materials.

THE CORROSION OF WATER JACKETS 281

3. Chemical Action. — At certain temperatures potassium
sulphate may attack iron, forming iron and potassium sulphate.

Possibly all three of these actions may take place at the same
time.

The remedy indicated is the addition of a little milk of lime
to the water. This will neutralize any acid reaction of the potas-
sium sulphate, and form a precipitate of calcium sulphate, which
will make a protective coating on the iron and prevent all three
of the actions above described.

James Douglas, New York, N. Y. (communication to the
Secretary l ) : The following extracts from correspondence with
Mr. Lee give some additional particulars concerning the corrosion
of the jackets:

Douglas, Ariz., June 12, 1906.

We have just taken out from one of our new furnaces a jacket,
which has been in use five months. I find the same trouble as
before, the inner sheet is very badly pitted. The outer sheet and
the angles that space the inner and outer sheets do not appear to
be attacked.

In connection with this, I wish to call attention to the fact
^fchat our boilers, which have been in use now for two years, have
just been inspected by the insurance company and have received
an almost perfect clearance. Apparently the corrosion is due
to some peculiar condition that exists with the fire on one side and
water on the other; and also that there is a difference between
this condition and that obtaining in the boilers. The steel plate
on the outside, which is much thinner to start with, and which is
air-cooled on one side, with hot water on the other, far outlasts
the thicker inner sheet, and the T-irons, or angles, which are
immersed in the water between the two sheets, seem to be very
little attacked.

In connection with this peculiar action, I would call attention
to the impellers in the rotary pumps which circulate the water
for the condensers in the power plant. These are made of cast-
iron; and we find that at certain points they are very badly pitted,
being eaten away to a depth of a quarter of an inch for a space
of one or two inches in area; and right next to this there will be
spaces that are apparently not attacked at all. We have had

1 Received September 28, 1907.

282 CORROSION AND PRESERVATION OF IRON AND STEEL

whole lengths of pipe, leading from the supply tanks to the power-
house, which were perforated, and lengths next to them apparently
very little attacked. The surface of the jacket seems to be more
uniformly attacked, but even on this there are smooth spots that
have apparently resisted this action.

You will see from the above that materials as different as
flange steel and cast iron are both attacked by the water; that
steel exposed to very hot water, such as exists in the boilers, is
apparently not attacked; that pipes handling water not over
75° F. are attacked; that pumps handling water both cold and
moderately heated are attacked; and that steel surfaces heated
on one side and with water on the opposite side of, say, 140 to
150° F., are badly attacked, while plates cooled on one side by the
air and exposed to the same water are very little attacked.

I suggested that samples of the inner shell and a stay-bolt
be sent for inspection by the members of the Institute, and that
the temperature of the water in contact with the inner and the
outer shell be taken. To this request the following reply, dated
December 11, 1906, was made:

Complying with your request, I am sending to-day, by express,*
a stay-bolt and piece of metal from a water jacket. The tem-
peratures which you suggested taking of the jacket water near
the inner sheet and near the outer sheet have been taken repeatedly,
and I enclose a statement showing the range of temperature.
You will note that, as the temperature increases, the variation
also increases. We have just now a report from the inspector
of the Hartford Boiler Insurance Company, in which we are given
an absolutely clear bill on the entire battery of boilers, six of
which have been in use 3-5- years.'

I believe I told you on your last trip here that we are now
making experiments by feeding oil into the water as it goes to
the jackets, with the hope of coating the inside of the jacket with
a film of oil, and possibly preventing the corrosive action of the
water.

THE CORROSION OF WATER JACKETS

283

Temperature of Water in Jackets Near the Inner or Fire Side and
Outer or Air Side.

Outside Degrees F,

Inside Degrees F.

Outside Degrees F.

Inside Degrees F.

102

106
107

108
112

104
106
108

116

115
115
118
130

In answer to a request as to the effect of this oil, Mr. Lee
wrote August 28, 1907, as follows:

I am in receipt of your wire of the 22d in regard to corrosion
of water jackets. Apparently the use of oil has been of some
benefit in reducing this corrosion, as the amount of jackets
renewed seems to be little more than it was, though the number
of jackets in use is considerably larger. I recently had occasion
to examine a jacket which had been taken out, and found a very
peculiar condition; namely, there was a place about 6 by 8 in.
right in the middle of the place which seemed to be quite smooth
and not pitted at all, while all around, it was very badly corroded.
This jacket had been in use about a year and a half.

I have received a very interesting letter from Mr. Beardsley,
who was formerly with the Mt. Lyell Company in Tasmania,
giving me a number of experiences that he had had of corrosion
of jackets, and, as far as he was able, the causes. In one instance
well-water was substituted for the former supply. The well-
water seemed to be highly charged with gas, and they experience
great difficulty from the jackets burning. This he attributed
to the formation of gas-bubbles on the fire-sheet. By mixing
the city water and this water the trouble was very much reduced
and a return to the city water stopped it entirely. This, of
course, is quite a different experience of ours, which is not one
of burning, but one of interior corrosion.

The following letter, dated London, September 17, 1907, is
from George M. Douglas, a member of our staff, who was an
engineer for some time on the White Star and other steamship
lines: On reading the correspondence you have received from
Mr. Lee at Douglas regarding the corrosion of the inner plates of

284 CORROSION AND PRESERVATION OF IRON AND STEEL

the furnace-jackets, I suggest that this might be caused by some
electrolytic action. This same corrosion takes place in the Scotch
type of marine boiler, particularly when the water contains a
little salt. This boiler is somewhat analogous in form with the
jacket, having a hot inner plate, a water space, and relatively
cool external shell.

This corrosion is prevented by hanging zinc plates on the stay-
rods between the spaces affected. It is also a practice to put
zinc plates near the water inlet, so that any free acid in the enter-
ing water may combine with the zinc and be neutralized.

Perhaps a similar application of zinc to the jacket shells at
Douglas might prove beneficial. I suggest applying it in the
following manner: The zinc plates should be about f in. thick,
and about 8 in. wide, and 16 in. long. Some authorities object
to the application of zinc direct to the iron, though it is customary
to do so in British engineering practice.

A suitable means would be to have brackets made of copper
strips, | in. by 1.5 in., placed about a foot above the bottom of the
jacket on the inner shell, into which brackets the plate could be
inserted from the top. 1 I also suggest that a plate be put where
the water jacket enters. Two plates of the size mentioned on
each side and one on each end of jacket (should there be any corro-
sion there) is enough. I do not know what the size of the furnaces
is. But the relative proportion of zinc surface to iron surface
should be about one to ten. A good contact should be insured
between the zinc and the copper and iron, or, if copper is not used,
between the zinc and iron.

It is not enough merely to place the plates in the water space;
they should be well fastened to the jacket.

Under the same date, Mr. Douglas, in response to an inquiry
from me, made, in substance, the following statement, which,
although not directly pertinent to the present discussion, may
be valuable as a contribution to the general question of the cor-
rosion of steel and iron:

With regard to the corrosion of stern-posts and plates in the
vicinity of propellers on ships, I should say that this action seems
to be well understood as due to the fact that the propeller-blades

1 This suggestion is not a good one, as the zinc would be destroyed in
protecting the copper instead of the iron. A. S. C.

THE CORROSION OF WATER JACKETS 285

etc., on one hand, and the stern-posts and plates, on the other
hand, are of different material, so that a galvanic action is set up,
the salt water acting as an electrolyte. If all these parts were of
exactly the same material, no corrosion would take place ; but this
is not the case in practice. The stern tube is usually of bronze,
and the propeller of bronze and steel, with blades of manganese
bronze. An interesting case is on record, of a vessel on which
iron propeller blades were replaced by blades of bronze. Imme-
diately upon this change, the corrosion of the stern post and sur-
rounding plates became so great they had to be renewed after
only one voyage to the Cape. They were afterwards protected
by zinc sheathing; and it is now the custom to protect such parts
by sheathing of zinc or some metal of similar electro-chemical
character.

Though these facts are interesting, I fear they will not help
you much in dealing with your jackets, since the conditions of
your problem are by no means the same as those of the marine
practice above stated, in which both the origin and the remedy
of the trouble seem to be clearly established.

With regard to the general question of the corrosion of steel or
iron plates, however, I may call your attention to one point
which may be worthy of consideration — namely, the electro-
chemical relations between metals and their oxides. According
to a leading author, 1 "every metal is electro-positive to its own
oxide." When steel or wrought-iron, with oxide scale upon it,
is placed in an oxidizing liquid, the conditions of active corrosion
are complete; and even without a specially oxidizing medium,
it is asserted that a galvanic action may be set up in the presence
of air and moisture between the metal and its scale. It is there-
fore regarded as very important that no "black oxide" should be
left on the plate; for, though in itself it tends to protect the sur-
face (the black or magnetic oxide of iron resisting ordinary oxidiz-
ing agents), yet if, in finishing, handling, or subsequently using
the plates, portions of it should be knocked off, the remaining
portions contribute to the corrosion of the exposed metal.

In 1879, Sir Nathan Burnaby declared, as the result of his

1 Metallic Structures: Corrosion and Fouling, and their Prevention, by
John Newman, p. 36. London: Spon & Chamberlain (1896). See also
Rustless Coatings, Corrosion and Electrolysis of Iron and Steel, by M. P.
Wood. London: Chapman & Hall (1904).

286 CORROSION AND PRESERVATION OF IRON AND STEEL

observation, that when mill-scale was left upon plates and angles
used in ships, its effect upon neighboring surfaces of bare metal
was as strong and continuous as that of copper.

Jn 1882, Mr. Farquharson conducted for the British Admiralty,
at different naval stations, exhaustive experiments as to the
action of mill-scale on ships' metal exposed to the conditions of
marine use, and found: (1) that no "pitting" occurred in mild
steel freed from all scale; (2) that the loss of weight by corrosion
was practically the same for clean mild steel and clean iron; and
(3) that the action of mill-scale in inducing corrosion is consider-
able and continuous — equal in these respects to that of an equal
amount of copper.

The Admiralty practice is to pickle all ships' metal, for the
removal of mill-scale. The scale may also be removed by sand-
blast, or by means of a gasoline-torch, followed with a scraper
and a wire brush. Pickling, however (in dilute sulphuric or hydro-
chloric acid), is probably more thoroughly effective.

The rivets should be of the same material as the plates. Iron
rivets in steel plates might cause trouble.

Hiram W. Hixon, Victoria Mines, Ontario, Can. (communica-
tion to the Secretary *) : I have had difficulties here similar to
those encountered at Douglas, and I found the cause to be the
carbonic acid given off when the water was warmed. All the
water in the streams in this country contains organic matter com-
ing from peat-bogs and muskegs. It is brown in color, and when
it strikes the fire-sheets of the jackets the carbonic acid is given
off and travels up along these fire-sheets because of the bosh in
the furnace. The lower side of the tuyeres was much pitted, and
they leaked badly until I had copper tubes put in in place of
iron ones.

The inner or fire-sheets were destroyed most rapidly opposite
the cold-water inlet, whare the greatest amount of carbonic
acid was given off. Our boilers are not affected and are perfectly
clear of scale. I think the acid is liberated in the feed-water
heater, in which there are copper tubes, and after it is in a gaseous
condition it does not attack iron, or at least the water is necessary
to make it destructive. The pipes leading from the feed-water
heater to the boilers are destroyed, but the boilers are not.

1 Received through Dr. Douglas, Sept. 28, 1907.

THE CORROSION OF WATER JACKETS 287

The Canadian Copper Company had a purifying plant for
the feed-water, and the pipe leading "from the purifier to the dif-
ferent boilers went over the boilers, and each lead to the boiler
came out of the bottom of the main pipe. Tests made of the
water to the different boilers showed that the water to the boilers
nearest the purifier was much less acid than the water to the boiler
at the end of the feed-pipe. The superintendent spoke to me
about it, and I suggested that the acidity of the water was due
to carbonic acid dissolved in the water, and that being a gas it
had a tendency to enrich the water in the top of the feed-pipe,
and, consequently, the water drawn off for the first boiler from
the bottom of the main contained less acid than the water which
went to the last boiler.

I think the trouble at Douglas is due to the water supply
coming from the deep wells containing carbonic acid, and this
acid is probably due to the source of the water being something
in the nature of a mineral spring, such as Saratoga, Manatau, or
Apollinaris. Ordinary chemical tests would fail to detect any
mineral acid, and the gas being small in quantity would escape
detection.

The remedy for the trouble is to use copper fire-sheets, or to
run the water through cooling towers and use it after the carbonic
acid has escaped.

C. D. Van Arsdale, New York, N. Y. (communication to the
Secretary *) : There are several explanations which present
themselves regarding the corrosion of the water jackets of the
Douglas furnaces. The most obvious of these — namely, that
the composition of the water is itself directly responsible — may
be dismissed as improbable. Analysis of the water shows that
it may be called a good boiler-water for this region, since it con-
tains very small amounts of incrusting solids and the non-incrust-
ing solids are not excessive; and this opinion is verified by its
causing practically no boiler troubles. Since no corrosion takes
place in the boilers, it is evident that the dissolved constituents
of the water do not alone afford sufficient explanation.

Granting that the water is itself non-corrosive, there is nothing
in the working of the ordinary water jacket to account for the
difficulty, otherwise such corrosion would be more or less gener-

1 Received October 31, 1907.

288 CORROSION AND PRESERVATION OF IRON AND STEEL

ally observed in other plants. It would, therefore, seem that the
only explanation left is electrolytic action; but it is not evident
what is the cause for electrolysis.

It is well known that lack of uniformity in the composition
of iron will cause corrosion on account of action due to minute
local galvanic couples. If this is the cause, then a suitable remedy
would be to hang zinc sheets inside the jackets, as has already
been suggested. Another cause of electrolytic corrosion may
be stray currents from some source. In the same way much
trouble has been experienced from corrosion of gas- and water-
mains in cities, due to stray electric currents passing along them.
A very small current has been found sufficient to cause a great
amount of trouble, but if this should be found to be what is caus-
ing the electrolytic action in the jackets, it should be quite simple
to put a stop to it.

The different temperatures observed in different parts of the
jackets might also be sufficient to cause some corrosion, since
electrical currents can be produced in an electrolyte by electrodes
of the same metal, portions of which are at different temperatures.
This could be obviated by a circulation of water in the jackets
sufficiently rapid to do away with any differences of temperature.

The fact that the jackets are much more corroded on the fire
side seems to indicate that the electrolytic action is due not to
lack of uniformity of the iron, but to one or both of the other
causes mentioned.

The Corrosion of Water Jackets of Copper Blast Furnaces. 1 —
C. D. Demond, Anaconda, Mont, (communication to the Secre-
tary): In order to throw some light on this interesting subject,
a series of experiments were made with strips of mild steel, con-
taining about 0.14 per cent, of C and 0.22 per cent, of Mn. These
strips were thoroughly cleaned and brightened before use, salts
were added to distilled water until it corresponded as nearly as
possible to the analysis given by Mr. Lee, and in this water at
different temperatures were placed four strips of steel; while four
other strips, at corresponding temperatures, were placed in the water
used at the Washoe Reduction Works. The latter plant has had
no trouble from corrosion of furnace jackets, pipelines or boilers.

1 (Transactions of the American Institute of Mining Engineers, xxxix,
806-817.) A discussion of the paper of Geo. B. Lee, Trans., xxxviii, 877 to
884 (1908).

THE CORROSION OF WATER JACKETS

289

The analysis of the water used at the Washoe works is:

Grains per 0. S.

Gallon

CaC0 3 2.92

CaS0 4 0.76

MgCU 0.64

MgC0 3 0.58

A1 2 3 0.12

Fe 2 3 0.12

SiOa 0.58

NaCl 0.35

Organic matter, etc. . 1.16

7.23

This water is neutral to litmus; while the artificial Copper
Queen water is very slightly alkaline, which Mr. Lee informs us
is also true of the water at Douglas.

Both waters were well aerated by pouring from beaker to
beaker, and by blowing in air from the experimenter's lungs.
The results of these tests, given in Table I, show that there is
no significant difference between the effects of the two waters.

Table I. — Oxidation of Ikon Per Square Inch of Surface* When
Immersed in Certain Waters.

Artificial Copper Queen Water

Average Temperatures

First 50 Hours

Second 50 Hours

Total

70° F.

Mg.
1.943

Mg.
1.360

Mg.

3.303

108

5.322

3.963

9.285

150

5.206

4.701

9.907

185

2.865

1.642

4.507

Washoe Water

70° F.

2.797

1.516

4.313

106

5.178

3.746

8.924

152

6.618 .

6.422

13.040

188

1.784

0.699

2.483

* These results were obtained by carefully rubbing all the rust from
the iron, dissolving it, and determining its quantity by titration.

At two temperatures the strips of steel in the Copper Queen

290 CORROSION AND PRESERVATION OF IRON AND STEEL

water showed less rust than those in the Washoe water, while at
the other two temperatures they showed more. Another experi-
ment indicated that distilled water would have given about the
same results. Hence, the corrosion of the Copper Queen jackets
must be due to some other cause than the quality of water. I
am informed, however, that the water is raised from artesian
wells by means of compressed air, and, after being used in the
furnace jackets, is passed over a cooling tower, and later goes
to the jackets again. Hence it probably contains an excessive
amount of dissolved oxygen. It will be noticed that rusting
increased with the temperature, up to 150° F., as is true of
chemical actions in general; but at 185° there was decidedly
less rust than at 150°, presumably due to the water at the higher
temperature being unable to hold enough oxygen in solution to
keep up the rate of oxidation.

The small differences of temperature (9° to 14° F.) do not
seem sufficient to account for the inner plates rusting more than
six times as fast as the outer; but is it not probable that the
temperature of the very surface of the inner plates is considerably
higher than is shown by inserting a thermometer in the water,
and that an unusual amount of dissolved oxygen at this tempera-
ture is the cause of the trouble? The jackets would probably
suffer much less if the water were discharged as near the boiling
point as practicable, and returned directly to the supply tanks;
the return pipe should dip well beneath the surface of the water
in this tank, since a plunging stream would entrain a fresh quan-
tity of air. This would lessen the amount of dissolved oxygen,
and rust cannot form without oxygen. Referring to the first two
paragraphs of Mr. Hixon's discussion, 1 it is the gases in solution
that are active, not those given off; and the probable reason for
his jackets corroding most rapidly opposite the cold-water inlet
is that the water at that point still holds the gases in solution.
The absence of rust in the Copper Queen boilers, and in those
at the Victoria mines, is perhaps due to the elimination of absorbed
gases immediately on entering the boilers and before the water
can touch the metal. It is certain that the temperature of the
water increases very rapidly upon entering the boilers.

We have experimented with several other remedies. The
well-known use of zinc is completely effective when the zinc is

1 Trans., xxxviii., 882 (1908).

THE CORROSION OF WATER JACKETS 291

properly distributed; but this metal corrodes so rapidly that it
has to be frequently renewed. Copper, on the other hand,
increases the rusting of the steel, these results being due to the
electro-chemical relations of the metals.

Knowing that solutions of alkalies and alkaline salts suffi-
ciently strong will prevent rusting, we tried the effect of lime, at
approximately the temperature of the water in the Copper Queen
jackets. An increase in the quantities of lime added up to 0.1
per cent, gave increased benefit, amounting to practically com-
plete prevention with the Washoe water, and reducing the rust
in the artificial Copper Queen water about 60 per cent. The
difference was evidently due to the reaction of the lime with the
sulphates and carbonates in the latter water. Larger amounts
of lime would probably complete the cure, but the precipitate
might have to be settled out.

Dr. A. S. Cushman 1 found, in certain experiments, that
potash bichromate completely prevented rust when dissolved in
water at the rate of 1 lb. or more in 1500 gallons. Our tests
indicates that, even in using the water over and over in the jackets,
the bichromate would be slowly exhausted, requiring fresh addi-
tions at intervals. We have not yet determined what the cost
of treatment would be ; but the protective effect is very striking.

The elimination of dissolved oxygen from the water is sug-
gested as the practical remedy in this particular case; but we may
take occasion here for some further remarks on the general sub-
ject of rusting.

The fact that some wrought-iron pipes at Douglas are prac-
tically unaffected, while others in the same line, as well as the
furnace jackets, are badly corroded, strongly suggests that there
is some trouble with the metal itself, and that it is not a question
of wrought iron versus steel. Indeed, the most recent and
reliable industrial and scientific investigations show that the long-
standing controversy as to the relative rust-resisting qualities
of wrought iron and steel is largely a beating of the air. The
real question seems to be one of care in the processes of produc-
tion. Cushman 2 very ingeniously demonstrates that an appar-

1 Bulletin No. 30, Office of Public Roads, U. S. Department of Agriculture,
and Proceedings of the American Society for Testing Materials, Vol. VII,
p. 211 (1907).

2 hoc. cit.

292 CORROSION AND PRESERVATION OF IRON AND STEEL

ently homogeneous piece of iron or steel carries a multiplicity of
positive and negative poles of an electrolytic system; and the
electrolysis between them increases the speed of rusting at the
positive pole, while preventing rust at the negative. He sug-
gests that this polarity is due to uneven distribution of certain
chemical constituents of the metal. In answer to the fact that
frequent investigations have failed to show this uneven distribu-
tion, he says 1 that "such extremely small differences in the
chemical composition as might easily escape detection in ordi-
nary chemical analysis are still sufficiently large to account for
slight differences of electrical potential.' 7 He might have added
that there may be considerable difference in composition which
chemical methods cannot detect, because it is impossible to
sample the segregations separately. Frank N. Speller, 2 on the
other hand, thinks that variations of density may sometimes
account for the polarity. However, there is no doubt as to the
existence of the polarity; and its effect was well shown in one of
our test strips, 4 in. long, after an exposure of 50 hours in water.
A large part of the surface shows perfectly bright, but other parts
are badly rusted. Mr. Lee says "the plates are pitted and eaten
away in some places more than in others"; also, that in one case
"there was a place about 6 by 8 in. right in the middle of the plate
which seemed to be quite smooth and not pitted at all, while
all around it was very badly corroded. " The greater the differ-
ence of potential the greater will be the action at particular
spots, which will increase the pitting and quickly produce holes
in one piece of metal, while another piece, of the same size and
thickness, long remains serviceable, even if yielding the same
amount of rust in a given time, because the action is uniformly
distributed. This action of iron is similar to that of zinc. The
latter, when pure, dissolves very slowly in acid, but w r hen impure
it dissolves readily, because of the electrolysis between the spots
of pure zinc and the spots of impurities.

The fuller details of this explanation accord with the present-
day theory of physical chemistry. Cushman's results confirm
earlier work by Whitney, and are independently verified by
Walker.

1 Farmers' Bulletin No. 239, U. S. Department of Agriculture, p. 20.

2 Applied Science, Proceedings of the Toronto Engineering Society, Jan-
uary, 1908, p. 125.

THE CORROSION OF WATER JACKETS 293

There may be many hundred independent local circuits on a
few square feet of surface. Cushman * says that, in almost all
modern steel woven-wire fences some wires will be found to far
outlast others. It is just this point of unevenness of lasting
quality in wires from successive heats in the same mill, which
have practically the same chemical composition, that is hard to
explain by any theory but that of galvanic or electrolytic action.

These ideas suggest the use of better steel for furnace jackets,
though the fact that the rapid corrosion has occurred with jackets
purchased from at least ten different manufacturers 2 makes it
appear that better metal is hard to obtain. Moreover, in view
of the satisfactory service of the outer plates, it may seem, at
first thought, that the contention for better steel is not well made.
But the key to this difficulty probably lies in the different his-
tories of the inner and outer plates in the steel mill.

For years the manufacturers have fully appreciated the prac-
tical benefit of an increased amount of work put upon steel, at
the proper temperature, in improving the physical properties.
The National Tube Company has found that the resistance to
rusting is greatly increased by similar treatment, and this com-
pany has developed a special method of applying it. 3 Now the
rolling of the outer plates of the water jackets to a thickness of
f in. necessarily requires more work than rolling the inner plates
to 0.5 in. and we may therefore expect the former to be more
resistant to rusting. The angle-iron stiffeners, which Mr. Lee
says do not rust seriously, probably received a good deal of work.
The fact that stay-bolts rusted badly at the ends next to the inner
plates, while being little affected at the other ends, is explained
by the electrolytic theory as follows: It is well known that rust,
once formed, increases the rate of corrosion. Mr. Speller 4 re-
ports finding a potential as great as 50 millivolts between one
clean iron and another rod exactly similar which had a very
slight coat of rust, both being immersed in water. When both
rods were clean, the potential was much less. The rust on the
inner plates of the water jackets will set up a current between

1 Farmers' Bulletin No. 239, TJ. S. Department of Agriculture, p. 21.

2 Private letter from Mr. Lee.

3 Private letter from Frank N. Speller of the National Tube Co.

4 Applied Science, Proceedings of the Toronto Engineering Society, Jan-
uary, 1908, p. 125.

294 CORROSION AND PRESERVATION OF IRON AND STEEL

itself and a stay-bolt, which will corrode the latter. This current,
in seeking the path of least resistance, will pass from the stay-
bolt through the water and back to the rust of the inner plate
without going as far as the outer plate. This effect will be de-
cidedly less in the case of angle-iron stiffeners, because the latter
do not have the intimate contact with the inner plate that the
stay-bolts have, and therefore the electric current meets more
resistance.

Speller found the voltage between steel and mill-scale to be
greater than between the metal and ordinary rust. It may be
that, for some peculiar reason, this scale is always worse on the
0.5-in. than on the f-in. plate, though this is hardly probable.
However, it is advisable to see that all scale and rust are thor-
oughly removed before the jackets are put together. I believe
that the main point is the working of the metal in the mill. The
steel manufacturers are making important investigations with
valuable practical results, but if they do not yet supply 0.5-in.
steel of suitable quality, it may be well to try jackets made wholly
of f-in. stock.

Arthur S. Dwight, New York, N. Y. (communication to the
Secretary l ) : Several times in the course of my smelter work
have I experienced trouble from corrosion of blast-furnace water
jackets, particularly of the wrought-iron or steel jackets com-
monly used in copper smelting, and in a manner very similar to
that described by Mr. Lee, but, in my case, always traceable in
the end to acid water.

The troubles at Douglas seem to be somewhat more aggra-
vated than one usually encounters, and the causes more obscure.
I understand that, like most of the smelters in the Southwest,
the Copper Queen works has two distinct systems of water pipes,
one for the regular high-pressure service, distributing fresh water
to all parts of the plant, and for fire protection; while the other,
at low pressure, circulates the jacket water between the blast-
furnaces and a system of cooling towers. The water in the jacket
circulatory system has numerous opportunities to pick up soluble
sulphates, principally in the vicinity of the blast furnaces (where
it is almost impossible to prevent copper flue-dust from sifting
into the launders and open places in the circuit) , or by flue-dust
being blown into the waters of the ponds and cooling tower.
1 Received September 7, 1908.

THE CORROSION OF WATER JACKETS 295

Although the tests for acidity in the jacket water which Mr.
Lee has made from time to time have failed to show the presence
of acid, and although we may perhaps properly suspect some
other cause to be the principal one, I am strongly inclined to
think, nevertheless, that the troubles are aggravated by the
presence of acid in the water, particularly as it seems to be
admitted that the corrosion is somewhat more evident in the
circulatory system than it is in the service system.

The fact, however, that some pitting and corrosion also occur
in the fresh-water line points to the presence of some peculiar
property in the water itself. The irregularity with which some
sections of pipe and some jacket sheets are corroded, while others
in the vicinity, or even adjacent parts of the same piece of metal,
are entirely free from attack, would point to faulty material, or
perhaps careless heat treatment in the manufacture of the steel.
But, granting all this, we are still far from a satisfactory answer
to the riddle. The Copper Queen smelter is built of the best
materials currently obtainable, purchased from various makers,
and, as it stands, represents better than average modern smelter
construction. The analysis of the water gives no clue; in fact,
it must be said to be unusually harmless looking. The theory
of electrolysis does not seem to meet the conditions. By the
process of eliminating those theories which fail to satisfy, we
are finally confronted with the question: can the fact that this
water is raised from the deep wells by compressed air have any
bearing on the problem? Is it not possible, that, by the intimate
commingling of the water and air at high pressure, such thorough
aeration might occur as would make the water an extraordinarily
active oxidizing agent? This could easily be determined by
experiment. Personally, I should consider this a surprising fact
if it proved to be the correct explanation, but the logic of the
situation seems to point to some such cause out of the ordinary.

In this connection, it may be pertinent to give some of my
own experiences with troubles of this kind, and the remedy that
was developed.

The most serious case of corrosion of pipes and water jackets
I have ever had occurred while I was in charge of the operations
of the Greene Consolidated Copper Company, at Cananea,
Sonora, Mexico, especially during the latter part of 1905. Cananea
is not more than sixty miles distant from Douglas, and there is

296 CORROSION AND PRESERVATION OF IRON AND STEEL

some similarity in the conditions of water supply to the furnaces
and in the character of the ground-water, etc.

Our troubles were made the subject of a long series of studies
by R. L. Lloyd, who was then superintendent of the smelting
department, and to him is due the credit for finding a practical
and satisfactory remedy. It seems to me proper, therefore, that
he should be allowed to tell it in his own way, and I insert the
following extract describing the episode in detail from a letter
written by him at my request, and which I have his permission
to include in this discussion: Complying with your request to give
you the details of the manner in which we worked out the troubles
at Cananea in connection with the serious corrosion of the water
jackets and pipes in the circulatory system of the blast furnaces,
I take pleasure in giving you the facts as I can recall them, though
I am at a disadvantage in not having my notes on the subject
accessible. I agree with you in thinking that the troubles which
Mr. Lee describes as being so serious at the Douglas plant are
almost identical with those which we experienced, and I think it
very probable that the same remedy may correct his difficulties
also.

At Cananea, the main part of the trouble occurred in the long
pipe-lines which extended from the water-cooling tower to the
jackets. There was also trouble in the jackets, particularly at
the joints around the tuyeres. It was noticed that wherever
iron and brass were in contact, as, for instance, at the brass
valves, the corrosive action was greatly intensified, presumably
on account of local electrolysis. The trouble at one time assumed
such proportions that it became very difficult indeed to keep
the pipe-lines and jackets in sufficiently good repair for steady
operation, and many expedients were tried in the attempts to
discover a remedy. Slaked lime was used to neutralize the acid
in the water, but the results were only partly successful, and were
attended with the serious disadvantages arising from the accumu-
lation of the lime in the tanks, and the formation of lime scales
on the smaller pipe-lines. We then tried crude sodium carbonate,
such as we had been using to prevent scale in the boilers, which
gave better results than the lime, but was still far from satisfac-
tory. Large slabs of zinc were also connected up in various
places, especially in the steel overflow-tank, in the attempt to
counteract electrolysis.

THE CORROSION OF WATER JACKETS 297

After much study and experiment we finally determined the
primary cause of the acid in the water to be the absorption of
fumes of S0 2 by the sprayed water in the cooling tower, which
was situated on the leeward side of the smelter with respect to
the prevailing winds, and on top of a hill, nearly on a level with
the top of the furnace building. The S0 2 gas in the smelter
fumes blowing through the cooling-tower was dissolved by the
falling water and slowl} r became oxidized, and eventually formed
a dilute solution of sulphuric acid. The amount of corrosive
sulphates was further augmented by the fact that more or less
flue-dust got into the water system by sifting into the launders
around the furnaces and under the feed floor, as is likely to happen
at any smelter plant, as usually constructed. The position of
the cooling tower was unfortunate, but on account of limitations
of space it could not have been avoided, even if the trouble due
to the smelter fumes had been anticipated.

While working hardest to correct the difficulty, which threat-
ened to be most serious, I read an article published in pamphlet
form by an author whose name I have among my notes, but
unfortunately not now accessible, which mentioned the effect
of arsenic salts in deterring the solubility of iron and steel in
acid solutions. I was at once very much interested to know if
this could possibly have any bearing on this problem, and I
proceeded to "doctor" the water system with commercial arsenic
oxide. The good effect of this addition was felt very quickly,
and the corrosion was practically ended. From that time on,
1 kg. of arsenic oxide was added to the water system each week,
and a portion of the water in the circulatory system was allowed
to run to waste, being replaced by fresh water, when analyses
showed that the amount of sulphates was getting high.

In this manner we were able to avoid the corrosion of the
jackets. It was noticed that even when the water became appre-
ciably acid there was little or no trouble from corrosion of the
pipe system and jackets, though we always endeavored to keep
the acid neutralized with commercial sodium carbonate.

It is to be regretted that Mr. Lloyd's citation of the article
from which he obtained the suggestion for trying arsenic oxide
cannot be made more concise at this time, but if it proves to be
a matter of special interest it can doubtless be supplied later;

298 CORROSION AND PRESERVATION OF IRON AND STEEL

meantime, perhaps the paper may be known to some members
of the Institute, who can complete the facts.

In this connection, it would be interesting to ascertain whether
there is as much trouble experienced from corrosion of pipes and
water jackets in smelting works treating large quantities of
arsenical copper ores, as there is in plants like Douglas and
Cananea, which treat ores exceptionally free from arsenic. It
is not impossible that the arsenic fumes might automatically
furnish the needed antidote for the acid in the water. Person-
ally, I have no comparison which I can present from my own
experience.

The following incident, which recently came to my attention,
presents what seems to be an interesting confirmation of the
deterrent action of arsenic in the corrosion of iron. A car-load
of commercial sulphuric acid was purchased by a western steel
works for pickling wire. The acid refused to work properly,
though chemical tests showed it to be of proper strength. The
steel company complained to the acid-makers, who sent an expert
to investigate. He looked over the situation, promptly sent the
lot back to the factory, and substituted a new lot of acid for it.
Though inclined to be reticent about the cause of the trouble,
the expert finally admitted to the chemist of the steel works
that this particular car-load of acid had been made from pyrites
containing considerable arsenic, and that there was an appreciable
amount of arsenic in the acid, stating, furthermore, that had his
company known the purpose for which the acid was to be used,
they never would have sent the kind they did.

Corrosion of jig screens and other iron work in wet-concen-
tration mills might also be averted by applying the arsenic remedy,
though I have never heard of its being tried.

J. A. Thomson, Pullman, Wash, (communication to the Sec-
retary l ) : In reference to this pitting and corrosion in the
water jackets of blast furnaces, to my mind there is no mystery
or fancied " electrolytic" action in the question. It is simply
an effect of the air carried by water, fed to the jackets to keep
them cool and the action is as follows: As soon as the cold water
comes in contact with the warm part of the jacket, it is heated
and compelled to give up its air, which, being in contact with the
plate, settles thereon. The circulation being sluggish, it is only
1 Received April 27, 1908.

THE CORROSION OF WATER JACKETS 299

when the bubbles have grown sufficiently large that they rise,
and this rise is hindered to some extent by the bosh. During the
period of rest, the air, containing both oxygen and carbonic acid,
will attack the iron, and when small irregularities have been
thus formed, subsequent bubbles find still better lodgment and
speedily effect the formation of pit-holes.

If the water is fed to the jacket near the bottom, and if it is
saturated with air, it can be shown that every square inch of
heated surface of the jacket generates about 4.5 cu. in. of air
per hour, equal to about one bubble i in. in diameter per second.

The great difference of temperature between the water and
the fire side of the jacket plate, with the consequent straining
of the grain of the plate, quickly loosens all rust as it is formed,
so that metallic iron is always exposed to the air given up by
the water.

The smooth parts mentioned as having escaped this pitting
may be accounted for in a number of ways. At the start, such
a spot may have had some adhering slag or other substance
which reduced the heat at that point, so that the first irregulari-
ties produced would be formed away from that spot, and the
bubbles would be more continuously produced where they found
good lodgment. Or, again, the spot may be a patch (of the kind
found in all plates) which has been clean-rolled in the making.
There are also various other causes. This sort of thing is a fre-
quent occurrence in the pitting by similar causes on the furnaces
of marine boilers.

The same reasoning, but reversed, will apply to the outside
sheets. There is no extra heat on the outer side of the sheet,
hence bubbles do not form there, but only on the hot sheet next
the furnace.

As to the impellers of the rotary pumps, this case is similar
to the action on a ship's propeller, in which the air is not driven
out by the heat, but is abstracted by the partial vacuum formed,
and in which, in spite of the high velocity of the water, there
seems to be sufficient time for the mischief to be done. Even
bronze blades are sometimes pitted in the same way.

My recommendation would be in the line of putting the water
under pressure before admitting it to the jackets; that is, let the
inlet to the jacket have a non-return valve like a large feed-
check, with a stem carrying a weight, so that the pump feeding

300 CORROSION AND PRESERVATION OF IRON AND STEEL

the jacket would have to force the water against, say, 20 or 30
pounds per square inch. From under this check a small (say,
f in.) pipe should lead back to the well. The water being
under pressure before entering the jacket, a large part of the
entrained air will escape through this pipe. I do not mean the
jacket to be under pressure, but only the pipe to the inlet. I
have no doubt this arrangement would give much relief. I have
used it many times as a remedy for pitting in marine boilers.

Of course, I presume there are no unknown acids in the water,
and that is just ordinary potable water, as seems to be proved
by the fact that the boilers are said not to have been attacked
in any way.

With regard to the statement of George M. Douglas, his
experience in the White Star boats must have been decidedly
limited, or he would not talk of a little salt water causing corro-
sion. It is a common practice to fill new boilers with sea-water
at first, for the purpose of preserving them; and when I went to
sea we had no evaporation to make fresh water, and if we ran
short, we made up with sea-water. I have run a ship from Han-
kow, China, to London, forty-three days' hard driving, without
opening a cock that was not open at the start, and not a pint
of fresh water to make up waste, only the sea to draw from, and
at the end of the voyage the boilers were in excellent condition.

If a satisfactory way cannot be devised to introduce the water
under pressure, I think that the use of nickel steel would solve
the whole problem.

METAL CORROSION AND PROTECTION 1

References to Books and Magazine Articles

The following abbreviations have been used: Diag., diagrams; Dr., drawings; ed.,
edition; 111., illustrations; n. d., no date; n. s., new series; no., number; p., page; pi., plate;
pt., part; v., volume; w., words.

BIBLIOGRAPHY AND INDEX

England — Patent office.

Abridgment of specifications, new series. 1855-1904. 111.

Class 20, Buildings and structures. 1906 p.

Class 41, Electrolysis. 430 p.

Class 123, Incrustation and corrosion, prevention and removing in steam generators,
water heating pipes and the like. 2781 p.

Class 95, Paints and compositions, antifouling. 496 p.

Class 113, Ships. 1661 p.

Abridgments of all British patents, arranged by subjects. Much information on pre-
servative coatings. Classifications for the periods 1617-1855 and 1904-1908 are in prepa-
ration.

Uberziehen des eisens mit anderen metallen. 3 p. 1908. (In Stahl und
eisen, Gesamt-inhaltsverzeichnis der jahrgange 1 bis 26, 1881-1906. 1908.
p. 242.)

Indexes carefully the contents of "Stahl und eiaen," giving rust prevention methods
under eleven different heads. Includes patents.

CORROSION

GENERAL AND THEORETICAL
Adie, R.

On the corrosion of metals. 10 p. 1845. (In Minutes of proceedings
of the Institution of Civil Engineers, v. 4, p. 323.)

Shows that saturated salt solutions are a great protection from corrosion.

Akerman, R.

Ueber das rosten des eisens. 4,200 w. 1882. (In Stahl und eisen,
v. 2, p. 417.)

Considers theory of rusting, especially of protective metal coatings, and of the influ-
ence of manganese in the rusting of steel.

Alford, H. Carroll.

Corrosion of iron and its prevention. 2,200 w. 1901. (In Proceed-
ings of the St. Louis Railway Club, v. 5, April 12, p. 9.)

Theory of rust formation and preventive measures.

1 Reprinted by permission from The Monthly Bulletin for July, 1909, issued by the Car-
negie Library, Pittsburgh, Pa; with additions and corrections to February, 1910.

301

302 CORROSION AND PRESERVATION OF IRON AND STEEL

American Society for Testing Materials. 1,800 w. 1906. (In Iron age,

v. 77, p. 2057.)

Abstracts of papers at ninth annual meeting of the society; corrosion of tube steel,
corrosion of wire fencing, electrolysis in structural steel, etc.

Andes, Louis Edgar.

Der eisenrost; seine bildung, gefahren und verhutung unter besonderer
beriicksichtigung der verwendung des eisens als bau- und constructions
material. 292 p. 111. 1898.

Treats very fully of rust formation and gives many methods of prevention, chiefly by
preservative paints.

Andrews, Thomas.

Effect of stress on the corrosion of metals. 6,000 w. 111. 1894. (In
Minutes of proceedings of the Institution of Civil Engineers, v. 118, p. 356.)

Bauer, 0.

Uber den einfluss der reihenfolge von zusatzen zum flusseisen auf die
widerstandsfahigkeit gegen verdiinnte schwefelsaure. 1,000 w. Diag. dr.
1905. (In Mitteilungen aus dem Koniglichen Material prufungsamt, v. 23,
p. 292.)

Considers the influence of aluminium and tungsten on the corrosion of steel in dilute
sulphuric acid.

Bradford, W. A.

Corrosion vs. so-called corrosion. 2,500 w. 4 ill. 1909. (In Metal

Worker, v. 72, October 2, p. 45.)

Effect of water and other destructive agents on various kinds of pipe.

Breuil, Pierre.

Corrosion tests on copper steels. 400 w. Dr. 1907. (In Journal of
the Iron and Steel Institute, v. 74, p. 41.)

Experiments using sulphuric acid as corrosive liquid "make copper steels "rank in value
with nickel steels in every respect of corrosion."

Breuil, Pierre.

Corrosion tests on the [copper] steels as rolled. 1,200 w. 1907. (In
Journal of the Iron and Steel Institute, v. 74, p. 60.)

Tests show corrosion to take place much more slowly with rolled steel.

Brown, A. Crum.

On the chemical processes involved in the rusting of iron. 1,200 w.
1888. (In Journal of Iron and Steel Institute, v. 33, p. 129.)

Discussion, 800 w.

Rusting caused by action of carbon dioxide and oxygen.

Bruhl, Paul.

On the preservation of instruments and machinery in Bengal. 10,000 w.
1903. (In Engineer, London, v. 96, p. 101, 125, 147.)

Effect of warm, moist climate, particularly on delicate instruments.

Buchanan, J. F.

Corrosion of metals. 2,200 w. 1904. (In Foundry, v. 24, p. 160.)
Briefly considers relative corrosion of the more useful metals and alloys.

Burgess, Charles F.

Corrosion of iron from the electrochemical standpoint. 32 p. Diag.

BIBLIOGRAPHY 303

Burgess, Charles F. — continued.

dr. ill. 1908. (In Transactions of the American Electrochemical Society,
v. 13, p. 17.)

Discussion, 6 p.

The same without discussion. (In Electrical review, New York, v. 53,
p. 371, 436.)

Considers the influence of strain and of inequalities of temperature on corrosion.

Burgess, Charles F. & Engle, S. G.

Observations on the corrosion of iron by acids. 3,000 w. 1906. (In
Transactions of the American Electrochemical Society, v. 9, p. 199.)

Effect of normal solutions of sulphuric and hydrochloric acids on electrolytic iron.

Calvert, F. Crace.

Experiments on the oxidation of iron. 1,000 w. 1871. (In Chemical
news, v. 23, p. 98.)

Paper before the Manchester Literary and Philosophical Society.

Indicates that "carbonic acid ia the agent which determines the oxidation of iron."

Corrosion and protection of metal surfaces. 9,500 w. 1897. (In Work-
shop receipts, v. 5, p. 283.)

Takes up copper, iron and steel, lead, silver and zinc.

Corrosion of iron. 4,700 w. 1907. (In Electrochemical and metallurgi-
cal industry, v. 5, p. 363.)

Gives in condensed form papers by Walker and Cushman.

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 corrosion and preservation of iron and steel"

See other formats