STP 534-1973

Copyright by ASTM Int'l (all rights reserved); Fri Jan 1 23:09:17 EST 2016
Downloaded/printed by
University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized.
MANUAL OF INDUSTRIAL
CORROSION STANDARDS
AND CONTROL
Sponsored by ASTM Committee G-1

on Corrosion of Metals
ASTM SPECIAL TECHNICAL PUBLICATION 534

F. H. Cocks, editor
List price $16.75

04-534000-27
Jt~[~ AMERICAN SOCIETY FOR TESTING AND MATERIALS

191 6 Race Street, Philadelphia, Pa. 191 03
University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized
(~) BY A M E R I C A N SOCIETY FOR TESTING AND MATERIALS 1973
Library of Congress Catalog Card N u m b e r : 73-75375
NOTE
The Society is not responsible, as a body,
for the statements and opinions
advanced in this publication.
Printed in Baltimore, Md.

November 1973
Foreword
The Manual of Industrial Corrosion Standards and Control has been

prepared and sponsored by the members of ASTM Committee G-1 on
Corrosion of Metals. Dr. Franklin H. Cocks was responsible for the
organization of this material.
University of Washington (University of Washington) pursuant to License Agreement. No further reproduction
Related
ASTM Publications
Metal Corrosion in the Atmosphere, STP 435 (1968),
$27.00 (04-435000-27)
Localized Corrosion--Cause of Metal Failure, STP
516 (1972), $22.50 (04-516000-27)
Stress Corrosion Cracking of Metals--A State of the
Art, STP 518 (1972), $11.75 (04-518000-27)
Contents
Introduction 1
Chapter 1. Introduction to Corrosion--F. H. COCKS 3
Chapter 2. Corrosion Standards and Control in the Petroleum Industry--
A. S. C O U P E R 42
Chapter 3. Corrosion Standards and Control in the Gas Industry--L M.
BULL 60
Chapter 4. Corrosion Standards and Control in the Automotive Industry--
C. O. D U R B I N 81
Chapter 5. Corrosion Standards and Control in the Pipeline Industry--
A. W . P E A B O D Y 89
Chapter 6. Corrosion Standards and Control in the Telephone Industry--
o. SCmCK 107
Chapter 7. Corrosion Standards and Control in the Marine Industry--
B. F. BROWN 133
Chapter 8. Corrosion Standards and Control in the Nuclear Power In-
dustry--w. E. BERRY 144
Chapter 9. Corrosion Standards and Control in the Chemical Industry--
L. W. GLEEKMAN 164
Chapter 10. Corrosion Standards and Control in the Nonferrous Metals
Industry--w. H. AILOR 194
Chapter 11. Corrosion Standards and Control in the Iron and Steel Industry
--H. P. LECKIE 209
Appendix A-1. Tabulated list of Current Corrosion Standards, Test Methods,
and Recommended Practices Issued by the American Society for
Testing and Materials (ASTM) and the National Association of
Corrosion Engineers (NACE) 236
Appendix A-2. Selected Tabulation of British, French, and German Stand-
ards Concerned with Corrosion Testing Methods and the Evaluation
of the Corrosion Resistance of Materials and Products 240
Appendix B. Selected ASTM Standards Referred to Frequently in Book:
A 279-63--Standard Method of Total Immersion Corrosion Test of
Stainless Steels. 245
B 117-73--Standard Method of Salt Spray (Fog) Testing. 253
G 1-72--Standard Recommended Practice for Preparing, Cleaning,
and Evaluating Corrosion Test Specimens. 261
G 4-68--Standard Recommended Practice for Conducting Plant
Corrosion Tests. 266
G 15-71--Standard Definitions of Terms Relating to Corrosion and
Corrosion Testing. 279
G 16-71--Standard Recommended Practice for Applying Statistics to
Analysis of Corrosion Data. 281
Frontispiece: Photograph of U.S. 35 Highway Bridge, Point Pleasant, W.Va. taken

after its collapse on 15 Dec. 1967. Courtesy National Transportation
Safety Board.
University of Washington (University of Washington) pursuant to License Agreement. No further reproductions auth
STP534-EB/Nov. 1973
Introduction
This manual is a working source book of procedures, equipment, and

standards currently being used to solve industrial testing and control prob-
lems. It is intended as a guide to those in university and government, as well
as in industrial laboratories, who are faced with combatting corrosion
problems or developing more corrosion resistant materials. The aim
throughout is to combine a brief discussion of fundamental principles with
clear descriptions of concomitant techniques and methods as well as the
types of problems to which these have been and are being applied.
Although corrosion problems are common to all industries, the test
methods and control procedures that have been developed to deal with them
are diverse. By combining descriptions of major corrosion problem areas
together with discussions of the approaches that have been evolved for
controlling them, more effective means for reducing corrosion losses may
be fostered. Thus, this manual is organized so that the first chapter pro-
vides a concise introduction to basic corrosion science, while subsequent
chapters, each written by a leader in his field, review the application of these
principles in practice. Emphasis is placed on the explanation of proven
methods and standards, as well as on suggestions for procedures which
might well become standards in the future. These chapters are followed by
two appendices. The first provides abstracts and sources for existing
corrosion standards, while the second appendix includes six ASTM stand-
ards referred to most frequently in the text.
Within the past decade it has become clear to an increasing number of
diverse scientific and industrial groups that more emphasis on the standardi-
zation of corrosion tests and the means for interpreting data derived from
them is both necessary and valuable. It is often difficult, however, when
faced with a specific corrosion problem, to know which of several different
testing procedures and standards should be utilized or where information
directly relevant to a particular situation might be obtained. It is hoped
that this manual will assist in resolving this difficulty.
Franklin H. Cocks
Duke University
School of Engineering
Durham, N.C. 27706
Copyright9 1973 by ASTM International www.astm.org
STP534-EB/Nov. 1973
Chapter 1
Introduction to Corrosion
F. H . C o c k s I
Webster [1] 2 defines corrosion as "the action or process of corrosive

chemical c h a n g e . . , a gradual wearing away or alteration by a chemical
or electrochemical essentially oxidizing process as in the atmospheric rust-
ing of iron." This definition does not restrict corrosion to any one class of
materials, nor to any one environment. It does, however, imply a degrada-
tion in properties through the reaction of a material with its surroundings.
This environment may be liquid, gaseous, or even solid as in the case of the
reaction of filaments of SiC with an aluminum matrix they are intended to
reinforce. Although many such new corrosion reactions are being en-
countered as more complex materials are applied in increasingly varied and
unusual situations, the problems associated with far more mundane and
widespread corrosion reactions have by no means been satisfactorily solved.
The formation of oxides on iron exposed to the atmosphere at both ambient
and elevated temperatures, for example, in automobile mufflers, year after
year continues to extract a cost of hundreds of millions of dollars. Con-
siderable progress has been and continues to be made, however, in reducing
these corrosion losses. It is to the further control and reduction of practical
and industrially important corrosion problems that this manual is directed.
Corrosion studies and the development of improved methods of cor-
rosion prevention and control are of enormous practical industrial im-
portance. It has been estimated that in the United States alone, the costs
attributable to corrosion amount to more than 10 billion dollars annually
[2]. While some corrosion losses may appear inevitable, the proper selection
of materials and the application of known principles and protection
methods can be expected to reduce these losses greatly.
In this introductory chapter, the basic principles of corrosion science are
reviewed as a guide to subsequent chapters which each provide a discussion
of how this knowledge can be applied in industrial practice to achieve the
desired goal--the minimization of the economic burden imposed by
corrosion. The unifying theme throughout these chapters is the use of
Duke University, School of Engineering, Durham, N.C. 27706.
Italic numbers in brackets refer to references hsted at the end of this chapter.
Copyright 9 1973 by ASTM International www.astm.org
4 INDUSTRIAL CORROSION STANDARDS AND CONTROL
standards which accurately detail the testing methods and control pro-
cedures now carried out in major industries. It is to be hoped that the
information provided will contribute not only to the more effective and
widespread use of available standards but to the development of additional
corrosion standard test methods and control procedures as well.
The attack on metals by their environment can take many forms, ranging
from uniform general attack and tarnishing to more complex reactions
such as pitting, filiform corrosion, corrosion fatigue, stress corrosion, and
other specific forms of damage discussed later in this chapter. The type of
property degradation that will occur depends not only on the nature of the
metallic material, and its physical state and conditions of use, but on the
composition of the environment as well. The specific chemical species
present in this environment, their concentration, and the temperature can
determine whether attack will be general or localized or whether it will be
fast or slow, accelerated or inhibited. The physical structure of many
metals of a given composition can be enormously altered by heat treatment
or cold working, and this structure in many cases will determine whether
attack will be catastrophic or relatively mild.
In evaluating and correcting an existing or potential corrosion situation
there are several fundamental choices to be considered. Does the metal or
alloy being considered represent an optimum choice both from the point of
view of economics as well as corrosion resistance? What will the environ-
mental conditions this alloy is exposed to be and is it feasible to consider
modifying this environment? What limits are imposed on the design of the
structure being considered and how can this design be changed to minimize
corrosive effects? Can protective coatings be used to isolate the whole
structure, or critical parts of it, from the environment? The design engineer,
too, can influence corrosion processes, not only directly through the speci-
fication of materials but also by providing material and environment
configurations that minimize corrosive effects. Such designs can only be
optimized if the processes that might lead to damage are understood.
While the range of possible corrosion situations is so large that a descrip-
tion of even a small fraction of them is not practical, a surprisingly few
basic principles are sufficient to understand the detailed mechanisms of each
case. Once the mechanism of damage is understood, the likelihood of making
the correct choice to eliminate or minimize this damage is greatly improved.
In the following section, these underlying principles of corrosion proc-
esses are described before going on to consider important special forms of
corrosion attack and methods of corrosion protection and control.
Basic Corrosion Principles

The conversion of elemental metals or alloys into ions in an electrolyte
(any electrically conducting solution, for example, seawater) is an essentially
electrochemical process. The electrochemical character of corrosion has
INTRODUCTION TO CORROSION 5
long been firmly established, and a concise review of the early experimental
proofs of the electrochemical basis of corrosive action is available [3].
When a metal is placed in an electrolyte it acquires an electrical potential
which is a measure of the tendency for that metal to dissolve as positive
ions in solution. Since the solution must remain electrically neutral, an
equivalent number of some other positive ions must be removed as the
metal corrodes. A sample of iron placed into a solution of copper sulfate,
for example, will begin to corrode (dissolve as iron ions) while at the same
time copper ions are plated out of solution forming copper metal on the
surface of the iron. The dissolution of the iron can be written as
Fe --~ Fe ++ q- 2e- (1)
and is said to be an anodic reaction because the solid iron (Fe) is being
increased in oxidation state to form iron ions (Fe++), by the removal o f two
electrons (2e-) per iron atom. The copper reaction can be written as
Cu ++ + 2e- ~ Cu (2)
and is said to be a cathodic reaction because copper ions are being reduced
in oxidation state through the gain of electrons, to form copper metal. The
combination of reactions 1 and 2 gives
Fe q- Cu ++ ~ Fe ++ -J- Cu (3)
as the overall electrochemical reaction. This corrosion reaction is self-
stifling, however, because the deposited copper acts as a barrier between
ZINC
H+
CI-
2e"
Zn "H" CI-
--DILUTE HYDROCHLORIC ACID
FIG. 1--Schematic drawing showing the corrosion of zinc in dilute hydrochloric acid.
6 INDUSTRIALCORROSION STANDARDS AND CONTROLS
the iron and the solution, thus preventing further reaction. In the case of
zinc immersed into acid solutions, it is hydrogen which is plated out from
solution in order to maintain electrical neutrality, as shown in Fig. 1. Here,
the electrons released by the zinc as it ionizes and goes into solution travel
through the remaining solid zinc to the points on the surface where hydro-
gen ions are neutralized to form hydrogen atoms. Two such neutralized
atoms must then combine to form a molecule of hydrogen gas. Since the
hydrogen gas can be removed as bubbles, the reaction is not a self-limiting
one, and the formation of zinc chloride is not stifled.
In both corrosion reactions just described, the flow of electrons occurs
within the specimen of corroding metal itself. This current flow could just
as well pass through an external wire to neutralize ions at some other point,
as for example, at a piece of copper immersed elsewhere in the solution as
shown in Fig. 2. In such a case, the corroding sample (zinc) is defined as the
anode and the copper sample, which does not corrode, as the cathode.
The tendency for zinc to enter the solution is dependent upon the concen-
211-
FIG. 2--Schematic drawing showing the separation of anodic and cathodic relations when
strips of zinc and copper in hydrochloric acid are electrically connected.
2e- >
POROUS
MEMBRANE
I- z..- -z.--I I-I
2e .
I-~176
. . . . . .
I It
2e ~
FIG. 3--Schematic drawing o f a metal-ion concentration corrosion cell.
tration of zinc ions already present in this solution. For example, one could
construct a corrosion cell as shown in Fig. 3, by placing two zinc specimens
in solutions containing different concentrations of zinc ions. In tl~,is case the
zinc sample which is immersed in the less concentrated zinc solution will
corrode while the zinc specimen immersed in the more concentrated zinc
solution will have additional zinc plated on it. This process is an example of
concentration cell corrosion and illustrates the point that corrosion can
occur even if the metals making up the anode and the cathode are identical.
The electrical potential reached by a metal immersed in an aqueous
solution thus depends on the concentration of its ions already present in
solution. The electromotive force series shownin Table 1 lists the potentials
acquired by different metals when each is in contact with an aqueous solu-
tion of its ions at unit activity (approximately 1 mole/1000 g of water at
25 C) [4]. The zero potential assigned to hydrogen is selected arbitrarily and
thus constitutes the reference potential against which the others have been
measured. Very reactive metals such as sodium and magnesium appear at
the negative or less noble end of the list, while inert metals such as platinum
or gold appear at the more noble or positive end.
University of Washington (University of Washington) pursuant to License Agreement. No further reproductions
8 INDUSTRIALCORROSION STANDARDS AND CONTROL
T A B L E 1--Standard electromotive force series (emf) at 25 C [4].
S t a n d a r d El e c t rode
Reaction Potential, volts
A u +++ + 3e- = Au 41.50 N o b l e ( m o r e c a t hodi c )

Ag + 4 le- = Ag 40.7991
Cu ++42e- = Cu 40.337
2H + 4 2e- = H~ 0.00
P b ++ 4 2 e - = Pb -0.126
Sn + + 4 2 e - = Sn -0.136
Ni + + 4 2 e - = Ni --0.250
Cd ++ 4 2c- = Cd -0.40
Fe ++ 4 2 e - = Fe -0.440
Cr +++ 4 3e- = Cr --0.74
Z n ++ 4 2e- = Zn --0.763
A1+++ + 3 e - = AI -1.66
M g +§ 4 2e- = Mg -2.37 Ac t i ve ( m o r e a nodi c )
As an example of how such a scale can be used, one can imagine a cor-
rosion cell constructed as shown in Fig. 4. Here one c o m p a r t m e n t contains
a specimen of zinc in a solution of zinc ions at unit activity (approximately
1 mole of zinc ions per 1000 g of water). The other c o m p a r t m e n t contains a
specimen of silver in a solution of silver ions also at unit activity. A volt-
meter connected between these two metal specimens would read 1.562 V as
would be expected from their relative position in Table 1. Then, when the
voltmeter is replaced by a copper wire, the more active zinc will be found to
corrode, while the less active silver is plated from solution. As this process
continues, the voltage measured between the zinc and silver specimens
would decrease as the concentration of zinc ions increased while that o f
silver ions decreased. Thus, corrosion cell potentials depend on both the
electrode material and the electrolyte composition.
In addition to the standard emf series of Table 1 it is also useful to know
cell potentials obtained using a single c o m m o n electrolyte. Such a listing is
called a galvanic series and the relative position shown by a group of metals
and alloys immersed in seawater as the standard electrolyte is shown in
Table 2. I f a pair of metals selected from this list are immersed in seawater
and connected together electrically, the metal lower on the list will be found
to corrode. The farther apart the metals of this pair are, the greater will
be the tendency for the lowermost one to corrode. It must be remem-
bered that this list applies only to a specific electrolyte--seawater--and a
much different sequence could result if some electrolyte other than seawater
were chosen.
As illustrated for the case of zinc in hydrochloric acid, corrosion reactions
can be divided into two parts. In the case of zinc in hydrochloric acid, the
anodic (corrosion) reaction is that involving zinc entering solution.
Anodic Reaction: Zn --~ Zn ++ q- 2e- (4)
University of Washington (University of Washington) pursuant to License Agreement. No further reproductions aut
1.56 VOLTS
/ / /
///
/// . . . .
/ / / -- -- _
- --Zn.H. -- /// Ag +
/,~'j
-- -UNIT -- f / / -- UNIT-- (D
-- --ACTIVITY- ///
/ / /
--ACTIVITY-
/ / /
/ / i
/ / / - - __ _ _
/ ' / /
/ / /
/ / J
i / /
i / /
/ / J
/ / J
ff[
FIG. 4--Schematic drawing showing the voltage developed between two standard half cells.
The second part is the cathodic reaction of the hydrogen required for
electrical neutrality of the solution.
Cathodic Reaction: 2H + -t- 2e- ~ H2 (5)
There are not many practical situations, however, in which metals are used
in sufficiently acid solutions that hydrogen gas evolution occurs. In many
service environments corrosion is decreased by the formation of a thin
film of hydrogen gas on the cathodic surfaces which decreases the current
flow and hence the corrosion rate. This situation is known as hydrogen
polarization. If this film of hydrogen is destroyed or prevented from form-
ing, the corrosion rate will be increased. The presence of dissolved oxygen
can lessen hydrogen polarization by shifting the potential to more active
values and reacting with the hydrogen to form water.
02 q- 2H~ (or 4H) ~ 2H20 (6)
University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authori
TABLE 2--Galvanic series of metals and alloys.
Noble (more cathodic) Platinum

Gold
Graphite
Silver
Chromium Nickel Stainless Steel Type 304 (passive)
Chromium Nickel Stainless Steel Type 316 (passive)
13 7o Chromium Steel Type 410 (passive)
Titanium
Monel
70-30 Cupro-Nickel
Silver Solder
Nickel (passive)
76Ni-16Cr-7Fe Alloy (passive)
Yellow Brass
Admiralty Brass
Aluminum Brass
Red Brass
Copper
Silicon Bronze
Nickel (Active)
76Ni-16Cr-7Fe Alloy (active)
Muntz Metal
Maganese Bronze
Naval Brass
Lead Tin Solders
Lead
Tin
Chromium Nickel Stainless Steel Type 304 (active)
Chromium Nickel Stainless Steel Type 316 (active)
Chromium Stainless Steel Type 410 (active)
Mild Steel
Wrought Iron
Cast Iron
Aluminum (2024)
Cadmium
Aluminum (6053)
Alclad
Zinc
Magnesium Alloys
Active (more anodic) Magnesium
It is also possible for dissolved oxygen to participate directly in the cathodic

reaction by being reduced to hydroxyl ions.
O~ q- 2H20 q- 4e- ~ 4 O H - (7)
In either case the presence of dissolved oxygen acts to depolarize the
cathodic reaction and leads to an increased rate of corrosion by increasing
the rate at which metal ions can enter the solution.
During corrosion, more than one oxidation process and more than one
reduction process may occur simultaneously. This situation would be
expected, for example, if the corroding metal were an alloy containing two
University of Washington (University of Washington) pursuant to License Agreement. No further repro
or more elements or if the solution environment contained more than one

reducible species. If, for example, the dilute acid in Fig. 1 also contained
dissolved oxygen, then both oxygen reduction as well as hydrogen reduction
could occur, leading to a higher corrosion rate for the zinc in oxygen-
containing acid than in deaerated acid. The anodic reaction, on the other
hand, would be increased if species were present which could form com-
plexes with the metal's ions, thus lowering the effective concentration of
such ions in solution. Conversely, inhibitors can act to slow the rate of
corrosion by interfering with the cathodic reaction, the anodic reaction, or
both, as discussed in Methods of Corrosion Prevention and Control of
this chapter.
In many practical corrosion situations in natural environments under
nearly neutral or alkaline pH conditions, the rate of corrosion is sub-
stantially determined by the concentration of oxygen. As was shown in
Fig. 3, corrosion can occur between two identical metals if the concentra-
tion of their ions in solution varies. Similarly, a corrosion cell will also be
formed if the concentration of dissolved oxygen varies, as illustrated in
Fig. 5. In this figure, the sample on the right is the cathode while the sample
on the left corrodes and is the anode, because of the difference in oxygen
concentration and the resultant ease with which the cathodic reaction
(Eq 7) can occur. There are many practical situations where such a dif-
ference in oxygen concentration can arise, as for example in the case of
crevice corrosion discussed in the next section where the oxygen deficient
conditions inside the crevice favor the anodic corrosion reaction. Oxygen
concentration cell corrosion is indeed a widespread form of attack. In a
tank that is only partially full of water, for example, the water at the top
will contain more oxygen than the rest, and the metal touching this oxygen-
ated water will be cathodic to the remainder of the tank. Similarly, scale,
rust, or other surface deposits can lead to oxygen concentration cell cor-
rosion by limiting the oxygen supply to specific local areas.
In addition to these effects, the relative area of metal on which the anodic
and cathodic reactions occur is also important in determining corrosion
rates. If, for example, the area in solution of the specimen of iron labeled B
in Fig. 5 were doubled relative to that of specimen A, the corrosion rate o f
specimen A would be increased. This increase would occur because the
greater area available for the cathodic reaction (Eq 7) would increase the
rate at which this oxygen reduction reaction could occur. Conversely, the
rate of corrosion would be reduced if the area of specimen B were decreased.
Effects such as this can be readily understood with reference to an Evans
diagram [5] as shown in Fig. 6. In this diagram, the changes in potential
which occur for both the anodie and cathodic reactions are shown as a
function of the current which flows between the anode and the cathode.
As may be seen, the potentials of each reaction approach each other as the
current increases. That is, each reaction becomes polarized as its rate
University of Washington (University of Washington) pursuant to License Agreement. No further reproduct
12 INDUSTRIALCORROSIONSTANDARDSAND CONTROL
Icorr
e--..>
N|TROGEN OXYGEN
A MEMBRANE : ~ l
/ / / ~-- -
i i i
: IT --
I I I
//'/
/ / /
/ / I
/ / I
- - - - I / / - -
FIG. 5--Schematic drawing o f an oxygen concentration corrosion cell
increases. In the case of the oxygen reduction reaction, this polarization

becomes particularly severe at relatively low currents because of the low
solubility of oxygen in solution. That is, at relatively low currents it begins
to require substantial changes in potential to produce slight increases in
cathodic current because the available dissolved oxygen at the cathode is
depleted (diffusion control). The corrosion rate, which is proportional to the
current flowing (il, i2, or i~) is fixed by the intersection of the anodic and
cathodic curves. As shown in the figure, increasing the area of the cathode
(or increasing the oxygen concentration) will increase the overall corrosion
rate by decreasing the degree of polarization of the cathodic reaction.
Similarly, the overall amount of corrosion would also be increased if the
area of the anode were increased although this increase would be relatively
small if, as shown, oxygen diffusion to the cathode were the limiting factor.
In the case just described, the corrosion reaction is said to be under
cathodic control since the greatest change in potential occurs in the cathodic
reduction reaction. In still other cases, the corrosion rate may be limited by
the electrical resistance of the electrolyte. In this latter case, the potentials
at which the anodic and cathodic reactions occur are not equal but differ
by the voltage drop which occurs through the electrolyte. Evans diagrams
University of Washington (University of Washington) pursuant to License Agreement. No further reproducti
I >2 >5
o~ INCREASING CATHODE
ELECTRODE AREA
4 )5 )6
INCREASING ANODIC
%. ELECTRODE AREA
I- 2 45j ~ ' J " [ METAL 9

z
bJ METALLIC IONS
I.-
o
0..
Z ~ 6
I-PLUS ELECTRONS
I
I
I
I
il 12 15
CURRENT ~ "
FIG. 6---An Evans diagram illustrating the effect o f increasing anodic or cathodic area
on corrosion where oxygen diffusion is the limiting factor.
illustrating these three situations are shown in Fig. 7. Such diagrams are
useful in interpreting many different corrosion effects and extended dis-
cussions of such uses are available [6,7,8].
The extremely important phenomenon of passivity can also be understood
by considering the way in which the rate of the anodic (corrosion) reaction
of certain metals varies with potential or, alternatively, with the oxidizing
power of the corrodent (corrosion solution).
Table 1, for example, shows that zinc is electrochemically much less
active than aluminum. Yet Table 2 shows that aluminum is cathodic to
zinc in seawater. This corrosion resistance of aluminum is due to the
presence of an adherent film of oxide on its surface. For metals such as
stainless steel this film may be extremely thin but will still give protection
in oxidizing environments. In reducing environments, however, this oxide
film is removed and the steel becomes active. The corrosion resistance of
titanium alloys depends similarly on the presence of protective, passive
films. There are, in fact, two distinct types of passive behavior. In the case
of lead in sulfuric acid, for example, a passive protective film is formed in
dilute solutions and the corrosion rate remains very low, until in more con-
centrated acid solution, the film becomes increasingly soluble and the
corrosion rate increases. For the case of iron in nitric acid solution, how-
ever, a different passive behavior is observed. In dilute nitric acid, iron
University of Washington (University of Washington) pursuant to License Agreement. No further reproduc
Er
_1
E
\
._1
I- I-
z
LU
I- I.-
0 E corr. 0
0. O.
J,
E I
',i Icorl: E a ~ Icorr ] Icorr.
CURRENT CURRENT CURRENT
(o) (b) (c)
FIG. 7--Evans diagrams showing corrosion reactions which are under (a) cathodic control,
(b) anodic control, and (c) solution resistance control.
corrodes at a high rate. As the concentration of acid is increased this

corrosion rate at first increases, as shown in Fig. 8. At a critical HNO3 con-
centration, however, a further increase in acid concentration causes a very
large drop in corrosion rate, due to the formation of a protective, passive
film on the iron. If the acid concentration is reduced to the initial dilute
condition the corrosion rate will remain low, because the passive film is
retained. However, this passive film is then unstable, and the original high
corrosion rate can be restored by scratching or tapping the iron sample.
1'
z
0
J
I-- ~:
PASSIVE
CORROSION RATE )
(CURRENT)
FIG. 8--Evans diagram showing the corrosion behavior o f iron in dilute and in concentrated
nitric acid, illustrating the onset o f passivity.
INTRODUCTION TO CORROSION ] 5
Passivity may thus be broadly defined as the decrease in corrosion sus-

ceptibility exhibited by certain metals and alloys brought about by the
generation of protective films or adsorbed layers in particular environ-
ments where they would be expected to corrode readily. The importance of
this phenomenon in determining the corrosion behavior of many imporant
ahoy systems, such as stainless steel and titanium alloys, cannot be over-
emphasized and has lead to a large number of investigations. Concise
reviews of this work and current theories on the nature of passive film alloys
are available [9,10].
The corrosion of iron, like that of all other metals, is strongly dependent
not only on potential but also on the pH of its solution environment.
From available thermodynamic and electrochemical data it is possible to
construct a diagram which shows the regions of potential and pH where
certain species are stable. These diagrams are usually referred to as Pourbaix
diagrams in honor of the man who first suggested their use. In using them,
it is to be emphasized that no rate information can be obtained and only
equilibrium data are involved. Figure 9 shows, for example, a simplified
Pourbaix diagram for iron in water [11]. In this diagram the only solid
substances considered are Fe, Fe304 and FelOn. A slightly different diagram
"I'L6 | l I I I l I I
Fe 0 4 - - 9
+1.2
Fe
+0.8
1'ui
+0.4
r
+0.2 Fe
--I
Z -0.2
hi
I"--
0
n
-0.6
-I.O Fe
HFeO--
- 1.4 I I I I I I I I
-2 0 +2. 4 6 8 I0 12 14 16
pH
FIG. 9--Simplified Pourbaix (potential-pH) diagram for Fe in H20, considering only Fe,
Fe30~, and Fe203 as solid phases.
would be obtained if Fe, Fe(OH)2 and Fe'(OH)3 were considered. The

potentials given are those which would be measured against a standard
hydrogen electrode.
In this diagram, when any reaction involves species other than O H - or
H +, such as Fe ++, a concentration of 10-6 moles/1 is assumed. Thus, the
horizontal line dividing the Fe and Fe ++ fields indicates that for potentials
more negative than - 0 . 6 2 V, iron will not corrode to f o r m a solution
containing more than 10-8 moles/1 of Fe ++ ions. Thus, iron is immune to
corrosion over the range of potentials and p H values where Fe is the stable
species. Conversely, iron will corrode in the range of potentials and p H
values where Fe ++, Fe +++, or HFeO2- are the stable species. N o informa-
tion is provided, however, on the rate of corrosion. In those regions where
solid Fe304 and Fe203 are formed, passive films can be formed, which m a y
give some protection against corrosion. It must also be remembered that
the diagram shown in Fig. 9 is for pure iron in water. A different diagram
would be needed if either an iron alloy or a solution containing a salt, such
as NaC1, were being considered. As data involving practical alloys and
c o m m o n environments become available, Pourbaix diagrams can be
expected to come into ever increasing use.
In this section we have shown how differences in both metal and solution
composition can give rise to the electrochemical potential differences
required to produce corrosion. In the next section we now go on to consider
some of the important special forms which this corrosive action can take.
Forms of Corrosion Attack

The previous section has outlined the basic electrochemical principles
which underlie corrosion processes. In this section we will describe some
o f the important specific forms which these corrosion processes can take in
aqueous, atmospheric, and soil environments, including a discussion o f
bacteriological influences and high temperature oxidation processes. This
will lead, in the last section, to an outline of the basic approaches which can
be used to minimize or prevent corrosion losses.
Uniform Attack
Corrosion which occurs uniformly over the surface of a material is the
most c o m m o n form of damage. It may proceed at a nearly constant rate i f
the reaction products are soluble or the attack may be self-stifling if these
products do not dissolve readily in the corrodent, as we have already seen
for the case of iron immersed in a copper sulfate solution. Similarly, in
corrosion of silver by a solution of iodine in chloroform, attack slowly
ceases as a film of insoluble silver iodide is built up. On the other hand, the
attack of unstressed Zn in dilute sulfuric acid also occurs over the entire
exposed surface of the zinc. Since in this case the reaction product, zinc
sulfate, is soluble, the rate of reaction of the zinc will be constant provided
University of Washington (University of Washington) pursuant to License Agreement. No further reprod
the sulfuric acid is present in excess. In other cases such as the rusting of
iron, the build-up of an oxide layer does not prevent further attack because
the porous form of the corrosion product does not exclude the environment.
Certain special grades of weathering steels now coming into use, however,
contain small amounts of alloying elements which lead to the formation of
protective oxides that stifle continuing attack. A typical composition for
such a steel would be (in weight percent) 0.12C-0.3Mn-0.1P-0.5Si-0.5Cu-
1.0Cr-0.5Ni-balance Fe. The way in which these elements influence the
corrosion process is still uncertain. It appears, however, to be related to the
combined influence of these alloying additions in providing a dense, adher-
ent oxide layer near the metal-oxide interface.
Most commonly, uniform attack occurs on metal surfaces which are
homogeneous in chemical composition or which have homogeneous micro-
structures. The access of the corrosive environment to the metal surface
must also usually be unrestricted. As we have seen, corrosion requires both
anodic and cathodic areas and on a specimen that is corroding uniformly
such areas may be visualized as fluctuating over the surface.
The rate of uniform attack can be evaluated in a straightforward manner,
using either weight loss or specimen thickness change measurements. It is
important to remember, however, that the rate of attack may vary with time
and so measurements should be made at more than one interval. An extreme
example of this is shown by the weathering steels mentioned previously
where the rates of attack may be initially quite high but continuously
decrease as the time of exposure increases. In the case of uniform attack this
rate can be expressed as milligrams per square decimeter per day (mdd),
inches per year (ipy), or other convenient units. Uniform corrosion attack
is quite common, but so too are other forms of corrosion which can make
the correct evaluation of corrosion damage more difficult.
Pitting Corrosion
One of the most troublesome forms of corrosion is the formation of pits
on metal surfaces. In pitting corrosion, attack is highly localized to specific
areas which develop into pits. Active metals such as Cr and A1, as well as
alloys which depend on Cr- or Al-rich passive oxide films for resistance to
corrosion are prone to this form of attack. Thus, stainless steels and alumi-
num alloys are particularly susceptible, especially in chloride containing
environments. These pits usually show well-defined boundaries at the
surface, but pit growth can often change direction as penetration progresses.
When solid corrosion products are produced the actual corrosion cavity
may be obscured but the phenomenon can still be recognized from the
well-defined nature of the corrosion product accumulations. Pitting cor-
rosion is usually the result of localized, autocatalytic corrosion cell action.
Thus, the corrosion conditions produced within the pit tend to accelerate
the corrosion process. As an example of how such autocatalysis works,
consider the pitting attack of aluminum in an oxygenated solution of

sodium chloride. Imagine that there exists a weak spot in the oxide film
covering the aluminum surface so that the corrosion process initiates at this
point. The local accumulation of A1+++ ions will lead to a local increase in
acidity due to the hydrolysis of these ions. That is, the hydrolysis of alumi-
num ions gives as the overall anodic reaction:
A1 + 3H20 --+ 3H + + AI(OH)~ + 3e-
If the cathodic oxygen reduction reaction, which produces alkali, occurs at a
region removed from this anodic reaction the localized corrosion of the
aluminum will produce at. accumulation of acid. This acid destroys the
protective oxide film and produces an increase in the rate of attack. In
addition, the accumulation of a positive charge in solution will cause the
migration of C1- ions to achieve solution neutrality. This increased C1-
concentration can then further increase the rate of attack. This process is
illustrated schematically in Fig. 10. Since the oxygen concentration within
the pit is low, the cathodic oxygen-reduction reaction occurs at the mouth
of the pit, thus limiting its lateral growth.
Pitting attack can also be initiated by metallurgical inhomogeneities.
Magnesium alloys, for example, are very sensitive to the presence of iron
particles sometimes imbedded in the surface during rolling. In chloride
environments, these iron particles give rise to pits which have pinnacles
in their centers, the iron particles resting on the topmost points of the
pinnacles. In this case, each iron particle provides a preferred site for the
cathodic oxygen reduction reaction and the pinnacle is associated with the
outward spread of alkali formed by this reaction.
In most cases pits tend to be randomly distributed and of varying depth
and size. The evaluation of pitting damage is difficult and weight loss meas-
urements usually give no indication of the true extent of damage. Measure-
ments of average pit depth can also be misleading because it is the deepest
pit which causes failure. Maximum pit depth information is therefore the
most useful in estimating equipment service life.
Crevice Corrosion
This form of localized attack occurs when crevices or other partially
shielded areas are exposed to corrosive environments. Attack usually
arises because of differences in the concentration either of ions or of dis-
solved gas (for example, oxygen). As we have seen, this difference in
solution composition can result in differences in electrical potential even
though the metal may be of uniform composition throughout. In general,
the region deep within the crevice corrodes while the cathodic reaction
takes place at the mouth of the crevice, which is not attacked. As in the
case of pitting corrosion, crevice corrosion may be autocatalytic because the
hydrolysis of the metal ions being formed within the crevice can lead to high
OXYGENATED _
NoCl SOLUTION
02 + H20 --~
_CI -
Na +
CI-
__ Na +_ Na + --
CI- __
__ __ 02+ H20
FIG. lO--Schematic drawing illustrating the autocatalytic nature o f pitting attack on

aluminum in oxygenated sodium chloride solution.
acidic conditions. The accumulation of positive charge in the solution

within the crevice will also lead to an increased concentration of anions
and, especially in the case of chloride-containing solutions, this accumula-
tion can lead to more aggressive corrosion conditions. Because of this
increased aggressiveness, severe corrosion can often occur at creviced
areas even though surrounding, smooth, uncreviced areas remain relatively
unattacked.
In the case of metals such as stainless steel, which are normally protected
by passive films, crevice corrosion conditions can be particularly dangerous.
This is true because the conditions of oxygen depletion existing within the
crevice can result in the removal of the protective oxide film. As seen in
Table 2, a sample of stainless steel without its protective film is chemically
more reactive than one still covered by such a film. A corrosion cell will
then be set up between the active region of the crevice interior and the still
passive regions outside. It should be noted that crevice corrosion conditions
can be brought about if the metal is partially covered or shielded with
either nonmetallic material or foreign matter and it is not necessary for the
crevice to be entirely metallic. For example, an elastic band placed around
a specimen of stainless steel in seawater will initiate severe corrosive attack
in the crevice formed between the rubber and the steel.
Galvanic Corrosion
As we have seen, an electrical potential difference will usually exist
between two dissimilar metals exposed to a corrosive solution. When these
two metals are electrically connected the more active meta't will become the
anode in the resulting corrosion cell, and its corrosion rate will be increased.
The extent of this increase in corrosion rate will depend upon several
factors. A high resistance in the electrical connection between the dis-
similar metals, for example, will tend to decrease the rate of attack. On the
other hand if a large area of the more noble metal is connected to a smaller
specimen of the more active metal, attack of the more active metal will be
greatly accelerated. This acceleration occurs because, as discussed for the
case shown in Fig. 5, the larger cathodic surface will not polarize readily.
If oxygen reduction, for example, is the cathodic reaction, a large area of the
more noble metal will enable this cathodic reaction to proceed easily. A
classic example of this situation would be the use of steel rivets to hold
copper plates together. The large area of the more noble (cathodic) copper
would lead to the rapid corrosion of the more active (anodic) steel. The
reverse situation, the use of copper rivets in steel plates, is not as damaging
because the corrosion is dispersed over the relatively large anodic (steel)
area, and only a small cathodic (copper) surface is available. Hence the rate
of corrosion of the steel will be under cathodic control, and the situation will
be that illustrated in Fig. 7a.
The conductivity of the corrosive medium will also affect both the rate
and the distribution of galvanic attack. In solutions of high conductivity
the corrosion of the more active alloy will be dispersed over a relatively
large area. In solutions having a low conductivity, on the other hand, most
of the galvanic attack will occur near the point of electrical contact between
the dissimilar metals. This latter situation is usually the case, for example,
under atmospheric corrosion conditions.
Not all galvanic corrosion is detrimental. Zinc coatings are used to
protect steel not because the zinc is resistant to corrosion, but because the
zinc corrodes preferentially and hence cathodically protects the steel by
making any exposed areas of steel into local cathodes. Magnesium and
zinc, which are anodic to steel, when electrically connected to buried steel
pipe make this pipe the cathode in the resulting corrosion circuit. Only the
sacrificial magnesium or zinc anode undergoes corrosion. A further dis-
cussion of cathodic protection as a means of controlling corrosion damage

is given in Methods of Corrosion Prevention and Control of this chapter.
Selective Leaching
As its name implies, selective leaching involves the preferential corrosion
and removal of one or more electrochemically active elements from an
alloy, with the less reactive elements remaining behind. The most common
example of this form of attack is dezincification or the selective removal
of zinc from brass. This dezincification can be either uniform or localized
(plug type). In either case, what remains is a porous residue of essentially
pure copper having little or no mechanical strength. Susceptibility to
dezincification tends to decrease with decreasing zinc content, and brasses
containing less than about 15 weight percent zinc (for example, red brass)
are substantially immune. Improved resistance to dezincification can also
be achieved through alloying, principally with tin (~-~1~o), arsenic, phos-
phorus, or antimony (~0.04 %), which inhibit the selective leaching process.
Other alloys are also susceptible to selective leaching. Buried grey cast
iron piping, for example, can sometimes become "graphitized" through the
selective corrosion of iron, leaving behind a porous mass of graphite
particles. Since graphite is very cathodic relative to iron, a galvanic cor-
rosion cell is established. As in the case of dezincification, the remaining
graphite sponge possesses almost no strength, even though the pipe may
appear to be relatively unattacked and its dimensions substantially un-
changed. Graphitization does not occur in nodular cast iron since the
graphite particles are discrete and do not remain as a porous residue.
White cast iron, which has effectively no free carbon, is also immune.
Potentially, any alloy which consists of elements widely separated in
electrochemical activity may be susceptible to selective leaching. The silver
in gold-silver alloys, for example, can be removed almost completely by
corrosion in dilute nitric acid leaving behind essentially pure gold.
lntergranular Corrosion
In many corrosive media, grain boundaries are anodic to grain interiors.
In most situations, the reactivity of such boundaries is not great enough,
however, to lead to significantly increased damage. The term intergranular
corrosion is therefore usually reserved for those particular cases where
corrosive attack shows a high degree of localization at grain boundaries in
preference to grain interiors, leading to a substantial degradation in
mechanical or other properties. This type of attack can occur, for example,
in improperly heat-treated stainless steels which do not contain special
stabilizing alloying additions. The corrosion resistance of stainless steels
depends to a great degree on their chromium content. When non-stabilized
stainless steels are heated to between 900 and 1500 F, the precipitation of
chromium carbides can occur. Grain boundaries are preferred nucleation
University of Washington (University of Washington) pursuant to License Agreement. No further reproductio
22 INDUSTRIAL CORROSION STANDARDS AND CONTROLS
sites for the precipitation of these carbides, and their preferential formation
at these boundaries therefore locally depletes the chromium content of the
steel. Since the grain interiors still regain a high chromium content, they
remain protected. The chromium-depleted zones at the grain boundaries
will thus be small anodic areas electrically connected to large cathodic
areas, and severe intergranular attack will occur. It is important to note
that sensitizing heat-treatment of stainless steel, which produces damaging
grain boundary precipitates, can also occur during welding. In this case
there will be an area near the weld where the temperature conditions of the
welding operation cause grain boundary precipitation of chromium car-
bides. This precipitation will lead during exposure to corrosive environ-
ments to the formation of localized bands of severe intergranular attack
(weld decay). Such zones can be avoided if the material is reheat treated after
welding to redissolve the carbide precipitates, thus restoring the chromium
to the alloy. To combat this problem of intergranular corrosion, stainless
steels have been developed which either contain very little carbon or which
contain small additions of elements such as columbium and titanium which
are strong carbide formers. In either case the effective carbide content of the
steel is lowered. The lack of available carbon prevents the formation of
FIG. 11--An electronmicrograph showing precipitate free zones along a grain boundary
margin of a sample of AI-4 wtTo Cu aged 20 h at 200 C.
FIG. 12--An electronmicrograph showing selective corrosive attack along three grain
boundaries in a sample o f AI-4 wt% Cu aged 20 h at 200 C and exposed to aerated NaCI
solution.
grain boundary chromium precipitates and hence prevents preferred grain-

boundary attack.
Grain boundary precipitates can also lead to intergranular attack in other
alloys besides stainless steels. In A1-Cu alloys, the CuA12precipitate particles
can be formed preferentially at grain boundaries, along with concomitant
precipitate free zones along the margins of these boundaries, as shown in
Fig. 11. These CuA12 precipitates are strongly cathodic relative to pure
aluminum. Hence, the preferential formation of these precipitates at grain
boundaries can lead to selective corrosive attack as shown in Fig. 12. In the
case of A1-Zn-Mg alloys, similar preferred precipitation at grain boundaries
can also occur, as shown in Fig. 13. In this case, however, the MgZn2
precipitates are strongly anodic relative to aluminum and are selectively
attacked as shown in Fig. 14. In both of these cases involving aluminum
alloys, intergranular corrosion is not as severe as in the case of sensitized
stainless steels. However, when tensile stress is combined with this selective
attack, it is possible for greatly increased damage to result from stress
corrosion, as discussed next.
FIG. 13--An electronmicrograph showing the preferred formation of MgZn~ precipitates

along a grain boundary in a specimen of AI-7.5 wt ~ Zn-2.4 wt ~o Mg alloy aged 72 h at 100 C.
Stress Corrosion
When the combination of tensile stress and corrosion acting together
produces greater damage than either applied separately, stress corrosion is
said to occur. It is important to note that the tensile stress can either be
residual or externally applied. This form of corrosion damage is par-
ticularly dangerous because failure can be catastrophic and occur without
warning. In general, stress corrosion is highly localized and occurs in the
form of cracks. Particularly in the case of high strength aluminum alloys
exposed to chloride-containing environments, these stress-corrosion cracks
proceed preferentially along grain boundaries. In other cases, however,
such as austenitic stainless steels in chloride-containing environments,
cracking occurs transgranularly. In still other cases, particularly copper
base alloys, cracking can occur either transgranularly or intergranularly
depending on the environment.
Susceptibility to stress corrosion is generally measured by the time
required to produce fracture after a stressed specimen is exposed to the
corrosive environment, and higher tensile stresses produce failure in shorter
times than lower tensile stresses. For most susceptible alloys there is usually
a lower stress level below which failure does not occur. Other tests have
FIG. 14--An eleetronmierograph showing the selective attack o f MgZn2 precipitates in a

sample of.4l-7.5 wtTo Zn-2 wt~o M g aged 89 h at 100 C and exposed to an aerated NaCl
solution.
been devised to separate the effects of stress and corrosion in materials

which are susceptible to stress corrosion [12]. These tests have proved
useful in evaluating the effectiveness of such surface treatments as shot-
peening, which are used to increase resistance to stress corrosion [13]. In
alloys which crack intergranularly for example, it can be shown that a
substantial part of the protective effect of shot-peening arises because of
surface grain boundary disruption, as well as from residual stress effects.
Whether cracking is intergranular or transgranular, cracks tend to grow
in the plane normal to that of the residual or applied tensile stress. In this
plane, the stress concentration at the head of the growing crack will be
highest and crack growth will be fastest. The resistance of high strength
materials to such crack propagation and the influence of corrosive en-
vironments on this resistance, can be evaluated by means of precracked
specimens [14]. By increasing the load on a specimen of suitable dimensions
containing a crack of known size, the stress intensity factor which causes the
crack to become unstable and extend can be determined. This factor then
gives the fracture toughness of the material under the environmental condi-
tions of the test. Thus, stress corrosi6n processes clearly involve both elec-
trochemical and metallurgical factors, and it is likely that the specific way
in which corrosion processes and tensile stresses interact will depend

critically on the particular alloy system and environmental condition
involved.
Hydrogen Ernbrittlement
As was shown in Figs. 1 and 2, during corrosion under acid conditions
the reduction of hydrogen ions to hydrogen atoms occurs along with the
production of metallic ions. These nascent hydrogen atoms can either
combine to form hydrogen gas or, especially in the case of titanium and
steel alloys, they can diffuse as hydrogen atoms into the metal. Certain
substances, such as hydrogen sulfide, arsenic, or phosphorus compounds
tend to prevent the formation of molecular hydrogen from nascent hydro-
gen atoms. These compounds thus tend to increase the number of nascent
hydrogen atoms present on the metal surface and hence increase the fraction
of the total amount of hydrogen produced by corrosion which dissolves into
the metal. Applied cathodic current can also tend to encourage the accu-
mulation of dissolved atomic hydrogen in metals. In any case, if this atomic
hydrogen diffuses to internal voids it can form trapped pockets of hydrogen
gas. Since molecular hydrogen cannot redissolve in the metal, a pressure of
hydrogen gas is built up. These pressures can easily become great enough to
rupture and distort even the strongest steel (hydrogen blistering). Even
worse, in very high strength steels, the presence of dissolved hydrogen can
lead to greatly reduced metal ductility (hydrogen embrittlement) and
concomitant cracking. Similarly, in titanium, brittle titanium hydrides may
be formed from dissolved hydrogen. These hydrides can give rise to similar
embrittlement and cracking effects. The outward appearance of specimens
which have cracked through hydrogen embrittlement is often very similar
to that of samples which have broken through stress corrosion. Whereas,
however, applied cathodic current can slow down or prevent stress cor-
rosion, such cathodic currents will tend to increase hydrogen embrittlement
by increasing the rate of hydrogen reduction.
Because it is accelerated by the presence of dissolved H2S, hydrogen
embritflement is often a severe problem in sour oil fields. Plating operations
which are generally carried out using strongly acid conditions, can also
sometimes give rise to hydrogen embrittlement in steel parts if excessive
plating current is applied.
Erosion Corrosion
This form of corrosion involves the acceleration and possible localization
of attack due to the relative movement of a fluid environment and a metal
surface. As in the case of stress corrosion, both mechanical and corrosive
processes are involved. Especially susceptible metals are stainless steels and
aluminum which rely for their corrosion resistance on the presence of
highly protective surface films. The liquid impinging on the surface causes
a wearing away of the protective film, exposing new reactive sites which are
anodic and surrounded by a relatively large cathodic area. Rapid, localized
corrosion of the exposed regions can then occur. Most other metals besides
stainless steels and aluminum are also susceptible. As mentioned already,
the resistance of lead to sulfuric acid, for example, depends on the formation
of mixed lead oxide-lead sulfate surface films. In situations where lead is
exposed to turbulent dilute sulfuric acid, rapid corrosion attack can occur.
In stagnant solutions of the same concentration, corrosion attack is mini-
mal. Similarly, in desalination tube bundles, erosion corrosion may occur
near the inlet end of the tubes, in the region of turbulence where the high
velocity water first enters the tube bundle. Aluminum brass (by weight
percent, 22Zn - 2A1 - 0.065As-balance Cu) is more resistant than admi-
ralty metal (24Zn-0.65As-balance Cu) because the presence of A1 contrib-
utes to the development of a more protective and adherent surface film.
Similar effects are observed for the addition of Fe to cupro-nickel. Con-
versely, erosion corrosion can be accelerated if the moving fluid contains
abrasive particles. Erosion corrosion processes can also occur in gaseous,
organic, or even liquid metal environments as well as under more familiar
aqueous conditions. Both gaseous and liquid environments can combine to
produce erosion corrosion. In cavitation damage, for example, large
pressure changes and rapid fluid flow cause the repeated formation and
collapse of bubbles at metal surfaces, thus destroying protective surface
films and giving rise to concentrated localized attack.
Corrosion Fatigue
Normal fatigue is the process by which metals fail under repeated cyclic
stressing, at loads which are substantially below the normal strength of the
metal. The fatigue limit is the highest stress which can be cyclically applied
an indefinite number of times without causing fracture. Corrosion fatigue
may be defined as the combination of corrosion and normal fatigue proc-
esses leading to a reduction in fatigue resistance. This behavior is illustrated
in Fig. 15, which shows the relationship between the level of applied stress
and the number of cycles required to produce failure for steel. Under
corrosion conditions the stress level which can be tolerated for a given
number of cycles is everywhere reduced, and there no longer exists a lower
stress below which failure will never occur.
As in the case of stress corrosion, corrosion fatigue processes are not well
understood and can be expected to differ substantially from one alloy and
corrosive envi.'onment to another. In general, however, corrosion fatigue
damage can be expected to be large if the corrosive environment is one that
can cause pitting. Any pits which are produced by corrosion will act as
stress concentrators and thereby locally increase the effective applied cyclic
stages.
T
O3
O9
ON
hi
n,"
I--
O3
Q
I,t.I
_.1
n
)-
_1
_1
f..)
.J
(.3
)..
(.3
I I I I I i
10 2 iO s 104 i0 ~ i0 e 107 i0 o
NUMBER OF C Y C L E S REQUIRE TO
PRODUCE FAILURE )
FIG. 15--Schematic drawing showing the normal fatigue and corrosion fatigue behavior
of steel.
Fretting Corrosion
This form of damage is usually denoted by surface discoloration and
wear, as well as deep pits, in regions of slight relative (vibratory) move-
ment between highly loaded surfaces. In fretting corrosion the slipping
movements at the interface of the contacting surfaces destroy the con-
tinuity of protective surface layers, thus allowing relatively rapid attack to
occur. This form of damage may be especially damaging because of re-
sultant seizing and galling or loss of close tolerance in machine parts.
Materials such as stainless steel or titanium alloys which depend critically
on protective films for corrosion resistance are especially susceptible to
fretting corrosion damage. Surprisingly small relative movements can give
rise to fretting damage. Tomlinson, who first used the term fretting cor-
rosion, showed that vibratory motions of as little as 8 • 10-8 cm could
produce fretting damage [15,16].
In the case of the fretting corrosion of steel on steel, it has been shown
that only oxygen and not moisture is required to produce damage [17].
Also, the rate of damage is decreased by moisture, an effect first noticed
from the difference in weight loss observed for tests made during winter and
summer. An aqueous corrosion process is therefore apparently not in-
volved. Instead, damage results from the localized abrasion of metal to
form oxide with subsequent acceleration of damage due to both the greater
volume of the oxide (relative to the metal from which it formed) and the
abrasive nature of the oxide particles. In this case, the effect of water in
decreasing damage may be due to a lubrication effect. As might be ex-

pected, fretting damage can be decreased through the use of either solid or
liquid lubricants as well as by the use of soft metal or other coatings which
can exclude oxygen from the faying surfaces. Although the mechanism of
fretting damage is not entirely understood, it would appear to be more
related to low temperature oxidation than aqueous corrosion processes.
Other oxidation processes can lead to corrosion damage, particularly at
high temperature as discussed next.
High Temperature Oxidation

The direct combination of a metal with oxidizing agents such as sulfur
dioxide or oxygen is termed high temperature oxidation or, alternatively,
dry corrosion. The forms which such attack can take are in many cases the
same as those which occur under aqueous conditions at ambient tempera-
tures. That is, attack may be uniform or localized and produce a variety of
morphological features, including pits, preferred grain boundary attack,
and selective leaching.
In high temperature oxidation, the physical and electrical properties of
the corrosion product films that are formed determine the severity and
extent of attack. If, for example, the oxide which forms is cracked or spalls,
so that access of the oxidizing agent to the metal is unimpeded, then cor-
rosion will continue at a constant rate. In a very early investigation of
oxidation corrosion, Pilling and Bedworth proposed that oxide protective-
ness was linked to the ratio of the relative volume of oxide produced to
that of metal consumed [18]
Md
R=
mDa
where a is the number of metal atoms per oxide molecule, M and m are the
molecular weights of the oxide and metal, respectively, and D and d are
their densities. If this value is either less than unity or substantially greater
than unity, then the oxide will be unprotective. This is so because if R is less
than unity, insufficient oxide volume will be produced to give complete
coverage while the case of R greater than unity will give rise to cracking or
spalling. In either case, the gaseous oxygen can continue to react with the
metal surface as shown in Fig. 16a. In general there is only qualitative
agreement with the Pilling-Bedworth rule, since other factors are important
as well. As was aqueous corrosion, high temperature oxidation is an electro-
chemical process. That is, to form the oxide, metal atoms (M) must be
increased in oxidation state while some other species, for example, 02, is
reduced in oxidation state. That is, the two partial reactions may be
written as
M ~ M "+ + ne-
and
0
2 c
m
(a) (c)
t%
o
~ o
0
Z
GROWINGOX~Di'
Z
0
Z
0
OXIDE OXIDE o
~ M+,,,>20 - ~-02 (b) ,M%<- 2 02 (d)

~e-J
GROWINGOX~'~IDE
FIG. 16--Schematic drawings showing regions o f oxide growth in cases where (a) the oxide f o r m e d in non protective, (b) the oxide is protective and
metal ion diffusion is rapid, (c) the oxide is protective and oxygen ion diffusion is rapid and (d) the oxide is protective and oxygen and metal ion
diffusion occurs at almost the same rate.
n
402 + n e - ~ 0 -2
which combines to give
M - ~ ~ O2---~ M 0
Thus as in aqueous corrosion, high temperature oxidation consists of an

oxidation reaction occurring together with a reduction reaction. For these
reactions to proceed both ionic and electronic migration through the oxide
film is required. As shown in Fig. 16b, if the rate of oxygen ion diffusion
through the oxide film is limiting, then oxide growth occurs near the oxide-
environment interface. If, on the other hand, metal ion diffusion is slow,
then oxide growth occurs near the metal-oxide interface, as shown in
Fig. 16c. The reaction site may also be inside the oxide film if neither metal
nor oxygen diffusion is limiting (Fig. 16d).
In all cases except that shown in Fig. 16a, the rate of oxidation will depend
upon both the electronic as well as the ionic conductivity of the growing
oxide film. Since the time required for both electrons and ions to pass
through the film will be proportional to the film thickness, the rate of film
growth in such a case will be inversely proportional to film thickness. That
is, the mass of the oxide layer will increase as the square root of exposure
time (parabolic growth). If the oxide film does not conduct electrons, ionic
diffusion will be inhibited, leading to a slower growth rate and an oxide
weight which increases with the logarithm of exposure time. A similar slow
growth rate situation occurs if the oxide being formed conducts electrons
but not ions.
To be protective, an oxide should be nonvolatile and nonreactive with its
environment. At high temperature, the oxides which form on tungsten, for
example, evaporate as they are being formed and so oxidation continues
unchecked. Accelerated or catastrophic oxidation can also occur through
the interaction of an oxide scale with contaminants in the oxidizing en-
vironment. The presence of vanadium in oil, for example, can lead to
greatly increased oxidation rates for steel in contact with the flue gas pro-
duced when this oil is burned. V205 forms a low melting (635 C) eutectic
with Fe~Os, whose melting point normally is 1565 C. In addition V205 is a
catalyst for converting SOs to SO~ and this can result in the incorporation
of damaging sulfate ions into the growing oxide scale.
Environmental control, alloying, and protective coating have all been
used to decrease corrosion losses through oxidation. Furnaces using
molybdenum windings, for example, may be used to produce temperatures
up to 1500 C or higher provided these windings are protected by an at-
mosphere of hydrogen. Iron-chromium-aluminum alloys may be heated for
long periods in air at up to 1300 C whereas normal low carbon steel will
University of Washington (University of Washington) pursuant to License Agreement. No further reproductions autho
oxidize at a rate of more than ten mdd at a temperature of less than 1000 C
[19]. The use of many refractory metal alloys in high temperature applica-
tions would not be possible without the use of coatings such as fused
silicides
Bacteriological Influences
Several types of bacteria are known which can cause or accelerate
corrosive attack on metals. In anerobic soils a type of bacteria called
Desulphovibrio desulphuricans can reduce SO4= ions to S= ions, with the
release of oxygen. This oxygen, as we have seen, can accelerate the cathodic
reaction. Alternatively, the S= ions can react with Fe ++ ions, thus also
depolarizing the anodic reaction in iron or steel corrosion. In this way,
corrosive attack instead of being slowed by the anerobic condition can
continue apace. The resulting corrosion product, rather than rust, is black
ferrous sulfide. This form of attack can often occur beneath asphaltic
coatings on pipeline and is particularly dangerous since the outer asphalt
layer shields the pipe from the applied cathodic protection current while
also providing anerobic and sulfur-rich conditions. Bacteriologists have
identified many different species within the genus desulphovibrio, some of
which are strictly limited to salt water, and reviews of their behavior in
corrosion situations are available [20,21]. Another form of microbial
corrosion involves the fungus Cladosporium resinae, which has the ability to
degrade the hydrocarbon found in jet fuel. Growth is controlled mainly by
temperature and the availability of water. These fungi produce a wide
variety of organic acids as waste products and very acidic conditions can
develop beneath growing colonies. In addition a highly anerobic condition
is to be expected beneath such a colony, and can lead to oxygen concen-
tration cell corrosion. This form of corrosive attack has only come into
importance with the replacement of piston powered aircraft by jet aircraft,
since these fungi grow preferentially in kerosene as opposed to gasoline.
Methods of Corrosion Prevention and Control

In the previous sections the basic electrochemical principles which
determine corrosion processes have been outlined and a discussion given
of some of the specific forms which these processes can take. This section
now reviews the principal general methods which can be taken to decrease
or eliminate corrosion damage.
There are many different approaches to the prevention of corrosion.
Substitute materials may be considered in place of originally chosen alloys
which cannot withstand environmental effects. Alternatively, the environ-
ment may be made less aggressive through the use of inhibitors, excluded
entirely by means of paint or other coatings, or altered in pH, dissolved
air content, or state of agitation. Equipment design can also be changed to
minimize crevice formation, water accumulation or other features which
may aggravate corrosive damage. Electrochemical methods too are available

which can either prevent corrosion entirely or greatly reduce its rate. In
what follows, the general principles of these basic approaches to corrosion
control will be outlined as an introduction to the discussion in subsequent
chapters of the detailed application of such methods in industrial situations.
Protective Coatings
The use of protective coatings is probably the most common means used
for retarding or preventing corrosion damage. In general, such coatings can
be classified into one of three groups: (1) organic and paint coatings;
(2) metallic and nonmetallic inorganic coatings; and (3') chemical conversion
and anodic coatings.
Organic coatings are used primarily to protect metal parts, equipment,
and structures from corrosion in the atmosphere, soil, or water. Their
principal action is as physical barriers to the environment. They may
contain, in addition, however, active pigments or other ingredients which
affect surface pH or which cause surface passivation. Such coatings
include paints, varnishes, enamels, and lacquers, as well as dipped, sprayed,
or baked-on plastic, rubber, or bituminous materials. Organic coatings may
often contain volatile ingredients which act simply as solvents and diluents.
The service life of such coatings depends principally on the durability of the
coating material itself and the adherence of this coating to the surface to be
protected. This latter factor can in turn depend critically on the method of
application as well as on the preparation given to the metal surface before
application. Surfaces to be coated should, of course, be as free as possible
from dirt, grease, scale, and initial corrosion products.
It is always advantageous to understand the true causes of corrosive
action when taking corrective measures. In galvanic corrosion, for
example, the intuitive approach would call for coating the obviously
corroding surface. If this is done, however, the result will be to
stimulate localized corrosive action at any holidays or other disconti-
nuities which may exist in this coating. This stimulation of corrosion occurs
because coating only the more active (less noble) surface produces a large
cathode--small anode corrosion cell situation. Concomitant accelerated
attack is therefore produced on any residual exposed anodic sites. It would
be far better to coat both surfaces or alternatively only the cathodic (more
noble) surface. Coating the more noble metal surface cathodically limits
corrosive cell action and in addition slows the overall rate of attack since
the available cathodic corrosion current is distributed over a large anodic
area.
Many paint or other organic coatings systems consist of multiple coating
layers each of which possesses a specialized function. Primer coatings, for
example, usually provide adhesion'to the metal surface for subsequent
finish coatings. This adhesion may be improved by prior chemical or
anodic surface treatments. Aluminum may be given a thin adherent

phosphate coating, as described below, which can greatly improve the
adhesion of the primer coat. Another important function of the primer
coat is as a vehicle for corrosion inhibiting agents such as red lead (PbaO~),
or lead and zinc chromates. The function of the top coat is principally
decorative and the provision of a barrier to weather and sunlight.
Metallic coatings can be applied to both ferrous and nonferrous alloys
to give increased resistance to corrosion. Such coatings can be applied by
electroplating, chemical reduction, hot dipping, cladding, metal spraying,
mechanical plating or other methods. Regardless of the method of applica-
tion, a continuous metal coating will serve as a physical barrier to the
environment until it is penetrated by corrosion or mechanical damage.
When the base alloy is exposed, however, the galvanic relationship of the
coating and the base alloy will determine the subsequent degree of pro-
tection provided by the coating. Coatings which are anodic to the base alloy
will give protection by sacrificial corrosion. More noble coatings will
accelerate corrosive action of the base metal at nicks and other holidays by
providing a large cathodic surface. Despite this possibility of enhanced
localized attack, many metal coatings are applied to more anodic base
metals. In the case of magnesium alloys, for example, virtually all metal
coatings are more noble than the base metal. In determining whether a
coating will be anodic or cathodic to the base metal, the influence of the
environment cannot be neglected. The electromotive force series (Table 1)
shows iron to be more active than cadmium. In seawater (Table 2), how-
ever, cadmium is seen to be less noble than iron. In seawater, therefore, a
thin coating of cadmium will give protection to iron exposed through
small pores or abrasions. In the case of tin coatings on steel, similar effects
occur. In solutions of mineral salts, tin is cathodic to iron. In most fruit
acids (for example, citric) tin forms complex anions which lower the
effective tin concentration. This increases tin activity so that tin becomes
anodic to iron. Therefore, in fruit acids, pinholes in tin coatings on "tin
cans" do not undergo the concentrated attack they would in mineral salt
solutions. Instead such pinholes receive protection through the sacrificial
corrosion of the thin coating. The steel is thus protected from perforation.
Because of the increased corrosion which occurs at pores in coatings when
a more noble metal coating is used, such noble metal coatings are usually
substantially thicker than coatings of less noble materials for which minor
perforation is not critical.
Nonmetallic inorganic coatings can also be applied to metals for increased
corrosion and wear resistance as well as for decorative purposes. Porcelain
enamel coatings, for example, are alkali-alumina borosilicate glass finishes
fused to the metal surface at temperatures high enough to liquify the
inorganic coating material. Most such coatings are applied to sheet metal
for use in such applications as kitchen appliances. The corrosion resistances
University of Washington (University of Washington) pursuant to License Agreement. No further reproductions a
of such coatings is usually very high, but they lack ductility and can be
chipped or cracked in service. In inorganic coatings of higher melting point,
more refractory ceramic materials can be applied by flame spraying. In
flame spraying, the coating material is first melted in a high temperature
flame or electric arc and then sprayed in droplet form onto the surface to be
coated. Both metals and high temperature refractory ceramics such as
boron nitride and hafnium carbide, as well as combinations of such ma-
terials have been applied by flame spraying processes.
Numerous other methods for the application of both metallic and
ceramic coatings are available. In diffusion coating (pack cementation
coatings) the coating material, in the form of a volatile compound (usually
a halide) reacts at elevated temperature with the metal surface to be coated.
The halide decomposes, releasing the coating metal which subsequently
diffuses into the surface. Diffusion coating processes are often called by
specific names such as calorizing (aluminizing), chromizing, or siliconizing.
A recent review of these and other and inorganic coatings and processes is
available [22].
Protective coatings and coating treatments which serve as a base for
paint or other layers can also be formed on many metals and alloys by
chemical and anodic methods. Chemical coatings are also termed chemical
conversion coatings because the metal surface is converted to a nonmetallic
compound as a result of the treatment. For example, aluminum can be
given a phosphate conversion coating by exposure to an acidic soluble
phosphate salt solution which contains a complexing agent (for example,
F-) for aluminum. The aluminum metal at the surface will dissolve pro-
ducing A1+++ ions which are complexed to form AIF4- ions by the fluoride.
Concurrently, the increase of pH by the reduction of H + ions causes the
precipitation of a basic phosphate salt (such as zinc phosphate) at the
aluminum surface, where the pH is more basic than in the bulk of the
liquid. If it were not for the fluoride ions, the aluminum ions would pre-
cipitate as AI(OH)3 producing a poorly adherent layer. Careful control of
the pH of the coating bath is, therefore, necessary because if the pH is too
low, no film at all will form, and if too high, the film will not adhere due to
the presence of AI(OH)~. Such coatings are often used, in the case of
aluminum alloys, as a base for subsequent paint layers.
Black oxide coatings can be produced on steel and iron parts by dipping
in an aqueous alkaline bath heated to 200 to 300 F. Such black oxide coat-
ings are chiefly used as bases for the application of oils and waxes. Chro-
mate conversion coatings can also be produced on zinc, aluminum, cad-
mium, and magnesium alloys. In this case adherent films containing
hydrated chromium oxides are produced, which give improved corrosion
resistance in a variety of environments.
Anodicfilms may be formed on aluminum, magnesium, zinc, copper, tin,
zirconium, and niobium alloys by electrolytic oxidation. In the case of
University of Washington (University of Washington) pursuant to License Agreement. No further repr
aluminum, for example, anodized coatings are produced by making the

aluminum specimen the anode in a bath of sulfuric acid. The oxide film
that forms naturally on aluminum is only about 25 to 100 ,~ thick. When
aluminufn is made the anode in sulfuric acid solution, however, this oxide
film can be increased to many mils in thickness, if desired. Such anodically
formed films in sulfuric acid are porous but can be sealed by exposure to
hot water, dichromate, or other solutions. The oxide coating itself may be
clear or tinted and can also be dyed before sealing. Aluminum alloys can
vary widely in composition, and these compositional differences exert a
considerable influence on the anodized coating. In general, lower concen-
trations of alloying ingredients will give rise to more transparent oxide
coatings. Alloys which contain silicon tend to assume a gun metal shade
while manganese as an alloying ingredient produces a brownish color due
to the presence of manganese dioxide. Besides the use of sulfuric acid as an
anodizing bath, anodizing processes for aluminum involving chromic and
oxalic acids are also available. Still other baths, such as boric acid or
phosphate solutions, are used to produce nonporous, high electrical
resistance films for use in electrolytic capacitor applications. As might be
expected from the number of practical applications, the literature dealing
with the anodization is immense; a critical summary of much of this work
is, however, available [23].
Designing for Minimum Corrosion

It is often more economical to achieve increased protection against cor-
rosion by improved structure design than by alternate material selection or
the use of protective coatings. In spite of the almost limitless number of
specific conditions of materials, structural arrangements, and corrosive
environments which may arise, the thorough application of a relatively few
basic principles can usually lead to substantial decreases in corrosive losses.
For example, sump or other areas where moisture may be trapped in
contact with metal should be eliminated either throught he use of non-
reentrant designs or the incorporation of adequate drain holes. Crevices
are an almost inevitable part of most engineering structures, but their
detrimental influence on corrosion resistance can be minimized by provision
for their drainage as well as by welding or the use of proper joint com-
pounds, for example, red lead in the case of steel crevices. The use of
dissimilar metals in structures may also be dictated by economic necessity,
but these should be electrically isolated if the alloys involved are widely
separated in electrochemical activity. Small anodic areas connected to large
cathodic areas are especially to be avoided. Corners and surface contours
should be as rounded as feasible to avoid conditions where liquids or solids
can collect. In welded structures, butt-welded joints should be given
preference over lap joints where possible, and the weld metal should be
slightly more noble than the base metal. In systems involving liquid trans-
port, turbulent flow and gas or air entrapment should be minimized.

Designs which minimize stress concentrations can be critical in avoiding
possible catastrophic failure through stress corrosion or corrosion fatigue
processes. In almost all eases intelligent design coupled with a basic
knowledge of corrosion processes can avoid or minimize many subsequent
corrosion difficulties.
Electrochemical Protection Methods

As we have seen, the corrosion of a metal requires the transfer of elec-
trons. Thus, when two dissimilar metals are electrically connected as shown
in Fig. 2, the more reactive metal passes into solution (becomes an anode),
and the electrons thus produced travel to the less reactive metal where they
are used up in the reduction of some substance in the electrolyte. If the
external electrical connection was broken, then both metals would begin to
corrode, and both anodic and cathodic processes would occur over the
surfaces of each. Thus, the electrical connection between a less noble and
more noble metal leads to the decreased corrosion (that is, the protection)
of the more noble sample at the expense of the increased corrosion of the
less noble metal, which becomes a sacrificial anode. This effect is the basis of
cathodic protection. If the ability of the less noble metal to supply electrons
is sufficiently great, then the corrosion of the less noble metal can be entirely
prevented. Instead of a less noble and a more noble metal electrically con-
nected together, an external direct current source can also be used to achieve
the same effect, as illustrated in Fig. 16. Here an external source of current
supplies electrons through an auxiliary anode to the buried pipe, shifting its
potential and preventing corrosion. These effects can be readily visualized by
considering Fig. 18. If no external current is applied, the submerged or
buried metal will have a potential Ez and be corroding at a rate given by
/~ .... If an external current of a magnitude I~ -- Lorr is applied (with the
polarity shown in Fig. 17), the potential of the corroding metal will shift to
less noble potentials, and the rate of the anodic (corrosion) reaction will
decrease to Icorr'. If a larger current, Iu, is applied the potential wilt shift to
Ea and the anodic (corrosion) reaction will be suppressed entirely. The
buried or submerged metal then cannot corrode as long as the external
current is kept applied. Even larger applied cathodic currents will not be of
benefit and may be harmful to amphoteric metals through the production of
high alkalinity at the metal surface. Where possible, cathodic protection is
usually combined with the use of protective coatings in order to decrease the
magnitude of the applied current density that must be maintained in order
to achieve complete protection.
The preferential use of either sacrificial anodes or an external current
source will depend on the details of each given cathodic protection applica-
tion. Sacrificial anodes such as special magnesium, zinc, or aluminum
alloys are commonly employed in the cathodic protection of ships and
~ J DIRECT CURRENT ! "~-
I
m
_SO,L OR O~-ER - - - - ~ ! ] ,.OOE I ~ - - :
ELECTROLYTE '~ . . : . ~ ' ~ . ~ .
PIPE(CATHODE) /
FIG. 17--Schematic diagram of the cathodic protection of a buried pipe.
Ec
I
..I
_ Ez . . . . . . . . . . . Z
I--
Z
bJ
0
O-
Ex -~.2"- ..... !. . . . . ~
9 I i \
Eo ___~ . . . . . . . . . . . . . L . . . .
t
\ v
L
I i
I
I
I I
l I
l~orr ' Ir Ix ly
CURRENT~
FIG. 18--Evans diagram illustrating the potential changes that occur during cathodic
protection.
buoys while impressed current methods are widely used to protect long
underground oil, gas, and water pipelines. In general, structures can only be
effectively cathodically protected if they are in open contact with an electro-
lyte. Thus, cathodic protection cannot readily be applied to the protection
of the interior of heat exchanger tubes, because of the high electrical re-
sistance of the electrolyte path, between the auxiliary anode and the
interior walls.
Anodic protection is a relatively new electrochemical protection method
which is based on the formation of protective films on certain metals by an
externally applied anodic current. It is only applicable to corrosion situa-
tions where passivity can occur, as shown for example in Fig. 8 in the case of
iron in nitric acid. Initially, with increasing potential the corrosion rate of
the iron increases. Beyond a certain critical potential, however, the corro-
sion rate falls drastically due to the development on the iron of a protective,
passive film. The applied anodic current density required to initiate passivity
may be quite high (mA/cm~). Once the passive state is achieved, however,
the current densities required to maintain protection can be extremely low
(uA/emZ). Since potential is the critical factor in anodic protection, a
reference electrode system is required along with an electronic device
(potentiostat) which can maintain the desired potential constant by auto-
matically increasing or decreasing the applied anodic current. Anodic pro-
tection is not as generally applicable as cathodic protection, because of the
special behavior required of the metal to be protected. Its low current re-
quirements, however, can give very great economic benefits in those
situations, such as stainless steel in H2SO4, where the required passive
behavior is observed.
Corrosion lnhibitors
Inhibitors may be defined as substances which slow down the rate of
corrosion reactions when added in relatively small quantities to the corro-
sive environment. Some inhibitors are effective because they form a pro-
tective deposit on the corroding metal. This deposit may increase the
effective electrical resistance of the corrosive environment as well as prevent
the access of the environment, particularly dissolved oxygen, to the metal
surface. The action of such inhibitors in decreasing the corrosion current
(rate) is similar to that shown in Fig. 7c. Other inhibitors may be termed
anodic or cathodic, according to whether they directly affect the anodic or
the cathodic corrosion reactions. Their effect on decreasing corrosion
current (rate), would be similar to that shown in Figs. 7a and 7b, respec-
tively. If an insufficient amount of an anodic corrosion inhibitor is added,
the effect may be to make corrosion worse rather than better. This can
happen, because, by decreasing the ratio of anodic to cathodic areas, cor-
rosive attack is concentrated at the remaining uninhibited anodic regions.
Cathodic inhibitors are safer to use because if added in a quantity insuffi-
3,0 iNDUSTRIAL CORROSION STANDARDS AND CONTROL
cient to achieve complete protection, acceleration of localized corrosive

attack does not occur. A reduction in the rate of attack can also be achieved
in some systems by means of passivating inhibitors or passivators. As shown
in Fig. 8 the corrosion rate of iron increases with increasing potential or
oxidizing power of the corrosive environment, until a certain critical value
is reached, above which the corrosion rate falls drastically to low levels.
Other metals and alloys, particularly stainless steels, can behave in a similar
fashion. Passivators act, in effect, to drive the potential of the corroding
specimen into the range in which passivity occurs as well as by adsorbing
onto the surface and making the onset of passivity easier to achieve. Passi-
vators such as chromate and nitrate anions, since they can be reduced, can
passivate steel even in the absence of dissolved oxygen while phosphates
and molybdates are only effective in the presence of oxygen.
Slushing compounds consist of greases, oils, or waxes which contain
organic polar compounds (such as amines) which adsorb onto the surface
of the metal to be protected. The adsorbed layers block the effective acccess
of the environment to the metal surface and help prevent metal dissolution
as well. Similar adsorbed layers account also for the action of inhibitors
used to prevent the corrosion of metal during pickling operations aimed at
removing mill or boiler scale. Typical pickling inhibitors for steel include
formaldehyde, propyl sulfide, and diamyl amine.
Volatile corrosion inhibitors are substances which sublime slowly at
normal temperature and inhibit atmospheric corrosion. Volatile corrosion
inhibitors have been principally developed to protect steel parts from rust-
ing during shipment and storage. They may have a detrimental effect,
however, particularly on zinc, cadmium, magnesium, and lead alloys
(including solders).
The concentration of inhibitor required to achieve a given level of pro-
tection is usually increased by the presence of chloride ions. Other dissolved
species can also affect the action of inhibitors, and a full knowledge of the
chemistry of the corrosive environment is necessary to insure the desired
effectiveness of inhibitors.
It has been possible here to summarize only in brief terms the methods
and techniques available for the control of corrosive damage. In the chap-
ters which follow, the specific measures currently used to minimize corrosion
losses in major industries are reviewed in depth. In each case particular
emphasis is placed on standard test and evaluation procedures which have
been developed to identify as well as to control corrosion damage.
References
[1] Webster's Third New International Dictionary, G and C Merriam Co., Springfield,
Mass., 1966, p. 512.
[2] Lichtenstein, S., "The Many Faces of Corrosion," National Bureau of Standards
Report STR-3454, U.S. Dept. of Commerce, Washington, D.C., Oct. 1966.
[3] Evans, U. R. in Corrosion Handbook, H. H. Uhlig, Ed., Wiley, New York, 1948,
pp. 3-10.
[4] Latimer, W. M., The Oxidation States of the Elements and Their Potentials In Aqueous
Solutions, 2nd ed., Prentice-Hall, Englewood Cliffs, N J., 1952, pp. 339-345.
[5] Evans, U. R., Journal of the Franklin Institute, Vol. 208, 1929, pp. 45-58.
[6] Evans, U. R., The Corrosion and Oxidation of Metals." Scientific Principles and Practical
Applications, Edward Arnold, London, 1960.
[7] Menzies, A. in Corrosion and Protection of Metals, American Elsevier, New York,
1965, pp. 1-36.
[8] Fontana, M. G. and Greene, N. D., Corrosion Engineering, McGraw-Hill, New York,
1967.
[9] Uhlig, H. H., Corrosion and Corrosion Control, 2nd ed., Wiley, New York, 1971,
pp. 60-9l.
[10] Tomashov, N. D. and Chernova, G. P., Passivity and Protection of Metals Against
Corrosion, Plenum, New York, 1967.
[11] Pourbaix, M., Atlas of Electrochemical Equilibria in Aqueous Solutions, Pergamon
Press, New York, 1966.
[12] Cocks, F. H., Materials Research and Standards, MTRSA, Vol. 9, No. 12, 1969, pp.
29-32.
[13] Cocks, F. H. and Bradspies, J., Corrosion, Vol. 28, No. 5, 1972, pp. 192-194.
[14] Brown, W. F., Jr. and Srawley, J. E., Plane Strain Crack Toughness Testing of High
Strength Metallic Materials, ASTM STP 410, American Society for Testing and
Materials, Philadelphia, 1967.
[15] Tomlinson, G. A., Proceedings of the Royal Society, Series A, Vol. 115, t927, p. 472.
[16] Tomlinson, G. A., Thorpe P. L., and Gough, H. J., Proceedings of the Institute of
Mechanical Engineering. Vol. 141, 1939, p. 223.
[L7] Feng, l-Ming and Uhlig, H. H., Journal of Applied Mechanics, Vol. 21, 1954, pp. 395-
400.
[18] Pilling, N. B. and Bedworth, R. E., Jr., Journal of the Institute of Metals, Vol. 29, 1923,
pp. 529-591.
[19] Austin, J. B. and Gurray, R. W. in Corrosion Handbook, H. H. Uhlig, Ed., John Wiley,
New York, 1948, pp. 630-638.
[20] Miller, J. D. A. and Tiller, A. K. in Microbial Aspects of Metallugry, J. D. A. Miller,
Ed., American Elsevier, New York, 1970, pp. 61-106.
[21] Foley, R. T., Electrochemical Technology, Vol. 5, No. 3-4, 1967, pp. 72-74.
[22] Murphy, J. A., Ed., Surface Preparation and Finishes For Metals, McGraw-Hill, New
York, 1971.
[23] Young, L., Anodic Oxide Films, Academic Press, New York, 1961.
STP534-EB/Nov. 1973
Chapter 2
Corrosion Standards and Control in the

Petroleum Industry
A. S. Couper 1
Petroleum refineries use water, steam, hydrogen, acids, and a variety of

other corrosive materials to convert crude oil into many different products,
including gasolines, lubricating oils, fuel oils, and chemical plant feed-
stocks. Hence, refineries have not only most of the corrosion problems
encountered in the utilities and chemical industries, but a few specific
problems as well.
A typical refinery consists of many large interconnected units that operate
continuously for up to five years without being shut down for maintenance.
Some of the units operate at such high pressures and temperatures that
unexpected corrosion failures can be dangerous as well as expensive.
Overall, corrosion is estimated to cost from 10 to 19 cents per 42-gallon
barrel of crude processed [1]. 2 In the United States, where refineries process
about 15 million barrels per day, these costs amount to at least 1.5 million
dollars per day. Worldwide, to process 42 million barrels/day the cost is
about 4.2 million dollars/day, distributed among alloying, maintenance,
inhibitors, inspection, equipment replacement, special safety precautions,
and added insurance payments because of uncertainties in predicting
corrosion. Such costs are inevitable, so a major objective of corrosion
control is to incur them in the least expensive way [2].
Selection of adequate materials of construction, adjustments in process
conditions, and addition of inhibitors are aimed at ensuring that each unit
will operate safely for much longer than its scheduled run. Materials are
selected on the basis of economic evaluations of the available alternatives
[3]. Such evaluations generally dictate a minimum life of ten years for each
piece of equipment, with the acceptable rate of corrosion dependent on the
process conditions involved. For example, a corrosion rate of 50 mils
(0.050 in.) per year may be acceptable for a low-pressure vessel that has
1 Research and Development Department, Amoco Oil Company, Whiting, Ind.

2 Italic numbers in brackets refer to references listed at end of this chapter.
,12
CORROSION IN THE PETROLEUM INDUSTRY 43
extra thick walls as a corrosion allowance, whereas the seating surfaces of

process control valves must have essentially no corrosion.
Refining conditions are tending to become more severe. The current need
to eliminate sulfur from most petroleum products by high-temperature,
high-pressure hydrogenation is an example. Consequently, work is con-
tinually under way to identify the major sources of corrosion and to
develop standards and procedures to eliminate them or to minimize their
effects. This chapter summarizes the present state of the art.
Sources of Corrosion
Corrosion in refineries results from a combination of sources related to
the composition of the crude oils and to the types of materials and operat-
ing conditions required in the various refining processes. Although the
thousands of available crude oils are predominantly noncorrosive hydro-
carbons, they also contain varying amounts of potentially corrosive sulfur
compounds, salts, water, acids, and oxygen. Figure 1 is a simplified flow
diagram of a refinery illustrating most of the corrosives and the process
units in which they can cause damage. Because each refinery has different
crude supplies and different processing schemes, the relative significance
of each corrodent varies. However, some common corrosion principles and
corresponding control techniques have been developed.
H2~
S FUEL
H~ -- 0 GASOUNE
HC~ H!S 0 JET ~UEE

KEROSENE
CO2
CRUDE
~) CRUDE OIL > H2~S GAS NAPHTHA H2
HEATING OIL
NAPHTHENIC
ACIDS
&
LUBRICATING OIL
I~ GREASE
NAPHTHA SO2 H2504 WAX
I~ COKE
I~ INDUSTRIAL FUEL
1-- I~ ASPHALT
= ORGANIC SULFUR COMPOUNDS AND H2S ABOVE ABOUT 500~
FIG. 1--Major corrodents in a petroleum refinery.
4~. INDUSTRIALCORROSION STANDARDS AND CONTROL
CRUDE NEUTRALIZER .'L

INHIBITOR CONDENSER
(T~
I~
WATER
I
PREHEATER
P[ROBE
--
WATER
FURNACE
~'1~ NAPHTHA
I
DESALTER
~ RIMARY
WATER _TOWER
,ll STEAM
TOVACUUM
DISTILLATION
FIG. 2--Typical crude distillation unit.
Crude distillation, the first step in refining, is outlined in Fig. 2. Most

crudes are washed with water to remove salts and possibly some acids. If
salts, particularly CaC12 and MgC12, are left in the crude, they will hydrolyze
to HC1 gas in the furnace that supplies heat for the distillation [4]. This
HC1, plus water vapor, accompanies the naphtha vapors emerging from the
top of the distillation tower. On condensation, the HC1 forms highly
corrosive hydrochloric acid with the water from the crude and from the
steam added to aid the distillation. Similar corrosive vapors and acid
condensates containing carbon dioxide, sulfur dioxide, and hydrogen
sulfide are encountered in other process unit distillation towers.
Sulfur compounds are especially troublesome because they not only
cause corrosion but also can poison some of the catalysts used in refining.
Hydrogen sulfide is sometimes a dissolved component of the crude oil
received at the refinery. In addition, it is formed in distillation furnaces, in
hydrodesulfurization units, and in hydrocrackers by decomposition of
organic sulfur compounds that are also present in the crude. Both the
organic sulfur [5] and the hydrogen sulfide [6] are corrosive to carbon steel
above 500 F. For sulfur removal, the products or the crudes are treated
catalytically with high-temperature, high-pressure hydrogen. The mixture
of hydrogen sulfide and hydrogen, which is becoming more prevalent as
pollution controls limit product sulfur levels, requires special precautions
in alloy selection.
At high temperatures, high-pressure hydrogen diffuses into metals and can
react with the carbon in steel to form methane. This reaction causes fissur-
ing and blistering as well as decarburization of the steel, with a consequent

loss in ductility. Then catastrophic failure can occur without warning and
with no visible deterioration of the metal.
Hydrogen is also produced by aqueous acid corrosion at low tempera-
tures. In this case, hydrogen atoms or ions form at the metal surface,
rapidly diffuse inward, and then form hydrogen molecules at defects or
dislocations within the metal. Because these molecules cannot diffuse out o f
the metal at low temperatures, pressures of several million psi can build up
and cause cracks, blisters, and serious ruptures [7]. Figure 3 is an example
of ruptured blisters on the inside walls of the tower of a vapor recovery
unit. Such damage is particularly severe in the fractionation section of fluid
catalytic cracking units or vapor recovery units where there are cyanides and
hydrogen sulfide, both of which accelerate the diffusion of hydrogen into
steel. Welds, which are often significantly stronger than the base metal, are
especially susceptible to cracking from this hydrogen.
The strong acids and bases required in many of the refining processes
can be particularly corrosive. These include the sulfuric and hydrofluoric
acids used as alkylation catalysts, the aluminum chloride plus hydrochloric
acid used as both an alkylation and an isomerization catalyst, the caustic
soda used to remove entrained acids from products, and the several bases
such as monoethanolamine and potassium carbonate used to remove
hydrogen sulfide and carbon dioxide from fuel gases. Although the bases
FIG. 3--Ruptured hydrogen blistem on the inside wall of a catalytic cracker absorber tower.
alone are noncorrosive, the dissolution and release of the acid gases can
cause severe corrosion, hydrogen blistering, and stress corrosion cracking.
The amount of acid gases dissolving in the water that condenses on storage
tank roofs or separates to the bottom of tanks usually determines the life
of such vessels and the amount of maintenance required.
Naphthenic acids cause the type of corrosion shown in Fig. 4, but the
source is often difficult to identify because the characteristic sharp-edged
pits and grooves are sometimes associated with severe high-temperature
sulfide corrosion [8]. Naphthenic acids are initially present in most crudes
at very low concentrations. However, in the fractionation process, suffi-
ciently high local concentrations of the acids can occur at 450 to 650 F to
corrode stainless steels that normally resist sulfide corrosion. Because the
metal naphthenates are soluble in oil, the scale on the corroded metal is
iron sulfide, which further complicates diagnosis. There is no clear con-
sensus on what constitutes a dangerous concentration of naphthenic acids,
but Derungs [9] suggests that corrosion will occur if the neutralization.
number is above 0.5 mg K O H / g of crude.
Perhaps the most pervasive sources of corrosion are the large amounts o f
water and steam required in refineries. Cooling water, both once-through
and circulating, is generally several times the volume of crude processed,
and is responsible for both corrosion and scaling of piping and exchangers.
FIG. 4--Naphthenic acM corrosion o f Type 410SS vacuum tower internals after 1 year at
about 700 F.
Coastal refineries face the added problems of sea water corrosion in coolers.
Steam is used to heat process streams, to remove gases from hydrocarbon
liquids, and to improve distillations. Low-pressure steam is usually adequate
for refinery process, but often high-pressure steam is also produced to
operate turbine-generators for electricity. The discharge steam goes to the
process units. Because very little steam condensate is returned to the
boilers, chemical treatment of the boiler feed water is more critical than in
many utilities. In addition, refineries burn fuels containing many by-
product oils that are too dirty or too corrosive to sell. Thus, furnaces and
flue gas venting systems require more alloying or chemical treatment than
typical steam generating plants.
Setting Corrosion Standards

The petroleum refining industry cooperates with several organizations
within which corrosion problems are discussed and classified to develop
acceptable standards and recommended practices. Composition and
physical properties of construction materials, as well as laboratory tests to
evaluate metals, refinery products, corrodents, and fabrication techniques
are defined to take advantage of the collective experience of many indi-
viduals and companies. The standards and recommended practices which
have the most influence on refinery corrosion control are listed in the
Appendix to this chapter.
The American Petroleum Institute (API) is a trade organization with
many working committees charged with establishing suitable standards and
recommended practices for all aspects of the petroleum industry. The
physical and chemical properties of most materials of construction are
defined in the API standards, and data for these standards are sometimes
developed by API-sponsored research programs such as the study on high-
temperature hydrogen attack at the University of Wisconsin [10]. Estab-
lishing and monitoring these research programs is primarily the responsi-
bility of the Division of Refining Committee on Refinery Equipment with
its subcommittees on Corrosion, Pressure Vessels and Tanks, Piping, and
Refinery Inspection Supervisors. The Inspection Supervisors Committee
is also responsible for an excellent inspection manual with a comprehensive
chapter on refinery corrosion [11]. The Division of Refining Mid-Year
Meeting in May has a number of symposia that include papers on the
latest research in refinery corrosion. Those papers are then published in the
meeting proceedings. API standards and publications are made available
to everyone at a nominal price.
The National Association of Corrosion Engineers (NACE) has a number
of technical committees concerned with petroleum corrosion problems.
These include: NACE T-l, Corrosion Control in Petroleum Production;
NACE T-5B, High Temperature Corrosion; and NACE T-8, Refining
Industry Corrosion. In recent years, these committees have also produced
48 iNDUSTRIALCORROSION STANDARDS AND CONTROl.
a number of significant standards and recommended practices. Several

more are in preparation. NACE also sponsors symposia on refinery cor-
rosion, and the papers are published in its magazines, Corrosion and
Materials Protection.
The American Society of Mechanical Engineers (ASME) Boiler and
Pressure Vessel Committee issues "codes" which define safe operating
conditions for materials. These codes apply to most refinery equipment
and must be considered in materials selection. New uses for materials and
new materials are evaluated by the committee. Specific recommendations
are published as "cases" which provide guidance for similar applications
elsewhere. Several unexpected failures have helped spur research programs
on the physical properties of materials sponsored jointly by ASME, the
Materials Properties Council, and the American Society for Testing and
Materials (ASTM).
ASTM is a comprehensive organization that covers all major industries
and establishes many standard tests, including some to define the quality of
both construction materials and refinery oil products. Standards of sand-
blasting for plant preparations have been defined and published, complete
with a portfolio of color photographs, by ASTM, NACE, and the Steel
Structures Painting Council (SSPC). ASTM also issues specifications for
paint evaluation in laboratory tests using the Salt Spray Cabinet and
Weatherometer, which are defined in ASTM Standards B 117-64 and
E 42-651, respectively. ASTM G 4-68, Recommended Practices for Con-
ducting Plant Corrosion Tests, is a basic guide for developing useful
corrosion control information.
Corrosion Testing and Control

Chemical analysis of process streams and metals play an important role
in corrosion control. An API survey lists 46 tests [12]. Some are general,
such as pH and the composition of process, cooling, or boiler-feed waters,
several are specific to petroleum refining.
Corrosion is measured in the laboratory and in operating units by
coupons and electrical resistance corrosion probes. Principles for corrosion
coupon exposures are outlined in ASTM (3 4-68. Corrosion probes and
other on-stream measurement techniques including ultrasonics [13],
radiography, and sentry holes are discussed in the Handbook on Corrosion
Testing and Evaluation [14]. Special couplants and instruments have been
developed to permit ultrasonic measurements at elevated temperatures and
to allow one man to make measurements at remote locations [15]. Linear
polarization techniques have also been attempted [16], but these require
careful interpretation because oil films on the electrodes can cause spurious
readings.
CORROSION iN THE PETROLEUM INDUSTRY 49
Crude Oils
Refiners can sometimes choose from many crudes. Therefore, a crude
corrosivity test is valuable to ensure that the refinery will not suffer excess
corrosion from a new crude. Crude oils are analyzed primarily for salt and
sulfur contents to estimate corrosivity. Because the correlations with chemi-
cal compositions have been relatively poor, tests have been devised to
correlate corrosion with HC1 and H2S evolution when the crudes are heated
at a controlled rate. Overhead corrosion in the crude fractionator correlates
quite well with HC1 evolution when the crude is heated to 650 F [17]. How-
ever, the test is extremely sensitive to the heating schedule.
The evolution of H2S from crudes heated rapidly to 850 F correlates well
with high-temperature sulfidic corrosion in crude units [18]. A similar test
measuring the cumulative H2S release when the crude is heated at 2 deg F /
rain [19] gives an overall estimate of relative corrosivity and also indicates
at which parts of the heating equipment corrosion is most likely to occur.
Several companies have constructed pressure equipment and pilot plants
to evaluate crude corrosivity directly by using coupons or electrical-
resistance corrosion probes. A simple pressure bomb with a 1-mil thick
carbon steel probe immersed in the crude has shown a good correlation
with experience and confirmed that sulfur content alone can be very mis-
leading as a corrosivity index for some crudes [18]. Figure 5 shows a more
elaborate pilot plant [20] that simulates continuous crude distillation. This
unit measures corrosivity at several locations and can be used to evaluate
the effects of special additives and crude pretreatments to minimize cor-
rosivity. It also helps select the chromium steel required to resist the sulfur
compounds in a particular crude. Another pilot plant has been used to
show that unexpected crude corrosivity can be due primarily to mercaptan
sulfur in the crude [21].
Process conditions can be modified with alkali to prevent acid corrosion.
Caustic added to crude oils before and after desalting helps suppress
chloride hydrolysis and reduces corrosivity by making the hot, salt-laden
desalter water alkaline. Caustic may also remove some of the naphthenic
acids. Ammonium hydroxide, ammonia, monoethanolamine, diethanol-
amine, and morpholine are added to the overhead vapors from distillation
towers to ensure that any condensed water has a pH of 5 or more. This
treatment reduces corrosion and provides conditions under which corrosion
inhibitors can completely stop corrosion provided the salts of the neutrali-
zation reaction do not deposit in the overhead condensers. The amount of
neutralizer must be carefully controlled to avoid caustic stress cracking of
steels and ammonia or amine stress corrosion cracking of copper alloys [4].
Refinery corrosion inhibitors are mostly solid, high-molecular-weight
nitrogen compounds (amides, amines, imidazolines and their salts with
fatty acids) dissolved in hydrocarbon solvents. They also are generally
COUPONHOLDERS
mmmm
||
I ....
yioNHos FEEDTANK
LEVELCONTROL
CAPACITANCE
SWITCH
COOLERI LEVELCONTROL
SOLENOIDVALVE
FIG. 5--Crude unit pilot plant.
surfactants and function by forming protective films on metal surfaces.

Because they are solids and because concentrated inhibitor solutions are
corrosive above about 250 F, special techniques are required to physically
disperse them into the mostly vaporized process streams. Inhibitor solutions
are diluted with a reflux stream to spray about 5 to 20 ppm by volume of the
inhibitor into the vapor [22]. The efficiency of the joint efforts of desalting,
neutralizing, and injecting corrosion inhibitor can be monitored with cor-
rosion probes or with analysis of any water condensate for dissolved
metals, particularly iron and copper [4].
Corrosion control can often be attained by simple changes in operating
conditions with little or no economic penalty. For example, the acid gases
(HCI, H2S, CO2, and SO2) are not corrosive at moderate temperatures pro-
vided there is no aqueous condensation. Consequently, raising the gas
temperature, lowering pressure, and installing insulation or steam tracing
equipment often can eliminate the danger that these gases will cause severe
corrosion or stress corrosion cracking. In equipment with little or no process
CORROSION IN THE PETROLEUMINDUSTRY 51
flow such as relief valves, bypass lines, and instrument lines, the accumula-
tion of corrosive gases can be prevented by a small flow (bleed) of hot
process gas or a dry purge gas.
Heat Transfer
Heat transfer has a major effect on corrosion by refinery processes during
both heating and cooling. Pilot plants often have low alloy heater tubes to
measure the effects of H2S release on corrosion. Simpler devices with cor-
rosion coupons heated with soldering irons [23] or small tubes heated with
small cartridge heaters are used for testing environments at moderate
temperatures and pressures.
The tubing tester shown in Fig. 6 is particularly simple and useful for
evaluating heat transfer problems in cooling water or in refinery process
streams. Scaling tendencies of waters and the conditions that may cause
dezincification of admiralty metal (70Cu, 29Zn 1Sn) can be evaluated
reproducibly with this new technique [24]. An electrical cartridge heater
inside the tube supplies the heat. Metal skin temperatures are adjusted by
controlling the heat input, the size of the annular space between the tube
and the metal cylinder, and the thickness of the partially insulating gauze.
The cylinder can also be used as an auxiliary electrode for electrochemical
studies.
An older test involves circulating a hot fluid through a piece of tubing
and noting the effects of such exposure [25]. This test has been used to
evaluate cooling water treatments [26], but it requires a cumbersome test
circuit with several automatic controls. Similar devices, with cooling fluids
in the tubes have been used to evaluate the corrosivities of condensates in
crude unit distillation overhead condensers [27] and in power station flue
gases. A somewhat more elaborate device uses a polished steel surface to
RECORDER J
I POTENTIOMETERI
CARTR,DGEHEATER / I I MANUAL I
/] I POTENTIOMETER J
/ ] ~ SATURATED
;:~...:.:.:<.!~ ~,,e,,,~-,~' ~ . . . CALOMELELECTRODE
~i! I V I HOT TUBE SPECIMEN
~ METAL CYLINDER
i I
FIG. 6---Heated tube corrosion tester.
evaluate "slushing compounds" that prevent corrosion from the con-

densation of moisture on steels in storage [28].
A standard heat exchanger to evaluate corrosion and fouling of steel
heat transfer surfaces has been proposed by NACE T-5C [29]. Water
passes at a controlled rate through the tubes while steam at a controlled
pressure on the shell provides heat and the desired metal skin temperatures.
The tubes are split after about a 30-day test and measurements made for
both corrosion and scale deposits.
High-temperature sulfidic corrosion is extremely sensitive to temperature
and so can be controlled to some extent by reducing distillation pressures
and metal skin temperatures in furnaces. Metal skin temperatures depend
both on heat distribution in the furnace and on the heat input required for
the distillation or reaction. Proper furnace design, appropriate burner
settings, and efficient combusion can eliminate most hot spots. Heat
recovery is crucial in refineries so streams usually pass through several
preheat exchangers before they get to the furnace. If the efficiency of these
exchangers is impaired, the heat duty and tube temperature in the furnace
must be increased to provide the desired process temperature. Solids
entrained in the process stream, salts precipitated as the stream evaporates,
or organic polymers formed in the heated oil, can foul and plug the ex-
changers and thereby raise metal temperatures in the exchangers as well as
in the furnaces.
Antifoulants that have recently been developed for refinery processes [30]
are primarily detergents and suspending agents that minimize deposit for-
mation and sometimes poison polymerization reactions. At 5 to 50 ppm in
process streams, they can minimize deposits in critical exchangers and
furnaces and extend operating periods between shutdowns, reduce heating
costs, and minimize corrosion. Although several laboratory tests are
available for screening the many compounds available [31], field tests are
still the only completely reliable method to select an additive for a par-
ticular fouling problem.
Process temperatures sometimes can be reduced to minimize corrosion,
but usually this approach is impractical. Instead, by varying the feed it is
often possible to locate the source of the corrodents and either modify feed
preparation or eliminate that particular crude from the refinery supply.
High-temperature sulfidic corrosion in hot process streams can be moni-
tored with corrosion probes [32]. Even if there is no practical remedy, at
least the probe measures corrosion without requiring that the unit be shut
down. The equipment life can be estimated and a resistant alloy selected
and fabricated to replace the corroded equipment quickly and economically
at a shutdown scheduled before the equipment becomes defective.
H2S-H~
Correlations have been developed to aid in selecting alloys to resist high
temperature sulfides, especially where the corrodent is primarily a mixture
10
1.0
:z
0.1
0.01
0.001
400 600 800 1000
TEMPERATURE,*F
FIG. 7--Carbon steel corrosion by HsS q- H2 in naphtha desulfurizers.
of H2S and hydrogen [33-36]. Corrosion rates for carbon steel in naphtha
desulfurizers are summarized in Fig. 7. As shown by these isocorrosion
curves, the mole percent H2S in the process stream and the operating tem-
perature define the expected corrosion rates. If the rate is too high for
carbon steel equipment to have a useful life, the appropriate alloy can be
selected by multiplying the carbon steel corrosion rate by the following
factors:
C, 1A Mo 1.0
1 ~o Cr, ~ Mo 0.957
2 ~ Cr, 1 Mo 0.906
5 Cr, a/~ Mo 0.804
7 Cr, 1 Mo 0.736
9 Cr, 1 Mo 0.675
Other alloys can be estimated by interpolations based on chromium alloy-
ing, which provides most of the corrosion resistance aganst high-tempera-
ture sulfides.
For gas oil desulfurizers and hydrocrackers, the naphtha desulfurizer
corrosion rate estimates are multiplied by 1.896.
Above specific temperatures and below specific H2S concentrations, the

corrosion reaction reverses, that is,
FeS -b H2 --~ Fe -1- H~S
Under these conditions, outlined by the "no corrosion" area in Fig. 7,

steels should not be corroded by H2S. However, because corrodents other
than H2S may still attack steels, caution is advisable when selecting ma-
terials for equipment operating in this no corrosion region. In particular,
alloying is often required to resist high-temperature hydrogen attack.
If corrosion is too severe for the low-alloy steels, Fig. 8 shows the appro-
priate stainless steels that can be selected as alternatives. In this case, the
isoeorrosion curves are for 18Cr, 8Ni austenitie stainless steels. Type 410
(12Cr) stainless steel corrodes approximately 6.026 times faster than the
18Cr, 8Ni alloys. Again, the corrosion rates for other high-alloy steels can
be estimated by interpolation primarily on the basis of chromium alloying.
There is essentially no difference between naphtha and gas oil processing
units in the corrosion rates of stainless steels.
10
1.0
z
0.1
0.01
0.001
400 600 800 1000
TEMPERATURE, *F
FIG. 8--18-8 stainless steel corrosion by H2S q- H2 -k hydrocarbons.
The data used to construct Figs. 7 and 8 were obtained by exposing

coupons for at least 500 h in operating units. Pilot plants also have been
used to relate corrosion to temperature, H2S concentration, and pressure
in mixtures of H2S and hydrogen [35-39] using coupons and electrical
resistance corrosion probes. Under a research project sponsored by API at
The Pennsylvania State University [40], another pilot plant continually
weighs the gain in weight of a corroding specimen to study the kinetics of
H2S corrosion. Similar effects of mercaptan sulfur in naphtha desulfurizer
feeds have been studied using a multiple-alloy corrosion prove in a pilot
plant at 650 to 750 F and 300 psig [41,42].
A novel method to study high-temperature H2S-H~ corrosion has been
developed by Dravnieks [43]. Hydrogen is bubbled through a bath of
molten sulfur at a controlled temperature, and all the sulfur vapor in the
hydrogen is converted in a catalytic reactor to a controlled concentration
of HsS in hydrogen. This system eliminates the possibility of segregation of
the gas mixture or loss of H~S concentration due to reaction with oxygen in
air or with the walls of a steel storage cylinder.
Naphthenic Acids
High-temperature sulfide corrosion is often blamed for equipment
failures that are caused by naphthenic acids [8]. These acids corrode not
only low chromium alloy steels but also Types 410 (12 Cr) and 304 (18 Cr,
8 Ni) stainless steels that normally resist sulfide corrosion. On the other
hand, Type 316 (18 Cr, 10 Ni, 2 Mo) stainless steel generally resists both
naphthenic acids and sulfides. Often these differences in corrosion re-
sistance are the only means of confirming that the corrodent is naphthenic
acids. Alloying is the only known way to resist naphthenic acids and sul-
fides at high temperatures, but the costs of the minimum alloying differ
so greatly that a proper diagnosis of the corrodent is critical.
Stress Corrosion Cracking

When austenitic stainless steels are used to resist sulfides or naphthenic
acids, special precautions must be taken to minimize the possibility of
stress corrosion cracking by both aqueous halides [44] and moist sulfide
scales containing polythionic acids [45]. Resistance of alloys to stress corro-
sion cracking is typically evaluated in boiling 42 percent MgC12 for chlo-
rides, and in cold water saturated with SOs and H~S for polythionic acids.
Chemical stabilization with small amounts of titanium or niobium in the
steel and thermal stabilization by specific heat treatments [45] can make
the austenitic stainless steels resistant to polythionic acids but not to halides.
Because stress corrosion cracking is primarily a problem when the process
units are shut down, special precautions can be taken as outlined in the
National Association of Corrosion Engineers Recommended Practice
RP-01-70. This practice involves washing with alkaline solutions to neu-
tralize the polythionic acids and adding nitrates to the solutions to inhibit
chloride cracking.
Hydrogen at High Temperature

Steels to resist high-pressure hydrogen at elevated temperature are
selected using the empirical "Nelson Curves" and the appropriate ASME
Codes. The Nelson Curves are issued as a recommended practice by the
API and are updated from time to time. The latest are in API Publication
961, July 1970. The ASME Boiler and Pressure Vessel Code, Section VIII,
Divisions 1 and 2, provides the design limits for steels in high-temperature,
high-pressure service.
Chemical analysis of metals for alloying constituents is recommended
before installation of critical equipment. Until better methods have been
developed for marking and segregating alloys, such a precaution will help
forestall serious equipment failures. In a recent refinery construction pro-
gram including several new units, 1 to 6 percent of the piping sections
analyzed were found to have been made of alloys other than those speci-
fied [46].
Because of its insidious nature, hydrogen attack is difficult to detect in
onstream equipment. The initial stages can only be detected by metallo-
graphic examination of samples cut from the equipment [47], whereas
severe damage can be detected onstream by ultrasonic measurements [48].
Hydrogen at Low Temperatures

Low-temperature hydrogen attack is equally insidious but can usually be
controlled by modifications in process conditions. Painted can tests and
hydrogen probes used to evaluate process streams are described in the
Handbook of Corrosion Testing and Evaluation [14]. Washing the proces
streams with water to remove some of the corrodents or adding corrosion
inhibitors can sometimes alleviate the problem.
Recommended practices have been developed by API and NACE to
minimize the susceptibility of welds to cracking from low-temperature
hydrogen attack. Experience indicates that a weld hardness of less than
Brinell 200 is satisfactory for refinery equipment in severe environments.
Welds over 200 Brinell should be given a tempering heat treatment. Re-
search sponsored by API is in progress at Battelle Memorial Institute to
better define the controlling parameters.
Other Corrodents
Refinery equipment that handles strong acids and bases requires special
corrosion protection, just as in the chemical industry. Similarly, cooling
water and boiler water treatments are basically the same as those in utilities
practice and so will not be discussed here.
Conclusion
In conclusion, refinery corrosion control, like corrosion, is expensive.
The development of effective controls requires the cooperation of m a n y
individuals with various skills. Although m a n y standards and r e c o m m e n d e d
practices have already been established, more are required. Several organi-
zations are now working to develop and report the necessary technology and
experience. Correctly diagnosing the causes of refinery corrosion and de-
vising practical cures can save money and help insure that equipment is
both safe and reliable.
APPENDIX
Corrosion Standards in the Petroleum Industry
API
Recommended Practice for Steels for Hydrogen Service at Elevated Tempera-
tures and Pressures in Petroleum Refineries and Petrochemical Plants, API
Publication 941, 1970.
Recommended Practice for Welded Plain Carbon Steel Refinery Equipment for
Environmental Cracking Service, API Publication 943, 1971.
NACE
NACE Standard TM-01-69, Laboratory Corrosion Testing of Metals for the
Process Industries.
Recommended Practice RP-01-69, Control of External Corrosion on Under-
ground or Submerged Metallic Piping Systems.
Recommended Practice RP-01-70, Protection of Austenitic Stainless Steel in
Refineries Against Stress Corrosion Cracking by Use of Neutralizing Solutions
during Shutdown.
NACE Standard TM-01-60, Visual Standards for Surfaces of New Steel Air-
Blast Cleaned with Sand Abrasive.
ASME
Boiler and Pressure Vessel Code
Section I, Power Boilers
Section V, Nondestructive Examination
Section VIII, Pressure Vessels--Divisions 1 and 2
Section IX, Welding Qualifications
B31-3, Petroleum Refinery Piping
ASTM
B 117-64, Salt Spray (Fog) Testing
D 130-68, Test for Detection of Copper Corrosion from Petroleum Products,
by the Copper Strip Tarnish Test
D 665-68, Test for Rust-Preventing Characteristics of Steam-Turbine Oil in
the Presence of Water
D 1261-68, Test for Effect of Grease on Copper
D 1275-67, Test for Corrosive Sulfur in Electrical Insulating Oils
D 1743-68, Test for Rust Preventive Properties of Lubricating Greases
D 1748-70, Test for Rust Protection by Metal Preservatives in the Humidity
Cabinet
University of Washington (University of Washington) pursuant to License Agreement. No further reproductions au
D 2200-67, Pictorial Surface Preparation Standards for Painting Steel Surfaces

D 2550-69T, Test for Water Separation Characteristics of Aviation Turbine
Fuels
G 4-68, Recommended Practices for Conducting Plant Corrosion Tests
G 23-69, Recommended Practices for Operating Light- and Water-Exposure
Apparatus (Carbon Arc Type) for Exposure of Nonmetallic Materials
SSPC
SP 8-63, Pickling
Vis 1-63T, Pictorial Surface Preparation Standards for Painting Steel Surfaces
References
[1] Guthrie, V. B., "Measuring Plant Corrosion Costs," Petroleum Processing, Vol. 9,
No. 5, 1954, p. 708.
[2] Hirschmann, W. B, "Some Economic Realities of Corrosion Control," Materials
Protection, Vol 4, No. 7, 1965, p. 8.
[3] Brauweiler, J. R., "Economics of Long- vs. Short-Life Materials," Chemical Engineer-
ing, Vol. 70, No. 2, 1963, p. 128.
[4] Couper, A. S., "Bothered by Corrosion of Your Crude-Unit Condensers?" Oil and
Gas Journal, Vol. 62, No. 29, 1964, p. 79.
[5] Husen, C., "High-Temperature Corrosion by Organic Sulfur Compounds," High-
Temperature Metallic Corrosion of Sulfur and Its Compounds, A~ S. Foroulis, Ed., The
Electrochemical Society, 1971.
[6] Couper, A. S. and Dravnieks, A., "High-Temperature Corrosion by Catalytically-
Formed Hydrogen Sulfide," Corrosion, Vol. 18, No. 8, 1962, p. 291t.
[7] Bonner, W. A., et al, "Prevention of Hydrogen Attack on Steel in Refinery Equipment,"
Proceedings AP1, Vol. 33, Section III, 1953, p. 255.
[8] Heller, J., "Corrosion of Refinery Equipment by Naphthenic Acid," Materials Protec-
tion, Vol. 2, No. 9, 1963, p. 90.
[9] Derungs, W. A., "Naphthenic Acid Corrosion--An Old Enemy of the Petroleum
Industry," Corrosion, Vol. 12, No. 12, 1956, p. 617t.
[10] Allen, R. E., et al, "Analysis of Probable Mechanisms of High-Temperature Hydrogen
Attack of Steel," Proceedings API, Vol. 42, Section III, 1962, p. 452.
[11] Guide for Inspection of Refinery Equipment, Chapter II, API Publication, "Conditions
Causing Deterioration or Failure," 1957
[12] Backensto, E. B., "API Panel Studies Corrosion Control Tests," Petroleum Refiner,
Vol. 38, No. 5, 1959, p. 181.
[13] Ostrofsky, B. and Parrish, C. B., "Ultrasonic Inspection of Boiler Tubes," Materials
Evaluation, Vol. 26, No. 6, 1968, p. 106.
[14] Ailor, W. H., Handbook on Corrosion Testing and Evaluation, Wiley, New York, 1971.
[15] "One Man Checks Corrosion with Ultrasonic Tester," Chemical Processing, Vol. 33,
No. 12, 1970, p. 64.
[16] Paul, R. and Shirley, W. L., "Determination of Refinery Corrosion Rates Using PAIR
Technique," Materials Protection, Vol. 8, No. 1, 1969, p. 25.
[17] Samuelson, G. J., "Hydrogen-Chloride Evolution from Crude Oils as a Function of
Salt Concentration," Proceedings API, Vol. 34, Section III, 1954, p. 50.
[18] Cataldi, H. A., Askevold, R. J., and Harnsberger, A. E., "Corrosivity of Crude Oils,"
Oil and Gas Journal, Vol. 52, No. II, 1953, p. 100.
[19] Piehl, R. J., "Correlation of Corrosion in a Crude Distillation Unit with Chemistry of
the Crudes," Corrosion, Vol. 16, No. 6, 1960, p. 139.
[20] Ereneta, V. G. and Couper, A. S., "Crude Unit Corrosion Control Based on Pilot
Plant Tests," Proceedings API, Vol. 50, Section III, 1970, p. 830.
[21] Easton, C. L. and Jameson, B. G., "High-Temperature Organic Sulfur Corrosion in
Crude Processing Units," Proceedings of the 25th NACE Conference, 1970, p. 572.
[22] Couper, A. S., "Plant-Wide Cooperation Yields Corrosion Failure Analysis," Oil and
Gas Journal, Vol. 66, No. 34, 1968, p. 82.
[23] Fisher, A. O. and Whitney, F. L., "Laboratory Methods for Determining Corrosion
Rates under Heat Flux Conditions," Corrosion, Vol. 15, No. 5, 1959, p. 257.
[24] Ereneta, V. G., "A New Thermogalvanic Method that Shows How the Dezincification
of Inhibited Admiralty Brass Occurs in Fresh Water," NACE Corrosion Research
Conference, March 1972.
[25] Backensto, E. B. and Gustafson, V. M., "Corrosion Rate Measurements in Refinery
Cooling Water Systems," Materials Protection, Vol. 6, No. 12, 1967, p. 25.
[26] Kerst, H., "Laboratory Investigation of Water-Side Scale and Corrosion in the
Presence of High Process-Side Temperatures," Corrosion, Vol. 16, No. 10, 1966, p. 523.
[27] "Water Cooled Corrosion Probe Extends Exchanger-Tube Life," Oil and Gas Journal,
Vol. 61, No. 35, 1963, p. 128.
[28] Dieman, E. A. and Gaynor, J. W., "Accelerated Condensation Corrosion Test for
Evaluating Rust Preventives," Corrosion, Vol. 14, No. 6, 1958, p. 302t.
[29] Hess, W. A., "Standard Heat Exchanger for Cooling Water Tests," Materials Pro-
tection, Vol. 4, No. 8, 1965, p. 70.
[30] Couper, A. S., "Process Side Antifoulants in Petroleum Refineries," Materials Pro-
tection, Vol. 9, No. 6, 1970, p. 29.
[31] Gillespie, B. G., "A New Process Antifoulant Test Correlates Better with Refinery
Experience," Materials Protection, Vol. 10, No. 8, 1971, p. 21.
[32] Moller, G. E. and Patrick, J. T., "Electrical Resistance Measuring Device Plus Sta-
tistical Analysis Yield Unique Approach to Corrosion Problems," Corrosion, Vol. 16,
No. 3, 1960, p. 155t.
[33] Sorell, G., "Compilation and Correlation of High-Temperature Catalytic Reformer
Corrosion Data," Corrosion, Vol. 14, No. 1, 1958, p. 15t.
[34] Dravnieks, A. and Samans, C. H , "Corrosion Control in Ultraforming," Proceedings
API, Vol. 37, Section III, 1957, p. 100.
[35] Backensto, E. B. and Sjoberg, J. W., "New Hydrogen Sulfide Corrosion Curves,"
Petroleum Refiner, Vol. 37, No. 12, 1958, p. 119.
[36] Couper, A. S. and Gorman, J. W., "Computer Correlations to Estimate High-Tem-
perature H~S Corrosion in Refinery Streams," Materials Protection, Vol. 10, No. 1,
1971, p. 31.
[37] Bruns, F. J., "Corrosion of Ni-Cr-A1-Fe Alloys by Hydrogen Sulfide at 1100 to 1800 F,"
Corrosion, Vol. 25, No. 3, 1969, p. l19t.
[38] Sharp, W. H. and Haycock, E. W., "Sulfide Scaling under Hydrorefining Conditions,"
Proceedings ,4PI, Vol 48, Section III, 1959, p 74.
[39] McCoy, J. D. and Hamel, F. B., "Effect of Hydrodesulfurization Process Variables on
Corrosion Rates," Materials Protection, Vol. 10, No. 4, 1971, p. 17.
[40] Zelouf, S. and Simkovich, G., "Kinetics of Sulfidation of Fe-Cr-X Alloys," Proceedings
AP1, Vol. 48, Section III, 1968, p. 176.
[41] Couper, A. S., "Laboratory Tests to Evaluate Corrosion Resistance to Sulfidic Media,"
Proceedings API, Vol. 41, Section III, 1961, p. 86.
[42] Couper, A. S., '+High-Temperature Mercaptan Corrosion of Steels," Corrosion, Vol.
19, No. 11, 1963, p. 396t.
[43] Dravnieks, A. and Samans, C. H., "Kinetics of Reactions of Steel with Hydrogen
Sulfide-Hydrogen Mixtures," Journal of the Electrochemical Society, Vol. 105, No. 4,
1958, p. 183.
[44] Couper, A. S., "Testing Austenitic Stainless Steels for Modern Refinery Applications,"
Materials Protection, Vol. 8, No. 8, 1969, p. 17.
[45] Samans, C. H., "Stress Corrosion Cracking Susceptibility of Stainless Steels and
Nickel Base Alloys in Polythionic Acids and Acid Copper Sulfate Solution," Corrosion,
Vol. 20, No. 8, 1964, p. 256t.
[46] Jacobson, J., "Metal-Sampling Program Finds Construction Errors," Oil and Gas
Journal, Vol. 69, No. 23, 1971, p. 60.
[47] Nelson, G. A. and Effinger, R. T., "Blistering and Embritflement of Pressure Vessel
Steels by Hydrogen," Welding Journal, Vol. 34, No. 1, 1965, p. 12.
[48] Bland, J., "Ultrasonic Inspection Used to Detect Hydrogen Attack," Petroleum
Refiner, Vol. 37, No. 7, 1958, p. 115.
STP534-EB/Nov. 1973
Chapter 3

Gas Industry
L. M . B u l l 1
The natural gas industry is concerned with the production, transmission,

distribution, storage, processing, and utilization of natural gas. High
efficiency and safety requirements result in the need for effective corrosion
control of these facilities. These facilities are subject to marine, under-
ground, and various atmospheric and process environments. A complete
description of materials and environments involved in the functions of the
gas industry are too numerous to completely describe herein.
The major capital investment in the industry is in the following functional
types of plants: (1) production and storage wells; (2) production and storage
field piping for gathering gas from the individual wells; (3) transmission
pipelines for carrying gas from production and storage sources to sales and
distribution facilities; (4) compressor stations and associated facilities for
maintaining or boosting pressure; (5) distribution pipelines used to deliver
gas to the individual residential, commercial, and industrial users; (6) meas-
urement and regulator stations used to measure the quantity of gas and
control or reduce pressure; (7) processing equipment in the form of scrub-
bers and dehydrators for hydrocarbon liquids and water removal. Addi-
tionally gas heating, cooling, and odorant injection equipment is common.
Production Materials and Environments

The materials and environments unique to the production of natural gas
are not substantially different from those in the Petroleum Industry. High
strength steel casing and tubing common to the production wells are of the
J-55, N-80 varieties and in conformance with the API 5A standards.
The most common internal corrosion problems are of the "sweet gas"
and "sour gas" variety, that is, hydrogen sulfide and carbon dioxide
induced.
1 Engineering Department, Columbia Gas System Service Corp., Columbus, Ohio.
60
CORROSION IN THE GAS INDUSTRY 61
Gas Storage Well Materials and Environments

Gas storage well materials are generally of the same steel casing and
tubing varieties described for production wells. The environment to which
the internal surfaces of the casing and tubing strings are exposed is basically
a clean natural gas stream in gas storage fields. Storage fields are most often
established in depleted gas reservoirs. Gas is injected into the field during
periods of low consumption, typically in the summer months. The gas is
withdrawn during peak consumption periods in the winter months.
The internal surface corrosion problems are not as common as in the
production wells. Occasionally residual cushion gases contribute CO2 or
H2S as do natural brine waters that frequently intrude during withdrawal of
gas at fast rates. The use of inhibitors is common to control such problems.
The corrosion problems on the external surfaces of the casing and tubing
strings are commonly those of the galvanic soil type and occasionally of the
stray current type.
Natural Gas Transmission Systems
Natural gas transmission is that part of the industry engaged in the
transport of natural gas, usually from the gas producing field to remote
distribution areas. The bulk of the producing areas in North America are
located in the Texas-Oklahoma Panhandle; Texas-Louisiana Gulf Coast,
and Offshore Gulf of Mexico; West Texas-Southeast New Mexico; Four
Corners area of Utah, Colorado, New Mexico, Arizona; Appalachian area;
California; Alberta Province, Canada; (and of present interest) North
Slope of Alaska. It should be noted that all major production areas are
remote from the industrial and population centers of the northeastern and
midwestern states. Figure 1 is pipeline map of the United States which
illustrates the production area and transmission pipelines.
The pipeline system usually includes a gathering system in the gas pro-
duction area which with compressor, metering and gas conditioning equip-
ment bring to the main pipeline arteries a gas suitable for transport and sale.
The mainline system includes one or more large diameter (16 to 36 in.)
pipelines with appropriate compressor stations at typically 80- to 100-mile
spacing along the pipeline length. A lateral system with metering, pressure
regulating and associated equipment completes delivery to the gas dis-
tributor.
A variety of gas dehydration, desulfurization, carbon dixoide, and solids
removal equipment, and natural gasoline plants will ordinarily be asso-
ciated with such systems. Compression equipment may include both
centrifugal and reciprocating compressors. Compressor prime movers
include two- and four-cycle engines, gas turbines and steam turbines with
the use of aircraft type turbine engines in recent years. The piping varies
from 0.25 in. through 36 in. with 48-in. diameter pipelines presently pro-
posed. Control of gas flow requires the use of a large number of valves,
~7
C
~t
o
! ]
i t ff
0
z
""- [ z
9x. I . I
0
z- t
FIG. 1 - - M a j o r United States gas fields and pipelines.
regulators, and meters, and an extensive communications system ranging

from private and leased telephone lines to microwave systems.
The prevalent corrosion problems are of the galvanic and differential
concentration cells common to ferrous metals in soils and natural waters
with resultant pitting type corrosion. A serious problem in certan areas is
the stray direct current interference problem from coal mine, street railway,
and industrial use of earth grounded direct current power. More recently
the development and proposed large scale use of High Voltage Direct
Current (HVDC) power transmission systems is of particularly serious
concern. Bacterial or microbiological corrosion although difficult to
validate in specific cases is generally thought to be a serious and common
contributor to the underground and natural water corrosion problems.
Stress corrosion cracking has occurred in a very few instances on external
surfaces of underground gas pipelines. The actual environmental com-
ponents have not been identified, however, investigators point to the
probability of nitrates, carbonate, bicarbonate mixtures, and caustic
environments (as part of the cathodic reactions, perhaps). Hydrogen stress
cracking has been reported to occur in a few cases on external surfaces of
pipelines in localized areas of extreme hardness, namely, so called "hard
spots" induced during manufacture of the pipe by accidental localized
quenching by water spills.
Internal blistering and cracks have occurred on production gathering
pipelines where the gas had a considerable hydrogen sulfide content.
The main methods of corrosion control on underground pipelines are
complete coating, cathodic protection, insulated fittings at strategic points,
and mitigation of stray currents by installing control bonds to the source of
the stray current. The National Association of Corrosion Engineers
Recommended Practice RP-01-69, "Control of External Corrosion on
Underground or Submerged Metallic Piping Systems," presents these
procedures and practices that are presently in general use.
The Department of Transportation, Office of Pipeline Safety has issued
minimum safety standards which contain minimum corrosion control
requirements for pipeline facilities carrying natural gas, and identified as
Subpart I of Part 192, Title 49, Code of Federal Regulations.
Corrosion problems in compressor station equipment is not a significant
problem since water treatment and inhibition of cooling waters is normal
practice.
Processing equipment such as dehydrators, gas heaters, desulfurization
plants and COs removal processes may be considered as a means of con-
trolling or eliminating internal corrosion to the pipeline system. Corrosion
problems in this equipment are controlled by the use of inhibitors. Cor-
rosion coupons and equipment inspection are used to monitor corrosion
rates in this equipment.
Atmospheric corrosion problems are significant in industrial areas and in
offshore and coastal facilities. Coatings, paints, and metallic coatings or

cladding are used to protect against atmospheric corrosion.
Coatings on the discharge piping of compressor stations (as well as
exhaust stacks) are subject to constant high temperature service. The under-
ground discharge piping is subject to temperatures ranging from 110 to
140 F typically, and occasionally as high as 160 F. These temperatures
along with soil stresses created by wetting and drying of surrounding soils
make a particularly tough environmental condition for the pipeline coatings
used historically and commonly result in cracking and flowing of the
coatings. Electrical shorts between carrier pipe and the metallic casing
used at road crossings is common although insulated spacers and casing
seals are used to prevent direct metallic contact and the intrusion of water
and silt. Both the high temperature service requirements of coatings and
the incidence of shorted casings result in extraordinary current require-
ments for cathodic protection.
Steel line pipe of the API 5L and 5LX specifications is the usual pipe
material. Some limited use of aluminum pipelines has been reported.
Natural Gas Distribution Systems

The transmission systems customers are largely gas distribution com-
panies. The distribution company sells to the individual residential, com-
mercial, and industrial user and operates a complex system of mains and
service lines that range from 0.75 to 20-in. diameter. Pressure regulators,
metering, odorizing equipment, along with dehydration, and gas heating
equipment is associated with these systems.
Pipe materials in service consist of wrought iron, cast iron, ductile iron,
copper tubing for service lines, steel, and plastic. Steel is the most common
material in service along with cast iron and ductile iron in certain areas.
Plastic pipe is becoming popular in the small diameters (up to 4 in.).
Coiled steel tubing has also had some use in recent years.
The common corrosion types are much the same as in transmission pipe-
lines. However,galvanic cells of the bimetal types are more prevalent due to
the use of different types of pipe metal in the distribution system itself, as
well as inadvertent connections to underground water piping and telephone
and electrical grounds. The congestion of utility facilities within service
easements and beneath city streets creates serious bimetal corrosion prob-
lems in that accidental contacts and low electrical resistance between
facilities often result.
Corrosion due to road deicing salts, cinder, and septic and sewer drainage
add to the natural soil corrosion problems. In mining areas the release and
drainage of acid mine waters creates very severe corrosion problems. The
pH of mine waters has been observed to be as low as 1.8 by the author with
a pH of 3 to 3.5 commonly observed.
Industry and Related Federal Standards

The natural gas industry is actively represented in and cooperates with
several committees on organizations devoted to setting standards or test
methods, and conducting research programs. Those standards and prac-
tices relating to the natural gas industry are listed in Appendix A of this
Chapter.
The American Gas Association, National Association of Corrosion
Engineers, and the American Society of Mechanical Engineers Gas Piping
Standards Committee and The American Petroleum Institute are the
organizations most involved in setting industry standards relating to
corrosion control. The NACE is the principal organization involved in
developing recommended practices, and both NACE and ASTM are
involved in arranging test programs and developing standard test methods.
Additionally, the American Petroleum Institute, the Steel Structures
Painting Council, the American Water Works Association and the National
Association of Pipe Coating Applicators are the source of certain pipe
material, coatings, and surface preparation specifications.
The Office of Pipeline Safety (OPS), Department of Transportation
(DOT) has issued minimum federal safety standards for the transportation
of natural gas and for pipeline facilities used for transportation. These are
contained in Part 192 in Title 49, Code of Federal Regulations, Subpart I,
that became effective 1 August 1971 and stipulates the minimum require-
ments for the protection of gas pipelines from internal and external
corrosion.
American Gas Association
The AGA Operating Section has a Corrosion Committee which is
responsible for assembling and disseminating information and investigating
problems pertaining to the control of above and below ground corrosion
on gas production, storage, transmission, and distribution systems. The
Corrosion Committee is not a standards writing organization but through
its normal functions acts to initiate development of corrosion standards
through organizations such as ASTM, NACE, and ANSI.
Additionally, the AGA sponsors industry research projects. Presently
the principal research programs in the corrosion control field are in study
of stress corrosion cracking, hydrogen stress cracking, and development of
new and better corrosion inspection and electrical survey methods on gas
pipelines.
National Association of Corrosion Engineers
The NACE has a number of Technical Committees involved in corrosion
control in the natural gas industry. The Technical Committees with major
involvement are as follows: T-1, "Corrosion Control in Petroleum Produc-
tion"; T-1J, "Corrosion Control for Storage Wells"; T-6, "Protective

Coating and Linings"; and T-10, "Underground Corrosion Control."
Group Committee T-10 is the most pertinent to the gas pipeline industry
since its scope involves engineering practices for the prevention and control
of corrosion on underground or submerged metallic structures. Within the
T-10 Group Committee are Unit Committees devoted to interference
problems, cathodic protection, protective coatings systems, materials of
construction, and internal corrosion of pipelines.
American Society of Mechanical Engineers (ASME)

The ASME Gas Piping Standards COmmittee publishes a Guide For
Gas Transmission and Distribution Piping Systems which includes the
Federal Gas Pipeline Safety Standards, together with the design require-
ments, material references, and recommended practices of the ASME Gas
Piping Standards Committee. This Committee has a subcommittee on
corrosion and includes recommended practices for internal, external, and
atmospheric corrosion control.
American Petroleum Institute (API)

The API has numerous working committees involved in all aspects of
the petroleum industry and includes corrosion subcommittees in the Pipe-
line Division, Production Division, and in the Refinery Division.
American Society for Testing and Materials (ASTM)

The ASTM has numerous committees that directly affect the natural gas
industry through corrosion test methods and material test methods. The
most pertinent are as follows: Committee G-1 Corrosion Of Metals,
Subcommittee G-01.10 Methods of Test in Soil Corrosion, Committee G-3
Deterioration of Non-Metallic Materials, Subcommittee G-3.06 Deteriora-
tion of Pipeline Coatings and Linings.
Appendix B is a compilation of test methods and recommended prac-
tices for coatings materials in the natural gas industry.
Selected Corrosion Problems

It has previously been stated that the corrosion problem on underground
gas piping is typically of the normal soil corrosion type. This is complicated
by the variation in materials used within the gas system, as well as inter-
connected and closely adjacent metals used in the electrical, water, sewage,
and telephone systems. Additional complications are introduced by the
various soil contaminants such as deicing salts, cinders, mine waters, septic
drainage, and fertilizers.
Soil itself is a heterogeneous mixture of individual components and
exhibits radical changes both with depth as well as with terrain, geological,
and vegetation changes. The corrosivity depends on many factors which
include resistivity, moisture content, acidity, pH, salt content and oxygen
content.
Extensive underground metal piping systems suffer to a large degree from
differential aereation cells (oxygen differential) along their length and is
characterized by local pitting. As a result corrosion rates determined on
individual specimens or coupons in specific soil types cannot provide a
realistic evaluation of the corrosion rate that exists on underground piping
passing through those particular soil types. The determination of corrosivity
polarization characteristics, and cathodic protection requirements are more
commonly made on sections of the route rather than at local areas.
The practice today is t o coat and cathodically protect all underground
metallic piping at the time of installation. This is in recognition that the
complex soil electrolyte corrosion rates are not readily determined on a
lengthy pipeline. The long period of service required of the piping makes it
unlikely that any such determinations would be of permanent value; that
so called low corrosion rates cannot be tolerated over this long service
period requirement; and the higher cost and impracticality of attempting
to delay the installation of cathodic protection until such time as the
corrosion loss has progressed to some maximum loss in wall thickness.
As a result the use of cathodic protection rectifiers has increased drasti-
cally in the past 20 years. This requires effective coordination and coopera-
tion between the cathodic protection engineers and technicians to minimize
and mitigate stray current corrosion effects that may result from these de
current sources. This is generally accomplished through regional and local
coordinating committees that are sponsored by NACE. These committees
vary in their formal organization and procedures however they arrange and
coordinate notification procedures and cathodic protection interference
testing between the various underground structure owners in their geo-
graphical area. A list of these committees is contained in Appendix C.
The test methods employed for the evaluation of the effectiveness of
coatings and cathodic protection, and for determining actively corroding
areas is of particular interest. The bulk of the facilities in the natural gas
industry are underground or marine installations of piping, wells, and
appurtenances that are not accessible for periodic inspection except at
great cost. This presents a need for improved methods of detecting active
corrosion, and improved methods of monitoring the effectiveness of
cathodic protection systems.
Pipe to soil potential readings at various points along the piping are the
normal method employed to determine if the cathodic protection level is up
to some appropriate criterion such as the -0.85 V to copper-copper sulfate
electrode that is most commonly used. This is generally effective and
accurate, however, specific situations occur which raise questions as to its
complete effectiveness in locating localized points where cathodic protec-
tion may not be effective. Marine pipelines and piping beneath concrete
68 iNDUSTRIAL CORROSION STANDARDS AND CONTROL
and asphalt roadways cannot be tested in a practical manner using this

method. In cases of coating disbondment it is postulated that the long
electrical paths that result are of sufficient electrical resistance to prevent
effective amounts of cathodic protection current from reaching the metallic
surface areas beneath the coating. There have also been questions raised as
to the validity of these potential readings made at the surface of the earth
directly over the center line of the pipe in cases of large diameter pipe,
namely, does this potential reading truly represent the effective potential
to earth of the full circumferential surface of the pipe or does it only
represent the potentials on the top quadrant of this circumference?
Existing piping that was installed previous to requiring coating and
cathodic protection at the time of installation presents the real corrosion
problems in that they are normally bare and may have various types of
pipe joint couplings that are not electrically conductive and indeed may be
intermittently conductive. This makes it extremely difficult to rely on
potential survey methods that are normally used to determine cathodic
and anodic areas on underground piping. The more electronegative areas
on unprotected pipelines are generally identified as the corroding areas and
cathodic protection is then provided for in such areas.
The above serves to point out the possible shortcomings in the present
practical test method in general use. Industry research is presently concen-
trated on various projects that hopefully will result in more effective and
accurate means of monitoring cathodic protection levels, and in detecting
active corrosion or corrosion damage whether propagated in the pitting,
general corrosion, or cracking mode.
Corrosion control through the use of coatings, cathodic protection, and
insulation has an admirable record to date and is probably 95 percent
effective in the gas industry. The above discussion is not meant to infer it is
not extremely effective, but to point out the areas of field testing that have
shortcomings.
Several commercial electronic tools are presently available from com-
mercial service companies that are designed to be run internally through
pipelines or well casing and detect corrosion damage. They employ eddy
current and or magnetic flux leakage methods to determine loss of wall
thickness or anomalous conditions on the steel surface and are capable o f
discriminating between external and internal corrosion and detect pits,
general corrosion, and localized hard spots. The pipeline tool for large
diameter can be run onstream at speeds of 5 to 10 mph and carries its own
power supply which provides sufficient power for an 8-h operation. These
tools are of great value, however, their shortcomings are that they provide
information only on past corrosion damage and do not provide information
on present corrosion rates or cathodic protection levels. They are expensive
to operate and are not practically applicable to all piping.
APPENDIX A-1
Regulations, Standards, and Specifications Related to the Natural Gas Industry
Title 49 of the Code of Federal Regulations, Part 192-Transportation of Natural
and Other Gas by Pipeline: Minimum Federal Safety Standards. (11 August 1970).
Title 18 of the Code of Federal Regulations, Part 2, Par. 2.69 and Part 157, Par.
157.14---Guidelines for Natural Gas Companies in Planning, Locating, Clearing
and Maintenance of Rights-of-Way and Construction of Aboveground Facilities
to Aid in Recreational Values: Exhibits. (10 July 1970).
National Association of Corrosion Engineers: RP-01-69
1. API Standard 5L, API Specification for Line Pipe, 1970 ed.
2. API Standard 5LS, API Specification for Spiral-Weld Line Pipe, 1970 ed.
3. API Standard 5LX, API Specification for High-Test Line Pipe, 1970 ed.
4. API Recommended Practice 5L1, API Recommended Practice for Railroad
Transportation of Line Pipe, 1967 ed.
5. API Standard 5A, API Specification tot Casing, Tubing, and Drill Pipe,
1968 ed.
6. API Standard 6A, Specification for Wellhead Equipment, 1970 ed.
7. API Standard 6D, Specification for Pipeline Valves, 1968 ed.
8. API Standard 1104, Standard for Welding Pipelines and Related Facilities,
1968 ed.
The American Society for Testing and Materials (ASTM)
1. ASTM Specification for Welded and Seamless Steel Pipe, A 53-71.
2. ASTM Specification for Welded Wrought-Iron Pipe, A 72-68.
3. ASTM Specification for Seamless Carbon Steel Pipe for High-Temperature
Service, A 106-68.
4. ASTM Specification for Electric-Fusion (Arc)-Welded Steel Plate Pipe
(Sizes 16 in. and Over), A 134-68.
5. ASTM Specification for Electric-Resistance-Welded Steel Pipe, A 135-69.
6. ASTM Specification for Electric-Fusion (Arc)-Welded Steel Pipe (Sizes 4 in.
and Over), A 139-71.
7. ASTM Specification for Electric-Fusion-Welded Steel Pipe tor High-Pressure
Service, A 155-71.
8. ASTM Specification for Spiral Welded Steel or Iron Pipe, A 211-68.
9. ASTM Specification for Seamless and Welded Steel Pipe for Low-Tem-
perature Service, A 333-67.
10. ASTM Specification for Metal-Arc-Welded Steel Pipe for High-Pressure
Transmission Service, A 381-71.
11. ASTM Specification for Electric-Resistance-Welded Coiled Steel Tubing for
Gas and Fuel Oil Lines, A 539-71.
12. ASTM Specification for Thermoplastic Gas Pressure Pipe, Tubing, and
Fittings, D 2513-70.
13. ASTM Specification for Reinforced Thermosetting Plastic Gas Pressure
Piping and Fittings, D 2517-67.
14. ASTM Specification for Carbon and Alloy Steel Forgings for Pressure
Vessel Shells, A 372-71.
15. ASTM Sampling of Natural Gas, D 1145-53.
16. ASTM Test for Calorific Value of Gaseous Fuels by the Water-Flow Calo-
rimeter, D 900-55.
The American National Standards Institute, Inc. (ANSI)

1. ANSI B16.1, Cast Iron Pipe Flanges and Flanged Fittings, (B16.1-1967).
2. ANSI B16.5, Steel Pipe Flanges and Flanged Fittings, (B16.5-1968).
3. ANSI B31.8, Gas Transmission and Distribution Piping Systems, (B31.8-
1968).
4. ANSI B36.10, Wrought-Steel and Wrought-Iron Pipe, (B36.10-1959).
5. ANSI Z21.30, Installation of Gas Appliances and Gas Piping, (Z21.30-
1964).
6. ANSI C1, National Electrical Code, 1968, (C1-1968).
The American Society of Mechanical Engineers (ASME)
1. ASME Boiler and Pressure Vessel Code, Section VIII, "Pressure Vessels,
Division 1," 1968.
2. ASME Boiler and Pressure Vessel Code, Section IX, "Welding Qualifica-
tions," 1968.
Manufacturer's Standardization Society of the Valve and Fittings Industry
1. MSS SP-25, Standard Marking System for Valves, Fittings, Flanges, and
Unions, 1964.
2. MSS SP-44, Steel Pipe Line Flanges, 1955 ed.
3. MSS SP-52, Cast Iron Pipe Line Valves, 1957 ed.
National Fire Protection Association (NFPA)
1. NFPA Standard 30, Flammable and Combustible Liquids Code, 1969 ed.
2. NFPA Standard 54, Installation of Gas Appliances, Gas Piping, 1969 ed.
3. NFPA Standard 58 Storage and Handling, Liquefield Petroleum Gases.
1969 ed.
4. NFPA Standard 59, LP Gases at Utility Gas Plants, 1969.
APPENDIX A-2
Regulations, Standards and Specifications Relating to the Natural Gas Industry
Structural Materials
Brass (rods and bars for structural use) ASTM B 21
Bronze (manganese bronze castings) ASTM B 132
Carbon-steel plates ASTM A 285
Cast iron (ordinary gray-iron castings) ASTM A 48
Chains ASTM A 56
Copper (wrought and copper alloy) rod, bar and shapes ASTM B 245
High-strength low-alloy structural manganese-vanadium steel ASTM A 441
High-tensile carbon-silicon steel plates ASTM A 212
Low-alloy structural steel ASTM A 242
Manganese-vanadium steel plates ASTM A 225
Malleable-iron castings ASTM A 47
Plates (carbon-steel with imporved transition properties) ASTM A 442
Springs, helical (for use on spring hangers ASTM A 125
Steel, structural ASTM A 7
Steel, structural ASTM A 36
Steel, structural bars ASTM A 29
Steel, structural (plates) ASTM A 283
Steel, structural plates, intermediate and high temperature ASTM A 515
service
Steel, structural plates, medium and low temperature service ASTM A 516
Steel, structural rivets ASTM A 502
Wrought iron (plates) ASTM A 42

Wrought iron (extra-refined bars) ASTM A 84
Fittings, Valves, and Flanges
Brass castings ASTM B 62
Bronze castings ASTM B 61
Cast-iron castings ASTM A 126
ANSI A 21.10
ANSI A 21.11
AWWA C 100
Cast nodular iron for pressure-containing parts for use at ASTM A 395
elevated temperatures
Ferritic nodular iron castings for valves, flanges, pipe fittings, ASTM A 445
and other piping components
Malleable iron for castings ASTM A 197
Plastic (thermoplastic) tubing pipe and fittings ASTM D 2513
Plastic (thermosetting) pipe and tubing ASTM D 2517
Steel (alloy castings) for high-temperature service ASTM A 217
Steel (low alloy) castings ASTM A 487
Steel (cast-carbon) for fusion welding for high-temperature ASTM A 216
service
Steel (forged or rolled) for high-temperature service ASTM A 105
Steel (forged or rolled) for general service ASTM A 181
Steel (forged or rolled alloy) tor high-temperature service ASTM A 82
Steel (forged or rolled) low temperature service ASTM A 350
Steel (factory-made w~ought carbon steel and ferritic-alloy ASTM A 234
steel welding fittings)
Steel (wrought carbon and alloy) low temperature service ASTM A 420
Steel pipe flanges AWWA C 207
Bolting
Steel (alloy) for high-temperature service ASTM A 193
Steel (alloy) bolting materials for low-temperature service ASTM A 320
Steel (carbon and alloy) for nuts ASTM A 194
Steel (carbon) bars ASTM A 107
Steel machine bolts and nuts (grade B) ASTM A 307
Steel (quenched-and-tempered alloy) bolts and studs ASTM A 354
with suitable nuts
Steel (quenched-and-tempered) bolts and studs ASTM A 449
Pipe and Tubing
Brass (seamless) pipe ASTM B 43
Carbon and alloy steel forgings for pressure vessel shells ASTM A 372
Cast-iron pressure pipe ASTM A 377
Cast-iron (centrifugally-case) pipe USAS A 21.7
USAS A 21.9
Cast-iron (pit-cast) pipe USAS A 21.3
Copper and copper alloy, pipe and tube ASTM A 251
Copper (seamless) pipe ASTM B 42
Copper (seamless) tubing ASTM B 75
Copper (seamless) bright-annealed tubing ASTM B 68
Copper (seamless) water tubing ASTM B 88
Ductile-iron (centrifugally cast) pipe ANSI A 21.52
Plastic (thermoplastic) tubing pipe and fittings ASTM D 2513

Plastic (thermosetting) pipe and tubing ASTM D 2517
Steel (electric-fusion-welded) 18-in. and larger pipe ASTM A 155
for high-temperature and high-pressure service
Steel (electric-resistance-welded) coiled tubing ASTM A 539
Steel (electric-resistance-welded) pipe ASTM A 135
Steel (electric-fusion-welded) pipe ASTM A 139
Steel (electric-fusion-welded) large-size pipe ASTM A 134
Steel (metal-arc-welded) pipe for high-pressure transmission ASTM A 381
service
Steel and iron (seamless and welded) line pipe API 5L
Steel (seamless and welded) high-test line pipe API 5LX
Steel (seamless and welded) casing, tubing and drill pipe API 5A
Steel (seamless) pipe for high-temperature service ASTM A 106
Steel, seamless and welded for low temperature service ASTM A 333
Steel (spiral-welded) line pipe API 5LS
Steel or iron (spiral-welded) pipe ASTM A 211
Steel (welded and seamless) pipe for ordinary uses ASTM A 120
Steel (welded and seamless) pipe for coiling and bending ASTM A 53
Wrought-iron (welded) pipe ASTM A 72
APPENDIX A-3
A S T M Specifications A 225-71 B 132-70
A 6-71 A 285-70a A 234-71 B 249-71a
A 20-71 A 307-68 A242-70a B 251-71
A 29-67 A 320-70 A 283-70a D 2513-70
A 36-70a A 333-70a D2517-67
A 42-66 A 350-65 M S S Standard Practices
A 47-68 A 354-66 SP-6-1963
A 48-64 A 372-71 SP-25-1964
A 53-71 A 377-66 SP-44-1955
A 56-68 A 381-71 SP-46-1955
A 72-68 A 395-70 SP-47-1956
A 84-68 A 420-71 SP-48-1969
A 105-68 A 441-70a SP-52-1957
A 106-68 A 442-71 SP-55-1961
A 120-69 A 445-70 SP-61-1961
A 125-65 A 449-68 SP-63-1967
A 126-66 A 487-71
A 134-68 A 502-65 American Insurance Association
A 135-69 A 515-71 SIB No. 294-1956
A 139-71 A 516-71 A N S I Standards
A 155-71 A 539-71 A21.1967
A 181-68 13 21-66a A21.7-1962
A 182-71 B 42-71 A21.9-1962
A 193-71 B 43-70 A21.10-1964
A 194-69 B 61-70 A21.11-1964
A 197-47 B 62-70 A21.50-1965
A 211-68 B 68-70 A21.52-1965
A 216-70a B 75-71 B1.1-1960
A 217-70a B 88-71 B2.1-1968
University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorize
B2.2-1968 A W W A Standard
B16.1-1967 AWWA C100-55
B16.3-1963 AWWA C207-55
B16.4-1963
B16.5-1968 NFPA Publications
B16.9-1964 No. 10-1962 + 1963 Adm.
B16.11-1966 No. 30-1963
B16.18a-1967 No. 58-1963
B16.20-1963 No. 59-1962 + 1963 Adm.
B16.24-1962
B16.25-1964 National Association o]
B18.2.1-1966 Corrosion Engineers
B18.2.2-1966 NACE RP-01-69
B31.1.0-1967
B31.4-1966 API Standards
B36.10-1959 5A 29th ed.-1968
C1-1968 5B 7th ed.-1968 and Suppl. 1, 1969
Z21.30-1964 5L 24th ed.-1969
A S M E Codes 5LS 4th ed.-1969
5LX 16th ed.-1969
ASME Boiler and Pressure Vessel 6A 6th ed.-1968 and Supp. 1, 1969
Code, 1968 6D 12th ed.-1968 and Supp. 1, 1969
A WS Standard 1104 llth ed.-1968
AWS A3.0-1969
APPENDIX B
Test Methods, Recommended Practices for Coatings, Materials in the Gas Industry
Standard for Coal-Tar Enamel Protective Coatings AWWA 8310-D Committee
for Steel Water Pipe-AWWA-C203
Coal-Tar Coatings for Underground Use NACE
Synthetic Resin Primer for Coal-Tar Enamel, U.S. Government
Research Report 8, U.S. Dept. of the Interior, Printing Office
Bureau of Reclamation
Asphalt Type Protective Coatings for Underground NACE
Pipelines-Mastic Systems
Asphalt Type Protective Coatings for Underground NACE
Pipelines-Wrapped Systems
Asphalt Protective Coatings for Pipelines-Construc- The Asphalt Institute
tion Series No. 96--Wrapped and Mastic Systems
Asphalt Type Protective Coatings tor Underground NACE
Pipelines
Hot-Applied Wax Type Protective Coatings and NACE
Wrappers for Underground Pipelines
Prefabricated Plastic Films for Pipeline Coatings NACE
"Control of Pipeline Corrosion," pp. 9-18, A. W. NACE
Peabody, Dec. 1967
Recommended Practices Associated with the Appli- NACE
cation of Organic Coatings to the External Surface
of Steel Pipe for Underground Use
Methods of Measuring Leakage Conducthnce of NACE
Coating on Buried or Submerged Pipelines
Inspection of Pipeline Coatings NACE

Nondestructive Measurement of Film Thickness of ASTM G 12-69T
Pipeline Coatings on Steel
Test for Cathodic Disbonding of Pipeline Coatings ASTM G 8-69T
Test for Water Penetration into Pipeline Coatings ASTM G 9-69T
Test for Rockwell Hardness of Plastics and Electric ASTM D 785-65
Insulating Materials
Test for Penetration ot Bituminous Materials ASTM D 5-65
Test for Indentation Hardness of Rubber and Plastics ASTM D 2240-68
by Means of a Durometer
Test for Shrinkage Factors of Soils ASTM D 427-61
Test for Resistance of Plastics to Chemical Reagents ASTM D 543-67
Test Method of Resistance of Plastics to Chemical General Services Adm.
Reagents, Federal Test Standard No. 406,
Method 7011
Thermal Evaluation of Rigid Electrical Insulating ASTM D 2304-68
Materials
Recommended Practice for Determining the Effect of ASTM D 2454-68
Overbaking an Organic Coatings
Test for Coatings Designed to be Resistant to Ele- ASTM D 2485-68
vated Temperatures During Their Service Life
Recommended Practice for Determining Resistance ASTM G 21-70
of Synthetic Polymeric Materials to Fungi
Test Method for Mildew Resistance of Plastics by General Services Adm.
Mixed Culture Method, Agar Medium, Federal
Test Standard No. 406, Method 6091
Military Specification and Test Method for Fungus U.S. Naval Publication
Resistance, MIL-F-8261A (WSAF)
Method of Test for Effects ot Outdoor Weathering ASTM G 11-69T
on Pipeline Coating
Test for Abrasion Resistance of Pipeline Coatings ASTM G 6--69T
Test for Bendability of Pipeline Coatings ASTM G 10-68T
Test for Adhesion of Organic Coatings ASTM D 2197-68
Test for Impact Resistance of Pipeline Coatings ASTM G 13-69T
(Limestone Drop Test)
Test for Impact Resistance of Pipeline Coatings ASTM G 14-69T
(Falling Weight Test)
Recommended Practice for Internal Coating of API RP5L2
Line Pipe
APPENDIX C
Corrosion Interference Coordinating Committees
Arizona Corrosion Correlating Council
Baltimore-Washington Electrolysis Committee
Birmingham Electrolysis Committee
Central California Cathodic Protection Committee
Central Ohio Corrosion Coordinating Committee
Chicago Area Joint Electrolysis Committee
Chicago Region Committee on Underground Corrosion
Cleveland Committee on Corrosion
Columbus and Central Ohio Committee on Corrosion
Connecticut Committee on Corrosion

Corpus Christi Coordinating Committee
Corrosion Subcommittee of Kentucky Gas Association
Dade County Utilities (Florida)
Dayton, Ohio Corrosion Committee
Denver Metropolitan Committee on Corrosion (not active)
Des Moines Electrolysis Committee
Detroit and Michigan Committee on Electrolysis
East Bay Electrolysis Coordinating Committee (Oakland, Calif.)
Eastern Montreal Electrolysis Committee
Eastern New York Corrosion Coordinating Committee
Eastern Ohio Corrosion Coordinating Committee
Eastern Pennsylvania Corrosion Committee
E1 Paso Area Corrosion Correlating Committee
Flagstaff, Arizona Underground Corrosion Correlating Committee
Greater Boston Electrolysis Committee
Greater Indiana Corrosion Committee
Greater New York Committee on Corrosion
Illinois-St. Louis Committee on Underground Corrosion
Indianapolis,Committee on Corrosion
Joint Committee for the Protection of Underground Structures in Alameda and
Contra Costa Counties (California)
Kentucky Corrosion Coordinating Committee (Kentucky Gas Association)
Lafayette, Louisiana Underground Corrosion Correlating Committee
Los Angeles, California Underground Corrosion Correlating Committee
Louisiana Coordinating Committee
Louisville Electrolysis Committee
Massachusetts Committee on Corrosion
Milwaukee Area Corrosion Committee
New Jersey Committee on Corrosion
Northeastern Ohio Corrosion Coordinating Committee
Northwest Electrolysis Coordinating Committee (San Francisco)
Northwest Electrolysis Coordinating Council (Oregon/Washington)
Northwest Pacific Electrolysis Coordinating Council (Vancouver, B.C.)
Northwest Pipe Line Operators (Oregon/(Washington)
Ok-Ark-La-Tex Corrosion Committee
Omaha and Council Bluffs Electrolysis Committee
Oregon Corrosion Committee, Dallas
Pacific Coast Gas Association Corrosion Mitigation Committee
Philadelphia Electrolysis Committee
Pittsburgh Public Service Coordination Committee
Public Utilities Commission Corrosion Committee (Ontario, Canada)
San Diego County Underground Corrosion Committee (California)
San Francisco Electrolysis Committee
Southern California Cathodic Protection Committee
Southern Idaho~Eastern Oregon Underground Corrosion Committee
Southern Ontario Council on Electrolysis Northern Technical Committee Western
and Central Committee
Southern West Virginia Corrosion Coordinating Committee
South Florida Corrosion Council
Southwest British Columbia Electrolysis Coordinating Council
St. Louis, Missouri Underground Corrosion Correlating Committee
Toledo and Northwestern Ohio Committee on Corrosion
Western Inter-Utility HVDC Committee for Earth Current and Inductive Coordi-
nation Studies
Western New York State Corrosion Committee
Western Ohio Corrosion Coordinating Committee
Western Pennsylvania Corrosion Coordinating Committee
Wyoming Underground Corrosion Coordinating Committee
APPENDIX D
Societies with Interests in Corrosion Control in tile Gas Industry
605 Third Ave.
New York, N.Y. 10016
American Institute of Chemical Engineers (AICHhE)
345 East 47th St.
American Institute of Chemists
79 Madison Ave.
American Institute of Consulting Engineers (AICE)
345 East 47th St.
American Institute of Industrial Engineers (AIIE)
345 East 47th St.
American Institute of Mining, Metallurgical, and Petroleum Engineers (AIME~
345 East 47th St.
1271 Avenue of the Americas
American Railway Engineering Association
59 East Van Buren St.
Chicago, Ill. 60605
American Society of Civil Engineers (ASCE)
Pipeline Division
345 East 47th St.
American Society of Safety Engineers (ASSE)
5 North Wabash Ave.
Chicago, Ill. 60602
American Society for Metals (ASM)
Metals Park, Ohio 44073
1916 Race St.
Philadelphia, Pa. 19103
American Water Works Association
2 Park Ave.
American Welding Society

345 East 47th St.
Asphalt Institute, The
University of Maryland
College Park, Md. 20742
British Association of Corrosion Engineers
London, England
British Cast Iron Research Association (BCIRA)
London, England
British Iron and Steel Research Association (BISRA)
London, England
Cast Iron Pipe Research Association (cIPRA)
Suite 3440, Prudential Plaza
Chicago, Ill. 60601
Copper Development Association, Inc.
405 Lexington Ave.
Electrochemical Society (Corrosion Division)
30 East 42nd St.
Federation of Societies for Paint Technology (FSPT)
121 South Broad St.
Philadelphia, Pa. 19107
Institute of Electrical and Electronics Engineers
Box A, Lenox Hill Station
Institute of Materials Research
National Bureau of Standards
Gaithersburg, Md. 20760
Manufacturers Standardization Society of the Valve and Fitting Industry
420 Lexington Ave.
National Association of Corrosion Engineers (NACE)
2400 West Loop South
Houston, Texas 77027
National Association of Pipe Coating Applicators
2504 Flournoy-Lucas Road
Shreveport, La. 71106
Office of Pipeline Safety
U.S. Department of Transportation
400 Sixth St., S.W.
Washington, D.C. 20024
Petroleum Industry Research Foundation
60 East 42nd St.
Pipe Line Contractors Association
National Bankers Life Building
202 South Ervay
Dallas, Texas 75201
Steel Structures Paint Council

4400 Fifth Ave.
Pittsburgh, Pa. 00000
Society of Consulting Corrosion Engineers
205-627 Eighth Ave.
Calgary 2, Canada
Society tor Non-Destructive Testing (SNT)
914 Chicago Ave.
Evanston, Ill. 60202
Society of Petroleum Engineers of A I M E
6300 North Central Expressway
Dallas, Texas 75206
Society of Plastics Engineers, Inc.
65 Prospect Street
Stamford, Conn. 06902
United States of America Standards Institute
(formerly American Standards Association)
10 East 40th St.
References
[1] Dueber, C. B., "The Present Status of Bacterial Corrosion Investigations in the United
States," Corrosion, Vol. 9, No. 3, March 1953. (Includes 42-item bibliography.)
[2] Hadley, R. F., "Studies in Microbiological Anaerobic Corrosion," AGA Proceedings,
1940, p. 1764.
[3] Oppenheimer, H., "The Microbial Corrosion of Iron," (Contracts) NONR-375(10),
NONR-840(21), March 1967.
[4] Kuznetson, S. I. and Pantskhava, E. S., "Increase of the Electrochemical Corrosion
Caused by Methane Generating Bacteria," Doklady Akaldemii, NAUK SSR, Vol. 139,
Feb. 1961. pp. 478-480.
[5] Muraoka J. S., "Effects of Marine Organisms," Machine Design, Vol. 40, No. 2, 1968,
p. 184.
[6] Iverson, W. P.,"Microbiological Corrosion," National Bureau of Standards, Pipes and
Pipeline Engineering Convention, 1968.
[7] Tripathi, B. N., "Microbiological Corrosion--Sulfate Reducing Bacteria and Corro-
sion Influence on Metals," Journal of Scientific and Indian Research, Vol. 24, No. 9,
1964, pp. 379-389.
[8] Iverson, W. P., "Microbial Corrosion," AD-670501, National Bureau of Standards,
April 1968.
[9] Iverson, W. P., "Anaerobic Corrosion of Mild Steel by Desulfovibrio," National
Bureau of Standards, 1969.
[10] Pomeroy, R. D., "The Role of Bacteria in Corrosion," NACE Conference, San
Diego, 1969.
[ll] Harris, J. O., "Bacteria, Oxygen and Soil Relationships in Corrosion," Collection of
Papers on Underground Corrosion, Vol. 8, Kansas State University, 1964, pp. 1-29.
[12] Costanzo, F. E.. "Weather Versus Cathodic Protection of Underground Pipelines,"
Corrosion, Vol. 14, Aug. 1958, pp. 363t-368t.
[13] Doremus, E. P., Doremus, G. L., and Parker, M. E., "Attenuation Equation Applied
to Cathodic Protection by Distributed Drainage," Corrosion, Vol. 5, No. 1, Jan. 1948,
p. 32.
[14] Doremus, E. P., Doremus, G. L., and Parker, M. E., "Engineering Aspects of Cathodic
Protection," Corrosion, Vol. 5, No. 9, Sept. 1949, p. 273.
[15] "Cathodic Protection with Zinc Anodes," Ebasco Services, Inc., Report to the Ameri-
can Zinc Institute, June 1953.
CORROSION IN THE GAS iNDUSTRY 79
[16] Ewing, S. P. and Hutchison, J. S., "Cathodic Protection Applied to Tank Bottoms,"
Corrosion, Vol. 9, No. 7, July 1953, p. 221.
[17] Gilbert, T. H., "Current Output of Galvo Pak Magnesium Anode," Gas, Vol. 26,
No. 8, 1950, p. 103.
[18] Gilbert, T. H., "Performance of Magnesium Anodes under Actual Service Condi-
tions," Gas Age, Vol. 106, 23 Nov. 1950, pp. 40-41.
[19] Osborn, O. and Robinson, H. A., "Performance of Magnesium Galvanic Anodes in
Underground Service," Corrosion, Vol. 8, No. 4, April 1952, pp. 114-149.
[20] Parker, M. E., "Attenuation Curves in Cathodic Protection," Oil and Gas Journal,
Vol. 49, No. 44, 15 March 1951.
[21] Parker, M. E., "Determination of Current Required for Cathodic Protection," World
Oil, Vol. 131, No. 6, 1950, p. 253.
[22] Parker, M. E., "Pipe-to-Soil Potentials as an Indication of Protection," Oil and Gas
Journal, Vol. 49, No. 43, 8 March 1951.
[23] Peabody, A. W., "Impressed Current Ground Beds for Cathodic Protection," Gas,
Vol. 27, No. 7, July 1951, p. 37.
[24] Peabody, A. W., "Use of Magnesium for Cathodic Protection of Pipe Lines in High
Resistivity Soil," Corrosion, Vol. 15, Sept. 1959, pp. 497t-502t.
[25] Shepard, E. R. and Graeser, H. J., "Design of Anode Systems for Cathodic Protection
of Underground and Water Submerged Metallic Structures," Corrosion, Vol. 6, No. 11,
Nov. 1950, p. 360.
[26] Standring, J. H., Corrosion, Vol. 3, No. 4, April 1947, pp. 151-154.
[27] Schwerdtfeger, W. J., "Current and Potential Relations for Cathodic Protection of
Steel in Salt Water," Journal of Research of the NBS, Vol. 60, March 1958, p. 153.
[28] Schwerdtfeger, W. J., "Current and Potential Relations for the Cathodic Protection of
Steel in a High Resistivity Environment," Journal of Research of the National Bureau
of Standards, Vol. 63C, Sept. 1959, p. 37.
[29] Dwight, H. B., "Calcu_lation of Resistance to Ground," Electrical Engineering, Vol. 55,
1936, pp. 1319-1328.
[30] Reinhold, Rudenberg, "Ground Principles and Practices," Electrical Engineering,
Vol. 64, 1945, pp. 1-13.
[31] Peabody, A. W., "Control of Pipeline Corrosion," National Association of Corrosion
Engineer& Dec. 1967, pp. 116-119 and 94-100.
[32] George, P. F., Newport, J. J., and Nichols, J. L., "A High Potential Magnesium
Anode," Corrosion, Vol. 12, Dec. 1966, pp. 627t-633t.
[33] Robinson, H. A. and George, P. F., "Effect of Alloying and Impurity Elements in
Magnesium Cast Anodes," Corrosion, June 1954.
[34] Final Report on Four Annual Anode Inspections, NACE Publication 2B156.
[35] Use of High Silicon Cast Iron for Anodes, NACE Publication 2B160.
[36] Doremus, G. L. and Davis, J. G., "Marine Anodes--The Old and New," Materials
Protection, 1967.
[37] Parker, M. E., Pipe Line Corrosion and Cathodic Protection, A Field Manual, Gulf
Publishing Co., 1962.
[38] Akimov, G. W., "Theory and Research Methods of Metallic Corrosion," Publishing
House of the Academy of Science, Moscow, 1945; translation of Chapters 2 and 3 by
R. B. Mears and J. D. Gat., Part I, Corrosion, Vol. 11, Nov. 1955, pp. 477t-486t;
Part II, Dec. 1955, pp 515t-534t.
[39] de Bethune, A. J., "Fundamental Concepts of Electrode Potentials," Corrosion, Vol. 9,
No. 10, Oct. 1953, p. 336.
[40] Brown, R. H., English, G. C., and Williams, R. D., "The Role of Polarization in
Electrochemical Corrosion," Corrosion, Vol. 6, No. 6, June 1950, p. 186.
[41] Glasstone, S., Introduction to Electro-Chemistry, D. van Nostrand Co., New York,
1942.
[42] Mears, R. B., "The Mechanism of Cathodic Protection," Metals et Corrosion, Vol. 23,
1948, p. 271.
[43] Pope, R., "Attenuation of Forced Drainage Effects on Long Uniform Structures,"
Corrosion, Vol. 2, 1946, p. 307.
[44] Reid, English, and Horst, "Elementary Mechanism of Cathodic Protection," AGA
Proceedings, p. 351.
[45] Schwerdtfeger, W. J. and McDorman, O. H., "Potential and Current Requirements for
the Cathodic Protection of Steel in Soils," Journal of Research, Vol. 47, Aug. 1951,
p. 104; reprinted in Corrosion, Vol. 8, No. 11, p. 381.
[46] Scott, G. N., "The Distribution of Soil Conductivities and Some Consequences,"
Corrosion, Vol. 14, 1958, pp. 396t-400t.
[47] Evans, S. R., Metallic Corrosion, Passivity and Protection, Longmans, Green and Co.,
New York, 1948.
[48] Ewing, Scott, Soil Corrosion and Pipe Line Protection, William J. Roth, New York,
1938. (Obtainable from American Gas Association.)
[49] Speller, F. J., Corrosion, Cause and Prevention, 2rid ed., McGraw-Hill, New York, 1935.
[50] "Underground Corrosion," Circ. 450, National Bureau of Standards, U.S. Govern-
ment Printing Office, Washington, D.C., 1945.
[51] Sunde, E. D., Earth Conduction Effects in Transmission Systems, D. van Nostrand Co.,
New York, 1948.
[52] "Cathodic Protection," National Association of Corrosion Engineers Symposium,
Houston, Texas, 1949.
[53] Uhlig, H. H., Corrosion Handbook, Wiley, New York, 1948.
[54] Schwerdtfeger, W. J., "Cathodic Protection of Copper in a Severely Corrosive Soil,"
Materials Protection, Vol. 7, Sept. 1968, p. 43.
[55] Schwerdtfeger, W. J., "Effects of Cathodic Current on the Corrosion of an Aluminum
Alloy," Journal of Research, National Bureau of Standards, Vol. 68c, Oct.-Dec. 1964,
p. 283.
[56] "Recommended Practice for Cathodic Protection," NACE Publication 2M263,
Materials Protection, Vol. 2, Oct. 1963, p. 106.
[57] Kuhn, R. J., "Cathodic Protection of Underground Pipelines Against Soil Corrosion,"
Proceedmgs, American Petroleum Institute (IV), Vol. 14, 1953, p. 153.
[58] Schwerdtfeger, W. J. and McDorman, O. N. "Potential and Current Requirements for
the Cathodic Protection of Steel in Soils," Journal of Research National Bureau of
Standards, Vol. 47, p. 104.
[59] Pearson J. M., "Electrical Instruments and Measurements in Cathodic Protection,"
Corrosion, Vol. 8 Nov. 1952, p. 391.
[60] Sudrabin, L. P. and Ringer, F. W., "Some Observations on Cathodic Protection
Criteria," NACE Publication 2C157, Corrosion, Vol. 13, Dec. 1957, p. 835t.
[61] Ewing, S. P., "Potential Measurements for Determination of Cathodic Protection
Requirements," Corrosion, Vol. 7, 1951, p. 410.
[62] Haycock, E. W., Corrosion, Vol. 13, 1957, p. 767.
[63] Pearson, J. M., "Contributions of J. M. Pearson to Mitigation of Underground
Corrosion," NACE Publication 56-12.
[64] McCollum, B. and Logan, K. H., National Bureau of Standards Technical Paper 351,
1927.
[65] Pearson, J. M., "Null Methods Applied to Corrosion Measurements," Transactions
of the Electrochemical Society, No. 81, 1942, p. 485.
[66] Kubit, R. W., "E-Log 1 Relationship to Polarization," NACE Conference, 1968.
[67] Doremus, E. P. and Canfield, T. L., "The Surface Potential Survey Can Detect Pipeline
Corrosion Damage," Materials Protection, Sept. 1967.
[68] "Underground Corrosion," Circ. 579, National Bureau of Standards, April 1957, pp.
180-186. Available from Clearinghouse, U.S. Dept. of Commerce, Springfield, Va.
22151.
STP534-EB/Nov. 1973
Chapter 4

Automobile Industry
Carl O. Durbin 1
Motor vehicles, passenger cars, and trucks are used in many parts of the
world. They are operated in all types of weather, parked outdoors when not
in use, and thus, subjected to extremes of weather; high temperature, low
temperature, rain, snow, high relative humidity, etc. They are driven over
roads which have been salted for deicing or dust control purposes and
parked in wet garages. Engine cooling systems, exhaust silencing and
related parts, hydraulic brake mechanisms, and various lubrication systems
are subjected to specific environments internally as well as the external
environments noted above.
The automobile is constructed primarily of steel and cast iron, but other
metals and alloys are used for specific parts. The general methods for
preventing corrosion are used, namely:
1. Selection of a metal or alloy resistant to a specific environment.
2. Modifying the environment by adding corrosion inhibitors or by
keeping metal surfaces dry.
3. Separating a corrodible metal from the environment with a pro-
tective coating such as paint or metallic coatings.
4. Use of sacrificial coatings or modification of electrode potential
with less noble metals.
Industry Standards
Although the standards for corrosion resistance of automobiles and
trucks are generally set by each manufacturer for his products, competitive
pressures and common suppliers have caused a similarity in choice of
materials and coatings for equivalent parts. Some differences do exist
because of differences in design. The corrosion resistance standards for a
specific company are based on the Society of Automotive Engineers
Information Reports, Recommended Practices, and Specifications tem-
pered by experiences of the engineers- of that company.
1 Chrysler Corporation, Highland Park, Mich.
81
"Prevention of Corrosion of Metals," SAE Information Report J477a

is a general discussion of corrosion and serves as a guide for automotive
designers. Other recommended practices or specifications such as tests for
" M o t o r Vehicle Lighting Devices and Components," SAE J575 and
" M o t o r Vehicle Seat Bdlt Assemblies," SAE J4 include corrosion re-
sistance requirements based on resistance to the neutral salt spray test.
These requirements are now included in M o t o r Vehicle Safety Standards
MVSS 108 [1] 2 and MVSS 209 [2] established by the National Highway
Traffic Safety Administration and are exceptions to the setting of corrosion
standards by each manufacturer for his products.
Corrosion Testing and Controls

Corrosion testing is similar to other test methods performed upon auto-
motive components in that the conditions expected to be encountered in
service are simulated. Normally this is difficult to accomplish in the labora-
tory because service environments for the automobile are complex and
differ for each automobile. Corrosion testing of automobiles and auto-
motive components, therefore, is based on an analysis of the environmental
factors which may be corrosive and the selection of those factors which will
affect the metals or metal systems being considered.
One of the simplest accelerated corrosion tests used in the automobile
industry is the neutral salt spray test which was first formally presented by
J. A. Capp in 1914 [3]. The salt spray test was widely used, even before the
method was adopted by the American Society for Testing and Materials as
Tentative Standard B 117 in 1939. This method of test has been refined and
the control of test parameters improved at various times since it was first
adopted. In 1944 the test solution was standardized at a concentration of
20 percent sodium chloride. The concentration was changed in 1953 to 5
percent sodium chloride and has remained at that percentage. The neutral
salt spray test has been useful as a first approximation for the corrosion
resistance of various assemblies such as exterior lamps and the protection
given to exterior automobile body surfaces by various combinations of
chemical surface treatments, primers, and color coatings. The test is also
considered suitable for passenger compartment hardware such as that used
for seat belts.
Prior to adoption as Tentative Standard B l l 7 in 1939, the neutral salt
spray test was used for various metallic coatings such as zinc, cadmium,
and decorative copper-nickel-chromium coatings, but the results of the test
did not correlate well with service. Copper-nickel-chromium electrodeposits
which failed early in the salt spray test would fail in service, but in many
instances coatings which were reistant to the salt spray test sometimes
failed in service in less time than in the salt spray test.
Italic numbers in brackets refer to references listed at the end of this chapter.
CORROSION IN THE AUTOMOBILE INDUSTRY 83
One of the earliest modifications of the neutral salt spray test was the
addition of acetic acid by C. F. Nixon [4]. This modification which was
adopted by ASTM B 287 Acetic Acid Salt Spray (Fog) Test produced
corrosion on copper-nickel-chromium plated zinc base die castings which
resembled that found on similar parts of cars operated in Detroit. The time
to produce this corrosion, however, was about 200 h, too long for use as a
quality control test.
Because neither the neutral salt spray test nor the acetic acid salt spray
test was considered satisfactory, the American Electroplaters' Society
Research Committee sponsored Project 15 to study accelerated corrosion
tests for electrodeposited metals. The method used for Project 15 illustrates
how suitable accelerated corrosion tests are developed to predict the service
life of automotive parts [5-11]. Project 15 was assigned to a committee com-
posed of representatives from the major automobile companies, suppliers of
plated parts, and suppliers of plating chemicals. The first phase of the
project was to establish goals, evaluate existing tests, and measure the
corrosion rate of decorative plating in Detroit by means of standard test
panels attached to taxicabs. The Detroit environment was chosen because
corrosion surveys indicated that automotive decorative plating deteriorated
in Detroit more rapidly than in most other locations in the United States.
The second phase of the project was to analyze the environment and to
determine which components were corrosive. Samples of slush were
collected from city streets in test collectors mounted on cars. Samples of
rainwater were also collected at stationary corrosion testing sites. Observers
also noted that corrosion took place more rapidly when decorative plated
parts were covered with road dirt and not washed frequently. Application
of a slurry of kaolin containing various salts permitted a systematic study of
the twenty metallic elements and chloride, nitrate, and sulfate anions found
in the slush and rainwater collections. Copper and iron salts were found to
be the most corrosive of the metallic elements present. The addition of
copper nitrate and ferric chloride to sodium and ammonium chloride in a
kaolin slurry resulted in Corrosion Testing of Decorative Chromium
Plating by the Corrodkote Procedure, ASTM B 380. Concentrations of
salts in the method were selected to produce, in 20 h, the degree of corrosion
observed in one year of atmospheric exposure on cars driven in Detroit.
The discovery that the addition of copper and iron salts to the kaolin
slurry would greatly accelerate the corrosion of decorative chromium
plated parts suggested that these ions would also accelerate the corrosion
produced in the salt spray test. The addition of copper salts was found to
greatly accelerate the acetic acid salt spray test and led to the development
of the Copper-Accelerated Acetic Acid-Salt Spray (Fog) Testing (CASS
Test) ASTM B 368. The CASS Test test produces, in 16 h, the same degree
of corrosion as the 20-h Corrodkote test. Both of these tests were instru-
mental in developing a plating procedure which greatly improved the
durability of decorative chromium plating on automobile parts [12,13].

A study of the mechanisms of corrosion of copper-nickel-chromium and
nickel-chromium coating on steel and zinc base die castings by R. L. Saur
[14] resulted in an accelerated electrolytic corrosion test (EC) which
permits predicting the durability of decorative chromium plating in min-
utes. This test is currently being studied by ASTM Committee B 8 on
Electrodeposited Coatings.
Both the CASS and Corrodkote tests have been tried on anodized
aluminum as well as stainless steel, automotive trim parts with unsatis-
factory results.
The F A C T (Ford Anodized Aluminum Corrosion Test) Testing Method,
ASTM B 538 and acid dissolution tests were developed for predicting the
durability of anodized aluminum bright metal trim. The Corrodkote test
does give an indication of the resistance to penetration of the anodic coating
resulting in white aluminum corrosion spots but does not correlate well
with service. An acid dissolution test [15] developed to correlate with the
formation of an opaque white stain on the surface of the anodic coating
consists of immersing a test sample cut from an anodized aluminum part in
a water solution containing 20 g of chromic acid (CrO~) and 35 ml of 85
percent orthophosphoric acid per liter at 100 deg F for 15 min. The re-
sistance of the anodic coating to dissolution in the acid solution is con-
sidered a measure of the resistance to blooming or white staining in auto-
motive service. This test is also being studied by ASTM Committee B 8 on
Electrodeposited Coatings. A similar test using a solution of 10 g per liter o f
anhydrous sodium sulfite per liter with the pH adjusted to 3.75 with acetic
acid and further reduced to 2.5 with sulfuric acid was also found to predict
the resistance of the anodic coating on aluminum to blooming [16,17].
Two accelerated corrosion test methods are available for stainless steel
bright metal decorative trim. Both tests use a solution containing 0.5 g o f
sodium sulfate, 0.25 g of sodium sulfite, 0.10 g of sodium thiosulfate, 52.2 g
of sodium chloride and 52.5 g of calcium chloride per liter of solution
adjusted to a pH of 9.3 4- 0.05. In one method, the dip dry test [18], the
samples of stainless steel are alternately immersed in the solution and
heated with infrared lamps for 90 s and the cycle repeated. This test requires
4 h to produce the type of corrosion observed on automotive stainless steel
trim parts. The other method [19] consists of soaking gelatin-coated
photographic paper in the test solution, application of the photographic
paper with the gelatin side in contact with the stainless steel test surface,
enclosing the stainless steel test sample with the photographic paper in a
polyethylene bag, and sealing the bag with a vapor tight seal. The sample is
heated in an oven to 215 -4- 5 deg F for 10 min. After removal from the oven
and the plastic bag, the photographic paper is immediately removed from
the test surface while still moist and immersed in a solution containing
10 g per liter of potassium ferricyanide. Corrosion products transferred to
the photographic paper are developed giving a record of the test results.
Both tests simulate the type and degree of rusting observed on cars operated
in western Pennsylvania where salt and cinders are used for deicing roads.
The two tests aided in improving the corrosion resistance of stainless steel
for automotive trim by additions of molybdenum and processing changes
to eliminate surface chromium depletion.
A number of accelerated corrosion tests are available for evaluating the
effectiveness of inhibitors in engine coolants. The simplest and most rapid
of these is the Corrosion Test for Engine Antifreeze in Glassware, ASTM
D 1384. This test is also the least reliable but is suitable for preliminary
screening of inhibitors. The other methods; Simulated Service Corrosion
Testing of Engine Antifreezes, ASTM D 2570, Recommended Practice for
Testing Engine Antifreezes by Engine Dynamometer, ASTM D 2758, and
Recommended Practice for Testing Engine Coolants in Vehicle Service,
ASTM D 2847 are more sophisticated, time consuming, and progressively
more reliable. Modifications of the glassware test" to produce erosion of
radiator brass and of the simulated service test to evaluate durability are
being used [20] but these modifications have not been adopted as ASTM
standard methods of tests.
Two specific corrosion tests; Test for Cavitation-Erosion Characteristics
of Aluminum in Engine Antifreeze Solutions using Ultrasonic Energy,
ASTM D 2966 and Test for Cavitation-Erosion Corrosion Characteristics
of Aluminum Automotive Water Pumps with Coolants, ASTM D 2809 are
useful in evaluating coolant formulations for use in engines having alumi-
num water pump components. The first of the above test methods is more
rapid but less reliable than the second.
Accelerated corrosion tests to evaluate inhibitors for various fluids and
lubricants used by the automobile industry simulate operation of the test
component under abnormal conditions which have been found by ex-
perience to be corrosive. ASTM STP 315 describes engine tests, one of
which, Sequence II, was designed to evaluate rusting and corrosion as well
as scuffing, wear and sludge and varnish deposition. In this test the coolant
temperature is purposely controlled to produce condensation of com-
bustion products which mix with the lubricant causing corrosion and
rusting of parts unless the lubricant contains suitable inhibitors. The
operating sequence simulates frequent short-trip type of operation.
Other standard tests used for evaluation of automotive lubricants are:
Test for Detection of Copper Corrosion from Petroleum Products by the
Copper Strip Tarnish Test, ASTM D 130; Test for Rust Preventing Char-
acteristics of Steam Turbine Oils in the Presence of Water, ASTM D 665;
Test for Rust Preventive Properties of Lubricating Greases, ASTM D 1743;
and Test for Rust Protection by Metal Preservatives in the Humidity
Cabinet. ASTM D 1748. ASTM D 1743 includes l-rain operation of a
roller bearing to distribute the lubricating grease on to the bearing surfaces
in a manner similar to that expected in an operating vehicle. The bearing is

then rinsed in water and stored over water in an airtight container to
simulate vehicle storage.
A form of humidity test which is frequently used to evaluate rust pre-
venting characteristics of automatic transmission and power steering fluids
is to coat steel test pieces having highly polished surfaces by immersion in
the test fluid, and to suspend the test piece over water in an airtight con-
tainer. Condensation is produced on the test piece by alternate heating and
cooling.
Tests for corrosion inhibiting characteristics of hydraulic brake fluids are
outlined in SAE Standards for Brake Fluid J1702 and J1703 published in
the SAE Handbook.
As noted earlier the Salt Spray (Fog) Testing Method, ASTM B 117, is
used for evaluation of automotive primers and paints. Other corrosion
tests used by the automotive industry for such coatings are: Water Fog
Testing of Organic Coating, ASTM D 1735; Test for Filiform Corrosion
Resistance of Organic Coating, ASTM D 2803; Testing Finishes on Primed
Metallic Substrates for Resistance to Humidity-Thermal Cycle Cracking
ASTM D 2246; and Testing of Coated Metal Specimens at 100 percent
Relative Humidity, ASTM D 2247.
Although not standardized, accelerated corrosion tests used to evaluate
exhaust pipe, muffler, and tail pipe material should be mentioned while
discussing automotive corrosion tests. Such tests are usually cyclic tests in
which the test pieces are periodically exposed to hot engine exhaust con-
densate (or a very dilute solution of sulfuric and hydrobromic acids) by
partial immersion followed by heating to a temperature in the range of 500
to 1000 deg F. The higher temperature is used for components located
close to the engine and the lower temperature for the components farthest
from the engine. The immersion in engine exhaust condensate simulates the
condensation which occurs during short-trip driving and exposure to high
temperature simulates the heat encountered during highway driving. The
test parameters are arbitrarily chosen tO suit the convenience of the labora-
tory doing the testing and have given reasonably good results when com-
paring various materials. The aforementioned test simulates the internal
environment of muffler and tailpipe components. Tests for external cor-
rosion of mufflers and tailpipes must include scaling tests conducted by
heating test specimens in a furnace to the maximum expected temperature
and periodically removing them from the furnace and spraying while still
hot with a dilute salt solution. This test simulates operation at high speeds
over salted and slushy roads.
Occasionally a test is required which will evaluate a particular property
of a metal, such as susceptibility to cracking when stressed during exposure
to a specific environment. An example is the observed stress corrosion
cracking of brass in an atmosphere containing ammonia or organic amines.

The test commonly used for evaluating the susceptibility of brass cracking
is the Mercuric Nitrate Test. This test consists of immersing the brass part
in a solution containing 1 percent nitric acid and 1 percent mercuric
nitrate for 1 h and then examining the part for cracking. Immersion in a
boiling solution of magnesium chloride is a similar type test used for
stainless steel.
Exposure of test panels on outdoor static and mobile sites, although not
an accelerated test, is used for evaluating exterior bright metal trim and
organic automotive finishes. Static sites are primarily located in Detroit,
near New York City, in Florida, and in Arizona. Mobile sites consist of
racks mounted on trucks or under the bumper on passenger cars. This type
of testing provides more reliable results than obtainable form accelerated
corrosion tests and are less expensive and more consistent than testing on
completed vehicles.
Electrochemical techniques, such as anodic and linear polarization
measurements have been used to study corrosion of automobile trim
materials [21,22] and to measure instantaneous corrosion rates in a simu-
lated engine cooling system [23]. As noted earlier, an EC test was devised
to test the durability of copper-nickel-chromium electrodeposits on bright
metal trim. This test has been proposed and is under consideration by
ASTM Committee B 8 on Electrodeposited and Related Coatings for
adoption as a standard method of test. Linear polarization measurements
can be used to evaluate test results in a similar manner to weight loss
measurements, thickness loss measurements, or degradation of appearance.
The final test for automobile corrosion is the test of the assembled
vehicle. There are probably as many variations of accelerated corrosion test
procedures for the complete motor vehicle as there are project engineers
supervising such tests, and variations may also be introduced because of
interest in a specific component. The test procedures usually contain
repeated cycles, each cycle consisting of various combinations of operation
over salted gravel roads, through splash troughs containing dilute salt
solutions, over rough roads, parked in humid locations, or exposed in
large salt spray rooms. Such cycles are ueful for evaluating body corrosion
[24,25] as well as the durability of various other components.
"An Appraisal of the Problems of Accelerated Testing for Atmospheric
Corrosion" prepared by the International Electrotechnical Commission
Technical Committee No. 50 Environmental Testing calls attention to
many of the problems which are associated with accelerated testing for
atmospheric corrosion. These problems also apply to accelerated corrosion
testing for automobiles which is further complicated by variations in
operation and usage. Corrosion surveys of rental cars and customers' cars
in parking lots conducted annually show how well the accelerated corrosion
tests predict durability.
References
[1] Federal Register, Vol. 37, No. 7, 12 Jan. 1972, p. 22905.
[2] Federal Register, Vol. 36, No. 232, 2 Dec. 1971, p. 22951.
[3] Proceedings, American Society for Testing and Materials, Vol. 14, Part II, 1914, p. 474.
[4] Nixon, C. F., American Electroplaters' Society Monthly Review, Vol. 32, No. 11, 1945,
p. 1105.
[5] LaQue, F. L., Plating, Vol. 39, 1952, p. 65.
[6] Pinner, W. L., Plating, Vol. 40, 1953, p. 1376.
[8] Pinner, W. L., Proceedings of the American Electroplaters' Society, Vol. 43, 1956, p. 50.
[10] Pinner, W. L., Proceedings of the American Electroplaters' Society, Vol. 47, 1761, p. 27.
[11] Snyder, W. and Saltonstall, R. B., Plating, Vol. 54, 1967, p. 270.
[12] Bigge, D. M., Proceedings of the American Electroplaters' Society, Vol. 46, 1959, p. 149.
[13] Nixon, C. F., Thomas, J. D., and Hardesty, D. W., Proceedings of the American
Electroplaters' Society, Vol. 46, 1959, p. 159.
[14] Saur, R. L. and Basco, R. P., Plating, Vol. 53, 1966, p. 35.
[15] Manhart, J. H. and Cochran, W. C., Plating, Vol. 58, 1971, p. 219.
[16] Kape, J. M., Metallndustry, Vol. 95, No. 6, 18 Sept. 1959, p. 115.
[17] Anderson, J. L., Metal Progress, Vol. 95, No. 6, June 1959, p. 92.
[18] Fowler, R. W. and Bishop, C. R., Metal Progress, Vol. 84, No. 88, Sept. 1963.
[19] Anon., "New Corrosion Test for Stainless Steel," Metal Finishing, Vol. 61, No. 67,
May 1963.
[20] Levy, G. G., "Development of a High Temperature-Pressure Test for Evaluating
the Life of Antifreezes," presented to the Chemical Specialities Manufacturers' Associa-
tion, 8 Dec. 1971.
[21] Walker, M. S. and Rowe, L. C., Corrosion, Vol. 25, Feb. 1969, p. 47.
[22] Baboian, Robert, "Clad Metals in Automotive Trim Applications," Society of Auto-
motive Engineers paper 710276, Automotive Engineering Congress, Detroit, Mich.
11-18 Jan. 1971.
[23] Walker, M. S. and France, W. D., Jr., Materials Protection, Vol. 8, No. 47, Sept. 1969.
[24] Durbin, C. O., "Manufacturing Process for Corrosion Control," Society of Auto-
motive Engineers, paper 535D, Summer Meeting, Atlantic City, N.J., 11-15 June, 1962.
[25] Wiegand, R. S. and Schrock, R. E.. "Analysis and Control of Automobile Body
Corrosion," paper 77, 1968 Conference, Cleveland, Ohio, National Association of
Corrosion Engineers.
STP534-EB/Nov. 1973
Chapter 5

Pipeline Industry
A. W. P e a b o d y I
The discussion in this chapter concerns itself with any pipeline used for
the transportation of materials. Materials involved can include, but are not
necessarily limited to, fuel gas, petroleum and petroleum products, water,
chemical solutions, effluents from manufacturing plants, and solids being
transported in slurry form. Our consideration of corrosion will concern
itself with external pipeline corrosion which may occur underground, in a
submerged condition, or exposed to the atmosphere. Internal corrosion
will be considered for those pipelines carrying material which can be
expected to cause internal corrosion under specific conditions. Discussion
of corrosion will concern itself with metallic pipelines as opposed to non-
metal pipelines (used in some instances) which are subject to deterioration
with time rather than corrosion with time, which is taken as a function of
metals only.
Upon consideration of the many materials being transported by pipeline
within the pipeline industry, it can be readily understood that the amount
of existing pipelines throughout the industry involves a tremendous invest-
ment in effort and capital. The magnitude of pipeline projects continues to
increase. As an example of this, a single specific project in the planning
stage, as this chapter is written, involves the projected construction of a
major high-pressure gas pipeline from the northern coast of Alaska to the
northern portion of Central United States. As an example of the magnitude
of this project, the amount of steel presently planned for installation in the
pipeline alone will involve on the order of 1 500 000 tons of steel. The final
cost of the overall project is currently estimated at $5 billion.
As the size of individual pipeline construction projects increases, and as
their location in inaccessible areas becomes more common (such as through
the Arctic wastes, or under deep marine conditions) the need for adequate
corrosion control becomes more and more important. This is associated
with the increasing cost of a corrosion failure in terms of cost of product
1 Ebasco Services Incorporated, N e w York, N.Y.
89
University of Washington (University of Washington) pursuant to License Agreement. No further reproductions author
lost and cost of effecting repairs. Further, if the failure is catastrophic in

nature, the cost of interruption of pipeline throughput will be a major
additional cost factor.
General
Types of Corrosion Problems Encountered
The pipeline industry is confronted with a wide range of corrosion prob-
lems. External surfaces of pipelines in contact with soils, waters, or an
atmospheric environment, and the interior surfaces of pipelines which are
in contact with a potentially corrosive material being carried by the pipe-
line, are subject to most of the basic corrosion processes and most of the
forms of corrosion attack which have been covered in detai I in Chapter 1
of this book. In addition to the general applicability of basic corrosion
processes and forms of attack, the pipeline industry is faced with corrosion
problems which are peculiar to the industry. Some of these are discussed in
detail in a later portion of this chapter.
Stray Current Corrosion--A form of corrosion attack which can occur on
pipelines, but which is not covered in detail in Chapter 1, is categorized as
stray current corrosion. By stray current, is meant a flow of direct current
in the earth from some outside source (as opposed to corrosion current
resulting from an environmental condition on the pipeline external or
internal surface) which results ha collection of the stray direct current by
the pipeline from the earth in one area, and discharge of the current back
to earth at some other area on the pipeline which may be closer to the
ultimate destination of the current flow pattern. Examples of stray current
sources include direct-current-operated transit systems; mining operations
utilizing direct current (such as hauling equipment and mining machines);
direct-current welding operations; industrial processes using direct current;
and cathodic-protection systems installed on other structures (including
pipelines of other ownership) for corrosion control but which may, if
improperly installed, cause stray current effects upon neighboring pipe-
lines. In addition to the man-made sources of stray current itemized, there
may, under certain conditions, be a corrosive effect exercised by so-called
telluric currents. Telluric currents, or earth currents, are a natural phe-
nomenon caused by disturbances of the earth's magnetic field which result
in induced direct currents on pipelines located in the areas where the
magnetic disturbances occur. No matter what the source of the stray
direct current may be, there is no damage, normally, where the current is
picked up from the earth by the pipeline, but where this same current is
discharged back to earth to continue its journey to its source, corrosion
attack does occur. The magnitude of currents involved under stray direct-
current conditions can easily be far greater than currents of natural origin
resulting from the corrosion processes discussed in Chapter 1, with the
University of Washington (University of Washington) pursuant to License Agreement. No further reprodu
CORROSION IN THE PIPELINE INDUSTRY 91
result that corrosion failures can occur within a relatively short period
of time.
The above discussion has been directed toward the corrosive effects of
stray direct current. Alternating current, as is associated with electric power
system transmission lines, can likewise become stray and use parallel pipe-
lines as a flow path. Alternating current at the usual commercial frequency
(60 Hz) does not have as great a corrosive effect per unit amount of current
as does direct current. For a given amount of alternating current, the
corrosive effect is but a small percentage of the corrosive effect of a like
amount of direct current. Nevertheless, with pipelines built on the same
right of way and closely parallel to high-voltage alternating-current trans-
mission lines, the amount of alternating current flowing on the parallel
pipelines may, under certain conditions, be so great that even a small
percentage factor in terms of direct-current equivalents can nevertheless
cause serious corrosion attack on the pipeline, and in addition, can con-
stitute a serious personnel hazard (if not properly provided for) because of
high pipeline-to-earth potentials often associated with the induced alter-
nating currents.
Within the past few years, there has been the development of the use of
high-voltage, direct-current, electric transmission line systems which have
been used for the first time in place of high-voltage, alternating-current
electric transmission systems. These direct-current transmission lines
(known as HVDC systems) involve an additional source of possible stray
current corrosion on pipeline systems. Under present concepts, HVDC
systems involve transmission of bulk electric power between terminals
which may be several hundred miles apart. At each of the terminals there
is a high-capacity grounding electrode through which unbalanced system
current can flow to or from the earth. Under conditions of unbalance on
the HVDC lines, the magnitude of this unbalance current can be quite
great, and it will be flowing, as stray current, through the earth along the
possibly several-hundred-mile-long path between terminals. The worst
condition develops during operation of the HVDC system under emergency
conditions with one overhead conductor completely inoperative; full load
current then flows through the earth path and through the remaining over-
head conductor. Pipeline systems in the vicinity of the terminals will be
subject to possible stray current pickup or discharge, depending on the
nature of the H V D C transmission line unbalance condition. With im-
properly designed terminal equipment, or for pipeline systems located too
close to HVDC system terminals, pipeline corrosion can be severe. Although
the use of HVDC electric transmission systems is presently quite limited,
the concept involved is getting greater acceptance with time, and inter-
ference from this type of system may ultimately become more prevalent.
Methods of Pipeline Corrosion Control

The material under this heading will be concerned with an outline dis-
cussion of the more c o m m o n means of corrosion-control procedures as
they apply to the pipeline industry.
Coatings--The first approach to corrosion control used in the pipeline
industry involves the use of coatings. It was early recognized that the
corrosion process was of electrochemical origin, and it was quite logically
reasoned that if the pipeline metal could be electrically isolated from its
environment, there could no longer be a flow of current between separate
points on the surface of the pipe and that, as a result, corrosion would be
eliminated. This would be perfectly true if an electrically insulating coating
could be installed that was 100 percent perfect at the time of installation,
and which could be maintained in a 100 percent perfect condition for the
life of the pipeline. This is not a practical possibility under the usual pipeline
construction conditions. As a result, there will be a certain number o f
coating defects where the steel will be exposed to the pipeline environment.
Although corrosion may be still stopped on better than 99 percent of the
pipeline surface, any corrosion current that does flow will be concentrated
at coating defects, and the rate of corrosion at these points may be such
that pipeline penetration will occur at an earlier date than would have
been the case had the pipeline been left bare (assuming coating to be the
only corrosion control method used).
Cathodic Protection--Once it was discovered that the use of coatings
would not provide the total answer to pipeline corrosion, the technique o f
cathodic protection was introduced. This is an electrical method of com-
bating corrosion in that any corrosive currents caused by contact between
the pipeline and an electrically conductive ionic environment are prevented
from discharging from the pipeline to the environment (with attendant
corrosion) by nullifying them with a superimposed direct-current flow from
an external source. When accomplished, as discussed in Chapter 1, the
entire metallic structure being protected is collecting direct current from its
environment. By so doing, the entire structure is forced to become a
cathode in the electrical circuit (hence, the name of the corrosion control
method) and, if the condition is fully satisfied, the corrosion will stop.
Although cathodic protection can be applied to bare pipelines, the a m o u n t
of current required may become so great in the ease of large-diameter long
pipelines that, in addition to the number and complexity of current sources
necessarily installed along the pipeline, there will be a possible problem
with stray current from the high-capacity cathodic protection systems
causing corrosion on adjacent underground metallic structures of other
ownership. As a result of this, time has proven that the best practical com-
bination for use in the pipeline industry under normal conditions is the use
of a combination of pipeline coating and supporting cathodic protection.
As was indicated earlier, a reasonably well-applied coating can be expected
to protect better than 99 percent of the pipeline surface. With the support-
ing cathodic protection, current from the cathodic protection system need
flow only to the less than 1 percent of the pipeline surface which involves
bare pipe metal contacting the adjacent earth. As a result, single cathodic
protection installations can, under normal circumstances, protect many
miles of big-inch coated pipeline with a minimal amount of current. The
combination of the two methods (coatings and cathodic protection) can
result in a very high degree of corrosion control on a given section of
pipeline.
Stray Current Control--The usual cathodic protection system may not
necessarily control high-intensity stray currents from man-made sources of
stray direct current. Other methods of control are often required. Since, as
discussed, the stray current is simply using the pipeline as a convenient link
on which to travel along its path between two points, the major amount of
corrosion damage is normally concentrated in the vicinity of the area where
the current leaves the pipeline to return to its source (such as the negative
bus of a transit-system direct-current substation). Where the distance
between the pipeline and the current source is not too great, a common
and convenient method of controlling the corrosion is to install a metallic
bond between the pipeline and the negative bus of the current source.
Current which is removed from a pipeline entirely by way of a metallic
path does not corrode the pipeline. Refinements to the simple bond approach
may be necessary, particularly in the case of transit systems where there
may be more than one direct-current substation involved. This is because,
as load conditions vary, a given substation may not continue to collect
current through the bond, but at times may tend to permit current to flow
from its negative bus, through the bond, back to the pipeline, and hence to
another area of discharge where the current can cause corrosion to occur.
Under such conditions it is necessary to install blocking devices which will
permit the current to flow in one direction only through the bond--and that
is toward the negative bus of the current source only. Many refinements of
the above basic procedure have been used but it serves to illustrate the
basic need.
In the case of HVDC electric transmission systems, discussed earlier, the
solution is not as readily arrived at as is the case with the more usual stray
current problem. This is because a given HVDC terminal may be either
discharging direct current or picking up direct current, depending upon the
system unbalance conditions at a given moment. This eliminates the
simple bond approach (with or without current blocking devices) which
often works effectively in the case of the other types of d-c interfering
systems discussed, where the polarity remains constant. By cooperative
efforts with the builders of the HVDC systems, installation of HVDC
terminal grounds should be located a sufficient distance from the nearest
pipeline systems to minimize the amount of stray-current pickup by the
pipelines, assuming that the terminal ground is properly designed. This

procedure is practical since the H V D C method is just becoming established,
and pipeline system operators are in a position to work with the H V D C
system designers to protect their interests. With sufficient separation
between the H V D C system terminal and the nearest pipelines, the a m o u n t
of stray-current pickup can be reduced to the point where it can be over-
come by normal cathodic protection systems. Once H V D C terminals are
established, no builder of a new segment of pipeline should permit the line
to be built so close to an H V D C terminal that a strong stray-current effect
exists.
The stray direct-current effects, resulting from disturbances of the earth's
magnetic field, are of such an erratic and unpredictable nature that there
have not been any hard and fast procedures established for overcoming
their effects. Experience has indicated that they are of a transitory nature
and that they do not necessarily occur in exactly the same location each
time the effect becomes apparent. Because of this, the amount of corrosion
damage which can be attributed to the phenomenon does not appear, in
most cases, to be established as a significant factor in the overall corrosion-
control program.
Internal Corrosion--Although external corrosion is a problem through-
out the length of any buried or submerged pipeline, internal corrosion is a
problem only if the material being carried by the pipeline is of a corrosive
nature. Where the material being carried by a pipeline is determined to be
of a corrosive nature, it may be possible to treat the material to inhibit the
corrosion properties. Where inhibitors are used, it is necessary to monitor
the effectiveness of the corrosion-control method by, for example, use of
internal coupons which can be examined at intervals in order to evaluate
the effectiveness of the treatment method.
In some instances where treatment is not possible, the use of an internal
coating system may be resorted to. In petroleum or petroleum products
lines, for example, it is possible to apply paint coatings in place on existing
pipelines. Although arJy paint or coating system applied internally may not
be 100 percent effective, it will nevertheless materially reduce the amount
of internal pipeline surface which is directly affected by the material being
carried. In the case of water pipelines, good experience has been obtained,
where internal corrosion is a significant problem, by the use of linings of
cement mortar which can be applied to the pipelines in place. In other
situations, where conditions warrant it, good experience has been obtained
with inserting plastic liners (of a type which are unaffected by the material
being carried in the pipeline) inside the original metallic pipe when it
approaches the point of becoming unserviceable because of internal
corrosion. This approach is normally applicable to lower-pressure pipeline
systems only.
Atmospheric Corrosion--Coatings are normally relied on for controlling
corrosion of external pipeline surfaces exposed to either aggressive industrial

atmospheres or marine atmospheres. The coatings selected must be suitable
for reasonably long-term performance under the environmental conditions
of exposure. Even though cathodic protection may be applied to the ex-
ternal surfaces of underground or submerged portions of the same pipeline,
this cathodic protection has no effect upon the part of the pipeline exposed
to atmosphere because there is no conducting medium surrounding the pipe
in atmosphere to conduct protective current to the pipe surface. It is for
this reason that coatings only must be relied upon. Coating maintenance
must be performed periodically in order to avoid progressive corrosion
damage at defects in the coating which will practically always develop
during the usual interval between maintenance inspections.
Selection of Materials During Design--In some instances it is possible to
eliminate corrosion problems in the design stages of a low-pressure pipeline
by eliminating metal as a material of construction. Where conditions are
known to be aggressively corrosive to metals, where a substitute material will
be adequate from the mechanical standpoint, and where the substitute
material will be competitive on an overall cost basis, the use of a non-
metallic material may be a good choice. Materials which are used in the
pipelining industry in lieu of metals include, asbestos-cement, reinforced
concrete, plastic, and filament wound reinforced plastic pipe which, with
development, is finding usage in larger and larger sizes and at higher and
higher pressures and temperatures.
Industry Standards and Sources of Information

In the pipeline industry, there are relatively few industry standards which
relate directly to corrosion. There are a number of associations and organi-
zations, as given below, through which information on pipeline corrosion
may be obtained. Under the following headings are given either the appro-
priate corrosion standards or the nature of information which can be
obtained from the organizations listed.
This organization is directly concerned with corrosion of metals in all
applications. There is considerable attention given to the pipelining ind.ustry
in NACE. It is suggested that since standards and guides are published at
intervals, getting, from NACE, an up-to-date list of information available
is desirable in the event of anyone's wishing to become familiar with infor-
mation available through NACE at any time following publication of this
manual.
Standards--The following standards have been published by N A C E
having an application to pipeline corrosion control.
RP-01-69 (Rev. l)--This is a recommended practice published by NACE
and titled "Control of External Corrosion on Underground or Submerged
Metallic Piping Systems." This recommended practice addresses itself to

recommended methods for pipeline corrosion control as well as criteria
and test methods by which the corrosion control system may be evaluated.
RP-05-72--This recommended practice titled, "Design, Installation,
Operation and Maintenance of Impressed Current Deep Groundbeds," is
related to one type of cathodic protection system ground bed (for use with
impressed current systems) which finds its major application on pipelines.
TM-01-72--This test method titled, "Antirust Properties of Petroleum
Products Pipeline Cargoes," is concerned with measurement of inhibitor
effectiveness in preventing internal surface corrosion in pipelines carrying
petroleum products.
Technical Committees--The following technical committees operating
within the framework of NACE are organized to direct their attention to
corrosion control matters relating to the pipeline industry.
Technical Practices Unit Committee T-3P--Internal Corrosion of Prod-
uct Pipelines and Tanks.
Technical Practices Committee T-6--Protective Coatings and Linings.
Technical Practices Committee T-10--Underground Corrosion Control.
Publications--The following publications of NACE are directed toward
pipeline corrosion control. Following publication of this manual, it is
recommended that the reader contact NACE for an up-to-date publication
list should he be interested in information of this nature.
Book--Control of Pipeline Corrosion.
American Society for Testing and Materials
Standards available through ASTM currently limited to those standards
relating to the testing of pipeline coating materials used for corrosion
control. These standards are as follows:
ASTM Designation Title
G 6-72 Standard Method of Test for Abrasion Resistance o f
Pipeline Coatings
G 7-69T Tentative Recommended Practice for Atmospheric
Environmental Exposure Testing of Non-Metallic
Materials
G 8-72 Standard Methods of Test for Cathodic Disbonding of"
Pipeline Coatings
G 9-72 Standard Method of Test for Water Penetration Into
Pipeline Coatings
G 10-72 Standard Method of Test for Bendability of Pipeline
Coatings
G 11-72 Standard Method of Test for Effects of Outdoor
Weathering on Pipeline Coatings
G 12-72 Standard Method for Non-Destructive Measurement

of Film Thickness of Pipeline Coatings on Steel
G 13-72 Standard Method of Test for Impact Resistance of
Pipeline Coatings (Limestone Drop Test)
G 14-72 Standard Method of Test for Impact Resistance of
Pipeline Coatings (Falling Weight Test)
G 17-71T Tentative Method of Test for Penetration Resistance of
Pipeline Coatings
G 18-71T Tentative Methods of Test for Joints, Fittings, and
Patches in Coated Pipelines
G 19-71T Tentative Method of Test for Disbonding Character-
istics of Pipeline Coatings by Direct Soil Burial
G 20-71T Tentative Method of Test for Chemical Resistance of
Pipeline Coatings
AGA does not issue pipeline corrosion control standards as such, but
does maintain an active corrosion committee through which information
relating to pipeline corrosion control may be obtained. Information may
be obtained from AGA by directing inquiries to the association head-
quarters to the attention of the Corrosion Committee Chairman.
American Society of Mechanical Engineers
ASME publishes the following codes which are, in part, related to cor-
rosion control in the pipeline industry.
A S M E CO DE
Boiler and Pressure Vessel Code
Section VII, Pressure Vessels--Division 1 and 2
Section IX, Welding Qualification
American Petroleum Institute
There are no known codes published by API which are directly related to
the control of pipeline corrosion. API may, however, be contacted for in-
formation pertaining to pipeline corrosion control in the petroleum industry.
American Water Works Association
The information published by American Water Works Association
pertains primarily to the application of coatings used in the water pipeline
industry.
National Association of Pipe Coating Applicators
This organization may be contacted for information relative to effective
application of pipeline coatings for optimum coating performance on
pipelines.
University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authoriz
Corrosion Coordinating Committees

Corrosion Coordinating Committees (also sometimes known as elec-
trolysis committees) are regional organizations which have been set up to
coordinate corrosion control problems (particularly on pipelines) involving
underground metallic structures of different ownership. These organiza-
tions normally consider each new installation of cathodic protection or
stray current drainage facilities. Information on each case directed to their
attention is disseminated to all members of the committee. The cases are
considered at periodic meetings of the committee and are closed when all
interested parties have indicated their satisfaction with cooperative tests
made with the owner of the installation necessitating the test. Should any
reader wish to obtain a current listing of active corrosion coordinating
committees, this information may be obtained through the National
Association of Corrosion Engineers.
Governmental Regulations
Federal Government
There are two sets of minimum federal safety standards applying to the
pipeline industry which contain subparts which apply to pipeline corrosion
control. Both of these standards are administered by the Office of Pipeline
Safety (OPS) of the Department of Transportation (DOT) of the Federal
Government. The applicable minimum Federal Safety Standards are as
follows:
1. Part 192, Title 49, Transportation, Code of Federal Regulations,
"Transportation of Natural and Other Gas by Pipeline: Minimum Federal
Safety Standards," that stipulates the minimum requirements relating to
the transportation of hazardous gases.
Subpart I is that portion of Part 192 which stipulates the minimum
requirements for gas pipeline corrosion control. Part 192 became effective
1 August 1971.
2. Part 195, Title 49, Transportation, Code of Federal Regulations,
"Transportation of Liquids by Pipeline." Sections 195.236, 195.238,
195.242, and 195.244 in Subpart D (Construction) pertain to pipeline
corrosion control. Sections 195.414, 195.416, and 195.418 in Subpart F
(Operation and Maintenance) also pertain to pipeline corrosion control.
State Governments
Any state may issue corrosion control regulations which supplement those
pertaining to pipelines as issued by the Federal Government. Any state is
free to issue regulations pertaining to pipeline corrosion control which
differ from the Federal regulations, provided that they in no way weaken
the provisions in the Federal regulations. Should the reader desire infor-
mation relative to the existence of pipeline corrosion control regulations
issued by any particular state, he can obtain that information by addressing

his inquiry to the Public Service Commission of the state involved.
Specific Pipeline Corrosion Control Problems
Under the following headings will be discussed some of the problems
which face pipeline corrosion engineers in applying satisfactory corrosion
control solutions on their systems.
Pipelines in Highly Congested Areas
All new pipelines which are to carry hazardous gases or liquids must be
coated and cathodically protected in order to comply with minimum
Federal Safety Standards. This applies as well to existing pipelines (except
that if existing pipelines are bare, they need not be excavated and coated).
Where cathodic protection is to be applied in areas where the pipelines to
be protected are closely adjacent to other underground structures, the
problem of getting sufficient cathodic protection current to the pipeline in
question can become rather acute. Whereas pipelines in open country
(particularly if well coated) can be protected with cathodic protection
installations at widely spaced intervals along the pipeline, this type of
installation is seldom effective in highly congested areas. In these areas, it
is usually necessary to install distributed anode systems for either impressed
current systems or galvanic anode networks. The distributed anodes need
to be placed at frequent intervals along the pipeline being protected through
the congested area and so placed that current from any one anode will be
able to reach all pipeline surfaces up to the midpoint of the pipe section to
the next adjacent anode on either side. Such installations of distributed
anodes are complicated as compared to the widely spaced concentrated
current source systems (used in open country). These complicated systems
involve careful maintenance and careful testing to make certain that current
from any anode is not interfering with (and causing possible damage to)
metallic structures adjacent to the pipeline being protected.
Protecting bare pipelines in highly congested areas is more difficult than
protecting well coated pipelines in the same areas. This is because the bare
pipeline requires a far greater amount of current. Compared to the well
coated pipe, many high current capacity sources are needed. Additionally,
there is a greater possibility of stray current from the anode system inter-
fering with and possibly causing damage to adjacent structures.
The design of a cathodic protection system for pipelines in highly con-
gested areas requires a high level of expertise and should be attempted only
by those who are qualified by experience and training in the design of such
systems.
Protection of lnternal Pipe Surfaces
Where pipelines are required to carry highly aggressive aqueous solutions,
some installations may prove to be most economical if steel pipe is used
provided the internal surface can be protected against corrosion as well as

the outside. Where the aggressiveness of the material carried is such that
coatings alone cannot be relied upon, it may be necessary to design cathodic
protection systems for the internal pipe surfaces. A combination of an
excellent coating properly applied together with a long life cathodic pro-
tection system can make it possible to utilize steel satisfactorily in applica-
tions of this nature. It should be noted that even though a pipeline in this
service may have cathodic protection applied to its external surfaces, this
external protection system has no beneficial effect upon the internal
surfaces of the same pipeline. For this reason it becomes necessary to
design a completely separate system for the internal surfaces. One solution
to the installation of internal cathodic protection in conjunction with
coatings is to use strip galvanic anodes of zinc which may be placed on the
inside of the pipe along the pipe bottom. Connections between anode and
pipe are made at periodic intervals (even as close as each welded pipe joint).
To use this type of installation, the current requirement of pipe steel coated
with the coating material being used and the particular aqueous environ-
ment must be known with a reasonable degree of accuracy. With this
information available, it is possible to design installations which will have
an expected life of the same order as the expected economic life of the
pipeline system on which it is being installed. Zinc is at present the pre-
ferred material for use in such installations because it has the highest
current efficiency making it possible to design for long life installations.
The use of internal galvanic anodes as described is the preferred solution,
where it can be shown to be effective, because once installed it requires the
least amount of maintenance and is least apt to become ineffective at any
time.
Another solution involves the use of impressed current systems. This
requires the installation of through-wall impressed current system anodes.
These anodes must be electrically insulated from the pipeline steel and must
be placed close enough together so that the areas protected by adjacent
anodes will overlap each other. Systems of this type are complicated (as
compared to galvanic anode systems described above) because of the need
for interconnecting cables to feed power from an impressed current power
source to each anode. Such installations require a high degree of mainte-
nance in that short circuits developing between anodes and pipe steel can
cause the power source to trip off the line. Trouble can also be experienced
with interconnecting cables which may corrode through defects in the
cable insulation or which may be broken by construction activities on
other facilities. The impressed current internal anode system, however, has
the advantage of higher levels of current output should it be needed in
connection with an internal pipe surface which has a relatively poor coating
or no coating at all.
Pipeline Protection in Frozen Earth

With the development of pipeline construction projects in Arctic areas,
the problem of providing satisfactory corrosion control for pipelines com-
pletely buried in frozen earth (permafrost) has arisen. Evidence indicates
that in some types of permafrost soils, corrosion can continue on steel in
the permafrost although at a slower rate than would be the case if the pipe-
lines were in the same material unfrozen. The situation is complicated by
the probability that any pipeline built in a permafrost region will pass
through areas of unfrozen soil inclusions such as at thermokarst lakes and
under rivers and streams where the water depth is sufficient to maintain an
unfrozen layer beneath the deep ice. Such situations create relatively small
anodic areas (at the unfrozen inclusions) which tend to discharge current
to large areas of cathodic pipe in frozen earth. This can lead to more
rapid corrosion of that portion of the pipeline in the unfrozen inclusion
because of the small anode--large cathode effect.
Corrosion control by application of a suitable coating for frozen condi-
tions plus cathodic protection appears to be the desirable protective combi-
nation for pipelines in permafrost areas. The coating used must be selected
for adequate handling and application characteristics under extreme low
temperatures as pipeline construction in certain types of permafrost areas
necessitates winter construction in order to permit adequate bearing for
construction equipment. The application of cathodic protection poses
problems with ground bed anodes for use with impressed current systems
for the pipeline in that the anodes may have to be buried in frozen earth.
Methods are being developed for coping with this situation. Where im-
pressed current cathodic protection system ground bed anodes can be
placed in nonfrozen earth (such as at the bottom of thermokarst lakes),
adequate protective current can be discharged from the anodes.
Pipelines which are installed in permafrost areas and which can be
allowed to freeze in place can be expected to require substantially less
cathodic protection current than would the same pipeline with the same
coating in unfrozen earth.
Shorted Casings
Where a pipeline passes through a casing at a road crossing, a railroad
crossing, or at other locations where required, good practice calls for having
the carrier pipe electrically insulated from the casing pipe. Should a short
circuit exist between the pipe and casing in any instance, this condition
makes it impossible for externally applied cathodic protection to reach the
carrier pipe inside the casing. This is because the shorted casing intercepts
the cathodic protection current and carries it to the pipe through the short
circuit connection.
Where shorted casings exist, steps should be taken to remove the short
I0"~ iNDUSTRIAL CORROSION STANDARDS AND CONTROL
circuit or to otherwise provide for full corrosion control of the carrier pipe
inside the casing.
Where the mechanical short circuit cannot be cleared, one procedure is to
remove any water from the annular space between pipe and casing and fill
this space with an inhibited casing filler such as an inhibited petroleum
jelly. This material prevents the entry of soil moisture and debris which
might otherwise cause corrosion of the carrier pipe.
A shorted casing which is poorly coated or bare on an otherwise well
coated pipeline, creates a load on any cathodic protection system on the
pipeline. Since the current required by one bare casing can easily require as
much current as many miles of coated pipeline, it is important that where
possible the short circuit be removed between casing and pipe rather than
using the casing filling technique as described above.
Coating Selection
Matter of coating selection and application is one of the most important
matters pertaining to adequate corrosion control systems on pipelines. It is
also one of the more abused corrosion control methods. Part of this stems
from inadequate selection of the best coating for a particular application
and part of it stems from improper application procedures.
There are many coatings available for use on pipelines and it is essential
that the coating selected for a particular application be compatible with
environmental conditions along the route of the pipeline. Although no
attempt will be made here to give any guidance on which coating to select,
it should be noted that the best coating for a particular application will be
that coating which is most stable throughout its useful life. By stability we
mean the obtaining of a reasonably high electrical resistance at time of
installation and retaining a high value of insulation throughout the pipe-
line life with the least practicable reduction in resistance with time.
No matter how carefully a pipeline coating has been selected, it will not
perform properly unless it is applied over well-prepared pipe surfaces and in
complete accordance with the manufacturer's recommendations. Applica-
tion procedures for any pipeline project should be complete in all respects
regarding pipe cleaning and coating application procedures. These appli-
cation specifications should be backed up by thorough inspection during
the pipeline construction project.
Industrial or Marine Atmospheric Exposure

Where pipelines come above ground in areas where they are subject to a
highly corrosive industrial or marine atmospheric environment, the pipeline
and its appurtenances may be subject to comparatively rapid corrosion
rates from the marine environment. Even though there may be a cathodic
protection system on the pipeline, the cathodic protection current will not
reach that part of the pipeline in atmosphere. Accordingly, reliance must
be placed upon protective paints or coatings. In less aggressive atmospheric

environments, paints may be used for esthetic reasons to give the pipeline
and its appurtenances a satisfactory appearance as far as the public is
concerned. In the severe atmospheric environments, however, more
rugged paints or thick film coatings are required which have been proved to
be resistant to the severe environmental conditions. Since there is no
cathodic protection back up, any such protective coatings used in these
applications must be carefully maintained on a periodic basis.
Controlling and Monitoring Internal Corrosion

Under a prior heading, the use of cathodic protection for internal pipe-
line surfaces was discussed. In that instance the exposure was to a pipeline
full of electrically conducting corrosive material. In other applications,
such as in the case of pipelines carrying gas or petroleum or petroleum
products, there may be corrosive components transported along with the
product which, usually in the presence of condensed water inside the pipe-
line, can create corrosion problems. Depending on the product being
transported and the corrosive elements involved, various means of corrosion
control can be used including the elimination of corrosive elements from
the product before it enters the pipeline, using various chemical inhibitors
to render the corrosive elements ineffective, or taking steps to prevent the
condensation of moisture inside the pipeline.
Where there is a possibility that there will be internal corrosion, it is good
practice (and required by regulations in many instances) that some type of
internal monitoring procedure be used. One common means of doing this
involves the use of coupons of the pipeline steel which are inserted in the
pipeline at test points so that the coupon material will encounter the product
stream and in those zones where corrosion is to be expected. These coupons
are removed periodically and inspected for corrosion attack and any
corrosion control program adjusted in accord with the results obtained
from the inspection program.
In addition to any monitoring program such as that described above,
should the pipeline be shutdown and opened for any reason, the internal
surfaces of the pipeline should be inspected for corrosion damage.
Electrically Discontinuous Pipelines and Cathodic Protection
The application of cathodic protection to metallic pipelines, particularly
existing pipelines, is often compounded by the presence of mechanical
joints in the pipeline rather than welded pipeline connections. Mechanical
joints, unless provided with a solid bond, can introduce an electrically
insulating point at each mechanical connection or, even if not completely
insulating, can introduce longitudinal resistance in the pipeline at these
mechanical joints. The presence of such insulation or resistance prevents or
retards the collection of cathodic protection current and flow of this pro-
tection current along the pipeline to a centrally located cathodic protection

installation.
Basically there are two alternative approaches to application of cathodic
protection to mechanicallyjointed pipelines. The first procedure involves the
uncovering and bonding of each mechanical joint in order that the pipeline
to be protected may be made electrically continuous. The problem of
locating the underground joints in existing lines and the cost of uncovering
and bonding them can be quite expensive, particularly in the case of lines
under paving. The other approach is to install separate cathodic protection,
usually with galvanic anodes of zinc or magnesium, on each individual
section of the pipeline between mechanical joints. Again this involves
problems in locating the individual pipeline sections on existing pipelines
and involves cost in installing the protection anodes. This method is, how-
ever, (depending on individual circumstances) usually the more reliable
procedure since failure of any one cathodic protection installation on an
individual pipe section will not endanger the cathodic protection on
adjacent sections. By contrast, the procedure of bonding pipeline joints and
supplying current from centrally located cathodic protection installations
can be made ineffective by breakage of any bond cable installed across a
mechanical joint because the one breakage will prevent cathodic protection
current from reaching all pipeline beyond that break in a direction away
from the cathodic protection unit.
Whichever type of procedure is used, periodic inspections must be made
to insure that adequate cathodic protection is being maintained.
Needed Standards
Under the following headings are discussed some of the standards which
it is felt would be of value for corrosion control programs associated with
the pipeline industry.
Standards for Pipeline Electrical Test Points
Pipelines equipped with cathodic protection systems are provided with
electrical test points which are installed at intervals along the pipeline to
permit periodic test measurement of the level of cathodic protection on the
pipeline. At the present time, each individual pipeline company has its own
established practices for the construction of such test points. There are
differences from company to company as to the number of wires installed
at each test point, the manner in which they are terminated, and the color
coding of the individual wires. It would appear desirable that the details of
test point installation be standardized, particularly insofar as color coding
is concerned, so that when cooperative tests are made between pipeline
systems, there will be no confusion as to the meaning of the color coding
which will then automatically indicate the nature of the test point. As
further justification, there are usually several types of test points used
along a pipeline. These may, among others, include test points at cased
crossings with wires to both casing and carrier pipe, normal potential test
points along the route of the line with wires directly to the pipeline for
potential measurement purposes, and test points with wires bridging
calibrated pipe spans so that the current flowing in the pipeline can be
measured. If the test point construction and color coding are standardized,
there would then be less possibility or confusion as to the type of test point
being worked with at any given location.
Standards for Monitoring Internal Corrosion
Minimum Federal standards for corrosion control on pipeline carrying
hazardous gases or liquids require that there be some form of monitoring
for internal corrosion in the ease of products containing corrosive elements.
The standards further provide that the internal monitoring provisions shall
be inspected at periodic intervals. The exact nature of the monitoring
devices is not set forth.
It would appear to be timely for the establishing of standards for moni-
toring internal corrosion on pipelines. These standards could be established
for various classes of service and could set forth accepted monitoring
devices or procedures applicable to these classes of service. It should be
noted that earlier reference has been made to NACE standard on internal
corrosion monitoring on petroleum products pipelines.
Standards for Pipeline External Corrosion Surveys
The minimum Federal standards for corrosion control on hazardous gas
and liquid pipelines provide that, where possible, electrical surveys be con-
ducted on pipeline systems to determine the status of corrosion control.
There are no provisions in the standards setting forth the details for such
electrical inspection. At the present time, various companies seeking to
comply with the minimum federal regulations are conducting pipeline
corrosion surveys with various requirements as to frequency of inspection
along the pipeline route and the manner in which the inspections are to
be made.
It would appear that it is timely for the establishment of standards for
pipeline corrosion surveys which will be used in complying with minimum
federal pipeline safety standards.
Summary
The pipeline industry is a very high investment industry. Fully imple-
mented corrosion control can effect high dollar savings by reducing prop-
erty loss, by reducing the cost of repairs, by avoiding catastrophic failures
with loss of property and life, and by conservation of natural resources as
the energy shortage crisis becomes ever more serious.
There are effective methods of corrosion control which are available for
both the external surfaces of a pipeline and for its internal surfaces where
they are required. At the present time, specific standards concerned with
corrosion and directly related to the pipeline industry are minimal. Govern-
mental regulations are now in effect relative to pipeline corrosion control
requirements for major segments of the pipeline industry. These regulations
will encourage the development of additional standards directly related to
pipeline corrosion control.
There are a number of associations through which information related to
pipeline corrosion control may be obtained as stated herein.
STP534-EB/Nov. 1973
Chapter 6

Telephone Industry
George Schick 1
The discussion of corrosion and corrosion protection in the telephone

industry requires the division of the telephone plant into two major areas:
outside plant and central office equipment. The outside plant, which in-
eludes all the cables, closures, hardware, radio towers, etc., requires
relatively advanced corrosive degradation before it stops functioning
properly. It is also the part of the plant which is the most exposed to the
corrosive environment. Central office equipment on the other hand is
always located in a building and exposed to a relatively controlled environ-
ment. However, technological advances have resulted in the development
of sophisticated components, usually small in size, and this has led to
closer component spacings, with separable electrical contacts having
lower contact forces and voltages than previously possible. In this type of
equipment even microgram quantities of corrosion products can result in
premature failure.
In view of these basic differences the corrosion problems and their solu-
tions are quite different. The amount of capital investment and cost of
repair or replacement, due to corrosion failure, is, by far, larger for the
outside plant than for the central office equipment. The major part of
this chapter will therefore discuss outside plant corrosion problems and
means of their protection.
Outside Plant
The outside plant is that part of the telephone plant which is located
between the subscriber's side of the main distribution frame in the central
office and the protector block on the subscriber's house. From the corrosion
standpoint the outside plant is subdivided into the following areas: (1) aerial
plant; (2) underground plant; (3) buried plant; and (4) submarine cable
systems.
The aerial plant is subjected to the corrosive effects of rain and dew. The
1 Bell Telephone Laboratories, Whippany, N.J. 07981.
| 07
areas considered to be the most corrosive are the sea coasts where the wind
blows salt-laden water, and the industrial areas where the air is polluted by
acidic fumes.
The underground plants are characterized by cables enclosed in conduits
and joined in manholes. Although the conduits provide substantial me-
chanical protection, they are not impervious to moisture and are often
flooded by soil waters. The manholes in many locations are partially or
completely flooded and polluted manhole water has a strongly corrosive
effect on cables and associated equipment. It is also possible that cables in
flooded duets or manholes pick-up or discharge stray d-c currents.
The buried plant is characterized by cables and splice closures directly
buried in the ground with access points brought above ground in pedestal
type terminals. This plant is exposed to both corrosion and physical and
biological degradation. Physical damages are inflicted by rock cuts or
lightning, biological degradation is caused by rodents and ants, and
corrosion is caused by soil waters which can be as acidic as pH 3 and as
alkaline at pH 10. Stray currents are also playing an important role in the
corrosion of buried plant.
Ocean cables and their accessories are exposed to one of the most
corrosive natural environments, the sea. In this environment the situation
can be further aggravated by the abrasive effect of coral and physical
damage caused by trawlers.
The protection against all these hazards is partially built into the com-
ponents of the plant and partially applied to the working plant. The
built-in protection is primarily based on the choice of materials.
Materials in the Outside Plant

Since about one third of all telephone plant investment is in cables, the
materials in cables and cable sheaths will be discussed first. The corrosion
protection of the cables is largely built into the cable sheaths. Therefore,
materials in the various sheath constructions will be emphasized.
Basically all cable designs are either multipair or coaxial. The former is
characterized by bundles of conductor pairs where the individual con-
ductors are surrounded by a dielectric material. Coaxial cable is made with
a single center conductor surrounded by a coaxial conductor tube (outer
conductor) and separated by polyethylene disks or solid polyethylene
dielectric (ocean cable).
The amount of built-in protection depends upon the channel carrying
capacity and future accessibility of the cable. Toll cables, which have large
channel carrying capacity over long distances, have more built-in protec-
tion than smaller distribution cables. Distribution cables in turn have more
built-in protection than service wires and drop wires.
CORROSION IN THE TELEPHONE INDUSTRY | 09
Multipair Cables and Cable Sheaths

The conductor material is either copper, tin-plated copper, or aluminum.
Insulation for the copper conductors is provided by paper pulp, paper
ribbon, low density polyethylene, or propylene copolymers. The latter two
are designated as PIC insulation. The aluminum conductors are insulated
with low density polyethylene or propylene copolymers.
Protection of the cable core against water ingress is achieved by dry air
pressurization (pulp cables) or by filling the core with polyethylene-
petroleum jelly mixtures (PIC cables).
Core or unit binders are used to hold together either the entire core or
part of the core (usually 25-pair units). The material of these binders is
either polypropylene copolymers or high density polyethylene.
Core wrap holds together the core in its assembled condition and size,
adds to the dielectric strength between the conductors and the shield and
protects the core from heat damage. In PIC cables the core wrap is made of
either polyethylene terephthalate (mylar), polypropylene or styrene-
butadiene rubber tape. The core wrap for pulp cables is paper.
The cable sheath protects the core from the environment mechanically,
electrically, and chemically. The sheath must perform effectively for 40
years or more and for those occasions when it fails, should be readily
repairable.
The standard sheath for telephone cables until the late 1940s was lead.
After the Second World War, this metal became scarce and expensive,
triggering the development of composite sheaths. Small quantities of lead
sheathed multipair cables are still produced for special installations, for
example, over steam locomotive tracks and where gasoline contamination
is anticipated.
The composite cable sheath may have some or all of the following com-
ponents; starting from the core wrap and progressing to the outer surface.
1. Adhesive coated aluminum--serves as diffusion barrier. The alumi-
num is EC grade and bonded with ethylene acrylic acid copolymer to the
inner jacket.
2. Inner jacket--provides liquid water block and isolates the core from
high potentials on the shield. The inner jacket is made of high density
polyethylene.
3. Shield--provides electrical shielding and interception of lightning and
po~ver-cross currents. The shield is made of EC aluminum, plastic-coated
EC aluminum, or copper.
4. Soldered steel shield--provides hermetic seal (for pulp cable), low
frequency electrical shielding, and armoring against rodents. This shield is
protected against corrosion with a bituminous thermoplastic flooding
compound. The shield material is mild steel or tin-coated steel. In some
1 10 INDUSTRIAl. CORROSION STANDARDS AND CONTROL
cables where hermetic seal is not needed but corrosion and rodent resistance
is essential the electric shielding and steel shield are combined in bimetallic
shields which can be copper and stainless steel or plastic-coated aluminum
and stainless steel.
5. Outer jacket--provides mechanical and environmental protection
for the underlying members. This j acket is made of low density polyethylene.
Multipair cables intended for use inside buildings (still part of the outside
plant) need fireproof insulation and polyeth~clene is unsuitable for this
purpose. In these cables the conductors are dual insulated with low density
polyethylene and polyvinylchloride. The core wrap is mylar (polyethylene
terephtalate) and the aluminum shield is coated with vinyl chloride-vinyl
acetate-maleic acid terpolymer and bonded to a polyvinylchloride jacket.
Outside Plant Wires and Cables
The outside plant wires are used in the aerial, and buried plants.
Drop wire is used in the aerial plant between the junction with an aerial
distribution cable and the customer's residence. The conductors are made of
copper-plated steel. The conductors are insulated with vulcanized styrene-
butadiene rubber. The insulation is reinforced with rayon servings and the
outer jacket is made of vulcanized polychloroprene (neoprene).
Service wire is used in the buried plant between the junction with a
buried distribution cable and the customer's residence. The conductors are
copper-coated steel insulated with high density polyethylene. The insulated
conductors are surrounded by a polyvinylchloride inner sheath. Electrical
shielding is provided by aluminum tape and the outer jacket is polyvinyl-
chloride.
Underground wire is used in the rural buried distribution plant between
junction with aerial or buried distribution cable and service wires or the
customer's residence. The conductors are made of copper insulated with
low density polyethylene. The conductors are protected against rodents
with galvanized steel tape and the outer jacket is polyvinylchloride.
Coaxial Cables are used as long haul toll cables. The center conductor is
copper, separated from the outer coaxial tube by slit disk, high density
polyethylene dielectric. The outer coaxial tube is both electric conductor
and strength member and made of copper tape and tin-plated steel, lami-
nated together with ethylene-acrylic acid copolymer. The coaxial units are
insulated with paper and surrounded by an inner low density polyethylene
sheath for dielectric strength. The polyethylene is covered with a paper
heat barrier on which the lead electric shield is extruded. The outer cor-
rosion protection is provided by a bituminous coating and a low density
polyethylene outer jacket. Since these long haul toll cables require a high
degree of reliability, pulp insulated copper conductor pairs are distributed
along the coaxial units to serve as water alarm circuits.
Armorless ocean cables are used in the deep ocean where protection
CORROSION IN THE TELEPHONE INDUSTRY 111
against trawlers is not necessary. In this cable the strength member is a high
strength steel strand located in the center. The inner conductor is a copper
jacket surrounding the steel strand. The dielectric separating the inner con-
ductor and the concentric copper tape outer conductor is medium density
polyethylene. The outer protective jacket is high density polyethylene.
Ocean cables laid on the continental shelf need more mechanical pro-
tection. This is achieved with helically applied galvanized high-strength
steel armor wires which are either coated with bituminous flooding com-
pound and jute or individually jacketed with neoprene.
Outside Plant Apparatus

The outside plant apparatus items are so numerous that their complete
description is clearly beyond the scope of this book. We are therefore
restricting ourselves to the discussion of the materials of some of the most
basic items.
Splice closures are the points where the cables are joined together.
Because the joints must have a high degree of electrical contact reliability,
these closures must be resistant to the corrosive environment.
Underground and buried plant uses hot dip galvanized cast iron splice
closures for multipair cables. The closure halves are tightened together
with type 304 stainless steel bolts and nuts. The inner cable clamp is also
type 304 stainless steel. The end plates, through which the cable enters
into the splice closures are made of a lead-bismuth alloy. Moisture-proofing
of the joined half closures is assured by a butyl rubber tape. Splice closures
in the buried plant are further protected with a bituminous mastic primer
and a hot applied tape (cotton fabric base saturated and coated on both
sides with a bituminous mastic). The splice closure for coaxial cables is
tin-plated steel tube with wiped-on lead alloy end plates. The outer cor-
rosion protection is the same as that of the galvanized cast iron buried
splice closures.
Splice closures and repeater houses for ocean cables are made of copper-
beryllium alloy since the environment is very corrosive and the system once
in place is virtually inaccessible.
Aerial plant uses mainly plastic splice closures in the distribution plant
which offers little corrosion protection since ready accessibility is its most
important feature. In such closures the joining elements are either made of
brass or the joint has its built-in encapsulant. Splice closures providing
corrosion protection in the aerial plants are made of cast aluminum.
Loading coil cases have two main types. The cover is either bituminous
hot melt coated steel or low density polyethylene. The latter contains
polyurethane inner space filler and a steel coil container. The loading coils
are made of permalloy or magnet wire (thermoplastic or thermosetting
coating on copper conductors). The steel covered type has nylon coil
1 12 INDUSTRIALCORROSION STANDARDS AND CONTROL
spacers and the coil encapsulant is silica powder-filled epoxy. The en-
capsulant in the polyethylene cover type loading coil case is polyurethane.
Protected terminal blocks are sometimes used in ready access terminals
of multipair cables. The shell of these terminals is made of 50 percent
acrylonitrile/butadiene styrene (ABS) and 50 percent polyvinylchloride
blend. The pigtail conductors (tinned copper) are insulated with polyvinyl
chloride, and the eneapsulant used is foamed polyurethane.
Coaxial terminals in the underground plant are red brass shells (85 ~o
Cu + 15 ~ Zn) with electroplated tin coating.
Hardware
In the aerial plant the telephone plant is located on telephone poles.
With the exception of the self-supporting cables, which have their own
built-in strand, all canes are attached to strands. The cables are secured to
the strand with lashing wires. At the wooden telephone poles, clamps,
hooks and fasteners are used to secure the strands, cables, closures, etc. to
the poles. The poles themselves may contain pole steps and are secured
against high winds with guys and anchors. With the exception of 400 series
stainless steel strands and lashing wires, used in corrosive areas, practically
all hardware items are made of galvanized steel. In some areas aluminum
coated steel hardwares are also used.
In the underground plant a large number of hardware items are used in
the manholes (racks, hooks, ladders, manhole steps, etc.). In the majority
of the manholes these items are made of galvanized steel. In particularly
corrosive areas where the manholes are flooded, the racks, hooks and their
fasteners are made of Monel.
Outside Plant Corrosion Standards
The materials described in the previous section provide substantial pro-
tection against corrosion. In fact, they are the only protective measures in
the aerial plant, but they do not solve all the corrosion problems en-
countered in the underground and buried plants. Bare metallic structures,
such as lead cable sheath, splice closures, and hardware are exposed to
corrosive soils or high water tables. The non-metallic protective coverings
are subject to physical damages, degradation by aging and lightning and
the underlying metallic layers are partially exposed to the corrosive media.
The underground and buried plants, besides the natural corrosion by
interaction with the environment, also are subjected to stray current
corrosion.
This section will discuss corrosion, corrosion surveys and corrosion
mitigation of underground and buried telephone plants both in non-stray
and stray current areas.
General Principles
The telephone plant corrodes where current leaves the sheath and flows
into the electrolyte (earth). Any current which leaves the outer metal
surfaces (primarily lead cable sheath) must have entered it at some other
point. (To simplify the following description, the term "cable sheath" or
"sheath" will also mean splice closures and associated hardware.) The
general attack on the buried and underground telephone plant corrosion
problem, therefore, consists of two equally important phases:
(a) Limiting as far as practicable the current which enters the cable
sheaths.
(b) Providing metallic paths through which the current may leave the
cable sheath without damaging it or other foreign buried metallic structures.
The range of conditions encountered in corrosion problems are extremely
wide. The currents involved may be manmade, otherwise called stray
currents: (1) dc transportation systems, including mining operations;
(2) dc power and lighting circuits; (3) d-c welding processes; and (4) cathodic
protection rectifiers on cables, pipe lines, ships in dock, steel piers, metal
framework of buildings, storage tanks either buried or above ground, and
gas, oil and water wells.
Currents may be caused by natural conditions (non-stray currents):
(1) dissimilar metal couples; (2) differences in the composition of metals
exposed to the electrolyte; (3) variations in earth resistivity; (4) variations
in soil composition; (5) differential aeration; and (6) sulfate reducing
bacteria.
The distance between the locations where the current is entering the
sheath and where it is leaving may range from minute fractions of an inch
to miles. The former are called local cells, the latter are long cells.
The determination of whether or not corrosion is occurring requires
careful measurements and the data obtained must be subjected to careful
analysis.
Protective arrangements should be such as to minimize the probability
of impressing current on plants owned by others. Since protective schemes
against corrosion mutally interact with other structures, cooperation
through local corrosion committees is a must.
Limitation of Current Pick-up
Two things must be done if the flow of current to the cable sheath is to
be at a minimum:
1. Except for interconnections which are specifically planned as part of a
coordinated protection scheme, the underground and buried plant should
be kept free from all connections to other grounded metallic structures.
2. The cable sheath should not be made more negative to earth in any
area than is necessary from the practical design standpoint.
Negative cable-to-earth voltages can be kept low by increasing the
number of drainage points and decreasing the current drained at each
point. Even with a carefully designed drainage system, there may be a
tendency toward high negative protentials in a few areas.
General Principles of Drainage

Current enters the cable sheath because a potential difference is im-
pressed by an external source or by self-generated electromotive forces.
All currents entering the sheath must leave it. The sheath can be prevented
from corroding by permitting the current to leave through a wire rather
than by leaking into the electrolyte.
If a drainage system is to approach the objective of providing protection
with minimum negative-to-earth voltage, minimum current flow on
sheaths, minimum reaction on plant owned by others, etc., it must be
based on a study of the driving potentials and currents in absence of
drainage. In a well-adjusted drainage system, the currents are permitted
to follow the paths indicated by the driving potentials as far as practicable,
the major difference between "drained" and "undrained" conditions being
that currents leave the sheath via a metallic connection rather than by
leaking from the sheath directly into the electrolyte. The drainage process
must change the cable potentials and hence the current distribution and
direction of flow somewhat, but these changes should be as few and as
small as practicable.
Where sheath currents are due to stray currents, it is usually practicable
to complete the circuit for the currents by means of wires between cables
and the source of the stray current.
In non-stray areas the current must be taken off the cable through a wire
and put into the earth through expendable anodes. In many cases a source
of d-c power must be used, in others the natural galvanic potentials between
the cable sheath material and certain anode materials, such as magnesium,
can be used in this process (cathodic protection). The anode locations
should be chosen so as to minimize the current to other metallic structures.
Tests
The objective of corrosion tests is to find out where and why current is
entering the sheath and in what amounts, where it is trying to go, and
where and how one can let it go there without causing corrosion. The
testing techniques are numerous, but whatever the test, its ultimate objective
is to tell something about current. Since there are no direct methods to
measure the currents being picked up or discharged in a short cable sheath
section, it is necessary to resort to various forms of indirect measurements.
Cable-to-earth voltages can indicate whether currents are picked up or
discharged but they do not indicate the rate of pick-up or discharge and
they do not always indicate whether the cable is picking up or discharging
current. An adequate corrosion investigation, therefore, uses many kinds
of tests.
Once an adequate protective system is established it is important to
watch its performance to see that it remains adequate. This is done in two
general ways.
CORROSION IN THE TELEPHONE INDUSTRY 1 15
(a) Periodic "routine" surveys.

(b) Use of "pilot" wires to central testing points so that checks can be
made at short intervals of conditions at "key" points.
Routine surveys are time consuming and expensive and usually made not
more than once a year. "Key" points with "pilot" wires can be checked
very frequently with a minimum of time and expense.
Stray Current Corrosion and Corrosion Protection
Voltage differences exist between different points on the running rails of
the d-c transportation systems due to IR drops in the rails caused by the
current discharged to the rails through railway propulsion motors, car
heaters, etc.
The earth, in which cable and piping systems are imbedded, forms a path
parallel to the rails. Part of the current delivered to the rails flows through
the paralleling earth path between points of higher and lower rail potentials.
In turn, part of the current in the earth flows between these areas over
cable and pipe systems. The higher the longitudinal resistance of the rails
and the lower the leakage resistance of the rails to earth, the larger will be
the proportion of the rail current which flows in the earth. The proportion
of the earth current carried by cable and piping systems is a direct function
of their leakance to earth and an inverse function of their longitudinal
resistance.
Measurements of Rail Voltages
In many cases adequate drainage systems can be designed without
knowing quantitatively the rail potential distributions. However, in some
cases quantitative information is desirable, that is, for explaining observed
cable-to-earth and cable-to-rail voltages, in studies to select the best
drainage points. Such measurements are: (a) direct measurements of rail
potentials using pilot wires and (b) measurements of rail gradients in the
field.
In the pilot wire scheme cable pairs are used to connect selected points
on the rails to a central testing location. Measurements from each of these
points to a single point (namely, rails at the substation) or between different
points can then be made using recording or indicating meters. This method
is used only where information extending over a considerable period of
time is desired. Some long pilot wires will have high resistance and to avoid
corrections high resistance meters are required. If a particular pilot wire is
common to two or more meter connections, the current taken by some of
the meters will react on the readings of the others because of the drop over
the common resistance. This effect can be avoided by high resistance meters
and low resistance pilot wires. If sufficiently low resistance cannot be
obtained the mutual effect can be excluded by using separate pilot wires for
all simultaneously connected meters.
1 16 INDUSTRIAL CORROSION STANDARDS AND CONTROL
General Principles in Design of Drainage System

There are many places in a rail system where the voltage gradients change
sharply. In some cases they change in such a way as to provide points where
the rails tend to pick-up current and undrained cables tend to discharge
current to earth.
An idealized approach is to ehminate all existing drainage and a careful
survey made to define areas in which the cables are picking up or discharg-
ing current, where the cables become positive to rail and what are the rail
potential differences, sheath current magnitudes and directions. At key
points the measurements should be made with recording instruments for
24 h and weekends. Based on the pictures thus obtained a drainage sys-
tem has to be designed to establish protective cable-to-earth potentials
with minimum current pick-up by the cables. This can be accomplished by
first installing drainage connections to rails as far away from the substation
as any significant positive cable-to-rail voltages are observed. Most of these
connections would require automatic reverse current switches. Next, addi-
tional drainage connections have to be installed closer in toward the sub-
station as needed to keep the cables from discharging current to earth at any
point. Automatic reverse current switches and current limiting resistors
should be installed in these drainage connections as needed. Drainage
connections normally should be made to rails rather than substation buses
in order to secure the best coordination between drainage current and
protection.
In the practical case it frequently is not possible to remove all drainage
from an urban cable plant while a complete survey and analysis are being
made. However, by analyzing the rail potential pattern and the cable
layout it is usually practicable to find the points where outlying drainage
connections might be made in the absence of existing drainage. Tests can
then be made at such locations with the existing drainage connections open
for a few minutes at a time. Any such outlying drainage connections found
to be practicable then can be installed and the drainage connections nearer
the substation rearranged if necessary.
Determining Whether Protection is Being Secured
In large parts of stray current areas voltage measurements between cable
sheaths and earth (usually in bottoms of manholes) are sufficient to indicate
whether current is being picked up or discharged; negative cable-to-earth
voltages indicating pick-up areas and positive voltages indicating discharge
areas. In some new steel reinforced concrete manholes the voltage is
measured 75 to 100 ft away from the manhole in either direction along the
cable route. Such indications are usually reliable where the voltages are
larger than those observed because of galvanic potentials (over 0.1 V).
The rate of current pick-up or discharge is not indicated by the voltages
since it also depends on the leakage resistance of the cables to earth.
Generally where the leakage resistance is low the cable-to-earth voltage

is usually lower and more steady than in adjoining areas. Hence, readings of
this type are likely to indicate danger areas or contacts between cables and
other grounded structures. Cable-to-earth voltages can thus be used to
block out the major pick-up and discharge areas.
In some areas of uncertainty sheath current measurements or measure-
ments used in nonstray current areas are used. In a small residue even
these measurements may not be completely indicative and probably the
best way to dispose of those areas is to increase slightly the drainage cur-
rents at nearby points.
Protection Against Cathodic Corrosion

Under some conditions cathodic corrosion of sheaths may occur where
the cable-to-earth potential is over a few tenths of a volt negative. In the
presence of salts this critical voltage may be as low as 0.2 V. Methods of
preventing cathodic corrosion are:
1. Keeping the negative cable-to-earth potentials as low as practicable.
2. Judicious use of insulating joints shunted, if necessary, by resistors
and/or capacitors.
3. Replacement of cables, failed by cathodic corrosion, with polyethylene
jacketed cables.
4. Periodic flushing of ducts to remove accumulated alkali.
Competing Drainage Systems

In many areas water or other piping systems are also drained. If the
potential of a piping system, having low leakage resistance to earth, is
reduced by drainage the earth potential in its vicinity tends to be reduced.
As a result, the tendency for current to be discharged from a nearby
telephone cable is increased. This may increase the difficulty of providing
adequate protection to the telephone plant particularly if the pipes are
drained directly to the negative bus at substations. However, by virtue of
the high leakage of pipes to earth the adverse effect of pipe drainage usually
does not extend over long distances except where the pipes are of unusually
low longitudinal resistance. If pipe drainage makes it difficult to provide
protection to the telephone cables, a coordinated drainage system, through
cooperation with other parties concerned, is the solution.
Surveys and Test Methods

Routine surveys involve a set of measurements made periodically to
check the corrosion conditions of the telephone cable plant. These surveys
determine: (a) areas where cable sheath is liable to damage from corrosion;
(b) important differences from any previous survey; and (c) what further
data may be required for determining necessary measures of protection.
1 18 INDUSTRIAL CORROSION STANDARDS AND CONTROL
The periodic measurements include: (a) potentials of cable sheath to

ground, nearby rails, and pipes; (b) current in cable sheath; measurements
made at approximately the same time as potential measurements; and
(e) overall check tests made at more frequent intervals than the above
potential and current measurements.
A cheek of any drainage wire fuses and automatic switches should be
made as a preliminary part of the routine surveys to avoid wrong eonelu-
sions. Underground dips in the aerial plant should be ineluded in the routine
survey.
In general routine surveys should be made yearly, however, the frequency
of the survey is influenced by local conditions to maintain satisfactory plant
operation. Where routine surveys are made less frequently, overall cheeks
are needed at approximate intervals. In the sections of the country where
the ground is frozen to appreciable depth, surveys during the winter months
should be avoided. In other sections, dry and wet seasons should be taken
into account.
Measurements of Cable Sheath Potential

Cable sheath to earth potentials are measured initially to have a qualita-
tive indication where the current is collected and where it is discharged.
Supplemental measurements are made of the cable voltage to adjaeent
rails, water, and gas pipes or other extensively grounded metallic structures.
When manholes are opened for routine survey measurements, these should
include measurements of cable sheath to electrified railway rail where the
rail runs closer than about five feet; when greater than five feet but on the
same street from every second or third manhole. Measurements to pipes
via house connection, hydrants, or gate boxes should usually be made at
every second or third manhole and also at points where the main pipe lines
themselves are aeeessible.
Since the potential will fluctuate in many cases, the voltmeter scale
should be observed a sufficient amount of time to obtain reliable average
value. In congested areas, readings should be taken every 30 s for 5 min,
taking the average positive indication and the average negative indication.
At manholes where sheath currents are also measured, these should be
correlated with the potential measurements. In less congested areas the
readings can be taken in 3 to 5 min intervals for half an hour or more. In
cases where there are both positive and negative potentials, an estimate
should be made of the percentage of time during which the cable is positive.
Under normal circumstances the potential of the cable to earth is best
represented by contact with damp earth about five feet distant from the
cable run. A copper/copper sulfate half eell or lead tipped rod or lead
ground plate should be employed for earth contact.
Current Measurements on Cable Sheaths

In routine surveys, data on amount and direction of sheath current are
obtained by potential drop measurements over a measured length of cable
sheath. A millivolt scale is used for this purpose. These current measure-
ments supplement the potential measurements to determine their sig-
nificance. A comparison of results along the cable run will give an approxi-
mate indication of the areas where current is collected and where it is
discharged.
It is desirable to make current measurements at or adjacent to each point
of change in conductivity or junction of cable routes including short spurs
or laterals. Between such points it will be desirable to make current meas-
urements at about half the manholes where potential measurements are
made or more often if consistent increases or decreases of current are noted.
For any particular manhole, current measurements should be made at
the same time as the voltage measurements. The current corresponding to
the millivolt drop is computed from the sheath geometry and conductivity.
Careful note should be made of the direction as well as the amount of
current. As in the case of voltage measurements, observations should
cover a sufficient period of time to give reliable average indication depend-
ing on the car or train headway.
Current measurement in drainage wires is made in a number of ways:
(a) insertion of an ammeter in series at the fuse terminals; (b) drop of
potential measurements over a section of the drainage wire; and (c) drop
of potential measurements from a "Central Testing Bureau" over a section
of the drainage wire.
The overall checks should include current measurements on jointly used
drainage wires. Tie bonds between cable sheaths of the telephone company
and other subsurface structures should be checked in the same manner as
other drainage wires about once a year.
Visual Inspection
In connection with routine surveys, notes should be made of the following
conditions.
(a) Evidence of corrosion.
(b) Any unsatisfactory conditions of cable bonding or racking in
manholes.
(c) Cables submerged in flooded manholes.
(d) Water running through ducts containing cables.
(e) Accidental or unauthorized contacts with pipes or other metallic
subsurface structures.
(f) Evidence of unusual amount of ground water in the vicinity ot the
duct line; for instance, springs or marshy ground.
(g) Evidence of cinder fills.
Duct Surveys
In some locations electrolysis measurements made at manholes do not
indicate the probability of corrosion of cable sheath in the ducts. This may
occur at stray current areas where the cable is very locally affected by
another underground structure crossing the duct line. It can also occur at
nonstray current areas where the potentials between cable and earth are
small, and relatively small variations may be important.
A duct survey consists of moving a lead slug through a spare duct and
measuring the cable to slug potential, the slug leakage current, and the slug
leakage resistance at regular intervals throughout the duct length. Since the
duet survey is more costly than manhole measurements it is warranted only
at the following locations: (a) where failures have occurred but where the
manhole measurements did not justify remedial measures; (b) where cor-
rosion is suspected or has been observed; and (e) where positive potentials
have consistently been found in the manhole measurements made in past
routine surveys.
Duct surveys can be made only in nonmetallic ducts. They are par-
tieularly well adapted to tile conduits.
A duet slug consists of a piece of 2-in. diameter lead cable sleeving cut to
about a 14-in. length. A 6-ft piece of 18-gage stranded wire with tough
rubber insulation is soldered to the inside of the sleeve at least 1 in. from
the end. The sleeve is then placed over a wooden duct rod so that the wire
will trail the slug when inserted in a duct, and the front end of the sleeve is
beaten down around the rod about 6 in. from the end. The sleeve is then
filled to about 1 in. from the open end with No. 1 pressure plug asphalt.
A protective band of lead is placed around the rod and wire and the open
end is beaten down around this band.
Where there are several duets available, a low duet should be used
because there is more chance to encounter silt, mud, or moisture. In some
instances upper ducts may be selected because a failure occurred or a
cable will soon be placed in it.
Of the three measurements to be made at any location the cable-to-slug
potential must be made first to avoid polarization from the other measure-
ments. The same considerations apply to this measurement as to the
cable-to-earth potentials at the manholes. However, because the resistance
to earth of the slug is likely to be high compared to that of a manhole
ground plate the use of a high sensitivity (200 000 ohms/V) center zero
voltmeter is advisable. With such an instrument, no correction for the
resistance of the slug is necessary. Where a relatively low sensitivity volt-
meter is used, the readings should be corrected wherever the slug resistance
exceeds about l0 percent of the meter resistance. The corrected potential
(Ec) is computed by
+
where E,~ = measured potential

R,~ = resistance of the meter on the scale used
R~ = resistance to earth of the slug
Slug leakage current is measured by connecting a low resistance milli-
ammeter between the cable and slug, in place of the voltmeter. This direct
measurement also takes into account the polarization of the slug which may
result from the leakage. This polarization may cause a drop of the current
value and the reading should be taken after 30 s to 1 min when the needle
becomes steady. The resistance of the meter should not be more than
0.5 to 1.5 ohms. This can be achieved by using shunts on low resistance
meters.
Because cable to earth resistance is small, the measurement of resistance
between the duct slug and the cables is very nearly the same as the duct slug
leakage resistance. Duct slug resistance can be measured by either a d-c or
a-c method. The d-c method may used 1.5-V external potential between the
cable and slug and measure the current. Direct readings can be made with a
volt-ohmmeter. Since an externally applied current rapidly polarizes the
slug, the maximum swing of the needle should be read. To eliminate the
effect of normal cable-to-slug potentials, two readings must be taken with
reversed polarities and the average used as the resistance value. To limit the
polarization effect, -the external potential should be applied only long
enough to make the readings.
An a-c method or a method employing rapidly reversing de is prefered to
the d-c method to exclude the polarization effect. An instrument called
Vibroground uses a vibrator for reversing the current and the reading is
obtained by adjusting a dial to get zero deflection on a meter. A few
1000-ohm scale on the instrument is necessary for duct surveys. Other
instruments are Direct-R G r o u n d Tester and Megger types.
In some cases duct surveys may be used to determine the effect of drain-
age. In this case two surveys are required, one with the drainage discon-
nected and one with the drainage operating. The drainage should be
disconnected several days before testing with drainage off.
In the interpretation of the measurements the duct slug can be con-
sidered as representing a small piece of the working cable sheath. In
general the same interpretation applies to the cable-to-slug potential as the
cable-to-earth potentials at the manholes. The measured potential may be
in error as much as 0.1 V due to the difference in the electrode potentials of
cable sheath (covered with corrosion products film) and the duct slug
(scraped to bare metal at least partially). This is an important consideration
at non-stray current areas. True cable conditions are more closely indicated
if a slug is left in the conduit undisturbed and bonded to the cable for two
to six months and readings taken after removing the bonds on the stabilized
undisturbed slug. Cable-to-slug potential may be subject to fluctuations
due to the operation of grounded d-c systems some distance away from the
test location. Under such conditions it may be desirable to take a 24-h

record of the potential with the slug at a critical point in the duct.
Positive cable-to-slug potential indicates the tendency for the cable to
corrode and negative potential indicates non-corroding cable. Where the
cable-to-earth resistance is high, very little corrosion may occur with
positive cable potentials of appreciable magnitude. Sudden changes in
potential, particularly isolated high readings, may indicate the presence of
foreign material, such as steel wire, in the duct. Such readings have little
significance from the corrosion standpoint.
Experience shows that leakage current from the slug in excess of 2 mA
indicate a need for corrosion protection and leakage current below 0.5 mA
does not warrant remedial measures. Current values between these two
need other considerations, such as the number and importance of the
cables, and the cost and maintenance of protective measures. Higher
leakage current can be tolerated from a completely submerged slug than
from a slug with limited area of contact with the electrolyte.
Duct slug resistance is affected by the moisture or liquid in the duct and
is frequently lower in the low duct than in the high duct. Rainfall or dry
weather may cause changes. It is, therefore, important to consider the
condition of the duct for the interpretation of the resistance readings. The
measurement of duct slug leakage resistance not only permits correction of
the potential and current readings, but also indicates the possibility of
corrosion, since points of low resistance to earth are more likely to be
corroded than points of high resistance.
Test Methods in Non-Stray Current Areas

Earth Gradient Measurements
Figure 1 shows a cross section of a cable lying in the earth. It is assumed
that the cable is discharging current radially into the earth. IR drop meas-
urements, through the earth between a point on the surface of the earth
directly over the cable A and points on each side of the cable B and C can
determine whether the earth near the cable is at a higher or lower potential,
than more remote earth. If point A is at higher potential than B and C,
current is flowing away from the cable. Measurements on each side of the
cable are made to establish whether or not there is a "transverse" current in
the earth, which produces IR drop across the surface of the earth but does
not involve the cable. Such current would be indicated if, for example,
point A were at a higher potential than B and lower than C. When an
indication of transverse current is obtained the determination of whether
the cable is picking up or discharging current is difficult.
Earth gradient measurements are made with electrodes whose potential is
not affected by the environment. As the ordinary lead plate is variable in
this respect, the best choice for measurements in the field is the copper/
copper-sulfate half cell. The contact resistance between this half-cell and
B A C
I I , I
FIG. 1--Cross section of cable discharging current into the earth.
the earth m a y be quite high, therefore, a meter having high resistance

(200 000 o h m s / V ) should be used.
The direction of current flow due to earth gradient can be determined
with three half cells in which case, prior to the measurements, the half cells
have to be calibrated on a glass tray containing a conductive solution, and
the differences used as correction factors. Another method using only one
half cell, illustrated in Fig. 2, does not necessitate prior calibration.
Analysis of earth gradient measurements is shown in Table 1. This table
can be m o r e readily understood if Fig. 2 is replaced with its idealized
electrical counterpart (Fig. 3). In the first example of Table 1, for instance,
cable to C = + 0 . 2 V, cable to A = + 0 . 2 5 V and cable to B = + 0 . 2 6 V.
So, C to A = 0.25 - 0.2 = + 0 . 0 5 V and C to B = 0.26 -- 0.2 = + 0 . 0 6 V
showing that point C is at higher potential than either A or B, indicating
current flowing away f r o m the cable. Due regard must be given, of course,
to the sign of the potential. In the third example of Table 1, C to A =
0 -- (--0.05) = + 0 . 0 5 V and C to B = 0.01 -- ( + 0.05) = 0.06V again
indicating current leaving the cable. These measurements need not be
HALFCELL LOCATED
SUCCESSIVELY OVER
CABLE AND ABOUT
4' EACH SIDE OF
CABLE
CABLE
MANHOLE
FIG. 2--Use of a single reference half cell for earth gradient measurement.
t 24 INDUSTRIALCORROSION STANDARDS AND CONTROL
CABLE
A C B
FIG. 3--Electrical counterpart of earth gradient measurement.
limited to locations near manholes. The tests can be repeated at intervals

along the cable.
In towns where streets over the cables are paved it is difficult to find
suitable electrode locations. Experience indicates that measurements with
high resistance meters can be m a d e through d a m p asphalt. When the
surface is dry a small a m o u n t of water is poured in suitably located de-
pressions, providing a good contact point. In some cases the use of a very
high input resistance vacuum tube voltmeter or potentiometer type volt-
meter is advantageous.
Centralized Testing
Centralized testing facilities comprise equipment at a central location
together with test leads or pilot wires to the various points concerned, f o r
making observations of the condition and performance of drainage wires,
automatic switches, and fuses. These facilities are used for relatively
frequent measurements of cable-to-earth potentials at key points to extend
the time intervals between general surveys. Test trunks normally assigned
for use in locating cable troubles or for general testing purposes m a y also
be used for centralized corrosion testing.
The resistance of the cable conductors used as a test lead will frequently
be several hundred ohms. The measurement of current by drop of potential
TABLE I--Interpretation of earth gradient measurements.

Cable to Earth Potential, volt
Over the Cable Left of Cable Right of Cable

(C) (A) (B) Interpretation
q-0.2 +0.25 +0.26 Current leaving cable

+0.2 +0.10 +0.11 Current picked up by cable
-0.5 0.0 +0.01 Current leaving cable
+0.01 +0.15 -0.10 Transverse current
method requires a high sensitivity millivoltmeter. A recording type milli-

voltmeter is preferable for this purpose since the measurements under
consideration should cover 15 min or more and at least once a year a period
of 24 h.
Permanent assignments are usually made for the relatively short leads
required for connection between the point under test and the nearest
central office. From the latter point, temporary connections as required
are made to trunks already assigned for general testing requirements, when
such are available, between the nearby central office and the cable location
test desk.
The best results may be expected from the use of separate pairs (from
nearby central office to the test points) for the measurements of drainage
wire current and cable-to-earth potential, and for fuse alarm circuits. Tests
of drainage wire current and operation of fuse alarms where such are
involved, may be accomplished over the same pair, provided the fuse is
included in the span of the drainage wire used for current measurement by
drop of potential method.
The preferred location for drainage wire fuses, switches and protective
equipment is on a pole adjacent to the point of underground attachment.
However, where the test leads are needed in a manhole, it is often practical
to pull a two-pair rubber insulated lead sheathed cable through a sub-
sidiary duct to a nearby terminal. Where this is not practical, a direct
connection can be made through a splice in the manhole. In this case pro-
vision should be made to insure against moisture leakage into the main
cable.
Drainage wire current measurements are important to detect deviations
from normal conditions. The following are examples where prompt
attention is needed:
1. Changes in layout, operation or condition of system.
2. Changes in layout or interconnection of telephone cable sheath.
3. Blowing of drainage wire fuses and conditions causing such operation.
4. Failure of drainage wire switches to operate properly.
5. Drainage wire in trouble (for example, corroding).
Although the necessary frequency of tests will be governed by local
conditions, they should be made once per month or oftener. At least once
a year a 24-h test should be made with a recording instrument.
The principle of the drainage wire current test is that the cable pair from
the point under test to the nearest central office will be bridged across a
span of the drainage wire, this span being so adjusted as to give direct
indication of current at the central testing point. Temporary connections
are made to such leads as may be available for this purpose from the
nearby central office to the central testing point. Assuming the availability
of test trunks from a central cable location test desk, the plugs standard
with these facilities will be satisfactory for the temporary connections.
There are two schemes for arranging the instruments for direct indication
of drainage wire current. Each involves potential drop measured over a
selected span of drainage wire. Scheme 1 provides for the same instrument
calibration for all of the drainage wires, giving full scale deflection with
maximum load on the drainage wire carrying the heaviest load. This, in
general, involves: (a) lowest reading scale of instrument, calibrated to give
direct reading with; (b) conductor leads of equal resistance (leads of lower
resistance padded to reach highest resistance); and (c) equal resistance of
drainage wire spans. Item b is determined by the cable conductor leads of
highest resistance. Item c is determined by maximum current on drainage
wire carrying heaviest load.
Scheme 2 provides for approximately full-scale deflection with each
drainage wire under condition of maximum load and correspondingly
different calibrations for the different drainage wires. In general this case
involves: (a) no adjustment of resistance of cable conductor leads; (b) ad-
justment of drainage wire span to give approximately full scale deflection
with maximum drainage current for the particular case in question, a margin
being allowed for abnormal conditions; and (c) different instrument cali-
bration for each drainage wire.
Drainage wire switches must also be kept in normal operating condition.
With manually operated corrosion switches, as at power houses, routine
checks with a voltmeter are of considerable value. Indication of proper
operation of automatic switches can be obtained from a study of the
sheath current data.
1. When the switches are operating properly, sheath current will be
indicated in one direction only, that is, from the cable to the negative return
system.
2. Reverse currents are indications of the switch failing to open as
required.
3. Currents of consistently negligible value during periods of the day
when drainage current would normally be expected is indication of the
switch failing to close .as required.
One of the overall checks from central testing points consists of periodic
measurements of cable-to-earth potentials at key points such as points near
drainage wire attachments. Data from such measurements give an indica-
tion of normal operation or departures there from, which may require
prompt attention. These tests of cable-to-earth potential, from a central
testing point, require the assignment of a pair of conductors exclusively
for this purpose between the point under test and the nearby central office.
As in the case of current measurements, this pair may be connected for the
time being to the regular test trunks to the central testing point. If prefered,
however, arrangements can be made for having these cable-to-earth po-
tential measurements made at the nearby central office. In the latter case
the conductors concerned are usually terminated at jacks in the local test set.
Remedial Measures
Design of Drainage Wires
To provide adequate mitigation without making the cables excessively
negative to earth requires careful selection of the points at which the bonds
are made as well as careful adjustment of the resistances of the bonds.
The major points of stray current discharge are generally near a sub-
station where the negative feeders are connected. In some cases negative
feeders are connected to the rail some distance away from the substation,
creating current discharge points away from the substation. Other points
of current discharge are main rail intersections and points where dis-
continuities, such as changes in number of tracks, occur. This geographical
information has to be supplemented with potential survey of cables with
respect to earth and rail and the magnitude and direction of current on the
cables. In general an anodic area, while it may extend for some distance
along the cable run, will be found to center around a point where the
cables approach the rails. This is the ideal point at which to make the
drainage connection if physical conditions are favorable. Other factors to
consider in choosing the drainage points are:
1. Cost (length of wire required, availability of spare ducts, type of
pavement).
2. Feasibility of installing reverse current switch if required (dry man-
hole, pole, substation basement).
3. Cable-to-earth voltage and cable-to-rail (or bus) potential should be
closely related in their variation with time (simultaneous readings of these
values for 24 h plotted against each other should closely follow a straight
line). Bad correlation indicates that the optimum point was not selected
and other points should be tested.
Installation of Drainage Wires

In general insulated stranded copper wires are the most satisfactory.
Where the wire is placed in ducts belonging to another organization there
is a chance that it may come in contact with other structures, thus altering
the drainage conditions. In this case lead sheathed conductors may be used.
The lead sheath has to be isolated from the telephone cable sheaths but
should be protected against corrosion by bonding, through some calcu-
lated resistance to the other protected structures.
Later checking is facilitated if solderless connections (that is, lugs bolted
together) are used and taped over. At a convenient point in the wire a
calibrated voltage drop should be provided for current measurements.
Short lengths of bonding ribbon soldered to the wire may be used. When
these terminals are soldered to stranded wire the solder has to be allowed to
run into the inner strands to avoid change in the "drop" with time.
Fuses in the drainage wire are used where there is equipment in the
drainage connection such as a reverse current relay. Other areas of using

fuses are where a small cable crosses an interurban trolley line and is
drained to the rails, or when the drainage wire is run aerially on trolley
feeders. Fuses should be installed to the cable end of the circuit. The fuses
have to be rated well above normal expected currents in order to reduce
maintenance.
Cathodic Protection Design from Experimental Reduction Test Data
After a consideration of all the factors in a corrosion situation indicates
that some form of forced drainage (cathodic protection) is likely to be the
most practicable remedial measure, a field survey should be made to
determine by experiment what arrangement will provide adequate and
economical protection. Information should be obtained to determine:
(a) The physical layout of all subsurface structures and rail systems in
the area involved, even at some distance from the cable plant.
(b) If a single drainage or multiple drainage will be required.
(c) The effects that may be expected per ampere of drainage current.
(d) The effects on other subsurface structures and rail systems of any
drainage proposed.
(e) Suitable anode locations; the size and resistance of each anode and
the ease of constructing these, unless use is to be made of an existing
structure such as an abandoned pipe, etc.
(f) Size of, available routes for, and methods of placing drainage wires.
(g) External power supply requirements.
Testing Procedures
Testing procedures will depend on local conditions. The same testing
methods are applied for single or multiple drainage systems. In the latter
case the overlapping effects of the different drainage points also have to be
established. Except in the case where a duct survey is included, the data
obtained for cables usually are restricted to manholes or test points.
Occasionally measurements can also be made on the ground surface over
the cable between manholes or test points. This can be done with a copper/
copper-sulfate half cell attached to a long insulated wire lead. The volt-
meter connected between the cable and half cell all should have a sensitivity
of 50 000 o h m s / V and preferably 200 000 o h m s / V or greater.
After a suitably located anode site has been selected a temporary anode
(made of ground rods) for delivering test current to earth has to be con-
structed. First sheath current and sheath potential values, with respect to
earth and other structures, have to be established. The reduction test
consists of determining the change in potentials of cable sheath to earth,
and to any neighboring metallic structures, per ampere of drainage current.
They have to be made for sufficient distances each side of the drainage
point to determine the extent of effect of the particular drainage. The power
for temporary drainage installations is obtained from storage batteries or

from a portable generator.
In making the reduction tests it is not necessary to increase the drainage
current to a value which will make the cable sheath negative to earth at all
locations. Sheath to earth voltage readings should be taken with and
without the experimental drainage operating, if there is any fluctuation in
the normal sheath to earth voltages, at least 10 readings should be taken
and the results averaged. Observations should be made at key points (points
of highest positive potential without drainage). F r o m the data taken, the
reduction of sheath to earth voltage per ampere of drainage current can be
determined for each point tested. Then the amount of drainage current to
eliminate all the positive conditions can be calculated. On plain lead cable
sheath the voltage change per ampere of drainage current is fairly constant
over a wide range of current. Polyethylene jacketed lead sheaths are not
constant in this respect and several values of test current should be used
and the results plotted to establish the final value of drainage current.
Where other metallic structures are within the area of influence of the
drainage, the reduction test data should include the change, per ampere of
drainage current, in the potential of such structures to earth and in the
current flowing in them.
Cathodic Protection Design by Estimation

The effects of cathodic protection on long uniform cables (intercity toll
cables) can be estimated from a few basic data. The information to be
obtained and the testing procedures in a field survey are similar to those in
the previous section.
1. "Megger" ground tester 3-point method.
2. Delivering IA = 200 mA d-c current to an electrode 50 ft to one side
of the cable. With the circuit first open and then closed, the sheath to earth
voltage change (Vo), to an electrode 150 ft to the opposite of the cable is
measured.
3. Draining IA amperes to a trial anode at least several hundred feet
from the cable and measuring the sheath to earth potential change (1Io) at
the drainage point. The electrode to measure/Io is placed as in 2 above.
In cases 1 and 2 the cable leakage resistance to earth is:
4(Vo~ 2
= (2)
where Rs is the series of longitudinal resistance of the cable.

Electrolysis Switch
The electrolysis switch is intended for use in drainage connections where
reversals in the direction of current are encountered and an automatic
switch is desired. It is a device that closes a low-resistance connection

(drainage bond wire) to drain stray currents from underground cables to a
railway substation ground (negative bus). It also opens the connection to
prevent the flow of current from the drainage point to the cable when the
potential reverses.
When the cable-to-bus potential is positive and high enough in value to
cause sheath corrosion, the switch closes the bond. The switch can be
adjusted to separate at any desired potential between q-0.15 to 0.40 V.
When the current decreases to zero or near zero, or when the potential
reverses, the switch is opened automatically.
Electrolytic Capacitor
In corrosion areas where insulating joints are employed to isolate under-
ground or buried cable sheath from aerial cable sheath, it is necessary to
leave a low impedance ground on the aerial cable to discharge fault cur-
rents and maintain satisfactory noise levels. This may be obtained by
bonding the aerial cable sheath to a multigrounded neutral of a power
distribution system or a metallic water pipe system.
Where aerial cables are not grounded to a power neutral conductor or a
water pipe, a capacitor may be bridged across the insulating joint to
provide a-c continuity for noise suppression or protection reasons.
There is a choice of two capacitors, 1000 uF and 10 000 uF type, both o f
them are 25 V dc, nonpolarized dry electrolytic capacitors. Both are
intended for operation in circuits where the d-c potential will not exceed
25 V and where the a-c potential will not exceed approximately 2.5 V rms
at 60 Hz.
These capacitors ma3~ be connected without regard to polarity of the
d-c voltage. The d-c blocking characteristic of the capacitors is dependent
upon the build-up of an oxide film on the electrodes. When the d-c potential
across the capacitor is suddenly reversed, direct current flows through it
for a short time; however, the capacitor will quickly recover and block this
current. In stable condition, leakage current of less than 2 mA is expected
with 25 V across the capacitor.
The electrolytic capacitors can provide a low impedance for bridging
insulating joints in all cables including carrier and video cables where
suppression of voice frequency or carrier frequency noise is necessary. At
carrier frequencies, if the length of the leads from the capacitor are more
than a few inches, they may have objectionable impedance even though
the d-c resistance is negligible.
Tests have indicated that to maintain circuit noise limits at frequencies
up to 4 M H z the capacitor leads must not exceed about 4 or 5 in. in length.
The electrolytic capacitor can withstand as much as 250 A ac for several
seconds; therefore, breakdown of the capacitor from contact with a power
circuit is not expected to be a problem.
Testing capacitors, with an ohmmeter, for an open circuit or a short

circuit should be necessary only where measurements indicate unsatis-
factory noise levels or where potential measurements show an increase in
the corrosion exposure of underground or buried cabIe.
Central Office Equipment
Central office equipments are primarily made up of small electronic
components. Although the fact that they are located in a building in more
or less controlled environment, does not exclude the possibility of corrosion.
It is important to keep in mind that in central office equipment even micro-
gram quantities of corrosion products can result in premature failure.
Experience has shown that the best and ultimately the least expensive time
to stop corrosion of electronic equipment is at the design stage. This means
that the corrosion mitigation of central office equipment is first and fore-
most based on the proper material selection and the exclusion of false
economy of applying less material than necessary.
Stress Corrosion Cracking
Some metal alloys are susceptible to stress corrosion cracking, under
applied tensile stresses or residual stresses developed during manufacture,
when exposed to a specific environmental contaminant or combination of
contaminants. An example is a nickel containing ferrous alloy used for
leads in glass sealed semiconductor devices. Both stressed unplated and
gold plated alloy leads fail rapidly in cycling temperature--high humidity
conditions. Stress corrosion cracking of this alloy can be avoided by
electropolishing the lead and plating with solder of sufficient thickness to
be nonporous. Complete isolation of the lead from the environment is
accomplished by applying silicon varnish to the lead where it enters the
glass.
Another example is nickel brass (nickel silver) in wire spring relays,
stress corrosion cracking by ammonium nitrate bearing dust where the
humidity is high enough to allow moisture absorption by the dust. Ex-
perience showed that cupronickel alloy is essentially immune to stress
corrosion cracking under similar conditions, therefore, a change in material
was necessary. Other means of control for existing relays with nickel brass
are air filtration and humidity control in central offices where failures
occurred.
Corrosion of Plated Metals Used for Electrical Contacts
Plated silver, either alone or as an underplating for gold, readily forms a
sulfide film with sulfur-bearing compounds in the air. Contact resistance
problems arise with this kind of plating if the contact forces and open
circuit voltages are low. When silver is used as an underplating for gold the
sulfide film develops at pore sites and ultimately creeps over the gold
surface.
Thin and porous or otherwise discontinuous gold plated over copper or a

copper alloy results in formation of copper oxides and sulfides, leading to
contact troubles.
The examples of silver and copper underplates illustrate that porous or
otherwise damaged gold plating leads to contact resistance trouble. If the
problem is to be solved with an underplating, this material cannot be less
noble than gold or the gold must be nonporous.
If silver is plated on conductor paths, or on contact fingers of printed
circuit boards of either plastic or ceramic material and d-c potential and
high relative humidity are present, the silver can migrate along the surface
to an adjacent conductor of opposite polarity or even through phenol fiber
insulators to create a dendritic growth. Silver plating under these conditions
cannot be used.
A good portion of the corrosion problems of central office equipment is
due to manufacturing, shipment and storage. High humidity and airborne
contaminants from nearby electroplating shops can cause corrosion.
Mechanical damage of platings can render them discontinuous. Cardboard
packing materials contain about 0.5 weight percent free sulfur which can
form sulfide film on silver and copper under porous gold. Packing in raw
wood may lead to corrosion by organic acids (generally acetic) of central
office equipment.
Some corrosion problems can be traced to the materials used within the
equipment. Adhesives may give off vapors corrosive to many metals found
in central office equipment.
A typical example was the corrosion of nickel underplating at the bases
of pores in a thin (~0.5 um) rhodium finish on printed wiring board con-
tact fingers. On the other hand a thicker (~3.0 urn) gold finish on a contact
spring on the same board protected the underlying metal from corrosion
because it was essentially pore free.
Other potentially corrosive materials commonly used in central office
equipment include resins, plastics, elastomers and organic finishes. One of
the most potentially dangerous, widely used materials is soldering fluxes,
containing activating agents, such as chlorides.
Such widely used metallic finishes as tin (maintain solderability), zinc and
cadmium (sacrificial corrosion protection), are prone to grow metallic
whiskers which can short out closely spaced circuits. Addition of lead to tin
and reflowing after electrodeposition appears to prevent whisker growth.
An exception to this latter preventive method is when the part is under
compressive load.
STP534-EB/Nov. 1973
Chapter 7

Marine Industry
B. F. B r o w n 1
The operation of a ship involves many of the technologies required for

the functioning of an urban society: generation of electric power; storage
and preparation of food; dispensing of health services; furnishing heating,
ventilation, and air conditioning; and in some instances operating nuclear
power plants and aircraft. Thus, the corrosion control measures needed in
the marine industry include many that have been developed for other in-
dustries unassociated with the marine environment. In addition, special
measures are required because of the special corrosive nature of sea air and
seawater. It is these specialized measures which form the subject of this
chapter. (For example, corrosion control measures for steam generation
are not discussed because the subject is included in the chapter on that
technology.)
The special corrosivity of seawater and sea air is due of course to the
presence of the chloride ion in high concentrations. This ion is small, it can
diffuse rapidly, it confers on the electrolyte high electrical conductivity,
and when involved in hydrolysis reactions it can provide localized acidity
and thereby oppose passivation of metal surfaces. The ratio of dissolved
solids in sea water is about the same regardless of geographic location, but
the concentration may differ considerably. There are not enough systematic
long term corrosion rate data to establish whether the corrosivity of sea
water in different locations differs to an important degree in an engineering
sense. Nevertheless there have been enough observations reported to lead
many corrosion engineers to conclude that waters with higher oxygen
content (such as the cold waters of the polar sea) are more corrosive to steel
and require more cathodic protection current than water lower in oxygen.
But metallurgical differences, such as as-extruded surfaces of an aluminum
alloy compared with as-rolled or machined surfaces of the same alloy,
influence corrosion behavior much more drastically than geographic
differences.
The corrosion characteristics of the steel hulls of modern merchant and
1 Consulting Engineer, Washington, D.C.
133
Copyright9 1973 by ASTM lntcrnational www.astm.org
naval ships do not differ importantly from those of the wrought iron of the
Monitor of Civil War days. What is comparatively new in this respect is
the realization that protection against corrosion roughening is important
to the operating costs of merchant ships, and similar protection is important
to the maintenance costs of naval ships. This realization has been associated
with new developments in cathodic protection technology since World
War II, and these developments have given the marine industry important
tools for the control of marine corrosion.
Military considerations have tended to cause metallurgical developments
in marine technology ahead of their need or adoption by merchant ships,
though this is not universally the case. It is not surprising therefore that
most of the standard specifications cited in this chapter are Military
Specifications, abbreviated as M I L - . . . . Some of these specifications are
appropriate for merchant ships as well as naval ships; in some cases these
specifications may be inappropriate for merchant ships because of economic
considerations. Reference 1 should be consulted as the classic treatise on
economic considerations in the selection of materials for marine
applications.
The structuring of the presentation of material in this chapter posed
problems because the lack of sufficient data and standards to organize
sections parallel in nature. One section treats cathodic protection especially
as it is applied to unalloyed or low alloy steel hulls because that is where the
emphasis has been placed in technology development. Another section
treats aluminum alloys because they are both important to marine tech-
nology and vulnerable to special corrosion hazards in seawater. Still another
section treats certain corrosion processes (corrosion fatigue and stray cur-
rent corrosion) because they are serious present-day marine corrosion
problems for which standard control methods are lacking.
In addition to the formal standards for the control of marine corrosion,
there is a body of knowledge of corrosion technology available in printed
form (see for example Refs 2 and 3), although this information is far
skimpier than is sometimes supposed and much of the data present diffi-
culties in interpretation. In addition to this body of knowledge in printed
form, there is a great deal of marine corrosion information which exists
only in the oral tradition, some of it exchanged annually at the unique
institution known as the Seahorse Institute of the Francis L. LaQue
Corrosion Laboratory of the International Nickel Company. The avail-
ability of this written and unwritten body of information is cited to empha-
size to any newcomer to this literature that although formal standards
for corrosion control are far from complete, there is much marine ma-
terials engineering knowledge available to supplement them.
Cathodic Protection
The past twenty years have seen significant advances in the technology
of cathodic protection of hulls and other underwater structures; they have
CORROSION IN THE MARINE INDUSTRY 135
also seen major increases in the utilization of this technology for naval and
merchant shipping and also for offshore oil well towers. During the same
period there have also been many advances in the technology of organic
coatings for marine service. The combination of modern organic coatings
and modern cathodic protection technology gives the marine corrosion
engineer much better control over the corrosion problem than was possible
prior to World War I1.
Many corrosion engineers believe that it is more than coincidence that
the paint systems on cathodically protected (but not over-protected)
surfaces perform better and last longer than on unprotected surfaces.
Certainly cathodic protection of large areas is not economically sound
unless these areas are coated. Thus there is a direct connection between
coating technology and cathodic protection technology. The technologies
of surface preparation, primer formulation, anticorrosive paint formula-
tion, and removal of fouling organisms from merchant hulls are in a state
of dynamic development. Reference 4 is a recent authoritative and ex-
haustive review of these technologies. It is to be understood that many
developments in these technologies are made quite some time before they
are accepted in the MIL specification and Federal specification system.
The following specifications relate to surface preparation, coating formula-
tion, and application (in the case of galvanizing).
Subject Specification
Abrasive materials for blasting MIL-A-21380B; MIL-S-22262
Solvent cleaning compound for grease M1L-C-20207C
removal
Pickling inhibitor for use with sulfuric Fed Spec 0-1-501B
acid
Primer pretreatment, Formula 117 for MIL-P-15328B
metals
Shipboard primer coating and anticorro- MIL-P-15929B
sive paint, vinyl-red lead for hot spray
Shipboard primer coating, vinyl-zinc MIL-P-15930B
chromate for hot spray
Steel ship maintenance primer coating, MIL-P-18994A
alkyd-red lead type
Alkyd zinc chromate primer Fed Spec TT-P-645
Zinc chromate anticorrosive MIL-P-15184
Hot plastic antifouling MIL-P-19452
Cold plastic antifouling MIL-P-19449A and 19451A
Polyisobutylene antifouling MIL-P-22299A
Hot dip galvanizing MIL-Z-17871
There was much confusion over the true electrochemical characteristics

of various galvanic anode alloys until experiments were conducted on full
size anodes working at current densities approximating reasonable service

conditions in full salt seawater. The results of these tests [5-7] showed the
shortcomings of small scale laboratory tests and led to the development of
MIL-A-18001 G and H specification for zinc anodes. Instead of high purity
with respect to all elements, this specification calls for the alloy to contain
0.025-0.15~o Cd, 0.10-0.50~o A1, and the following maximum limits:
0.006 ~o Pb, 0.005 ~o Fe, 0.005 ~o Cu, and 0.125 7o Si, balance Zn. Within this
specification there are various classes and types which specify the presence
or absence of cores, and various anode shapes and sizes. This specification
has been conspicuously successful in producing anodes which have nearly
100 percent theoretical efficiency plus dependable freedom from going
passive. These anodes are used on both steel and aluminum hulls.
MIL-A-21412 specifies the composition of the alloy for magnesium base
anodes: 5-7~o A1, 2-4~o Zn, 0.15~o Mn (minimum), and the following
maximum limits: 0.3~o Si, 0.1~o Cu, 0.003~o Fe, 0.003~o Ni, and 0 . 3 ~o
other, with the balance being Mg. As with zinc anodes, there are numerous
types which specify size, shape, and the presence or absence of a coating.
At one time there were experimental installations of magnesium anodes on
bilge keels, with some of the anodes left coated at the time of undocking
in order to extend the life of the installations; these coatings would be cut
off by a diver as other anodes became consumed. A dielectric shield is
recommended extending 2 ft around each magnesium anode to avoid unduly
high currents in the immediate vicinity of the anode. This shield should be
a coal tar epoxy coating conforming to MIL-P-23236 (Type I, Class II).
Magnesium anodes are preferred over zinc or aluminum anodes for high
resistivity brackish or fresh water.
Both zinc and magnesium anodes have been used for partial cathodic
protection systems (around propellers) and for full protection of the entire
hull, and the use of both has been demonstrated to save maintenance
costs. Aluminum base anode alloys are attractive because of the high
theoretical capacity as well as possibly desirable electrochemical potential
characteristics. Indeed, numerous offshore structures employ aluminum
alloy anodes,though whether these alloys are functioning with the expected
electrochemical efficiency in all installations is unknown. At this writing
there are proprietary aluminum alloy anodes [8,9] which appear to have
desirable characteristics, and a military specification is presently being
prepared for this family of alloys.
Galvanic anodes are recommended for mounting within fairings of strut
bearings and stern tubes to the extent of space available. Because of
hydrodynamic considerations they are not attached to propellers, shafts,
rudders, or strut barrels.
Zinc anodes are recommended for bilges which are wet more than half
the time. Other applications of galvanic anode systems are discussed in
conjunction with condensers and salt water piping.
One of the perennial problems with the use of galvanic anodes around
ships is the tendency for paint crews to paint the anodes, even sometimes
when the admonition DO N O T P A I N T appears on the surface of the
anode casting. Positive action such as taping a protective layer of paper
over the anodes can keep paint off. The paper will come off the anode upon
undocking if no one remembers to remove it after painting.
Impressed Current Systems
Galvanic anodes have the advantage of low capitalization and simplicity
(they cannot be connected backwards, a distinct advantage where small
boats are maintained by a small crew with scant training and rapid turn-
over). If they are procured according to the specifications given above,
experience has shown that they will not go passive; if they are attached to a
metal structure in sufficient number and if there is a sufficiently low re-
sistance electrical path to that structure, they are effective. Navy data have
shown that their use saves money. However, if shore or shipboard power is
available, and if well trained electrical crew members are available, as on
larger ships, impressed current systems are preferred.
During the 1950s impressed current cathodic protection systems were
developed to control the potential of a ship hull at a predetermined value.
It is the potential of a metal surface, not the current density flowing into it,
which determines whether the metal corrodes or not. These potential
control systems--gigantic potentiostats--have been used largely for either
unalloyed steel hulls or for hulls low in alloying elements, although at least
one small Navy hydrofoil having an aluminum hull has been fitted with a
small scale model of the same type of controller, Potential control units
suitable for small boats are available commercially.
The potential prescribed for steel hulls in seawater is 0.85-1.0 V negative
to Ag/AgC1 (in seawater essentially identical to SCE). In high resistivity
water the allowed range is expanded to 0.75-1.0 V.
On active ships the shipboard power is rectified and conducted to anodes,
usually platinized titanium (though platinum, lead-platinum, lead-silver,
tantalum, and niobium have been used). A dielectric shield of coal tar
epoxy (MIL-P-23236, Class 2) 22-mils thick is laid on the steel hull 4 ft
around an anode with up to 12 V between anode and hull, or 6 ft if the
voltage exceeds 12. The preferred potential is - 0 . 8 5 ~ 0.02 V (Ag/AgC1),
and a good system should provide this at speeds up to 25 knots.
The reference electrode used to monitor the potentials of active ship
hulls is the Ag/AgC1 electrode (MIL-E-23919). Two or four such hull-
mounted electrodes are used depending upon the size of the ship, located
port and starboard at least 5 ft below the waterline. The Ag/AgC1 reference
electrode is for use only in seawater.
On active ships the propeller shafts are grounded to the hull using a
silver-graphite brush and a hard silver plate. The rudder is also grounded
to the hull.
Potential control impressed current cathodic systems have also been used
for laid-up ships with great success (except for an unfortunate instance of
reversed polarity). In this case shore power is rectified with selenium
rectifiers and is led to anodes which, unlike the ease of active ships, are
suspended in the water around the ship but at some distance from the hull.
These anodes may be of silverized lead (MIL-A-23871) if the site is essen-
tially full salt seawater. It is essential to have a minimum current density of
1 A / f t 2 coming out of the silverized lead anode surface to preserve the
brown peroxide coating necessary to maintain the integrity of the anodes.
Lead-antimony-silver alloys have also been used. Typically the lead-rich
anodes are 0.5 in. in diameter by 72 in. long (maximum current 8 A) or
0.75 in. in diameter by 72 in. long (maximum current 15 A).
Graphite anodes (MIL-A-18279) are used for laid-up ships in both salt
water and fresh water. The maximum current density of a 3- by 60-in.
cylindrical graphite anode is specified to be l0 A in salt water, and 5 A in
fresh water. (Graphite anodes have been used on at least one active ship,
but mechanical breakage was a problem.) At fresh water sites Ag/AgC1
reference electrodes are not used, but Cu/CuSO4 electrodes are used
instead.
Perhaps it should be noted in passing that the leading aluminum-base
anode contains traces of mercury, and that the ecological effect(s), if any,
of this mercury, of the cadmium in the standard Navy zinc anode alloy,
and of the lead in the silverized lead alloys may ultimately cause a shift in
the use of standard alloys.
Internal Salt Water Circuit
Salt Water Piping Systems
Zinc anodes are recommended for protecting sea chests. If the sea chest
is steel and the valve is nonferrous, a waster sleeve of mild steel is recom-
mended as additional protection. In the specified iron-bearing cupronickel
piping (either 70-30 or 90-10) the maximum flow rate is limited to 15 ft/s.
Corrosion and erosion-corrosion are minimized by having minimum flow
rate and by eliminating air. During idle periods (one week or more in
duration) the system should either be drained or else operated daily. Pro-
tective spools are neither specified nor desired for either ferrous or non-
ferrous systems.
Mixed nonferrous and ferrous systems should be designed to include a
24-in. (12-in. minimum) waster piece of extra heavy galvanized steel on
either side of the nonferrous section, connected in such a way as to afford
easy removal. There is no problem in mixing 70-30 with 90-10 cupronickel,
but 90-10 should not be used to replace 70-30 in main condensers equipped
with solder-coated water boxes. In general, however, 90-10 is preferred to
70-30 because of cost differential except as noted above and where the
additional strength of 70-30 is needed.
For protection against galvanic corrosion in stagnant ( < 5 knots) areas
in such systems as bilge pump strainers and valves, involving mostly
stagnant conditions and various combinations of steel, stainless steel, and
bronze, etc., one rule of thumb is to attach a zinc anode to provide 1 ft 2 of
zinc for each 50 ft 2 of bare metal to be protected.
Flooding due to rapidly corroding ferrous plugs in copper-base piping
continues to occur.
Condensers and Other Salt Water Heat Exchangers

As with so many components and structures, corrosion control measures
involve design, material selection, fabrication, and maintenance. Designs
avoid erosion-corrosion by attention to the configuration of waterboxes and
injection piping. General specifications require designs to limit the flow
rate of either seawater or brine to 15 ft/s for both 70-30 and 90-10 cupro-
nickel to minimize erosion-corrosion.
The waterboxes on combatant ships are of MoneF alloy 400 with the
water side coated with solder (two-thirds lead, one-third tin) to minimize
galvanic attack on the tube sheets and tubes-themselves. Cast iron or steel
waterboxes on noncombatant ships are abrasive blasted and coated with a
coal tar epoxy equivalent to MIL-P-23236, Class 1, 2, or 4.
The cupronickels used for salt water piping contain iron which is thought
to improve the integrity of the corrosion product layer which controls the
corrosion of the underlying metal. There are a few seawater condensers
into which iron compounds are deliberately introduced periodically to
assist in maintaining a protective film. It is believed that initially operating
a cupronickel condenser in badly polluted water seriously hazards the
development of a satisfactory protective coating. Indeed standard practice
for a condenser of any age is to avoid operating in polluted water if at all
possible.
During shutdown there is a tendency for seawater to pocket, especially
in sagging tubes. Such pocketing causes serious corrosion which can be
avoided by washing with fresh water and then drying out (using an air
lance to empty sagging tubes, for example). If the condenser cannot be
completely dried out during shutdown, it should be filled completely with
fresh water.
Stones, pieces of wood, shell, etc., lodging in condenser tubes cause
pitting corrosion. They are removed periodically by use of a water lance,
soft rubber plugs driven through by an air gun, or a rotating bristle brush.
Wire brushing and abrasive cleaning of condenser tubes are specifically
prohibited, since that would remove all protective coating.
Monel is a registeredtrademark of the InternationalNickel Company.
Attention to the upkeep of the ship's electrical system should eliminate

serious condenser tube corrosion due to stray currents.
Zinc anodes were formerly mandatory on the seawater side of con-
densers, but many such installations have now been eliminated. These
"protector plates" were sized to give a zinc area Z = 0.078 (0.75D 2 4_
6Nd 2) where D is the diameter of the tube sheet, N is the number of tube
ends exposed, and d is the inside diameter of the tubes. The zinc when used
must conform in composition to MIL-A-18001. It is recommended that
the zinc surfaces be inspected at 90-day intervals to ensure that they are
active, though it is unlikely that they will be otherwise if the material
actually conforms to MIL-A-18001. If the zinc becomes filmed and in-
active, it should be wire brushed. The cathodic protection afforded by these
zinc anodes does not extend down the tubes more than a few tube diameters
if the seawater is flowing, but it is of some help in mitigating the erosion-
corrosion near the tube inlet end, though incompletely so. Plastic inserts
were tried as protectors in this area, but their use has been largely dis-
continued.
Oil coolers when out of service more than 24 h should have the seawater
drained, and should then be flushed with fresh water and dried.
Special Forms of Marine Corrosion
Stray Current Corrosion
This form of corrosion is probably more widespread than is commonly
realized because of the difficulty in identifying it except in the more severe
cases. It is recognized as a hazard around cathodically protected structures
in seawater and also especially if welding is done where the welding gen-
erator is not mounted directly on the ship or structure being welded.
These hazards have led to the following standard practices.
No unprotected ship should be moored in a group of protected ships.
The reason for this rule is that the unprotected ship represents a low
resistance path in a higher resistance electrolyte carrying an electric current.
Kirchoff's law is obeyed, and part of the cathodic protection current enters
one part of the ship (causing no corrosion) and exiting at another point
(causing much corrosion). The unprotected ship then functions as an
intermediate electrode. For exactly the same reason a steel camel should be
electrically bonded to an adjoining protected ship.
Welding generators should be mounted directly on the ship or structure
being welded. The reason is that if the generator (d-c rectifier) is located on
shore, an enormous d-c current will run from ship to shore during welding,
tar too large a current for a cathodic protection system to overcome.
Running a d-c grounding cable ashore is of little value, since the total
resistance of the large cross section electrolytic path is small compared to
anything less than an enormously large cable, and again Kirchoff's law is
obeyed.
At the present time we do not know whether a-c fields accelerate

corrosion of the various structural alloys in seawater.
Corrosion Fatigue
The most c o m m o n form of fracture failure in ships is corrosion fatigue.
One might wonder therefore why there are no standard corrosion fatigue
tests and why the designer apparently pays little attention to corrosion
fatigue data. The reason is that the designer seldom knows the algebraic
sum of residual plus working stresses, and he does not know whether the
order of merit in one corrosion fatigue test is the same as in a different test.
A typical engineering solution to a corrosion fatigue problem is exemplified
in the following procedure for avoiding corrosion fatigue in propeller
shafts: the steel is protected against the electrolytic action from the bronze
propeller and bronze bearing journal sleeves by a rubber or plastic sheath
over the steel. Seals and rust preventive compounds are used to prevent
the entry of seawater under the sheath. The ends of shafts are plugged to
prevent internal corrosion.
Corrosion fatigue is not the only serious consequence of corrosion lacking
a formally accepted test procedure. The same is also true of stress corrosion
cracking, erosion-corrosion, hydrogen embrittlement, crevice corrosion,
pitting, and even general corrosion in seawater. Neither are there clear cut
procedures for interpreting the results of most of the marine corrosion
tests now in use for purposes of design of fundamentally new structures or
of using fundamentally new alloys. Unfortunately there are very few
laboratories so sited and staffed as to be able to correct the foregoing
deficiencies in the near future.
Aluminum Alloys
Although aluminum-hulled boats were built before the end of the 19th
century, aluminum continues to give serious corrosion problems in marine
service, even with the ailoys which are resistant to stress corrosion cracking
and to exfoliation. The reason is that the metal is fundamentally active,
and the protective oxide which coats it is more easily broken down chemi-
cally and is less readily repaired than the oxide coating on say titanium.
The common alloys for boat hulls are 5086 or 5456. For piping and
railings 6061 is used. 5083 and 7039 have been used in limited quantities
for atmospheric service. The alloy 5086 is bought to an interim Federal
specification (QQ-A-00250/19 of 11 Dec. 1968) which requires each pro-
duction lot to be checked metallographically for evidence of susceptibility
to exfoliation. If the candidate lot is predominantly free of a continuous
grain boundary network and the microstructure is equivalent to or "better
t h a n " a reference standard, no further testing is required. If the candidate
lot fails this check, then specimens from the lot must successfully pass the
standard salt water acetic acid test ( " S W A A T " ) in order to qualify. The
same specification applies to the 5456 alloy. It should be emphasized that

this is a dynamic technology area.
One of the first rules in using aluminum successfully in seawater is to
avoid machining the as-rolled or as-extruded surface if at all possible. The
machined surface corrodes far worse than the unmachined. The second
rule is to avoid using steel wool or wire brushes (other than stainless) on
aluminum; sanding may be done if needed, but avoid using abrasive disks
or other nonmetallic scouring pads which h a v e b e e n used to remove paints
which contained copper or mercury antifouling compounds. Wood or
plastic scrapers are to be preferred.
The most common causes of galvanic corrosion problems in aluminum
in ships are placing an aluminum deckhouse on a steel support and attach-
ing steel or copper-base alloy fittings to the hull or to the piping, or both.
These more noble metal fittings should be electrically insulated from the
aluminum; even so, waster-plate practice is recommended. Cathodic
protection is recommended where copper alloys are involved even though
the copper alloy is electrically insulated from the aluminum, for there is a
degrading effect from simple proximity of the copper alloy to the alumi-
num alloy.
Unlike steel, aluminum cannot be made thermodynamically stable in
seawater by cathodic protection methods. Nevertheless, cathodic pro-
tection is highly effective because it counteracts any tendency toward
localized hydrolytic acidification with its attendant breakdown of passivity.
Only zinc (MIL-A-18001) and certain aluminum-base galvanic anodes are
permitted, magnesium anodes and mercury-bearing aluminum anodes
being prohibited for aluminum hulls.
Cleanliness is always important, as both solid particles and grease marks
invite localized hydrolytic acidification and consequent local film break-
down. Regular fresh water wash down of decks and bilges is helpful in
removing solid debris and also in keeping down the concentration of
chloride which plays an essential role in hydrolytic acidification.
Paint coatings play an important role in controlling hull corrosion. It is
standard practice to have a 2-mil primer, followed by a 2-mil antifouling
paint containing neither copper nor mercury. One of the organo-tin com-
pounds is presently the preferred toxicant in antifouling paint for alumi-
num hulls.
Where there are threaded bolts or fittings of aluminum exposed to sea-
water, the use of 50 percent zinc dust in petrolatum may be used on the
threads as an anti-seize lubricant. Graphite, lead, or tin bearing lubricants
must be avoided. Stainless steel lock nuts and washers are usually accept-
able on such aluminum components.
Standard stipulations for aluminum alloys for marine atmosphere service
call for "stress-corrosion resistant alloys" without, however, specifying the
test. Where cadmium-plated steel bolts or nuts are used with aluminum
alloy parts, they are preferably isolated from the aluminum with aluminum
alloy washers under the bolt head and nut; but cadmium-plated washers
m a y be used for bolts under high tension. The usual precautions of
minimizing stress concentration, using shot peening and stress relieving, are
r e c o m m e n d e d where feasible to minimize stress corrosion problems. In
highly textured alloys a rule of t h u m b is to keep working stresses in the
longitudinal direction less than 50 percent of the yield, 35 percent in the
long transverse direction, and 25 percent in the short transverse direction.
Faying surfaces should be filled with sealing c o m p o u n d . Aluminum hulled
boats should be provided with nonconducting mooring lines when next to
steel ships and steel piers, and insulating camels should be used.
Marine Structures Other Than Ships

The corrosion control technology of offshore structures tends to m a k e
use of developments in corrosion control for ships. The potential-control
impressed current cathodic protection controller for protecting the tower
known as "Argus Island" near Bermuda was derived f r o m the controller
originally developed for N a v y ship hulls. M a n y offshore towers and
ancillary components are protected by galvanic anodes previously proved
out on ship hulls.
Note
The military specifications applying to the marine environment are
developed by the U.S. N a v y and m a y be requested f r o m the Naval Publi-
cation and Distribution Center, 5801 T a b o r Ave., Philadelphia, Pa.19120.
References
[1] LaQue, F. L. and TuthiU, A. H., Transactions of the Society of Naval Architects and
Marine Engineers, Vol. 69, 1961, p. 619.
[2] Tuthill, A. H. and Schillmoller, C. M., Guidelines for Selection of Marine Materials, The
International Nickel Co., New York, 1966.
[3] Rogers, T. H., Marine Corrosion, Newnes, London, 1968.
[4] Saroyan, J. R. in Handbook of Ocean and Underwater Engineering, J. J. Myers, C. H.
Holm, and R. F. McAllister, Eds., McGraw-Hill, New York, 1969, p. 7-37.
[5] Reichard, E. C. and Lennox, T. J., Jr., Corrosion, Vol. 13, 1957, p. 410t.
[6] Waldron, L. J. and Peterson, M. H., Corrosion, Vol. 16, 1960. p. 375t.
[7] Lennox, T. J., Jr., Materials Protection, Vol. 1, 1962, p. 37.
[8] Lennox, T. J., Jr., Groover, R. E., and Peterson, M. H., Materials Protection, Vol. 7,
1968, p. 33.
[9] Lennox, T. J., Jr., Groover, R. E., and Peterson, M. H., Materials Protection and
Performance, Vol. 10, 1971, p. 39.
STP534-EB/Nov. 1973
Chapter 8

Nuclear Power Industry
W. E. B e r r y I
The first commercial nuclear fuel power plant in the United States was
the Shippingport Atomic Power Station (PWR) that commenced operations
late in 1957. Despite the brief existence of this industry, there have been a
number of standards promulgated and adapted to meet the industry
requirements. The need for standards in this field is emphasized by the high
performance requirements and the necessity for overdesign to insure safety
and prevent ecological damage.
In the areas of corrosion control, two standards have been issued that
relate directly to corrosion testing (ASTM G 2-67 and NACE TM-01-71)
and a third is about to be issued (NACE). However, a number of standards
have been issued with special corrosion requirements or have been adapted
to contain these requirements. In addition, many existing standards on
corrosion have been adopted in toto for nuclear-industry applications.
The primary sources of the standards are: (1) U.S. Atomic Energy Com-
mission Division, Division of Reactor Development and Technology
(RDT); (2) American Society for Testing and Materials (ASTM); (3) Na-
tional Association of Corrosion Engineers (NACE); (4) Military and
Federal Standards and Specification (MIL, FED); and (5) American
Society of Mechanical Engineers Pressure Vessel Code Material Specifi-
cations (ASME).
The ASTM recommended practices are well known in the scientific
community. The NACE has recently begun to issue standards to insure that
the field of corrosion is adequately covered. The issuance of standards by
the AEC-RDT to provide assistance and guidelines is a natural consequence
of the AEC being a tax-supported regulatory body. The military and other
branches of the Federal Government have long issued specifications and
standards as necessary to the conduct of their operations. The ASME ma-
terial specifications recognize corrosion and are more indirectly applicable
to corrosion in the nuclear power industry than are the other sources.
1 Battelle Memorial Institute, Columbus Laboratories, Columbus, Ohio.
144
CORROSION IN THE NUCLEAR POWER INDUSTRY 145
Background
Before embarking on a discussion of the use of standards in the nuclear-
power industry, it is imperative that a brief description be given of the
characteristics and problems of the industry. Since the industry is less than
two decades old, new problems are continually arising as more stations
are put in service and experience is gained with older stations. Thus, the
critical corrosion problems of today may be superceded by other critical
problems in the future as the current problems are solved and new ones
emerge.
At present, the nuclear power industry is based on the water-cooled
reactor concept. These are low-temperature (~260 C) units that generate
steam at pressures less than 1000 psi. They employ thermal neutrons
( < 1 MeV) to sustain nuclear fission and consume fissionable material. A
concerted effort is now under way to develop a Iiquid metal (sodium)
cooled breeder reactor that will generate steam at higher temperatures
(~540 C) and possibly higher pressures. This reactor will employ fast
neutrons ( > 1 MeV) and will breed fissionable material. Sufficient ex-
perience has been gained with the liquid-metal cooled reactors that many
of the corrosion problems are well known, and corrosion control standards
have been issued or are being issued.
Both the water and liquid-metal cooled reactors rely heavily on stainless
steel for piping, valves, pumps, vessel linings, and heat exchangers. Other
materials in common may include zirconium alloys, uranium-oxide fuel,
and low alloy or carbon steels. Thus, many corrosion problems that arise
as the result of cleaning, pickling, fabrication, and plant erection are
common to the two types of reactors.
A current problem in reactor construction is the possibility of inter-
granular attack or stress-corrosion cracking of sensitized austenitic stainless
steel. The sensitization occurs in stainless steel components attached to
carbon steel vessels that must be stress relieved to meet code requirements.
The stress-relieving temperature is ~650 C which results in heavy chro-
mium-carbide precipitation at grain boundaries and sensitization of the
stainless steel. Laboratory studies have shown that moisture and the
fluoride ions from fumes or spatter produced by welding with coated elec-
trodes can cause intergranular attack of unstressed specimens and stress-
corrosion cracking of stressed specimens of sensitized stainless steel [1].2
Nickel-chromium-iron alloys do not exhibit this behavior. Furthermore,
it has been shown that sensitized austenitic stainless steel (Type 304) will
exhibit stress-corrosion cracking in a marine atmosphere [2]. Since both of
the above conditions may be present at a nuclear power plant construction
site, measures must be taken to either eliminate or avoid the problem.
Sensitization is also a problem if the component is subsequently cleaned or
pickled in strong acids such as HNO3-HF. Intergranular attack many mils

deep has been produced under these conditions and the attack led to stress-
corrosion cracking failure in service [3].
Heavily sensitized stainless steel components have failed by stress-
corrosion cracking in boiling-water reactors particularly at high stress
loadings and when the oxygen content of the water was abnormally high
[4-7]. Presumably highly stressed sensitized stainless steel would also
exhibit stress-corrosion cracking in pressurized water reactors if there was
a prolonged incident of high oxygen in the coolant.
Chlorides can also cause cracking of carbon steel boiler components
under certain conditions. Ferric chloride has produced transgranular
cracking of mild steel in aqueous solutions at 316 C [8]. Ferric chloride can
be produced by the periodic introduction of air into boiler water that con-
tains chloride (by such practices as uncontrolled shutdown of boilers over
weekends).
Carbon- and low-alloy steel components are also susceptible to rusting
and pitting corrosion during transfer and storage unless they are purged
and contain a desiccant to maintain low humidity and are sealed to prevent
ingress of high humidity air. Failure to remove rust prior to operation can
lead to pitting when the unit is operated with water or steam at high
temperatures. During down time, air must be excluded from carbon steel
units if they remain wet to avoid pitting.
The principal corrosion problems associated with the operation of
water-cooled reactors are discussed in Ref 9 and include stress-corrosion
cracking of sensitized stainless steel (see preceding paragraphs), chloride or
caustic cracking of rolled in heat-exchanger tubes, localized attack of
Zircaloy cladding due to 'fluoride contamination or localized concentration
of caustic materials used to treat the primary coolant, the corrosion of con-
denser tubes, and the formation and transport of crud (stainless steel
corrosion products). The latter can be serious because the transported crud
eventually deposits in low velocity areas and can affect flow characteristics
and component operations (such as valves). The crud also is activated as it
passes through the neutron flux in the core and when it subsequently
deposits in areas remote from the core, raises the radioactivity level in that
area. This is a major problem in maintaining and repairing reactor com-
ponents. Not only is the activity level high but cutting and welding opera-
tions produce airborne particles that might be ingested into the body.
The major corrosion problems associated with the operation of sodium-
cooled reactors are somewhat different from those with water-cooled
reactors. The same problems exist on the steam side, that is, pitting of steel
components and chloride or caustic stress-corrosion cracking of stainless
steel components. Sodium is a reactive metal and readily reacts with
oxygen, nitrogen, and hydrogen and will pick up carbon. As the dissolved
oxygen content in sodium increases, the corrosion rate of iron-base alloys
also increases [10]. Nitrogen may cause nitriding under some conditions
[11]. Hydrogen does not appear to be a problem because of its low content
in sodium and the fact that most hydride-former metals (such as zirconium)
react with the oxygen in the sodium to form a protective oxide surface
layer [11]. (At temperatures > 6 0 0 C, zirconium will dissolve its own
oxide [11].) Carbon in sodium can lead to carburization of stainless steel
[11,12]. In binary metallic systems, carbon may be transported from one
material to the other. As an example, low-alloy steels in the cooler zone of a
sodium circuit become decarburized while austenitic stainless steels in the
hotter zone of the same circuit become carburized [13].
The removal and steam cleaning of components from a sodium system
and their reinsertion into the sodium can be a potential problem if crevices
or other areas of entrapment exist. When moisture contacts the sodium,
sodium hydroxide is formed and hydrogen is evolved. On exposed surfaces,
the N a O H is washed away or is slowly converted to Na~CO3 by reaction
with CO~ in the air. However, in crevices, the rate of formation of Na2CO8
is slow because of diffusion of CO2. Thus, under these conditions, a com-
ponent could be returned to the sodium coolant with N a O H in the crevice.
If there are high stresses in this area, cracking can occur because studies
have shown that N a O H or N a O H / N a will produce cracking in stressed
stainless steel at temperatures on the order of 450 C [14-16]. For this
reason, care must be exercised that moisture not enter a vessel that has
been drained of sodium and is subsequently to be refilled with sodium and
heated to high temperature.
The slow leakage of sodium into the atmosphere can also produce inter-
granular attack of stainless steel at ambient temperature [15]. Presumably
Na20 is the corrodent because concentrated N a O H and Na~CO3 are
routinely handled in steel or stainless steel at ambient temperature with no
corrosion problems.
A water-steam jet into sodium (as at a leak in a heat exchanger) produces
rapid metal wastage of steel, stainless steel, and nickel-base alloys [17].
However, this is not a normal operating problem in sodium-cooled reactors.
Application of Standards
The standards that apply directly or indirectly to materials in, and the
operation of, nuclear power plants are presented in Table 1. R d a t e d
standards will be mentioned in the text and in the references at the end of
the chapter. Except where stated, the standards apply to both water-cooled
and sodium-cooled reactors. In the following sections, only the general
features of the standards are discussed and for the detailed procedures one
must consult the standard itself.
Material Requirements
Standards on corrosion criteria for steel, stainless steel, and nickel-base
alloy sheet, plate, tubing, and piping are included in M I L SPECS and R D T
University of Washington (University of Washington) pursuant to License Agreement. No furthe
O0
TABLE 1--Standards and specifications relating to corrosion in the nuclear power industry.
Tc
Standard Number Date Title IO
C
American Society of Mechanical Engineers m_

SA-155 1970a Specification for Electric-Fusion Welded Steel Pipe for High Temperature Service
SA-240 1968 - Specification for Chromium and Chromium-Nickel Stainless Steel Plate, Sheet, and Strip for Fusion- r
Welded Unfired Pressure Vessels m
O
SA-312 1968 ~ Specification for Seamless and Welded Austenitic Stainless Steel Pipe
SB-407 1968 a Specification for Nickel-Iron-Chromium Alloy Seamless Pipe and Tube
SB-409 1968 a Specification for Nickel-Iron-Chromium Alloy Plates, Sheet, and Strip z
American Society for Testing and Materials
ASTM A 155-71 1971 Specification for Electric-Fusion Welded Steel Pipe for High-Pressure Service t~z
ASTM A 240-71 1971 Specification for Stainless and Heat-Resisting Chromium and Chromium-Nickel Steel Plate, Sheet, and ~"
Strip for Fusion-Welded Unfired Pressure Vessels
ASTM A 312-71 1971 Specification for Seamless and Welded Austenitic Stainless Steel Pipe >
ASTM A 262-70 1970 Recommended Practices for Detecting Susceptibility to Intergranular Attack in Stainless Steels x
ASTM A 393-63 1963 Recommended Practice for Conducting Acidified Copper Sulfate Test for Intergranular Attack in Austenitic
Stainless Steel O
ASTM B 407-71 1971 Specification for Nickel-Iron-Chromium Alloy Seamless Pipe and Tube
ASTM B 409-70 1970 Specification for Nickel-Iron-Chromium Alloy Plate, Sheet, and Strip
ASTM G 2-67 1967 Recommended Practice for Aqueous Corrosion Testing of Samples of Zirconium and Zirconium Alloys
ASTM G 28-71 1971 Method of Detecting Susceptibility to Intergranular Attack in Wrought Nickel-Rich Chromium-Bearing
Alloys
NACE Standard 1971 Test Method Autoclave Corrosion Testing of Metals in High-Temperature Water
TM-01-71
U.S. Atomic Energy Commission, Division of Reactor Development and Technology
RDT AI-IT Oct. 1971 Coolant Composition in Pressurized Water Reactors
R D T A1-2 in prep. Coolant Composition in Boiling Water Reactors
R D T A1-3 in prep. Coolant Composition in Light Water Cooled Test, Research, and Experimental Reactors
RDT A1-5 in prep. Purity Specifications for Large Operating Sodium Systems
RDT C8-5T Jan. 1972 Electrochemical Oxygen Meter for Sodium Service
RDT C8-6T Jan. 1972 Hydrogen Meter for Sodium Service
RDT C8-7T Jan. 1972 Diffusion Type Carbon Meter for Sodium Service
RDT C8-8 Jan. 1972 Equilibration Device for Measuring Nonmetallic Impurity Activities in Sodium
RDT E4-1T Dec. 1971 Steam Generator for Pressurized Water Reactors
RDT E4-5T Dec. 1970 Forced Circulation Cold Trap Assembly for Removal of Sodium Impurities
Amendment 1 Sept. 1971
RDT E4-6T April 1971 Intermediate Heat Exchanger for Liquid Metals
Amendment 1 June 1971
RDT E4-7T June 1971 Sodium to Air Heat Exchanger
Amendment 1 Oct. 1971
RDT E4-17T July 1971 Heat Exchanger, Class 2, Water-to-Water, Straight or U-Tube
t3
RDT E5-1T Dec. 1970 Pressurizer for Pressurized Water Reactors O
RDT E8-16 to be prepared Environmental Conditioning Stations for Refueling Operations
RDT E13-15T July 1971 Fuel Assemblies for Pressurized Water Reactors O
t~
RDT E15-2T July 1971 Requirements for Nuclear Components
Amendment 1 July 1971 z
RDT F3-6T July 1971 Nondestructive Examination
RDT F5-1T March 1969 Cleaning and Cleanliness Requirements for Nuclear Reactor Components
RDT F5-4 to be prepared Cleaning Components Contaminated with Sodium (Nonradioactive)
RDT F7-2T Feb. 1969 Preparations for Sealing, Packaging, Packing, and Marking Components for Shipment and Storage Z
C
Amendment 1 Oct. 1971
RDT M3-6T May 1969 Austenitic Stainless Steel Pipe N
g
RDT M3-9T July 1971 Nickel-Iron-Chromium Alloy Seamless Pipe and Tubing
RDT M3-11T July 1971 Carbon and Low Alloy Steel Welded Pipe
RDT M5-1T July 1971 Stainless Steel Plate, Sheet, and Strip
RDT M5-7T July 1971 Nickel-Iron-Chromium Alloys Plate, Sheet, and Strip N
RDT M12-1T Feb. 1969 Test Requirements for Thermal Insulating Materials for Use on Austenitic Stainless Steels
RDT M13-1T June 1970 Reactor Grade Sodium-Purchase Specifications
Amendment I Oct. 1971
Amendment 2 Dec. 1971
4~
~O
o
Standard Number Date Title
U.S. Government Federal and Military Specifications and Standards

Fed. Spec. QQ-P-35a 20 June 1968 Passivation Treatments for Austenitic, Ferritic, and Martensite Corrosion-Resisting Steel (Fastening
Devices)
Federal Test Method 6 M a y 1959 lntergranular-Corrosion Test for Corrosion Resistant Austenitic Steels
STD No. 151a
Method 821.1
NAVSHIPS 250-1500-1 Jan. 1968 " Standard for Welding of Reactor Coolant and Associated Systems and Components for Naval Nuclear
Advance Change Power Plants (Pressurized Water-Cooled Systems)
Notice No. 2
MIL-B-3180B 20 Feb. 1967 Boilers and Related Equipment, Packaging of
MIL-S-23196A 10 June 1966 Military Specification, Steel Plate, Sheet and Strip; Corrosion Resistant
M IL-T-23226A 26 Oct. 1964 Military Specification. Tube and Pipe, Corrosion-Resistant Steel, Seamless
a ASME Boiler and Pressure Vessel Code Section II, 1968, Material Specs, Part A, Ferrous, and Winter and Summer, 1970 Addenda.
Standards. Included in the MIL SPECS are MIL-S-23196A (SHIPS) and

MIL-T-23226A (SHIPS) for corrosion-resistant steel sheet and strip and
tube and pipe. The standards require that the material be free of grain
boundary carbide precipitates which can cause intergranular corrosion.
(It should be noted that the operating temperature of some components in a
sodium-cooled reactor may be in excess of 540 C and this temperature is
sufficiently high to cause carbide precipitation at grain boundaries in
austenitic stainless steels.) A hot-acid-etch test is specified according to
Method 321 of FED-STD-151. This test consists of a 48-h exposure of
the subject material to a boiling solution of 100 g of copper sulfate
(CuSO4.5H~O) and 100 ml of sulfuric acid (H~SO4, sp gr 1.84) and sufficient
distilled water to make 1000 ml of solution. Upon completion of the 48-h
exposure, the specimens are bent and examined for cracks and fissures that
are indications of intergranular attack. Flat specimens are bent through
180-deg over a diameter equal to the thickness (IT) of the specimen. Tube
specimens are flattened as prescribed in ASTM A 370 [18]. However,
FED-STD-151 has recently been modified and Method 321 now is identical
with ASTM 393. The latter test contains the same provisons as those in
Method 321 except that the exposure period in the boiling copper-sulfate-
sulfuric acid is 72 h instead of 48 h. As an interesting side light, there is
increasing evidence that the copper sulfate-sulfuric acid test is "too mild"
and does not detect many cases of sensitization and in particular, where the
time at temperature is short and the carbon content of the stainless steel is
low (0.04 to 0.05 percent). Research performed on the problem has revealed
that the test can be made more sensitive by contacting the specimen with
metallic copper during its exposure to the copper sulfate-sulfuric acid
solution [19-22]. Accordingly, ASTM A 262 has recently been changed and
now includes the copper-copper sulfate-sulfuric acid test. The procedures
are essentially he same as those described in ASTM A 393 and Method 321
of FED-STD-151 except that the specimen is covered with copper shot or
turnings and the exposure time is 24 h.
There are also standards for detecting susceptibility to intergranular
attack in nickel-base alloys. ASTM G 28 specifies exposure to a boiling
solution of 25-g reagent grade ferric sulfate [Fe2(SO03] and 236 ml reagent
grade 95 to 98 percent sulfuric acid (H2SO4) added to 400 ml distilled water.
Alloy Testing Time, h

Hastelloy Alloy C 24
Hastelloy Alloy C-276 24
Hastelloy Alloy G 120
Carpenter Stainless No. 20 Cb-3 120
Inconel Alloy 600 24
Incoloy Alloy 800 120
Incoloy Alloy 825 120
The presence of intergranular attack is usually determined by comparing

the calculated corrosion rate to that for properly annealed material.
Metallographic examination may also be used to determine the degree of
intergranular attack.
The RDT standards pertaining to materials include RDT's M3-11T,
M5-1T, M3-6T, M5-7T, and M3-9T. These are not complete standards but
are additional requirements to ASME SA-155, SA-240, SA-312, SB-409,
and SB-407, respectively, which, in turn, are identical with ASTM A 155,
A 240, A 312, B 409, and B 407 respectively. The ASTM and ASME stand-
ards specify that stainless steel shall be heat treated to meet the mechanical
property requirements of the specification and be capable of meeting the
test for resistance to intergranular corrosion (ASTM A 393). However, the
intergranular corrosion test is not required unless it is specified on the
purchase order. No intergranular test is required for the nickel-iron-
chromium alloys although the specified heat treatments are as hot-finished,
annealed, or solution-annealed.
From a corrosion standpoint, the principal additional requirements in the
RDT's specify cleanliness and the limits for chloride and sulfur contamina-
tion in handling stainless steel. The contamination requirements include"
(1) Marking materials shall not contain sulfur, chlorine, or other halogens
in amounts greater than 200 ppm; (2) All processing compounds, degreasing
agents, cleaning solutions, and foreign materials shall be completely
removed at any stage of processing prior to any elevated-temperature
treatment; and (3) Any pickling or descaling in a bath containing chlorides
shall be followed immediately by a nitric acid pickle, followed by immediate
rinsing with hot water containing not more than 20 ppm chloride.
The above three procedures are designed to minimize the threat of stress-
corrosion cracking of sensitized stainless steel and in the ease of Item 2, the
cracking of solution-treated material.
Manufacturing Requirements
Corrosion-related standards that pertain to manufacturing procedures
are included in RDT Standards F5-1T, E4-6T, E4-7T, E4-17T, ES-1T, and
E13-15T. The first of these deals with cleaning requirements and contains
the following pertinent requirements:
1. Grades of water (A, B, or C) to be used in hydrostatic testing and
rinsing of components.
2. Precautions against the use of oxidizing agents such as nitric acid or
the exposure to halide-bearing environments such as salt air of stainless
steels that have been sensitized by heating in the range 425 to 870 C.
3. Prohibition of the following where crevices are present (may be used
if there are no crevices):
a. Vapor degreasing
b. Trisodium-phosphate degreasing
c. Acid cleaning
4. Prohibition of halogenated cleaning agents and solvents on austenitic
stainless steels.
5. Prohibition of the use of materials containing lead or sulfur on
nickel-base alloys (to prevent embrittlement if the component should be
subsequently heated to high temperatures).
6. Prohibition of aluminum in contact with stainless steel or nickel-base
alloys or zinc in contact with stainless steel.
.7. Prohibition of mercury or mercury compounds in any equipment.
8. Prohibition of corrosion inhibitors from materials intended for
coolant surfaces. This restriction is probably related to a ease of inter-
granular pitt'ng of heat-exchanger tubes where one of the proposed causes
of the attack was the degradation of an inhibitor.
9. Clean room or clean area requirements starting with that stage of
fabrication where critical surfaces that will be exposed to reactor coolant
will no longer be accessible for cleaning or inspection.
10. The removal of rust from critical surfaces that will be exposed to
reactor coolants. Rust contains ferric ions, frequently retains the corrodent
(such as chlorides), and may be hygroscopic so that failure to remove rust
could result in localized attack and possibly stress-corrosion cracking if
the component is highly stressed.
11. Protection of components and materials from the general shop
atmosphere or other contaminated atmospheres, such as salt air or blowing
dust, where possible during fabrication and storage.
Surface finish and cleaning are specified in R D T E4-17T, E5-1T, and
E13-15T. The latter repeats many of the restrictions described above and
prohibits mercury, lead, phosphorus, zinc, cadmium, tin, antimony,
bismuth, mischmetal, and other similar alloys during fabrication, cleaning,
testing, or final assembly of components. R D T E5-1T prohibits electro-
plating on surfaces in contact with pressurized water or steam presumably
because under certain conditions, electroplates, such as nickel corrode
and spall and are transported by suspension or solution.
Federal Specification QQ-P-35 is often followed to passivate stainless
steel after fabrication. The passivation treatment consists of immersion for
10 to 30 min at 70 to 155 F (depending upon the alloy) in a 20 to 25 volume
percent nitric acid (HNO8 at sp gr 1.42) solution containing 2 to 3 weight
percent sodium dichromate (Na2Cr~OT.2H20). An alternate procedure is
30 to 60 rain in 20 to 50 volume percent HNO3 (sp gr 1.42) at r o o m tem-
perature. This is followed by a hot-water rinse, and for ferritic or martensitic
stainless steels, a 30-rain immersion in 4 to 6 weight percent Na2Cr~OT. 2H~O
at 140 to 160 F followed by a hot-water rinse. A passivated surface must
pass a 24-h exposure to 100 percent humidity at 100 F as prescribed in
Method 101 of MIL STD 753 [23]. Many believe that the function of the
passivation treatment is not to form a protective film but to remove con-
tamination (particularly iron) from the surface and thereby prevent the
localized breakdown of the already existing protective film during sub-
sequent service.
Ineluded in R D T E4-6T and E4-7T are requirements that crevices be
minimized and corrosion allowances be included for liquid metal com-
ponents. The corrosion allowance is to compensate for loss of material due
to erosion, corrosion, carbon transport, sodium-water reaction effects and
other surface wastage effects. The crevices are minimized to avoid retention
of Na20 or N a O H that might form if the component is removed from
service, exposed to moist conditions, and then put back in service.
Packaging Transportation and Storage Requirements

The cleanliness and contamination requirements described in the previous
two sections also apply to packaging, transportation, and storage. These are
reiterated in R D T F7-2T plus additional prohibitions on materials con-
taining fluorides and copper. (The intergranular attack of sensitized stain-
less steel by fluorides has been covered in the Discussion section.) Vapor-
phase inhibitors are also prohibited.
Carbon steel components may rust during shipment and storage. R D T
F7-2T specifies that small carbon steel components (and other materials
that do not resist atmospheric corrosion) be packaged in moisture-vapor
proof envelopes that have been purged with a dry, inert gas such as nitro-
gen, argon, or helium free of dirt,'dust, oil, or halogens, and with a dewpoint
of - 4 0 F or lower. A bagged desiccant that is nondeliquescent, nondusting,
nonhalogenated, and chemically inert is to be inserted in the envelope. For
large steel components, all openings are to be plugged and sealed with
approved materials, the vessel is to be purged with dry, inert gas (see above)
in a prescribed manner, a desiccant is to be inserted in the vessel, and a
humidity card is to be included near a clear plastic (nonhalogenated) cap so
that it can be monitored without admitting air to the unit. The amount of
desiccant per total internal volume is specified.
R D T E4-1T specifies that a desiceant and humidity chart shall be placed
on the steam side of all steam generators.
MIL-B-3180B describes the packaging of the waterside of boilers which
is pertinent to surfaces that do not come in contact with the nuclear-reactor
coolant. The waterside is drained and dried and is entirely coated with a
preservative that, when applied to steel, must be capable of passing a 30-day
humidity test and a 6-month outdoor storage in a louvered shed as de-
scribed in ASTM 1748 [24]. All openings are then sealed with tape, plastic
caps, or barrier material depending on their size. To put the unit back in
operation, it is cleaned with a detergent conforming to MIL-D-16791
Type I [25] and then rinsed.
Prior to putting carbon steel boilers into operation, they are frequently
acid cleaned with acetic acid or inhibited HC1 [26] and are then boiled out
with a passivator solution to build up a protective film. MIL-P-24138

(SHIPS) [27] specifies the composition of the dry passivator compound for
Navy boilers to be 50 weight percent sodium nitrite (NaNO2), 25 weight
percent sodium dihydrogen phosphate (NaH~PO4), and 25 weight percent
disodium monohydrogen phosphate (Na2HPO4).
On-Site Erection Requirements

There appear to be no standards that relate directly to the on-site erection
of reactors. However, conscientious contractors would abide by the
cleanliness and contamination requirements described in the preceding
sections. Of particular importance on-site are the exclusion from sensitized
stainless steel of marine air and weld fumes and spatter from coated elec-
trodes because the attendant chlorides and fluorides can cause inter-
granular attack or stress-corrosion cracking.
Of particular importance in reactor-plant erection is the type of insulation
used on heated stainless steel pipes and vessels. Mineral-base thermal
insulation materials, cements, and adhesives can contain leachable chlorides.
If water drips onto the insulation, these chlorides can be carried to the hot
surface below and cause stress-corrosion cracking of the stainless steel
component. Thus, low chloride insulation materials are specified. In the
event that the water dripping onto the insulation may itself contain chlor-
ides, some mineral insulation materials are treated with sodium silicate that
affords some inhibition to stress corrosion.
RDT MI2-1T Specifies a corrosion test that a mineral insulation must
pass if it is to be used on austentic stainless steel. The test consists of
attaching four sensitized Type 304 U-bend specimens snugly onto an
Inconel pipe so that the compression side of the U-bend contacts the
surface of the pipe. (See Fig. 1.) The insulation is then fit snugly onto the
tension side of the U-bends and is placed in a stainless steel container so
that the portion of the insulation but not the U-bends will be immersed
when the container is filled with distilled water. The interior of the pipe is
heated so that the water wicking up through the insulation evaporates on
the hot specimens. The test period is 28 days of continuous exposure at a
prescribed rate of evaporation of water. Upon completion of the test, the
specimens are cleaned and examined at a magnification of 30 for evidence
of cracking at the U-bend or below the bolt holes. The insulation is con-
sidered rejectable if two or more of the four specimens contain cracks.
The acceptance criteria for the above test is open to question in view of
the small sample size and the known erratic failure pattern of stress-corro-
sion samples. In the usual tests for stress corrosion, if one of four specimens
exhibits cracking in a finite exposure period, then the conditions would
generally be considered to be conducive to initiating stress-corrosion
cracking.
7111 / ' N i-Cr-Fe ALLOY- ,=,..o.oov=.

/~.0 625 AISl 500 TYPE
III
II / / / / II EXPANDED METAL 1
tt I I I /
INSULATION UNDER I I I
"FAO(WEL EST I ~ THICK ~ i r J
3"O P A N ~ - ~ ~-.~/.-~2]~:'~.-~ I~ APPROXIMATE I1~.._.~_~. ,, I I-I Irtl~ L2~..._. ,' i. 'LL....~
l-~r~
~ , \ / ~ ~ ' ~ i I~ c16 *~
t
~ ! ! ',, I-~-14-iH--I,i
'_o_L. . . . /._~_l '
I ' ~!l
\. ~ ' -
0
L.,5." 9 " "" ~3 OVER FLOW I

,IT i I // OiAM.I II 1 t,~" ii
WATER LEVEL STAINLESS STEEL PAN I
MIN.~ ~
/I.~l ~I-PINT VOLUME
L_z'_
--
....... _j
- ........
NI-Cr Fe ALLOY PIPE,
"~" Z+P.S. ASTM B-167
FIG. 1--Schematic diagram of apparatus for conducting the wet insulation test (as repro-
duced from RDT M12-1T).
Reactor Operation
Water-Cooled Reactors
Coolant C o m p o s i t i o n s - - R D T Standard AI-IT specifies the coolant com-
position in pressurized-water reactors and mentions that materials in
contact with the primary coolant may include austenitic stainless steels,
nickel-chromium-iron alloys (such as Inconel 600), and Zircaloy 2 or
Zircaloy 4. The make-up water quality before chemical additions is to be
pH 6.0 to 8.0 (25 C), specific conductivity 1.0 ~mhos/cm max. (25 C),
chloride 0.15 ppm max., fluoride, 0.10 ppm max., and total suspended
solids 1.0 ppm max. For neutral reactor coolants the water should meet
the above specifications except that the allowable specific conductivity is
5.0 ~zmhos/cm (max), dissolved hydrogen should be 10 cc (STP)/kg H20
min (15 to 60 cc/kg normal), and dissolved oxygen should be 0.10 ppm
max. The hydrogen is added to promote radiolytic recombination with
oxygen. A low oxygen level is desired to minimize the amount of crud
(corrosion products) that are released to the system. Hydrogen also pre-
vents the formation of nitric acid which occurs under nuclear radiation
when oxygen and nitrogen are present in the coolant, pH control may be
used to minimize the formation, transport, and deposition of crud. In the
absence of boric acid, the following concentrations of alkaline agents are
recommended:
Ammonium hydroxide (NH~OH) 10 to 40 ppm NH3, pH 9.9 to 10.3 (25 C)
Lithium hydroxide (LiOH) 0.2 to 2.2 ppm Li, pH 9.5 to 10.5 (25 C)
Potassium hydroxide (KOH) 1 to 12 ppm K, pH 9.5 to 10.5 (25 C)
CORROSION IN THE NUCLEAR POWER INDUSTRY 157'
LiOH and KOH should be avoided in systems containing heated crevices

or undrainable stagnant spaces where the additiveg could concentrate and
cause localized corrosion (such as pitting or stress-corrosion cracking). No
chemical additive should contain more than 30-ppm halide in its most
concentrated commercially available form. Specifications are also given
for reactor coolants that use boric acid for reactivity control either with
or without alkaline (NH4OH, LiOH, or KOH) additions for pH control.
RDT standards are being prepared but have not yet been issued for
coolant compositions of boiling-water reactors (RDT A1-2) and of light
water test, research, and experimental reactors (RDT A1-3).
Testing M e t h o d s - - A S T M G 2 and NACE TM-01-71 specify test methods
for evaluating materials in water-cooled reactor environments. The former
deals with the testing of zirconium alloys in 360 C water and 399 C steam
FIG. 2--Photograph of principal autoclave components. Left to right: main closure nut,
autoclave head and corrosion coupons, autoclave body, and gasket.
158 INDUSTRIAL C O R R O S I O N STANDARDS AND CONTROL
while the latter deals with structural and containment materials (primarily
stainless steels, nickel-base alloys, and steels). N A C E is currently (1972)
balloting on a standard for dynamic corrosion testing in water. ASTM
B 356 [28] contains a brief description of the 399 C-1500 psi steam test for
zirconium and zirconium alloys which is now covered in greater detail in
ASTM G 2.
Examples of the types of set ups used in these tests are shown in Figs. 2
and 3. Many variables must be considered in testing samples in simulated
water-reactor coolants and these are discussed in detail in Ref 9. Many of
these variables have been considered in ASTM G 2 and N A C E TM-01-71.
For example ASTM G 2 requires that all new or reworked autoclave
components be subjected to 399 C-1500 psi steam for at least 1 day to
oxidize the parts and insure that the autoclave is clean. To meet the latter
requirement, water of initial specific resistivity of 500,000 ohm-cm shall be
at least 100,000 ohm-cm after test. Oxidizing the surfaces minimizes
galvanic effects, and under most conditions bare specimens can be sup-
ported on preoxidized hooks with no adverse galvanic effects. The proce-
dures for preparing the zirconium alloy surfaces are carefully outlined to
prevent retention of the HNO3-HF pickling solution on the specimen
surface and subsequent accelerated corrosion when exposed to the high-
temperature water. ASTM G 2 also specifies a noncorrosion-resistant
control coupon for the 399 C-1500 psi steam test because experience has
shown that steam from water containing small quantities (to 6 ppm) of
Pressurizer
f
~ meter
r~fice
Flow
:ontf
volv(
,i JJ I ~llFdter
,el ~ ~
~Centrlfugol pump ]l I Woter feed

i pump
Immersion
heoter S
J
FIG. 3--Schematic diagram of dynamic corrosion loop.
calcium, chloride, or sodium ions has an inhibiting effect on the corrosion

of Zircaloy [9,29]. On the other hand, 0.12 to 1.3 ppm silicon in the water
will cause accelerated attack of Zircaloy in the steam test depending on the
heat treatment of the Zircaloy [30].
In the above tests, the zirconium alloys are evaluated on weight gain and
appearance. G o o d quality material should exhibit a lustrous black film
after test. Within segments of the industry, reference sets of panels are used
to evaluate the appearance of the specimens. N A C E TM-01-71 specifies
that the specimens are to be descaled by procedures described in N A C E
Publication 7D167 [31]. ASTM G 1 [32] also describes one of the electro-
lytic methods that is used extensively to descale corrosion test coupons.
Inspection Methods--Inspection of the primary water side of reactor
components for evidence of corrosion is often required where corrosion has
occurred or is suspected (as in the possibility of stress-corrosion cracking
of sensitized stainless steel). This inspection is most frequently accom-
plished with dye penetrant because of its simplicity and ease of handling.
The requirements for nondestructive testing by this technique are presented
in R D T F3-6T, R D T E15-2T, and NAVSHIPS 250-1500-1. Among other
things, the R D T standards describe the size, type, and distribution of
acceptable and unacceptable defects as well as a maximum allowable
halogen or sulfur level in the cleaner, penetrant, or developer of 0.5 percent.
NAVSHIPS 250-1500-1 allows a maximum of 1 percent halogens or sulfur.
These tolerances appear to be too liberal when it is considered that they
represent concentration levels of 5000 to 10,000 ppm and the amount of
chloride in a chloride stress-corrosion crack in stainless steel is often on the
order of several thousand ppm. Thus, retention of the penetrant in a crack
or crevice might lead to additional cracking because of the retained chloride
ion. Sulfur would not be expected to be a problem with stainless steel, but
might conceivably cause problems with nickel-base alloys.
MIL-STD-271D(SHIPS) and MIL-I-23135C(ASG) also describe pene-
trant inspection methods [33,34]. They are not nuclear oriented, apply to
many alloy systems, and do not limit halogens or sulfur in the cleaner,
penetrant, or developer and in fact, MIL-STD-271D permits the use of
trichloroethylene and perchloroethylene to clean components prior to
inspection.
Sodium- Cooled Reactors
Sodium Composition--RDT M 13-1T specifies that sodium purchased for
use in liquid metal fast breeder reactors shall have a minimum sodium con-
tent ot 99.9 weight percent as determined by total alkalinity. In nonradio-
active environments, the allowable sodium impurities are:
calcium 10 ppm
carbon 30 ppm
potassium 1000 ppm
halogens 50 ppm
The same limits apply for radioactive environments plus the additional
limits listed below.
boron 5 ppm
cadmium 2 ppm
indium 10 ppm
lithium 5 ppm
silver 5 ppm
sulfur 15 ppm
One of the main contributors to corrosion, oxygen, is not specified. Pre-
sumably, the oxygen content is expected to be reduced by cold trapping or
hot trapping during the initial stages of plant shakedown operations.
RDT M13-1T also specifies that equipment used in contact with sodium
shall be free of rust, dirt, oil, moisture or other material that might con-
taminate the sodium. Rust will be reduced by sodium with a resultant
increase in the oxygen content and probably the hydrogen content of the
sodium. Dirt and oil can lead to carbon pickup and an increase in carburiz-
ing potential of the sodium. Moisture can lead to oxygen and hydrogen
pickup or the formation of sodium hydroxide with attendant increases
in the corrosivity of the sodium.
Operating Conditions--RDT E4-7T specifies that the use of steels in
sodium shall be based on the mechanical properties, metallurgical stability,
and sodium compatibility required for the sodium system at design tem-
peratures. In general, carbon steels are limited to <370 C, chromium-
molybdenum steels ( 2 ~ C r - l M o , 5 C r - ~ M o , 9Cr-lMo) are limited to
540 C, and the austenitic stainless steels may be used at higher temperatures.
RDT 4-5T describes the requirements for a forced circulation coldtrap
assembly for the removal of sodium impurities. These are essential to
reactor operation to maintain low levels of impurities particularly oxygen
and carbon, which can lead to corrosion and carburization of components.
RDT Standards C8-5T, C8-6T, C8-7T, and C8-8T deal with meters or
devices for measuring oxygen, hydrogen, carbon, and nonmetallic impurity
activities in sodium. These standards describe the limits that the meter will
detect, but do not specify the minimum acceptable levels for these impurities.
ANL/ST-6 describes analytical techniques for measuring metallic and
nonmetallic impurities in sodium [35]. The range of detection of the non-
metallic impurities by these techniques are:
Impurity Detection Range, ppm Accuracy

Oxygen
Total Consumption 3 to 100 4-30 percent
Extrusion Method <5 • 1 ppm
Vanadium Wire 0.1 to 14 -4-10 percent
Hydrogen 0.5 to 25 -4-10 percent at 0.5 to 5.0 ppm
Carbon 0.5 to 1000 0.5 ppm at 0.5 to 10 ppm
Nitrogen 0.1 to 100 +15 percent at 0.5 to 10 ppm

Chlorine Limit is 0.5 4-10 to 15 percent at 5 ppm
RDT E8-16 and F5-4 have been designated for environmental condition-
ing during refueling operations and for cleaning sodium components, but
the preparation of these standards has not yet been started. They are
expected to contain provisions that will take into account the corrosive
environments (NasO and NaOH) that can be produced when moisture
contacts sodium.
RDT A1-5 on purity specifications for large operating sodium systems
has been prepared and is being revised, but has not yet been issued.
Inspection--Liquid penetrant inspection is also used extensively on
components in sodium service (see RDT E4-5T, E4-6T, E4-7T). The
halide limitation (described under Inspection-Water-Cooled Reactors) also
applies to sodium-cooled reactors particularly on steam surfaces and
external surfaces that are exposed to ambient air during shutdown. The
sulfur limitations are particularly pertinent to sodium service if nickel
alloys are used in the high temperature portion of the system, because
sulfur embrittlement is a potential problem.
Areas Requiring Additional Attention

Although the nuclear power industry is still in its early stages of growth,
major strides have been made in providing corrosion standards to guide its
operation. As to be expected, there are still some gaps that have not yet
been filled. In particular, attention should be givento on-site plant erection
practices. Of major importance is the protection of both primary and
external surfaces of stainless steels from construction debris and dirt,
marine atmosphere, and welding fumes. These can lead to intergranular
attack or stress-corrosion cracking under the conditions described in the
background section.
As more reactors are placed in operation, repair and inspection practices
will require standardization. In particular, are the problems of reducing the
radiation levels in the vicinity of the repair site and the further reduction of
the allowable halides in the liquid penetrants used in inspections.
In sodium technology, the principal problem appears to be a standard
means for assuring that steam cleaning or other moisture contact on sodium
covered component will not result in the retention of NaOH or Na20 that
can lead to localized attack when the component is reintroduced to hot
sodium. Perhaps, the introduction of CO2 during these operations could be
used to form sodium carbonate which is less corrosive.
In water-cooled reactors, the principal corrosion problem areas requiring
attention continue to be transported corrosion products (crud) and
sensitized stainless steels. Significant success with crud has been attained
by pH control in pressurized-water reactors. However, it would appear that
e v e n t u a l l y the water-cooled reactors will have to be desealed to reduce the

erud i n v e n t o r i e s a n d s t a n d a r d s will be needed to avoid c o r r o s i o n problems.
T h e sensitization p r o b l e m is n o w receiving a t t e n t i o n . Y e t to be resolved are
w h a t degrees of sensitization are h a r m f u l a n d what simple tests can be
a p p l i e d to detect these degrees of sensitization.
References
[1] Ward, C. T., Mathis, D. L., and Staehle, R. W., "Research in Progress: Intergranular
Attack of Sensitized Austenitic Stainless Steel by Water Containing Fluoride Ions,"
Corrosion, Vol. 25, No. 9, Sept. 1969, pp. 394-396.
[2] Copson, H. R., "An Appraisal of the Resistance to Stress-Corrosion Cracking of Iron-
Nickd-Chromium Alloys in Pressurized Water and Related Environments," paper
presented at the Polytechnic Institute of Brooklyn Seminar on Corrosion in Nuclear
and Conventional Power Plant Systems, New York, May 1969.
[3] Rideout, S. P., "Stress-Corrosion Cracking of Type 304 Stainless Steel in High Purity,
Heavy Water," Second International Congress on Metallic Corrosion,National Associa-
tion of Corrosion Engineers, 1966, pp. 159-171.
[4] Burghard, H. C., Norris, E. B., and Wylie, R. D., "Examination of Upper Liquid
Level Line from Elk River Reactor," USAEC Report SWRI-1228-P9-16, 27 Feb. 1969.
[5] "LaCrosse Boiling Water Reactor Primary System Sensitized Nozzle Safe End Re-
placement Program," United Nuclear Corporation Preliminary Summary Report,
to be published.
[6] "Reactor Primary Systems Investigation at Nine-Mile Point Nuclear Station," Niagara
Mohawk Report dated 1 May 1970.
[7] "Reactor Primary System Investigation at Nine-Mile Point Nuclear Station Report
No. 2," Niagara Mohawk Report dated 11 May 1970.
[8] Strauss, M. B. and Bloom, M. C., "Cracking of Low Carbon Steel by Ferric Chloride
Solutions," Corrosion, Vol. 16, N6. 11, Nov. 1960, pp. 553t-556t.
[9] Berry, W. E., "Testing Nuclear Materials in Aqueous Environments," Handbook on
Corrosion Testing and Evaluation, W. H. Ailor, Ed., Wiley, New York, 1971. pp.
379-403.
[10] Tyzack, C., "Application of Sodium Chemistry in Fast Reactors," Alkali Metals, an
International Symposium held at Nottingham, 19-22 July 1966, Special Publication
No. 22, The Chemical Society, London, 1967, pp. 236-259,
[11] Berry, W. E. in Corrosion in Nuclear Applications, Wiley, New York, 1971, pp. 242-245.
[12] Natesan, K. and Kassner, T. F., "Calculations for the Thermodynamic Driving Force
for Carbon Transport in Sodium-Steel Systems," USAEC Report ANL-7646, Dec.
1969.
[13] Goldman, K. and Minushkin, B., "Sodium Technology" in Reactor Technology, L. E.
Link, Ed., Argonne National Laboratory, Selected Reviews, 1965, p. 31 ; and USAEC
Report TID-8541.
[14] Brush, E. G., "Behavior of Type 347 Stainless Steel in Sodium Hydroxide at Elevated
Temperatures," USAEC Report KAPL-M-EGB-22, 12 July 1956.
[15] Comprelli, F. A., Hetzler, F. J., and Lauritzen, T. A., "Clad Compatibility with
Oxide Fuel and Sodium," USAEC Report ANL-7120, Proceedings of the Conference
on Safety, Fuels, and Core Design in Large Fast Power Reactors, Argonne National
Laboratory, 11-14 Oct. 1965, pp. 355-372.
[16] Moberly, J. W., Barlow, M., Garrison, M. C., and Planting, P. J., "Interaction of
Liquid Sodium with 304 Stainless Steel," USAEC Report TID-24602, 31 Jan. 1968.
[17] Chamberlain, H. V., Kanamori, A. J., and Lindsey, P. S., "Evaluation of Materials
Wastage Due to Reactions of Water in Sodium," USAEC Report APDA-227, June
1969.
[18] ASTM A 370-71, Methods and Definitions for Mechanical Testing of Steel Products,
Annual Book of A S T M Standards, Parts 1-4, American Society for Testing and Ma-
terials.
[19] Rochal H. J., discussion of paper by E. Braims and G. Pier, Stahl w Eisen, Vol. 75,
1955, pp. 579-586.
[20] Scharfstein, L. R. and Eisenbrown, C. M., "An Evaluation of Accelerated Strauss

Testing," Second International Congress on Metallic Corrosion, National Association
of Corrosion Engineers, 1966, pp. 497-501.
[21] Streicher, M. A., "General and Intergranular Corrosion of Austenitic Stainless Steels
in Acids," Journal of the Electrochemical Society, Vol. 106, 1959, pp. 161-180.
[22] Cihal, V. and Prazak, M., "A Contribution to the Explanation of Intergranular
Corrosion of Chromium-Nickel Steel," Corrosion, Vol. 16, No. 10, Oct. 1960, pp.
530t-532t.
[23] MIL-STD-753A, Method 101, High Humidity Test, 17 July 1963.
[24] ASTM D 1748-70, Test for Rust Protection by Metal Preservatives in the Humidity
Cabinet, Annual Book of A S T M Standards, Part 17, American Society for Testing
and Materials, 1970.
[25] MIL-D-16791E, 11 Aug. 1964, and Amendment 3, 24 May 1968, Military Specifica-
tion, Detergents, General Purpose (Liquid, Nonionic).
[26] Bell, W. E. and Rice, J. K., "Corrosion from Repeated Acid Cleaning, Citric vs.
Hydrochloric," Combustion, Vol. 38, No. 3, Sept. 1966, pp. 35-42.
[27] MIL-P-24138 (SHIPS), 11 March 1965, "Passivator Compound, Navy Boiler."
[28] ASTM B 356-67, Specification for Zirconium and Zirconium Alloy Forgings and
Extrusions for Nuclear Application, Annual Book of A S T M Standards, Part 7, Ameri-
can Society for Testing and Materials, 1967.
[29] Goodwin, J. G., Gray, R. D., and Collins, L. F., "Sensitized Coupons in Steam
Acceptance Tests," Handbook on Corrosion Testing and Evaluation, W. H. Ailor, Ed.,
Wiley, New York, 1971, pp. 435-452.
[30] Callahan, E. J. and Kabat, J. F., "The Effects of Silica in Autoclave Test Water on the
Steam Corrosion of Zircaloy," USAEC Report KAPL-M-6748, 30 Jan. 1968.
[31] NACE Publication 7D167, "Procedures for Quantitative Removal of Oxide Scales
Formed in High Temperature Water and Steam," Materials Protection, Vol. 6, No. 7,
July 1967, pp. 69-72.
[32] ASTM G 1-67, Recommended Practice for Preparing, Cleaning, and Evaluating
Corrosion-Test Specimens, Annual Book of A S T M Standards, Part 31, American
Society for Testing and Materials, 1967.
[33] MIL-25135C (ASG) 21, Oct. 1959, and Amendment 3, 1 June 1964, Military Specifi-
cation, Inspection Materials, Penetrant.
[34] MIL-STD-271D (SHIPS), 11 March 1965, Military Standard, Nondestructive Testing
Requirements for Metals.
[35] Burris, L., Cafasso, F. A., Meyer, R. J., Barsky, M H , and Edwards, H S., "Interim
Methods for the Analysis of Sodium and Cover Gas," USAEC Report ANL/ST-6,
Jan. 1971.
STP534-EB/Nov. 1973
Chapter 9

Chemical Industry
L. W . G l e e k m a n I
The chemical engineer in the chemical industry, responsible for the

selection of materials of construction, currently must be more than a
corrosion engineer; he must be in the broadest sense a materials engineer.
In view of the fact that materials other than metals are widely used in the
chemical industry, this engineer must know more than the metallurgy and
electrochemistry of the destruction of metals (commonly called corrosion).
He must know the technical and scientific information with regard to the
deterioration of plastics, graphite, wood, rubber, concrete, roofing ma-
terials, coatings, glass, brick, mortars, among many materials.
By chemical industry is not meant that segment of business uniquely
concerned with the production and manufacture of chemicals. It is but an
extension of chemical engineering when one considers that the fertilizer
industry, the pulp and paper industry, the petroleum refinery, the mining
and metal extraction, the petrochemical industry--all these are variations
of the chemical industry. It is interesting to speculate what these industries
have in common and thus to establish why corrosion control in these
industries is basically similar. Such similarities include the handling of
liquids, solids, and gases, separately and in combination. There is also the
handling a n d / o r production of acids, bases, and salts, both inorganic and
organic. There are power requirements for the transmission and movement
of process fluids as well as for the generation of heat to increase the rate of
many of the chemical reactions. There is a need, in most cases in the
industries mentioned, for purity of product which is one aspect of no, or
low, corrosion. As with all industry, there is a need for raw materials and,
of course, an outlet for the finished materials as well as control of waste or
by-products. The present day economy demands maximum life of proces-
sing equipment at minimum cost, with minimum labor requirements both
in the operation of the plant to produce the chemicals and in the mainte-
nance of the plant.
1 Consultant, Materials and Corrosion Engineering Services, Southfield, Mich.
164
Copyright9 1973 by ASTM International www.astm.org
CORROSION IN THE CHEMICAL INDUSTRY 165
These various factors have had a major influence on the materials of con-
struction used in these industries and, indeed, on the philosophy of the
selection of these materials of construction. In the early days of the chemical
industry, before the development of joining techniques such as welding and
brazing and before the development of what is now commonly called
corrosion-resistant alloys, the materials then used were those which were
readily available either naturally, or within the production techniques of
the then existing industries. For example, wood was (and still is to a lesser
degree) a widely used material of construction for vessels, tanks, and pipes.
Natural rubber in sheet form was also an extensively utilized material.
Of the metals, cast iron was perhaps the most widely used early metallic sub-
stance because of its low cost, relative ease of formation (low melting point
and tolerance for impurities), and fair strength in compression. Copper and
brass and bronze were used where their special properties, such as heat
transfer or rotating load-carrying capacity, as in bearings, could be justified.
All these early materials had their limitations, and it is hard to say whether
the limitations of the materials brought on the development of new com-
peting materials, or whether the competing materials would have developed
in and of themselves, irrespective of the limitations of the early materials.
The intricacies and size of equipment in the early chemical industry, the
absence of refinements in process controls, inexpensive labor, and readily
available land for plant sites, were among factors which led to the extensive
use of cast iron. It was the common situation, at least in the chlor-alkali
phase of the chemical industry, for virtually every major manufacturing
company or plant to have its own foundry. In the early days of the industry,
not only were there no standards in corrosion testing, but standards were
lacking for dimensional values on pumps, valves, flanges, etc. For such
reasons, each company developed their own internal standards in manu-
facturing their process equipment; this was in preference to farming out the
work to jobbing foundries. It was not uncommon, within so short a period
as the early 1960s, to find certain chemical companies producing cast iron
caustic fusion pots in which 73 percent caustic soda was concentrated to
close to 100 percent. These pots were approximately 10 ft in diameter and
weighed 10 tons with an approximate 1-in. wall thickness. It should also be
pointed out that in the early days of the chemical industry, as was virtually
true in every industry, each company had what it felt were certain manu-
facturing and processing secrets which they did not want to divulge to their
competitors. These included design details of the equipment. For this
reason, it was then justifiable in the eyes of the engineers involved to
manufacture this equipment within the confines of the company rather than
to have it manufactured by a concern whose shop was open to chemical
competitors. In a soda ash plant, it is common today to see the carbonating
towers and the still units, both of which are large in diameter and quite tall
(up to 9 to 12 ft in diameter and 100 ft tall) made from cast iron sections.
This is true not only for the older plants but even for plants built in 1950.
Part of this may be due to the desire for interchangeability of new equip-
ment with old, but the fact remains that in this segment of the chemical
process industry, cast iron is still a strong contender in spite of the proven
corrosion resistance, lesser wall thickness, and improved ease of fabrication
of such materials as aluminum, stainless steel, and titanium.
It is interesting to note that the older industries fabricating the materials
previously mentioned, (cast iron, wood, rubber, and glass) not only have
not disappeared from the scene, but are trying to recoup the losses which
they had made to the newer competing materials. A well qualified materials
engineer today will not merely specify "cast iron": he will specify "ma-
terials to be ASTM A 48-64 Class 20." Here he indicates the strength of
the cast iron to be used, or he may indicate the alloying elements to be
added to the cast iron to give it a particular corrosion resistance in a given
environment. This is but one phase of the revitalization of cast iron as a
material of construction in the chemical process industry.
Fabricators of wood equipment have formed an organization known as
the Wood Institute where presumably they pool certain experiences and
developments to further the use of their materials in other using industries.
Certain large size applications such as storage tanks still continue to be
made in wood in certain industries. Because of the improvements that have
occurred in wood technology, such as impregnation of wood with chemi-
cally resistant resins (phenolic and furane in particular), the impregnated
material offers many corrosion advantages that cannot be achieved with
unimpregnated wood. While natural rubber, in many cases, has given way
to synthetic elastomers which have better life and improved oxidization and
temperature resistance, nonetheless, natural rubber is still considered a
standard elastomeric material of construction usually as a lining rather
than a self-supporting material. Glass, of course, has been the favored
material for use by chemists since perhaps the early work of the alchemists
in the Middle Ages; while it has size limitations and also a fragility factor,
it is used quite extensively for certain shapes, such as cylinders, in the form
of pipe, etc. As will be seen subsequently, techniques have been developed
to overcome the fragility aspects of glass and also the limitations on joining.
Because of weight considerations, formability, and the resulting ability to
fabricate a vessel that will have maximum through-put with minimum metal
involved, wrought metals are more widely used in the chemical industry
than cast metals or other materials. Corrosion Data Survey, published by
the National Association of Corrosion Engineers, has information on the
following classes of metals; iron-base, copper-base, nickel-base, aluminum-
base, precious metals, reactive metals, and others. In each one of these
categories, there is considerable variation in corrosion resistance and also
in price. The question that the corrosion and materials engineer faces is--
what material to select for a given piece of equipment in a given process and
how to know that the material he has selected will have the combination of
properties including costs and long life desired.
Corrosion Testing
One comes then, to the matter of corrosion standards and control for
process (immersion) conditions as opposed to what will be discussed
subsequently, corrosion standards and control for atmospheric conditions.
The early work on the procedures to systematically select materials of con-
struction for process conditions was that done by the Research and De-
velopment Department of the International Nickel Company sometime
prior to 1930. This led ultimately to the development of the INCO Corro-
sion Test Spool which, in spite of the several standards which have been
written and adopted by both ASTM and NACE, in particular, is still
popularly known by that name rather than by association standard num-
ber? The reader who wishes more detailed information on this subject
should go to the Edgar Marburg lecture of 1951 by Francis LaQue on
"Corrosion Testing" [1]. 3 He will find there a discussion not only for process
conditions but virtually every other aspect of corrosion testing. The popu-
larity of the INCO Test Spool is seen by the fact that since the early 30s,
approximately 6000 of these assemblies have been distributed by Inter-
national Nickel Company to those interested in determining the corrosion
rate of metals and alloys. The INCO Test Spool has served to revolutionize
corrosion testing under process conditions, in more than several ways.
1. It took from the hands of the chemists who, in establishing process
reactions in the early days, determined materials of construction other than
laboratory glassware. The habit of the chemist from his training was that of
expressing everything as a percent. Thus, early corrosion data may have
indicated that Metal A had a 1 percent weight loss while Metal B had a
3-percent weight loss under presumably identical process conditions;
therefore, Metal A was superior to Metal B.
2. The INCO Test Spool standardized the size and shape of test speci-
mens; this facilitated calculating the area of the samples which was a
necessary step in the determination of weight loss per unit area per unit time.
3. The INCO Test Spool standardized on the number of identical speci-
mens to be exposed to achieve a consistent statistical value. Samples of a
given metal in a given heat-treated condition were used. Usually only two
identical samples were exposed to increase the total number of specimens
of different materials exposed to the process environment.
4. The INCO Test Spool with its determination of weight loss per unit
area per unit time as milligrams per square decimeter per day (todd) led
2 Recommended Practice for Conducting Plant Corrosion Tests, ASTM G 4-68; Labora-
tory Corrosion Testing of Metals for the Process Industries, NACE TM-01-69.
3 Italic numbers in brackets refer to references listed at end of this chapter.
directly into the concept of corrosion allowance, as so many inches per year
of attack. This was achieved by the simple procedure of dividing todd by
density in consistent units and converting to the desired units of corrosion
allowance.
5. The INCO Corrosion Test Spool tested other than the general corro-
sion rate based only on weight loss. The crevice between the spacers used to
separate samples served for many alloys to indicate their susceptibility to
crevice corrosion; the stamping used to identify the specimens by code
served to indicate the susceptibility of certain metals and alloys to stress
corrosion cracking. The size of the sample was ideal for heat treatment so
that sensitization effects could be imparted to the sample to simulate
various conditions of welding. If desired, welded specimens could be
formed into the shape used on INCO Test Spools. Also, where desired,
specimens could be couPled together to achieve galvanic effects that might
inadvertently be encountered in service. About the only thing that the
INCO Test Spool could not do was to subject the samples to differential
velocity conditions on one test spool or to different conditions of heat
transfer on the same test spool. However, two or more spools with the same
samples could be subjected to different velocity or heat transfer conditions,
and so partially bring forth the desired answer. In addition, the shape of the
INCO Test Spool did not allow for testing under combined liquid and
vapor conditions with measured weight loss for each condition. Floating
supports for test spools were devised which suspended a test spool accu-
rately at the interface so that half of the specimen was in the liquid and half
in vapor. Here, however, one had to rely on measurement of thickness
changes (before and after exposure) or the depth of pitting in each phase to
come up with meaningful results. On the other hand, it was again fairly
simple to use two INCO Test Spools, one in the liquid phase, and one in the
vapor phase, and possibly even a third under condensing conditions from
the vapor phase to get weight loss figures which could be quantitative for
corrosion allowance, along with the measurement of pitting tendencies.
ASTM has through the years worked out standards and practices for
corrosion testing. This currently manifests itself in Recommended Practice
For Conducting Plant Corrosion Tests which has the designation G 4-68.
This replaces Standard A 224-46, in itself an outgrowth of the INCO Test
Spool. Among other things that Recommended Practice G 4 does is to
indicate the data that can be derived from a test spool, most of which falls
in with the above. One additional important consideration is that of
metasomatic corrosion and its determination. This is the condition in which
one constituent is selectively removed from an alloy in what used to be
called by the most c o m m o n phenomenon, dezincification. Metasomatic
corrosion is generally not determined by weight loss but by bending the
sample after the exposure period and checking the microstructure of the
surface to determine the selective corrosion of one or more constituents.
The INCO Test Spool does not always appear in the form normally
supplied by INCO and outlined in Recommended Practice G 4. Modifica-
tions of the test spool have been developed for use in pipes and inside tubes.
Figure 1 shows a modified INCO Test Spool arranged for insertion and
removal into a process vessel while the vessel is operating. The samples are
mounted on a corrosion resistant rod which allows the samples to be with-
drawn from the vessel through a gate valve into a chamber which withstands
the process temperature and pressure conditions of the vessel. Thereafter,
the gate valve is closed and the samples then can be removed without the
necessity of a process shut-down of the unit. Oftentimes, samples are
suspended on plastic cord such as Saran or Teflon woven braid. One
fundamental aspect of this immersion or " dunk'" corrosion testing, as it
is often called by electrochemists, is that the samples must be retrievable at
the end of the test. There is nothing more discouraging or embarrassing to a
corrosion engineer than to find that the samples thought to be on the end of
the test cord have disappeared because the cord holding the samples broke.
6. Part of the problem of standard specimen for weight loss corrosion
data is the removal of any corrosion products on samples in a standard
manner. Such corrosion products must be removed since generally they
correspond to an increase in weight and thus, with corrosion rate being
determined by weight loss per unit area, could be misleading. Again,
ASTM has established Recommended Practice for Preparing, Cleaning
and Evaluating Corrosion Test Specimens, G 1-67.
RETRACTABLE CORROSION SPOOL
V
E
S
S
E
L
t[ ~ATE VALVE
VALVE BONNET
ROD
FIG. 1--Schematic drawing o f modified 1NCO test spool arranged for use in an operating
process vessel.
Stress Corrosion Testing

This is a complicated topic as seen by the fact that many symposia have
been organized around this theme to bring together the procedures and
experiences that those in the field have been following. While many fine
test methods have been proposed, as of this writing there is no one standard
adopted by any group for stress corrosion testing. The reason for this is due
to the variety of stress modes to which metal can be subjected while under-
going corrosive conditions. The old adage, "there is no test like a service
test," would certainly be applicable here if one could afford the time, the
equipment and duplication of process conditions. Since this chapter is not
uniquely devoted to stress corrosion testing, there will be no attempt made
to synthesize in detail the developments which appear in the literature.
Instead the reader is referred to Effect of Stress on Corrosion by Dr. J. F.
Bates of the Applied Research Laboratory of U.S. Steel Corporation which,
in addition to being a fine discussion, also contains an excellent bibliog-
raphy [2]. Dr. Bates points out two widely used tests in the U.S. Steel
organization for stress corrosion; the cup-and-circular-weld specimen and
the tuning fork. The first uses the Ericksen cup tester (nominally used to
measure the ductility and drawing characteristics of sheet materials) plus a
weld using the same material as the sheet; the combination imposes me-
chanical and thermal stresses and metallurgical structural changes. In
addition, stamped identification, punched and drilled holes, and sheared
edges impose conditions similar to actual fabrication variables. The
tuning-fork samples can have as many variables and, in addition, allow
predetermined values of stress to be imposed.
Not all tests can be, or should be, service tests or simulated service tests.
To bring about standard methodology in laboratory testing, NACE,
through its Technical Practices Committee in 1969, issued Standard Test
Method TM-01-69 on "Laboratory Corrosion Testing of Metals for the
Process Industries." In this test method, the factors which influence
laboratory tests were discussed. These factors include specimen preparation,
apparatus, test conditions (solution composition, temperature, velocity,
aeration, volume, method of supporting specimens, and duration of the
tests), methods of cleaning specimens, evaluation of results and calculation
of corrosion rates. Since this is a laboratory method, the techniques of
handling the specimens are slightly different than those involved in plant
corrosion test G 4. In this method, a so-called resin flask is used as the
container with a reflux condenser to prevent the loss of corrodent by
evaporation. Four specimens are exposed at one time in this test compared
to the plant corrosion testing. This standard also includes a reprint of the
planned interval test of Wachter and Treseder [3] with regard to determin-
ing the effect of exposure period on the corrosion of metals and also the
effect of the corrosivity of the environment. This procedure does not
CORROSION IN 1"HE CHEMICAL iNDUSTRY 171
require the removal of solid corrosion products between exposure periods.

NACE has a unit committee, T-SE, working on stress corrosion cracking
(SCC) of all metallic materials from austenitic stainless steel through, and
including, aluminum and low alloy steels. So far as is known at the present
time, there have been no standards on corrosion testing coming out of this
committee, though liasion is maintained with ASTM, the Metals Prop-
erties Council, and others. This group is particularly interested in environ-
mental effects in stress corrosion cracking. A similar group which is very
active in this regard is ASTM Committee G 01.06. Section I of this Com-
mittee is investigating specimens such as C-rings, U-bends, bent-beams, and
use of direct tension. Section 2 of this ASTM committee is concerned with
testing environments and may be expected to produce a recommended
practice shortly for stress corrosion tests for titanium alloys, for aluminum
alloy 7039, and for certain copper base alloys.
Another fine reference for stress corrosion testing is the chapter "Corro-
sion Testing" by Dr. M. G. Fontana in the NACE, Proceedings of the Short
Course on Process Industry Corrosion, Sept. 1960 [4]. This 400 plus page
loose-bound book still is an excellent compendium on many aspects of
chemical industry corrosion problems and solutions. It is regrettable that it
never appeared in hard-cover form. However, Dr. Fontana's joint book
with Dr. N. D. Greene, Corrosion Engineering, contains much of his work
from his 1960 lecture plus much more, drawing, as it does, on his many
productive years with the duPont Company and his years of teaching,
research, and consulting at Ohio State University [5].
Accelerated Corrosion Testing

Experiences in the chemical industry have shown that accelerated corro-
sion tests are often misleading in comparison with the procedures outlined
for the plant corrosion test and, as indicated above, for the laboratory
corrosion test. As a result, relatively little emphasis in the chemical process
industry, and its related adjuncts, is placed on accelerated tests. However,
as a reference for those who may be interested in standards which may exist,
four ASTM standards are listed below which give reliable results in acceler-
ated corrosion testing for certain materials in certain simulated environ-
ments. These include:
B 117-64, Salt Spray (Fog) Testing
B 287-68, Acetic Acid-Salt Spray (Fog) Testing
B 368,-68, Copper-Accelerated Acetic Acid Salt Spray (Fog) Testing
(CASS Test)
B 380-65, Corrosion Testing of Decorative Chromium Plating by the
Corrodkote Procedure
In this connection it is considered wise to quote, hopefully not out of
context, the words of Francis LaQue, given at a talk many years ago to a
17"~ INDUSTRIAL CORROSION STANDARDS AND CONTROL
group of chemical industry and automotive industry engineers, when

asked what he thought of the salt spray test, Mr. LaQue stated, "Materials
that do well in the salt spray test should be used to build salt spray test
cabinets!"
Galvanic Corrosion Testing

This was one of the earliest topics of corrosion to be investigated for the
very simple reason that the corrosion behavior of two metals coupled
together serves to explain in a theoretical way the behavior of parts of the
same metal corroding at different rates. The concept of anodes and cathodes
is most easily drawn by using one metal whose behavior in the electromotive
force series or the galvanic series is known to be more active than a second
metal to which it was coupled.
Again, it was the early work of LaQue of International Nickel Company
who pointed out the importance of the relative size of the anode and
cathode being coupled. It was not sufficient merely to put two pieces o f
metal in contact, irrespective of the size of the materials. One had to
recognize, as was developed by the early work of LaQue, that a small
anode coupled to a large cathode would corrode at an extremely high rate
compared to a large anode coupled to a small cathode. All testing for
galvanic effects in corrosion has to take into account the relative anode-
cathode size. In addition, the corrosion products from the anode must be
considered in terms of the possibility they have in increasing the electrical
resistance between the two specimens coupled together. It is for this reason
in particular that the INCO Test Spool lends itself to the testing of galvanic
corrosion behavior where the two specimens are placed in contact rather
than being separated by plastic insulators. Where design conditions are
such that galvanic coupling of dissimilar metals cannot be avoided, it is
best in testing to simulate these design conditions to establish the relative
corrosivity of the couple, taking into account size effects. LaQue discusses
this with sketches of equipment in his Marburg Lecture [1].
Velocity Conditions on Corrosion

Early quantitative corrosion experimenters established what have
become the standards for determining the effect of velocity on corrosion
rate. Among these were Dr. Mars G. F o n t a n a at Ohio State University
whose writings on this go back to his early regular monthly publications on
corrosion in Industrial and Engineering Chemistry, back in the late '40s and
continuing into the '50s. Whitney and Fisher at Monsanto in St. Louis were
also prominent in this work. In essence, velocity corrosion testing consists
on having a pump drawing the corrosive medium from a reservoir and
moving it through pipes which often serve as the corrosion specimen in a
closed loop back to the reservoir. To simultaneously test the effect of the
temperature, there is usually a source of heat input to the circuit. Often-
times the test specimen is the equivalent of a one-tube heat exchanger or

possibly a multi-tube heat exchanger where the inside diameter of the
tubing governs the velocity of the process fluid in the test unit. Most of the
data which is now available for the limiting velocity on corrosion in surface
condensers, heat exchangers and other tubular heat transfer devices come
from procedures of the sort mentioned. The Navy Department through
the work of W. Lee Williams at the Naval Experimental Station in Annapolis
deserves much credit for organizing the work and procedure along these
lines.
Committee T-5A of the National Association of Corrosion Engineers,
dealing with corrosion in the chemical industry, has for years attempted to
standardize a procedure on velocity corrosion testing, first through a series
of round-robin tests of process fluids in different apparatus, and subse-
quently through the development of an actual testing procedure. This is now
manifesting itself in a test method which is in the final stages of publication
by the corrosion engineers. In the meantime, NACE has developed Stand-
ard Test Method TM-02-70, Method of Conducting Controlled Velocity
Laboratory Corrosion Tests. In this standard, a corrosive solution is
moved at a known tangential velocity across the face of one or more
corrosion test specimens. Velocities of 8 ft/s or 15 ft/s are obtainable
depending on the agitator drive used. This work derives in major part from
the original work of R. J. Landrum of duPont and may be found in U.S.
Patent 3,228,236 (1969). Landrum's paper, "Evaluation of Structural
Materials for Corrosion Resistance," not only contains an outstanding
discussion of corrosion testing in general but velocity corrosion testing in
particular [6].
Figure 2 shows a laboratory velocity tester used to simulate expected con-
ditions in a chemical plant. As indicated earlier, this consists of a reservoir
from which the process fluid is drawn by a pump, sent through a heat ex-
changer and then through a series of pipes whose corrosion resistance is to
be evaluated. The writer made use of this technique in a pilot plant for a
chemical company about to build a large size unit. The circulating tester
simulated the velocity conditions expected to be encountered in the main
plant, testing a series of reinforced plastic pipes as well as several different
metallic pipes. Evaluation was carried out by sectioning the pipes longi-
tudinally at the end of the test and determining the loss in thickness for the
plastic pipes and metallurgical changes such as the metasomatic corrosion
of the metallic specimens. For a test of this sort to be meaningtul, the ma-
terial that has the best corrosion resistance must also have good fabricating
capabilities. In the case of plastic pipes, this means the ability to attach
flanges and to have fittmgs and valves available in the same material with
proven corrosion resistance of all components. Thus, if the pipe is filament
wound and the flange press-molded, the flange and its adhesive must with-
stand the corrodents equally well with the pipe proper.
FIG. 2--Photographs o f corrosion test facility used to simulate expected conditions in

a chemical plant; (a) bench scale unit (b) 40-gal unit.
Corrosion under Heat Transfer Conditions

It was mentioned in the previous section that, in testing for velocity
effects, a source of heat is introduced into the process test fluid. The
method of introducing the heat and its effect on the surface through which
heat was transferred gave rise to the early w o r k of corrosion under heat
transfer conditions. One looks to the early w o r k of the Atomic Energy
personnel at the H a n f o r d works of General Electric, citing such names as
Scharfstein, Groves, Eisenbrown, and others, in addition to F o n t a n a ' s
publications in " I and E C " on this subject and the work by Fisher and
Whitney in N A C E publications.
In the process industry, work with heat exchangers established at an early
date that the surface through which heat was being transferred to bring the
temperature of the solution to a given value corroded at a more rapid rate
than the surface of the same metal merely exposed in the process fluid.
Since heat exchangers and similar heat transfer devices represented not only
large investment on the part of manufacturers of chemicals, but also
critical items where the failure of such an item would shut down the
process, it became necessary to know this influence.
The old rule of physical chemistry about doubling the reaction rate for
every 10 deg C increase in temperature still holds true; this applies to the
difference in corrosion between a sample at one temperature and a sample
at another temperature. This is not, however, an accurate rule for predicting
the corrosion rate under heat transfer conditions. The reason for this lies
with the nature of heat transfer to a process fluid; in this transfer process the
surface temperature of the metal through which heat is being transferred is
higher than the solution temperature due to the presence of surface films on
the process fluid side which serve to act as barriers or resistances to that
flow. The rate of heat input to the metal surface, and the rate of heat
removal f r o m the metal surface, are i m p o r t a n t variables, governing to a
large extent the temperature of the surface where the heat is being trans-
ferred.
Currently, National Association of Corrosion Engineers, through C o m -
mittee T-5A have established a test procedure 4 which allows these variables
to be evaluated. The test procedures used by the early experimenters at
A E C and Ohio State University, in particular, are the basis for this corrosion
testing standard. The usual technique is to use a high wattage soldering iron
whose heating surface is a fiat disk coupled to an aluminum block, used for
good heat transfer and for the measurement of the rate of heat transfer,
and ultimately coupled to a corrosion specimen which serves as the cover
to an opening in the test vessel. The use of thin-gage thermocouples serves
to measure both the rate of heat transfer and to approximate the tempera-
4 Laboratory Corrosion Testing of Metals Under Heat Transfer Conditions, (Number not
yet assigned by NACE Technical Practices Committee T-5A-6b).
ture of the metal surface to which the heat is being transferred. In most test
methods, a duplicate specimen is used at the interface between the liquid
and vapor phases to transfer heat, with a third duplicate specimen attached
to a cold finger to remove heat. Proper gasketing is necessary and agitation
of the process fluid is also required. Corrosion rate can be determined by
weight loss for the area of the specimen exposed to the heating fluid. In
addition, pit depth can be quantitatively measured, as can any other
variables relative to corrosion, such as intergranular corrosion, sensitivation
effects, crevice corrosion, etc.
Welding Effects on Corrosion in Process Fluids

It had been mentioned earlier that welded specimens can be used as part
of the INCO Test Spool, and indeed this has been done in many cases.
However, because the specimen is made up partially of unwelded and
partially of welded material, it has been difficult to establish an absolute
quantitative corrosion rate for the welded portion only. Furthermore, it
should be recognized that there are variations metallurgically across a weld.
This becomes evident when one considers that the weld bead represents
molten metal at the time of welding, and the temperatures in welding
therefore go all the way from the melting point of the material being
welded (or the weld rod) to the base metal away from the bead which stays
at ambient temperature. The distance over which the welding operation has
a metallurgical influence is a function of the material being welded and the
welding rod, the heat input and such factors as thermal conductivity, mass
of material, etc.
Work with 18-8 austenitic stainless steels relaUve to their then unpre-
dictable corrosion behavior when welded, brought about the early Strauss
test [7] (copper sulfate as a corrodent) and the Huey test [8] (boiling nitric
acid) to determine the susceptibility to intergranular corrosion. Currently,
ASTM A 262-70, Recommended Practices for Detecting Susceptibility to
Intergranular Attack in Stainless Steels, covers not only the boiling nitric
acid test, but also the electrolytic oxalic test. The extensive use of stainless
steel in the largest chemical company in the country, namely, E. f. duPont
de Nemours, led to much work on the part of researchers at duPont on
tests that would be shorter and more definitive than the Huey and Strauss
tests. This has culminated particularly in the work by Michael Streicher and
the oxalic etching test [9]. This test allows a purchaser to determine the
relative corrosion resistance of welds of the material in various media
before fabricating the material and putting it in service.
The less refined method to determine the behavior of stainless steel in its
welded conditions in a quantitative fashion, is to sensitize the sample
before exposing it to the corrosive environment on an INCO Test Spool.
For an austenitic stainless steel, this consists of heating the sample in the
800 to 1200 F range where carbides precipitate and where, therefore, the
matrix is deprived of the chromium and nickel necessary for corrosion

resistance. The entire specimen behaves as if it were the critical part of
weld, namely, the sensitized zone of carbide precipitation in the weld.
In titanium and zirconium in certain environments, a galvanic effect
between the large grains of the weld bead and the small grain size of the
base metal has been observed. The large grain size is preferentially cor-
roded in these environments. Since this is a galvanic effect, it would have to
be tested in a fashion described above for galvanic corrosion.
Electrochemical Techniques
While the resistance probe technique is truly not an electrochemical
method for determining corrosion, it serves essentially as an intermediate
step between weight loss specimens and true electrochemical techniques
and, as such, will be briefly described. The resistance probe works on the
basis that the electrical resistance of a thin metal specimen varies directly
with the thickness of that section: consequently, as the thin cross-sectional
area is corroded, the resistance of the probe changes. This resistance can be
read indirectly without weighing the specimen via the standard Wheatstone
Bridge circuit and correlated with loss in thickness per unit time. This
means, therefore, that resistance probes can be read while they are com-
pletely immersed in a process stream without the necessity for removal from
the stream. Furthermore, the change in resistance can be fed to a recording
device and the behavior of the process stream, relative to one or more
resistance probes, can be recorded for a continuing view of process vari-
ables. Such resistance probes, therefore, work to determine variations in
process streams, such as absence of inhibitor, presence of oxygen, upset
conditions. There has been great difficulty on the part of researchers in the
process industries, however, to correlate the corrosion rate as indicated by
the resistance probe in its fully immersed conditions in the process stream
with the rate determined by the behavior of the walls of a vessel or a pipe.
There is generally good correlation between a resistance probe and a
sample on an INCO Test Spool since both are immersed in the bulk solution
and do not represent conditions of heat transfer through the specimens. F o r
extremely sensitive conditions, such as the presence of water in what is
required to be an anhydrous solution (for example, dry hydrogen chloride),
a resistance probe of the right metal serves as an excellent warning device
of leakage of moisture through a pump seal or a failure of a heat exchanger
tube, etc. Probes can be made by individuals, or they are procurable from
at least two major sources in the United States.
The true electrochemical techniques are those involving the potentiostat
and the galvanostat, both of which essentially represent determinations of
current versus potential or potential versus current of a given metal in a
given environment. For several years both of these techniques were labora-
tory oriented and were quite useful in determining fundamentals of corro-
sion behavior. In fact, the work of Stern and Wissenberg which led to the
development of the 0.2-percent palladium alloy of titanium was done along
these lines [10]. Much has been written in the literature on the exacting
techniques involved; in view of the fact that there are numerous models in
the market, many of which differ primarily in their electronics, a detailed
discussion will not be presented.
Fontana and Greene have a chapter dealing with " M o d e r n Theory-
Applications" which discusses electrochemical methods used to determine
corrosion rate [5]. This should serve as an excellent reference for those
who are interested in this approach. They summarize the advantages o f
electrochemical corrosion rate measurements, particularly linear-polariza-
tion techniques, as
1. They permit rapid corrosion rate measurements and can be used to monitor
corrosion rate in various process streams.
2. These techniques can be used to accurately measure very low corrosion rate
(less than 0.1 mpy) which are both difficult and tedious to perform with
conventional weight loss or chemical analytical techniques. The measure-
ment of low corrosion rates is particularly important in nuclear, pharma-
ceutical, and food processing industries, where trace impurities are problems.
3. Electrochemical corrosion rate measurements may be used to measure the
corrosion rate of structures which cannot be visually inspected or subjected
to weight loss tests. Underground tanks and pipes and large chemical plant
components are examples.
The American Society for Testing and Materials has established two
recommended practices that pertain to electrochemical techniques. G 3-68
is the Recommended Practice for Conventions Applicable to Electro-
chemical Measurements in Corrosion Testing; this deals with reporting
and displaying electrochemical measurements. G 5-71 is the Recom-
mended Practice for a Standard Reference M e t h o d for Making Potentio-
static and Potentiodynamic Anodic Polarization Measurements; this
covers experimental procedures that lead to repetitive results when con-
ducted by different researchers in different laboratories.
The widest use industrially of electrochemical principals of corrosion has
been that of anodic protection. This is a procedure first commercially
developed by researchers at Continental Oil C o m p a n y and since reported
by many others; however, it was a paper by C. Edeleanu of England in 1954
that first called attention to anodic protection [11]. Its techniques are not
applicable to all solutions; it has found widest use involving steel in concen-
trated sulfuric acid. While anodic protection is beginning to play an im-
portant role in certain segments of the chemical process industry, its
somewhat limited involvement has not yet resulted in any standards being
developed for its use. Contrariwise, the more widely used technique of
cathodic protection, both applied to underground structures such as pipe-
lines and tanks and also to process fluids and their containers, has been
evident in many standards, particularly by the National Association of
Corrosion Engineers and from the work of N A C E by the Department of
Transportation and other regulatory bodies who are concerned with

corrosion on underground units transporting dangerous fluids.
NACE Standard, RP-01-69, is the recommended practice for Control of
External Corrosion on Underground or Submerged Metallic Systems.
Much of this has now appeared in an American National Standards Insti-
tute Specification (ANSI) B31.4-1971 for liquid petroleum transportation.
An earlier standard on gas transmission and distribution piping systems,
USASI B31.8-1968, also has sections dealing with standard methods of
determining cathodic protection requirements and the measurement of this
protection.
Elevated Temperature Corrosion

The INCO Test Spool cannot be used for temperatures above 500 F
because of the breakdown of the plastic insulators between specimens.
This is the maximum temperature if Teflon is used; if phenolic spacers are
used, a lower temperature limit of 300 F must be established. However, it is
possible to adapt the INCO Test Spool by using either ceramic insulators or
by ignoring the concept of insulators entirely and by recognizing that at
temperatures of over 800 F oxidation effects will be more critical than any
electrolytic dissimilar metal effects. Thus, where aqueous or conductive
solutions are not a factor, INCO Test Spools have been prepared for use at
elevated temperatures and have given excellent results. The techniques of
evaluation are a variation on those in the conventional spool in the sense
that with oxidation effects there is usually a weight gain rather than weight
loss. Thus, it is oftentimes required to do a metallurgical cross-section
study of the depth of oxidation, the amount of metal remaining unaffected,
and therefore the load-carrying capacity at the end of the test. In practice,
ultrasonic and eddy current techniques can be used with certain alloys to
determine the amount of the material affected by the high temperature.
These nondestructive techniques, it should be pointed out, can also be used
in other phases of corrosion control. Particularly, eddy current techniques,
as exemplified by the Probolog or the Introview, 5 are excellent in determin-
ing localized corrosion of alloying elements as in dezincification, and
general corrosion loss from equipment such as pipe and tubing. Not to be
forgotten also are calipers that have been developed for inside wall tubing
measurement, though in a sense, this represents physical dimensioning and
thus falls into the category of corrosion rate determined by any dimensional
change in a given period of time.
Section 3.4.6 of Recommended Practice G 4 (Conducting Plant Corrosion
Tests) deals with the use of a bayonnet heater to determine the hot wall
effect in plant corrosion tests. Recommended Practice for Aqueous Corro-
5Probolog is an instrument manufacturedand sold by Branson Instruments Co., Stam-

ford, Conn. Introviewis an instrument manufactured and sold by Sperry GyroscopeCo.
Ltd., London, England.
sion Testing of Samples of Zirconium and Zirconium Alloys, ASTM

G 2-67, contains procedures for the evaluation of these materials at elevated
temperatures and pressures including a pickling procedure for preparation
of the specimens. N A C E Standard TM-01-71 deals with Autoclave Corro-
sion Testing of Metals in High Temperature Water.
Localized Corrosion
Localized corrosion exists in many forms such as pitting, crevice corro-
sion, filiform (underfilm), exfoliation, fatigue, and intergranular corrosion.
In most cases, the detection of one or more of these forms of corrosion is
best done visually, though in certain cases magnification from an optical
microscope is necessary. Crevice corrosion is often called contact or con-
centration cell, deposit, differential, aeration, gasket, interface, poultice,
water line, or wedge corrosion. It is a form of localized attack that occurs at
shielded areas on metal surfaces exposed to particular solutions. Usually
this corrosion occurs because of the design. Examples include spot-welded
lap joints, threaded or riveted connections, gasket fittings, porous welds,
valve seats, coiled or stacked sheet metal, marine or debris deposits, or
meniscus at a water line. The cure of such corrosion is often based in
changes in design.
Two recent symposia have taken place on this subject. One was the
Symposium on Localized Corrosion at the ASTM annual meeting at
Atlantic City, June 1971, and the other the U. R. Evans International
Conference on Localized Corrosion held at Williamsburg, Virginia, in
December 1971. A paper by W. D. France, Jr. [12] on Crevice Corrosion
of Metals was presented at the first conference, dealing specifically with
crevice corrosion of metals but touching also on other types of localized
corrosion. This serves as an excellent bibliography on most metal systems
that have been discussed in this field.
F. L. LaQue presented the keynote speech at the Evans Conference on
" T h e Problem of Localized Corrosion: What is it, its Extent, its Causes, its
Cures and What Needs to be Done." It suffices to say at this time that there
are no accepted standards of testing for localized corrosion. Examples
have been given earlier in this paper; both references cited above have
further examples.
Low Temperature Measurement

Since the corrosion rate generally, as was stated earlier, is doubled for
every 10-deg C rise in temperature, one would therefore not have to worry
about corrosion at low temperatures, other than for the effects of condensa-
tion on the surface, plus any possible metallurgical changes. However, the
corrosion engineer working with cryogenic materials must be cognizant of
the impact strength loss at low (below room) temperatures with certain
alloys. While this is not basically a corrosion phenomenon, it is a concern
of the materials engineer in the chemical industry. Condensation effects or

the corrosion on a cold surface can be determined, as was discussed earlier,
in a laboratory hot wall tester by the use of a cold finger on a metal test
surface allowing condensation of the corrodents to occur on that surface
and determining the corrosion in the conventional fashion. Embrittlement
may be determined by use of an Izod or Charpy pendulum impact tester.
The metallurgical changes in a weld must also be evaluated for impact
behavior at low temperatures. Certain plastics also become impact-sensitive
at lower than ambient temperatures so that their use at lower than normal
temperatures becomes delicate. In addition, the greater coefficient of con-
traction (and expansion, on heating) of many unreinforced thermoplastics
requires proper design (expansion joints or loops for long straight runs) to
avoid mechanical failures.
Process Evaluation of Other Materials

Materials other than metals do not corrode in the accepted electro-
chemical sense, since they do not haye metallic ions which can go into
solution; rather they deteriorate and are no longer usable under certain
conditions. The types of deterioration can be similar to corrosion; for
example, certain plastic materials stress crack in the presence of certain
organic solutions. There can be the equivalent of metasomatic corrosion
with certain parts of the composite nonmetallic materials being leached out.
The testing of nonmetallic materials for their corrosion resistance can
largely follow that used for metals in terms of Recommended Practice G 4
on Conducting Plant Corrosion Tests, except that the evaluation of the
nonmetallic materials is along a different path than that of metallic ma-
terials. Since many nonmetallic materials do not lose weight in certain
corrodents, but gain weight, obviously then, the use of mdd or ipy as a way
of reporting results has no great meaning. The preferred technique of
evaluating nonmetallic materials, irrespective of how they are exposed, is
the change of physical properties. Such properties can include tensile
strength, yield strength, flexural strength, hardness, elasticity, etc. The
American Society of Testing and Materials has many standards that
pertain to the measurement of these properties for a given material and
obviously such standards and procedures should be followed.
Perhaps one of the more widely used properties because of its simplicity
and rapidity is the flexural strength evaluation. This test is quite similar to
that for cast iron in A 48-48 in that the sample is flexed at the center by the
application of a load; the sample is supported as a simple beam at two
points equidistant from the application of the load. However, in the case
of plastics and other nonmetallics, conditioning after removal from the
environment and before testing, is critical in terms of establishing equi-
librium with the given temperature and humidity conditions.
This matter of testing reinforced plastics has now manifested itself as a
National Bureau of Standards Voluntary Product Standard, P-15-69,

called Custom Contact Molded Reinforced Polyester Chemical Resistance
Process Equipment. Section 10 of that standard deals with chemical
resistance and describes part of ASTM C 581-68, Test for Chemical
Resistance of Thermosetting Resins Used in Glass Fiber Reinforced
Structures. This method is based on a test procedure developed by the
Reinforced Plastic Corrosion Resistant Structures subcommittee of the
Society of the Plastics Industry, and stems from the early work at duPont
of Harvey Atkinson and Robert Webster. ASTM Committee C-3 on
Nonmetallic Materials has been active in this area.
C 581-68 requires that the test laminate be cured at room temperature
for 16 h to a Barcol hardness equal to the resin manufacturer's minimum
specified hardness for the cured resin. Tests are to be conducted under one
or more specified temperatures as well as the reflux temperature and the
required service temperature. Twenty-three corrodents are suggested for use
in obtaining general comparative chemical resistance data under non-
agitated, static conditions. The properties to be evaluated are determined
on specimens immersed in the test solutions for 30, 90, 180 days and 1 year
for one set of control specimens immediately following the cure period. In
addition, another set of specimens which have been aged in air at the test
temperature for the total test period are used as controls. The properties to
be determined are the change in thickness, change in Barcol hardness,
change in flexural strength and flexural modulus, and change in appear-
ance. These are determined at each time interval, with appearance observ-
ance including any surface changes, color changes, obvious softening or
hardening, crazing, lamination, exposure of fibers, or other effects indicative
of complete degradation or potential failure.
The flexural test in the specification of the Bureau of Standards can also
be used on other nonmetallics such as impregnated graphite, impregnated
.wood, rubber, other plastics, and under some controlled conditions,
concrete.
Tests for Lining Materials

A lining material is usually an organic or metallic thin film applied by one
of several means on a substrate. Since the substrate will not be exposed to
the environment in service, with the unlined (outside) side of the lined
substrate usually exposed to ambient condition, then it is not proper to test
lined steel or other metallic panels by full immersion or dunking. Such
dunk testing does not take into account the heat transfer through the
specimen to the surroundings and also the flexural variations that occur
because of temperature differentials both inside and outside the solution.
It was for this reason that the National Association of Corrosion Engineers
through Unit Committee T-6 has established a proposed standard test
method entitled L a b o r a t o r y Method for Evaluating Protective Coating for
CORROSION iN THE CHEMICAL INDUSTRY 183
Use as Lining Material in Immersion Service (no number assigned yet).

This method consists of exposing one size of a coated panel to environ-
mental conditions closely approximating those which are encountered
under actual service conditions. This test method, unlike other tests, closely
simulates the phenomenon found in service involving temperature differ-
entials between the external and the internal surfaces of the coating: these
differences may accelerate the permeation of the coating by the corrosive
medium. This work was originally presented in a technical journal in the
article, " A n Improved Method for Evaluating of Tank-lining Systems" by
A. F. Torres and S. S. Feuer, then of Atlas Chemical Industries [13].
Basically, the test cell normally consists of a 6 in. in diameter open
cylinder of Pyrex with a 4-in. minimum length and several side connections;
steel plates coated on one side with the lining to be tested are used to close
the open ends of the cylinder. The solution that serves as the environment
is introduced and heated internally, with a condenser attached to one of the
cylinder side connections to prevent evaporation of solution. Exposure of
the lining is for one face (the lining), exactly as in an actual tank. By proper
volume of solution, size of cylinder and size of plate, the ratio of the
surface area of the lining to the volume of the solution in the test cell can be
made to approximate that found in service conditions in a tank, drum, tank
truck, tank car, barge, etc.
Not only is this test used for lining materials but it is appropriate to
evaluate any material used as a vessel. This includes resin and fiberglass and
metals. It should be recognized, in the case of testing metals in this fashion,
that weight loss considerations would not be valid; instead, it is necessary
to check for pitting, crevice corrosion, loss in thickness, and more par-
ticularly, pickup of metallic ions from the metal surface in the corrodent.
Oftentimes, in the chemical industry for determining the proper material
for a storage vessel, it is necessary to check the change in color of the solu-
tion under standard exposure conditions. In this regard, it should be
recognized that slight corrosion which is not measureable in conventional
techniques, can cause a catalytic reaction of certain organic materials
resulting in their deviation from water-white appearance. The use of color
testing comparators, such as the American Public Health Association
Colorimeter, is often used in this regard.
Atmospheric Effects
In LaQue's paper on corrosion testing referred to earlier, there is a fine
historical discussion of the early work of ASTM over the controversy re-
garding the relative merits of different kinds of iron and steel in resisting
atmospheric corrosion. Around the turn of the century, Committee U
(later to become Committee A-5) reported on some of their early tests of
the different kinds of iron and steels exposed to an accelerated acid test, as
well as exposed to industrial and marine atmospheres. This early work led
to a round-robin test work on the part of steel and other metal producers,
as well as metal users. The late C. P. Larrabee, of the then Carnegie-Illinois
Steel Corporation, has a chapter in Uhlig's Corrosion Handbook dealing
with this matter [14]. Larrabee indicates that most corrosion investigators
in the USA followed the procedures of ASTM Committee A-5 and ex-
posed specimens at an angle of 30 deg to the horizontal facing south.
Four by six inch specimens were nominally used, mounted on porcelain
insulators to prevent galvanic effects. Conventionally, 70 specimens are
accommodated in an area roughly 68 in. long and 38 in. wide. Larrabee
indicated that nonferrous metals are exposed vertically by ASTM Com-
mittee B-3 while many tests of painted specimens are exposed at a 45-deg
angle.
R. K. Swandby, formerly of International Nickel and Wyandotte
Chemical Corporation, in a chapter in LaQue and Copson's book indicates
that "the degree that a metal or alloy is exposed to the atmosphere also
plays an important part in the severity of attack [15]. Many materials ex-
posed directly to the atmosphere are not attacked as severely as the materials
partially sheltered. The probable reason for this is that the partially sheltered
materials remain wet from dew for longer periods of time. They also do not
receive the beneficial effects of washing by rain."
Of particular significance which should be mentioned in regard to at-
mospheric corrosion testing and the results thereof is the famous atmos-
pheric test site of International Nickel at Kure Beach, North Carolina. One
of two locations is 80 ft from the Atlantic Ocean water line and the other is
800 ft. Duplicate specimens are exposed in both lots to get comparative
effects of the severity of salt spray and moisture on the various materials.
A direct offshoot of this testing of atmospheric materials is the develop-
ment of the low-alloy, high-strength steels to ASTM Standard A 242,
popularly known by trade names such as Cor-ten and Mayari-R 6 among
others. These materials are widely used in many atmospheric exposures,
including chemical plants, in their unpainted, natural conditions. Not only
are there savings in paint costs but also the weight (cross-section) can be
reduced to take advantage of the higher strength. Over the years, countless
atmospheric tests have been made at many locations around the world, so
that comparative results now are available both with regard to the ma-
terials and to the environments.
Other committees of ASTM, as well as NACE and the Federation of
Paint Societies, have evaluated the atmospheric resistance of paints and
protective coatings; for chemical plant environments, the most widely
accepted test specimen is the KTA panel. This was devised by the late
Kenneth Tator, the leading industrial paint testing consultant in the USA,
and exists in several variations. Essentially, a KTA panel consists of a piece
6 Cor-ten is a trade mark of the United States Steel Corporation. Mayari-R is a trade
mark of the Bethlehem Steel Corporation.
of 1/~-in. steel, approximately 4 by 6 in., to which is welded a piece of

channel iron. The weld purposely is a skip-weld and is rough; all sharp
edges are kept as such and not rounded. Surface preparation is varied to
meet industrial test requirements. After the panel is coated, it is scribed
down to the surface as well as being impacted. Thus, one ends up with steel
that has most of the common industrial deficiencies, so far as design for
protective coatings is concerned--skip welds, sharp edges, pockets, impact
area, scribe, corners, pits, planes, etc. Rating of such panels is done on a
comparative basis of 0 to 10 for each potential location of failure as a
function of time of exposure. Size of panels and construction can vary
depending on space requirements and the number of systems to be tested.
The ultimate in such testing was Texas-style; individual 55-gallon drums
blasted and coated with individual paint systems and arranged in a l~-acre
field. Inspection of the finished paint system before exposure is requisite to
avoid a lack of correlation between laboratory preparation of panels and
industrial application of paint on structures.
Corrosion Control
Many years ago, one of the early undergraduate texts which dealt with
materials indicated that there were three ways of stopping corrosion.
These were:
1. Change the materials
2. Change the environment
3. Protect the material
As advances have been made in corrosion science and engineering, one
should also add several other factors in controlling or stopping corrosion
that include the following:
1. Improve the design
2. Develop improved testing methods
3. Tighten existing specifications
4. Improve inspection of equipment
5. Share common corrosion experiences
6. Join appropriate technical societies to disseminate the above
knowledge.
Certain of these items are self-evident; others need more detailed ex-
planation. For example, with regard to changing the materials as a way of
stopping corrosion, this is usually the step that a novice would undertake
in his lack of corrosion fundamentals. There are still those who consider
that upgrading to a stainless steel is a solution to all corrosion problems!
Obviously, this is not the case; otherwise, there would not be so many
materials of construction on the market. However, by upgrading to the
proper material of construction, one can often avoid, or at least minimize,
the effect of a given environment. Since it is not the purpose of this chapter
to go into all the ramifications of alternate materials of construction, the

matter cannot be properly pursued. It suffices to say that the gamut of
materials of construction, as indicated earlier, range all the way from steel
and cast iron through stainlesses, high nickel alloys, copper-base, aluminum-
base, nickel-base, titanium, zirconium, precious metals, and ultimately into
the entire field of nonmetallics which can include glass. Not that glass is the
ultimate material of construction, but it does have a very wide range of
chemical resistance to all but strongly alkaline and acid conditions involv-
ing H F and HC1. As indicated earlier, the fragility problem with glass had
been a problem until Coming Glass Company, in particular, developed
their armourized material which uses an epoxy-fiberglass coating on the
exterior. In addition, they were very ingenious in developing a ball joint
that allowed flanges to be slightly misaligned and still not put undue stress
and incipient failure on the part.
The matter of changing the environment is one that leads automatically
into the field in inhibitors. It is not always possible to change major process
conditions, but there is often the feasibility of adding a small amount of an
independent material which would minimize corrosion. Environmental
changes which can take place to reduce corrosion include change in tem-
perature and pressure, agitation, aeration, etc. One of the many variables
which can be controlled to change the environment is that of the addition
of one or more compounds which change the surface characteristics of the
metal exposed to the environr~ent. Such compounds are usually called
inhibitors in the sense that they reduce the corrosion rate of a metal in a
given environment. These inhibitors can be organic or inorganic in nature
and, as indicated, are usually specific for a given metal in a given compound.
There are several fine sources of the behavior of specific compounds and the
metals which they inhibit to be found in the literature; these include the
data in Uhlig's Corrosion Handbook [14] as well as in Fontana and Greene
[5] and in the Russian book called Metallic Corrosion Inhibitors by Putilova,
Balezin, and Barannik [16]. Inhibitors have been widely used in the past in
cooling water circuits because of the closed nature and therefore fixed
amount of water to be inhibited. Under conditions of a closed loop, as
opposed to an open or continuous flow-through circuit, the economics of
inhibitors are well established. One generally speaks in terms of a tenth of a
percent by weight or less of inhibitor relative to the concentration of the
major corrodent. Even at this low concentration in a once-through circuit,
inhibitors are generally too expensive to be used. Until the recent upsurge of
interst in the toxic effects of chromates on plant and water life, sodium
chromate was a very widely used inhibitor to prevent attack of aerated
water on iron and steel. Other materials have been substituted for chromate
such as nitrites, nitrates, phosphates and silicates, sometimes singly and
sometimes in combination. It should be noted that what is an inhibitor for
one material is oftentimes an accelerator for another material. Thus,
amines are relatively effective on steel but because of their ammonia com-
plex, under certain conditions of temperature, concentration, and design,
they can cause stress corrosion cracking of copper-base materials.
The organic, the physical, and the electrochemists have had great oppor-
tunities, and have taken advantage of these opportunities, to investigate
corrosion mechanisms and develop corrosion inhibitors, particularly using
the technique of the potentiostat for this work. Noteworthy among these
investigators have been N o r m a n Hackerman, formerly of the University of
Texas, and his students and colleagues in this field. Certain organic com-
pounds have been discovered and investigated which make effective
inhibitors; these include thiourea and certain benzoates. There are many
companies who specialize in the development and manufacture of in-
hibitors for particular metals in specific environments. These materials
lend themselves widely to certain production operations.
There was a class of inhibitors first discovered and developed by Wachter
and colleagues at Shell Development Corporation that worked in the vapor
phase, rather than in the liquid phase, in preventing the formation of oxida-
tion and rust on c o m m o n metals [17]. These vapor phase inhibitors are
generally organic in nature and have in common a relatively high volatility
from the crystalline solid material. The component is generally an amine in
nature and can be applied by any one of several techniques. Crystals of the
VPI (Vapor Phase Inhibitor, used by Shell Development as a trade name)
can be placed in an open container in a closed vessel; it will be found that
the interior surfaces of this vessel will not develop rust or oxide even if
water vapor or oxygen enter the vessel. As such, vapor phase inhibitors
have been used to maintain standby equipment such as boilers, turbines,
etc. in prime condition without the necessity of having to do any cleaning
when these units are put back on the line. The vapor phase inhibitor can also
be applied by fogging a solution containing the crystals into the container
where the solid inhibitor crystallizes and puts a mono-molecular film on the
surface of the container. Other techniques have involved the impregnation
of paper with the inhibitor so that a delicate part, such as a ball bearing, can
be wrapped in the inhibited impregnated paper and sealed in paper or a
plastic film that does not allow the vapor phase inhibitor to rapidly dissipate
to the atmosphere. Such parts can then be stored for quite some time in
humid conditions without rusting occurring. Quite obviously, the military
is interested in such inhibitors for the storage and maintenance of critical
equipment under humid conditions, such as is tound in many parts of the
world.
Where it is impractical, for any one of several reasons which may include
economics as well as ease of fabrication, to change the material or to add an
inhibitor, then the materials engineer may choose to use a protective coat-
ing on an inexpensive substrate. The purpose of the coating is to act as a
nonreactive barrier between the environment and substrate. F o r this
reason, the coating preferably should be impervious to the environment, it

should be continuous (free of voids or pin holes), and it should be suffi-
ciently durable to withstand the temperature and process conditions of the
environment. A differentiation is made in corrosion circles between a
coating which is used for atmospheric exposure, and a lining which is used
for immersion conditions. While it is true that coatings and linings have
much in common and may often be applied by the same skilled trades,
nonetheless, there is sufficient difference in terms of the severity of corrosive
conditions between a coating and a lining that special requirements exist for
both, including application procedures. There are many committees of the
National Association of Corrosion Engineers that have been involved in
these problems; basically, Technical Practices Committee T-6 with its
various subcommittees has dealt with coatings for atmospheric conditions
and linings for immersion conditions including all the ramifications of
surface preparation. It should be recognized that linings can differ from
coatings in the sense that linings include not only material applied as a
liquid which converts to a solid in o n e of several fashions, but also solid
materials themselves. For example, one would speak of a sheet lining of
rubber or plastic, and one could also speak of a lining that was made up of a
solid membrane plus chemically resistant brick and mortar construction.
Irrespective of whether the lining is applied as a solution and then con-
verted to a solid (this is done by any one combination of three techniques
that include drying by oxidation, drying by solvent evaporation, and drying
by catalysis), it is requisite that the resulting lining material be resistant
chemically to the environment to which it is in contact. Since most conven-
tional linings are organic in nature, one has to be careful of their use with
organic solvents. The adage to be applied in this case is "when in doubt,
test." Of course, testing should be carried out, as indicated earlier, so as to
avoid any obvious erroneous conclusions due to faulty testing techniques.
One lining worthy of mention because of its wide applicability is that based
on glass flakes disp'ersed usually in a polyester resin; this can be applied as a
spray with a high pressure pump gun or can be troweled on the surface. In a
case of the spray the catalyst is applied simultaneously with the polyester
resin glass flake mixture, while in the case of troweling the catalyst is mixed
in with the viscous mixture of flake and resin in a batch fashion. Depending
upon the size of flakes and the finishing operation the flakes may or may not
be rolled as they are applied to the surface (currently, the larger size glass
flakes are rolled after application to take on an overlapping parallel position
to the substrate like fish scales; smaller flakes are not rolled and exist in
their r a n d o m position). The resistance of this material is quite outstanding
particularly to many inorganic environments. There are other combinations
of reinforcing materials and resin binders such as glass mat and glass cloth
with epoxy resins, polyester resins and even furan resins; much of the
technology with these materials is similar to that of the monolithic rein-
forced plastic construction, except that when these materials are used as a
lining, the difference in expansion coefficient of the reinforced plastic lining
and the substrate must be taken into account if there are sizable temperature
extremes. A material that offers superb corrosion resistance is not par-
ticularly good if it cracks because its expansion was greater than that of the
vessel in which it was installed--a failure is a failure, irrespective of whether
it was due to corrosion or due to mechanical problems such as thermal
expansion or contraction.
With regard to control of corrosion by coatings, it should be indicated
that the present experience in the chemical industry, as well as in variations
of the chemical industry, such as, petrochemical, pulp and paper, fertilizer,
etc., has established that a zinc-rich primer with protective top coats offers
the longest maintenance-free life provided there is continuous inspection and
patch repair. These zinc-rich primers are either organic or inorganic in
nature. The inorganic primers are generally based on silicates of one form or
another, with a very finely divided zinc powder at approximately a 90 per-
cent by weight basis. This material must be applied on blasted steel (free of
mill scale and rust). Depending upon the silicate the curing often takes place
under moist conditions. The organic zinc-rich primers are generally based
on chlorinated rubber or catalyzed epoxy resins; a blasted surface, free of
moisture, must be used. It is important that all the solvent be removed from
the primer and that the primer be completely dry before being topcoated;
otherwise, blisters and mudcracking may form. The topcoats for inorganic
zinc primer and organic zinc primer are usually chosen from those based on
vinyl resins, catalyzed epoxy resins, or chlorinated rubber resins, often
depending upon the environment to which the coated steel will be subjected
and to the temperature. There are many advantages to each of these top-
coats and the materials engineer will do well to balance ease of application,
foolproof application, ease of topcoating with a second coat, ease of repair,
maintenance of gloss and cost, taking all these factors into account. The use
of the zinc-rich primers is based on the cathodic protection given to localized
breaks in the primer and the bare steel. While this is true in theory, nonethe-
less one must balance the aggressiveness or the conductivity of the environ-
ment against the particular zinc primer used and the frequency of repair.
Let it be categorically stated that in an aggressive environment, even the
best protective coating system requires continuous inspection and patch
repair; the better the system, the less frequent and smaller the amount of
repairs to be made. A protective coating system in a strong chemical
environment should not be expected to have an indefinite life. A long life,
yes, but an indefinite life, no.
Changing the material obviously can be used to control corrosion in the
sense that for most environments, even the most aggressive, there can be
found a material perhaps at a great cost which is resistant to that environ-
ment. Thus, in the melting of glass to form fiberglass, where temperatures
in excess of 3000 F are involved with oxidization, the material of construc-

tion of the container for the molten glass is platinum. Here one needs
resistance to oxidation and maintenance of hole size with resistance to
abrasion of the fowing glass. While this may be an extreme, the fiberglass
industry would not use platinum if less expensive materials would do the
job. Long continuous maintenance-free life is required and platinum has
been found to be the one material that will do this job. Advances have been
made with other of the high melting point materials such as tantalum,
columbium, hafnium, and tungsten, such that in many specific corrosive
environments these materials may be used often as a deposit on a less ex-
pensive and less resistant substrate. Techniques have been developed for
the deposition of these materials usually from fused salts so as to achieve
the chemical resistance of the expensive high density materials without the
cost of these materials in their solid form. Periodically, interest revives in
deposits on nickel on steel, usually when there is a critical nickel shortage.
Historically, one looks back to the story of the uranium hexafluoride
diffusion plant at Oak Ridge for the widespread use of the electrodeposited
nickel on the inside of steel tubes to provide the corrosion resistance of
nickel at a lesser cost and greater availability than solid nickel. At one time,
there was a major producer in the United States who made sheet steel
coated with 10 or 20 mils of electrodeposited nickel of one or both sides;
this steel could then be rolled and fabricated (including welding) in the
technique quite similar to that of sheet steel. Other developments that have
taken place over the years included the development of the electrode-less
nickel conventionally known by the General American Transportation
Corporation trade name, Kanigen. This material held great promise for the
caustic soda industry until problems began to develop in the adhesion of the
electrode-less deposited nickel on other than sandblasted surfaces. In
addition, repair techniques were somewhat difficult for large vessels. The
best source for authoritative and quick information on alternate materials
of construction for a given chemical environment is that found in the book
published by the National Association of Corrosion Engineers, entitled
Corrosion Data Survey. Here, for a given environment (environments are
thoroughly referenced by the chemical radical involved), one finds the
famous Nelson chart, correlating temperature and concentration of
environment and the corresponding symbol for corrosion rate of approxi-
mately 2 0 + metals. Volume V of Corrosion Data Survey will be in two
parts; one concerned with metals and the other concerned with nonmetals.
As indicated earlier, the gamut of metals surveyed in this book is quite
thorough, since it represents the work of contributors across the industries
of the United States and Canada.
Electrochemical Methods
In the section on Standards and Testing, both cathodic and anodic pro-
tection were discussed as being applied in many industrial situations; they
represent a variation of protection of the substrate by other than a lining or
a coating. While cathodic protection has been most widely applied to
buried tanks, hydraulic elevator shafts, and pipes and related underground
structures, applying the principles of cathodic protection can allow its use
to tank interiors and even in certain circumstances the interior of pipes. At
the 4th International Corrosion Conference on Metallic Corrosion in
Amsterdam, there was a report on the application of anodic protection to
the interior of titanium heat exchanger tubes handling a reducing viscose
solution that was most corrosive to titanium in the absence of the impressed
electrical current [18]. In addition, the solution was so corrosive that no
other metal was practical for this service; difficulties had been found in
impregnated graphite tubing because of fouling and fragility. Changing the
potential of a surface from a region in which the surface is active to a region
in which it is passive can be used under many conditions to effectively
control corrosion.
Major Trends and Problems

The corrosion engineer in the chemical industry is continuously faced
with a problem that has ever bothered all engineers, irrespective of the
industry in which they are located; namely, how to update engineering
knowledge with the scientific developments that have been brought forth by
others. One has only to look at the pages of the NACE publication,
Corrosion, or the section in the ASM-AIME publication, Metallurgical
Transactions dealing with transport phenomena, to realize that monthly, if
not daily, great steps are being made in the field of materials science. More
than ever before, it is necessary to coordinate the activities of the materials
engineer with those of the materials scientist. Several years ago, such an
attempt was made by personnel at the National Bureau of Standards and
personnel from the Office of Naval Research when a conference bridging
the gap between theoretical and applied corros_:on was held in Washington
with great success [19].
Most noteworthy, however, have been the developments in electro-
chemistry as applied to theoretical and ultimately to practical corrosion
problems. It behooves the corrosion and materials engineer to remain
alert to the theoretical developments and incorporate them in his practical
thinking and doing. The very fact that in many universities and engineering
schools there has been a decided trend towards a Department of Materials
Science and Engineering is indicative of the trend taking place toward
making the materials engineer a well-rounded man. In many universities,
the materials scientist and materials engineer are grounded in the basic and
current theory of metallurgy, polymers, and ceramics. It may be assumed
then, that the future corrosion engineer will be at home in organic polymer
chemistry, in metallurgy of all alloys for the complete gamut of tempera-
ture and physical requirements, and in nonmetallics, such as glass, ceramics
and graphite, along with knowledge of electrochemistry as it relates to the
corrosion phenomenon and as it relates to cathodic and anodic polarization.
Today, the fact that engineering students in the United States can now
take more than one course at a university in the field of corrosion is quite an
achievement. So far as is known at the present time, a student would not
receive a degree in corrosion engineering but perhaps would receive his
degree in metallurgical engineering with a specialty in corrosion, or more
particularly, a degree in materials engineering with a specialty in corrosion.
In fields such as engineering the demand on a student's time for funda-
mental and theoretical courses in all engineering disciplines, and fields such
as humanities, mathematics, and physics, are such that it is often difficult
for such a student to get a thorough grounding in materials or corrosion
engineering by actual undergraduate course work. The work that the
National Association of Corrosion Engineers has done in publishing a
lecture and home study course on basic corrosion engineering [20], as well
as the course which Dr. F o n t a n a has written and given for the American
Society of Metals on Corrosion [21], is a technique by which the graduate
engineer and others may grasp the fundamentals of corrosion and ma-
terials engineering. That, along with special graduate courses, is a procedure
the practicing engineer can use to improve his knowledge of the field.
References
[1] LaQue, F. L. in Proceedings of the American Soeiety lor Testing and Materials, Vol. 51,
1951, pp. 495-582.
[2] Bates, J. F., Industrial and Engineering Chemistry, Vol. 58, No. 2, 1966, pp. 19-29.
[3] Wachter, A. and Treseder, R. S., Chemical Engineering Progress, Vol. 43, June 1947,
pp. 315-326.
[4] Fontana, M. G. in Proceedings oJ Short Course on Process Industry Corrosion, presented
by Department of Metallurgical Engineering, Ohio State University and Technical
Group Committee T-5, NACE, Houston, Sept. 1960.
[5] Fontana, M. G. and Greene, N. D., Corrosion Engineering, McGraw-Hill, New York,
1967.
[6] Landrum, R. J., Metals Engineering Quarterly, Vol. 1, No. 2, May 1961, pp. 45-57.
[7] Scharfstein, L. R. and Eisenbrown, C. M., Advances in the Technology of Stainless
Steels and Related Alloys, STP 369, American Society for Testing and Materials, 1963,
pp. 235-239.
[8] Huey, W. R. in Transactions of the American Society of Steel Treating, Vol. 18, 1930,
p. 1126.
[9] Streicher, M. A., ASTM Bulletin 188, American Society for Testing and Materials,
Feb. 1953, p. 35.
[10] Stern, M. and Wissenberg, H. J., Journal of the Electrochemical Society, Vol. 106,
1959, p. 759.
[11] Edeleanu, C., Metallurgia, Vol. 50, 1954, p. 113.
[12] France, W. B., Jr., "Crevice Corrosion of Metals" presented at the Symposium on
Localized Corrosion, ASTM Annual Meeting, Atlantic City, N.J., June 1971.
[13] Torres, A. F. and Feuer, S. S., Materials Protection, Vol. 2, Jan. 1963, p. 23.
[14] Larrabee, C. P. in Corrosion Handbook, H H. Uhlig, Ed., Wiley, New York, 1948,
pp. 120-125.
[15] Swandby, R. K. in Corrosion Resistance of Metals andAlloys, 2nd ed., F. L. LaQue and
H. R. Copson, Eds., Reinhold, New York, 1963, Chap. 2, pp. 45-65.
[16] Putilova, I. N., Balezin, S. A., and Barannik, V. P., Metallic Corrosion Inhibitors,
Pergamon Press, New York, 1960.
[17] Wachter, A., Skei, T., and Stillman, N., Corrosion, Vol. 7, No. 9, Sept. 1951, pp. 1-12.
[18] Evans, L. S., Hayfield, P. C. S., and Morris, M. C. in Proceedings of 4th International
Metallurgical Corrosion Congress, NACE, Houston, Texas, 1972, pp. 625-635.
[19] Symposium on Coupling of Basic and Applied Corrosion Research, ONR, NRL, and
NBS, 21-22 March 1966; published by NACE, Houston, Texas, 1969.
[20] NACEBasic Corrosion Course, National Association of Corrosion Engineers, Houston,
Texas, 1970.
[21] Metals Engineering Institute Corrosion Course, American Society for Metals, Metals
Park, Ohio, 1961.
STP534-EB/Nov. 1973
Chapter 10

Nonferrous Metals Industry
W . H. A i l o r ~
Industrial Corrosion Test Standards

Nonferrous metals are used in a wide variety of applications: archi-
tectural and structural uses, as coatings for other metals or materials, as
electrical conductors and protective anodes, as packaging foils, and in
composite products. Corrosion testing, therefore, is diversified and includes
laboratory, field, and service requirements in natural as well as artificial
environments. Typical tests range from stress corrosion methods to dura-
bility of coatings and include effects of antifoulants, heat treatments,
anodized coatings, and weldments. Weathering characteristics are important
for not only corrosion considerations but also for reasons of esthetics.
Many of the commonly used corrosion tests may be used for evaluation
of any of the nonferrous metals. In addition, there are specialized tests for
single alloy systems, ir~ addition to those for different alloys within a
system [1].
Listings of corrosion test specifications for individual metals and alloys
are to be found in the series of Military Standardization Handbooks which
include:
MIL-HDBK-694 (MR) Aluminum and Aluminum Alloys
MIL-HDBK-698 (MR) Copper and Copper Alloys
MIL-HDBK-693 (MR)Magnesium and Magnesium Alloys
MIL-HDBK-697 (MR) Titanium and Titanium Alloys
An additional useful source of reference to specifications is Military
Handbook MIL-HDBK-H1, "Cross Index of Chemically Equivalent
Specifications and Identification Code (Ferrous and Nonferrous Alloys)."
Included in this handbook are listings of specifications issued by the
following agencies: General Services Administration (Federal); Depart-
ment of Defense (MIL and JAN); American Iron and Steel Institute;
Society of Aeronautical Engineers (AMS); Aluminum Association;
1 Reynolds Metals Company, Richmond, Va.

194
CORROSION IN THE NONFERROUS METALS INDUSTRY 195
Society of Automotive Engineers (SAE); and American Society for Testing

and Materials [2-6].
Standardization
The standardization of industrial corrosion test methods is accomplished
through several channels. Trade associations for both nonferrous metal
producers and users, professional organizations involving metallurgists and
engineers, government and national standardization groups, and the Iflter-
national Standards Organization (ISO) are among those attempting to
coordinate and establish acceptable and useful test methods. Such groups
include the Aluminum Association (U.S.), the British Non-Ferrous Metals
Research Association, the Aluminium Federation (British), the Inter-
national Lead-Zinc Research Organization, the Copper Development
Association, the American National Standards Institute, and the Inter-
national Standards Institute.
Societies devoting strong leadership in the establishment of industrial
corrosion standards are the American Society for Testing and Materials
(Committee G-1 on Corrosion of Metals), the National Association of
Corrosion Engineers, the American Electroplaters Society, and the Metal
Properties Council (Subcommittee VII1 on Corrosion).
Nonferrous Metals
Nonferrous metals include aluminum, copper, lead, magnesium, nickel,
titanium, zinc, beryllium, and refractory materials, such as molybdenum,
tantalum, and tungsten. In addition, certain metals are used for plating--
among these are cadmium, chromium, tin, gold, and silver.
These nonferrous metals vary widely in chemical activity, from the
generally active magnesium, aluminum, and zinc to the relatively inert,
or passive, tantalum, silver, and gold. In addition to these differences, there
are differences in activity among the alloys within a metal system such as
aluminum.
Associated with the wide range of chemical activities for the nonferrous
metals and alloys is a similar variation in corrosion resistances. The
corrosion may take one or more forms of the following types of corrosive
attack: uniform corrosion, pitting attack, intergranular attack, erosion
corrosion, impingement attack, cavitation corrosion, fatigue corrosion,
stress-corrosion cracking, filiform corrosion, dezincification, graphitization,
fretting corrosion, high-temperature oxidation, biological attack, galvanic
corrosion, and concentration-cell attack. These types of corrosion attack
are discussed in Chapter 1 of this book [7-9].
The corrosion of nonferrous metals may be usefully grouped into the
several metal systems. A discussion of a number of nonferrous metals and
their alloys follows.
Aluminum
A l t h o u g h a l u m i n u m is one of the highly active metals, the n o r m a l oxide
film present on the surface is relatively inert and acts as a corrosion barrier.
W h e n corrosion does occur on aluminum, the m o s t c o m m o n f o r m is
pitting attack, a l t h o u g h a n u m b e r o f other f o r m s m a y be present. The
corrosion resistance o f a l u m i n u m alloys varies considerably depending o n
the m a j o r alloying elements. Careful alloy selection a n d proper corrosion
testing are essential, therefore, to insure satisfactory service life in a par-
ticular environment.
The designations o f the various alloy g r o u p s u n d e r the A l u m i n u m
Association classification are associated with the m a j o r alloying elements
and are shown in Table 1 ( w r o u g h t alloys) and T a b l e 2 (casting alloys).
TABLE 1--Designations for wrought aluminum alloy groups.
Alloy
Number
Aluminum--99.00 7o minimum and greater lxxx~

Major Alloying Element
Aluminum Alloys Grouped by Copper 2xxx
Major Alloying Elements Manganese 3xxx
Silicon 4xxx
Magnesium 5xxx
Magnesium and Silicon 6xxx
Zinc 7xxx
Other Element 8xxx
Unused Series 9xxx
a The last two digits identify the aluminum alloy or indicate the aluminum purity. The
second digit indicates modifications of the original alloy or impurity limits.
TABLE 2--Designations for aluminum casting alloy groups.
Designation
Number
Aluminum--99.00 ~o minimum and greater lxx. x

Major Alloying Element
Aluminum Alloys Copper 2xx. x
Grouped by Major Silicon, with added Copper and/or Magnesium 3xx.x
Alloying Element Silicon 4xx. x
Magnesium 5xx. x
Zinc 7xx. x
Tin 8xx. x
Unused Series 6xx. x
Other Major Alloying Elements 9xx. x
CORROSION IN THE NONFERROUSMETALS INDUSTRY 197
Alloys h a v i n g silicon, m a g n e s i u m , or m a g n e s i u m a n d silicon as m a j o r

a l l o y i n g elements h a v e c o r r o s i o n resistances r a n k i n g with t h a t for p u r e
aluminum.
The h i g h - s t r e n g t h c o p p e r - b e a r i n g alloys m a y develop i n t e r g r a n u l a r
attack i n corrosive e n v i r o n m e n t s , a n d stress-corrosion c r a c k i n g m a y result
f r o m e x p o s u r e o f alloys h a v i n g m a g n e s i u m c o n t e n t in excess o f 3.5 p e r c e n t
in severe e n v i r o n m e n t s .
T a b l e 3 lists the c o r r o s i o n resistances of m a n y a l u m i n u m alloys a n d
shows typical applications.
TABLE 3--Typical characteristics, applications, and resistance to corrosion of aluminum

alloys.
Resistance to Corrosion
Stress
Corrosion
Alloy and Temper Generala Crackingb Typical Applications
EC A A Electrical Conductors
1060 A A Chemical Equipment, Railroad Tank Cars
1100 A A Sheet Metal Work, Spun Hollowware,
Fin Stock
2011-T3 Dc D ScrewMachine Products
T4, T451 Dc D
T8 D A
2014-T3 D" C Truck Frames, Aircraft Structures
T6, T651 D C
2017-T4 D, C ScrewMachine Products Fittings
2018-T61 Aircraft Engine Cylinders,
Heads and Pistons
2024-T4, T3 D~ C
T361 D~ C Truck Wheels
T6 D B Screw Machine Products
T861, T81, T8511 D A Aircraft Structures
2025-T6 D C Forgings, Aircraft Propellers
2117-T4 C A
2218-T61 D C Jet Engine Impellers and Rings
2219-T31 Dc C Structural Uses at High Temperatures (to
T37 Dc C 600 F), High Strength Weldments
T81 D A
T87 D A
2618-1"61 D C Aircraft Engines
3003 A A Cooking Utensils, Chemical Equipment,
Pressure Vessels, Sheet Metal Work
Builder's Hardware, Storage Tanks
3004 A A Sheet Metal Work, Storage Tanks
4032-]'6 C B Pistons
5005 A A Appliances, Utensils, Architectural,
Electrical Conductor
5050 A A Builder's Hardware, Refrigerator Trim,
Coiled Tubes
5052 A A Sheet Metal Work, Hydraulic Tube,
Appliances
TABLE 3--Continued.
Stress
Corrosion
Alloy and Temper General~ Cracking b Typical Applications
50564) Aa Be
H111 Ae Be Cable Sheathing
H12, H32 Ae Be Rivets for Magnesium
H14, H34 Ae Be Screen Wire
H18, H38 Aa Ca Zippers
5083 Ae Be Unfired, Welded Pressure Vessels
5086-0 Ae Ae
H32 Ae Ad Marine, Auto Aircraft
H34 Ae Be Cryogenics, TV Towers
H36 Aa Be Drilling Rigs
H38 Aa Be Transportation Equipment, Missile Com-
ponents
5154 Ae Ad Welded Structures, Storage Tanks, Pressure
Vessels, Salt Water Service
5254 Ad Ad Hydrogen Peroxide and Chemical Storage
Vessels
5252 A A Automotive and Appliance Trim
5454 A A Welded Structures, Pressure Vessels,
Marine Service
5456 A~ Ba High Strength Welded Structures, Storage
Tanks, Pressure Vessels, Marine Appli-
cations
5457 A A
5652 A A Hydrogen Peroxide and Chemical Storage
Vessels
5657 A A Anodized Auto and Appliance Trim
6053 A A Wire and Rod for Rivets
6061--0 B A Heavy-Duty Structures Requiring Good
T4 B B Corrosion Resistance, Truck and Marine,
T6 B A Railroad Cars, Furniture, Pipelines
6063 A A Pipe Railing, Furniture, Architectural
Extrusions
Copper
C o p p e r and c o p p e r alloys are g e n e r a ll y classed as being c o r r o s i o n

resistant. A surface c o a t i n g or p a t i n a (basic c o p p e r sulfate and o t h er
c o m p o u n d s ) p r o t e c t s the m e t a l f r o m f u r t h e r attack. A q u e o u s a m m o n i a ,
s o l u t i o n s o f cyanides, o x id iz in g salts a n d acids, and acids or salts in the
presence o f oxidizing agents p r e v e n t th e f o r m a t i o n o f p r o t e c t i v e films.
C o r r o s i o n a t t a c k on c o p p e r and c o p p e r alloys m a y take the f o r m o f
ge n eral c o r r o s i o n , pitting, dezincification, stress c o r r o s i o n , c o r r o s i o n
fatigue, an d intercrystalline a t t a c k [6].
T h e g r o u p i n g s o f c o p p e r and its alloys u n d e r the C o p p e r D e v e l o p m e n t
A s s o c i a t i o n system is s h o w n in T a b l e 4 [10].
TABLE 3---Continued.
Stress
Corrosion
Alloy and Temper Generala Cracking b Typical Applications
6066-0 C A Forgings and Extrusions for Welded

T4 C B Structures
T6 C B
6070 B B Heavy Duty Welded Structures, Pipelines
6101 A A High Strength Bus Conductors
6151-T6, T652 Moderate Strength Intricate Forgings for
Machine and Auto Parts
6201-T81 A A High Strength Electric Conductor Wire
6262 B A Screw Machine Products
6463 A A Extruded Architectural and Trim Sections
7001 Cc C High Strength Structures
7039 B C Welded Cryogenic and Missile Applications
7075-0
T6 Cc C Aircraft and Other Structures
T73 C A
7079-T6 Cc C Structural Parts for Aircraft
7178-T6 Cc C Aircraft and Other Structures
a Ratings A through E are relative ratings in decreasing order of merit, based on ex-
posures to sodium chloride solution by intermittent spraying or immersion. Alloys with A
and B ratings can be used in industrial and seacoast atmospheres without protection.
Alloys with C, D, and E ratings generally should be protected at least on faying surfaces.
b Stress-corrosion cracking ratings are based on service experience and on laboratory
tests of specimens exposed to the 3.5 percent sodium chloride alternate immersion test.
A = No known instance of failure in service or in laboratory tests.
B = No known instance of failure in service; limited failures in laboratory tests of
short transverse specimens.
C = Service failures with sustained tension stress acting in short transverse direction
relative to grain structure; limited failures in laboratory tests of long transverse
specimens.
D = Limited service failures with sustained longitudinal or long transverse stress.
c In relatively thick sections the rating would be E.
d This rating may be different for material held at elevated temperature for long periods.
F o r commercial p u r e c o p p e r small v a r i a t i o n s are n o t significant as to t h e

c o r r o s i o n resistance. R e d brass is m o s t c o r r o s i o n - r e s i s t a n t o f the brasses
a nd resists d ezi n c i f ic a t io n a n d u n c o n t a m i n a t e d fresh waters. T h e b r o n z e s
are also s t r o n g l y resistant to fresh w a t e r attack. C o p p e r - t i n a n d c o p p e r -
nickel alloys are very resistant to clean seawater. In c o n t a m i n a t e d s e a w a t e r
the c u p r o - n i c k e I alloys, a l u m i n u m bronzes, a n d straight b r o n z e s m a y be
used. T h e 70-30 c o p p e r - n i c k e l alloys resist s t r e s s - c o r r o s i o n c r a c k i n g a n d
i m p i n g e m e n t attack.
Lead
L e a d is c o r r o s i o n resistant in h a r d a n d soft waters, m o s t a t m o s p h e r e s a n d
m a n y chemicals. T h e p r o t e c t i v e c o a t i n g s f o r m e d on lead are i n er t in m a n y
University of Washington (University of Washington) pursuant to License Agreement. No furt
TABLE 4--Designations for wrought copper alloys.
Copper or Typical
Copper Alloy Alloy
Number Number Previous Trade Name General Classification
100to 150 Copper (99.3 7o) Coppers

160 to 200 170, 172 Beryllium Copper
(Cu 99.5 ~o-Be 1.7) High Copper Alloys
175 Beryllium Copper
(Cu 99.5 ~o-Co 2.5)
200to 300 210 Gilding, 95 ~o
220 Commercial Bronze, 90 7o Copper-Zinc Alloys
230 Red Brass, 85~o (Brasses)
240 Low Brass, 80 ~o
260 Cartridge Brass, 70 ~o
280 Muntz Metal
300 to 400 342 High-Leaded Brass Copper-Zinc-Lead Alloys
(Leaded Brasses)
400 to 500 465 Naval Brass, Arsenical Copper-Zinc-Tin Alloys
(Tin Brasses)
500to 530 510 Phosphor Bronze, 5 ~o (A) Copper-Tin Alloys (Phosphor
Bronzes)
530 to 645 614 Aluminum Bronze Copper-Tin-Lead Alloys
(Leaded Phosphor Bronzes)
645 to 665 655 High Silicon Bronze (A) Copper-Silicon Alloys
(Silicon Bronzes)
665 to 700 675 Manganese Bronze (A) Miscellaneous Copper-Zinc
Alloys
700to 735 Copper Nickel (All Grades) Copper-Nickel Alloys
735 to 800 745,754, 770 Nickel Silver, 65-18 Copper-Nickel-Zinc Alloys
(Nickel Silvers)
environments. Alkalies attack lead as do soils having organic acids f r o m

wood. Lead-tin coatings on steel are effective if bare points can be avoided.
Table 5 lists some of the c o m m o n commercial lead and lead alloys and
their uses.
TABLE 5--Designations for lead or lead alloys.
Alloy Composition Typical Uses
Chemical Lead 99.90 ~o min. Chemical Industry

Corroding Lead 99.73 ~o Batteries, paint, cable
Calcium Lead 0.028 ~o calcium Cable and pipe
Antimony Lead 1-9 ~o antimony Cable and batteries
Soft Solders 5-50 ~o tin Solders
Lead-base Babbitt Antimony (10-15 ~o)-tin (5-10 7o) Bearings
Type Metal Antimony (4-16 ~o)-tin (4-8 ~) Type
Magnesium
Magnesium is one of the very active metals. In most atmospheres, an
oxide film formed on the surface protects the magnesium. This film tends to
break down in salt environments [6].
Magnesium alloys are subject to general corrosion attack, pitting, stress
corrosion, corrosion fatigue, galvanic corrosion, and intergranular attack.
Pure magnesium has considerably better corrosion resistance than its
alloys, but the alloys are stronger.
Atmospheric corrosion rates are determined by the alloying elements and
the type of environment. High humidity intensifies the corrosion. All
magnesium alloys corrode in seawater and, to a lesser extent, in fresh
water. Table 6 gives typical magnesium alloys.
TABLE 6--Typical magnesium alloys.
Designation Forms Available Main Alloying Elements
AM100A Sand and mold castings A1-1070, Mn-0.270 min.

AZ92A Sand and mold castings A1-970, Zn-2 70, Mn-0.2 70 min.
M1A Extrusions Mn-1.2 70 rnin.
AZ31B Castings, sheet, extrusions AI-3 70, Zn-1 70
HK31A Sand castings and sheet Th-3.3 70, Zr-0.7 70
ZE10 Sheet and plate Zn-1.3 70, rare earths---0.17 70
Nickel
Nickel and its alloys are very resistant to corrosion and oxidizing agents
are necessary for corrosion to take place. Monel has excellent resistance to
minerals, acids, and salts. Inconel is usually more resistant to atmospheres
than Monel or nickel. Nickel is a c o m m o n plating material for steels in
atmospheric exposures. Nickel resists corrosion by all fresh waters but pits
under barnacles in seawater. High nickel alloys m a y be subject to inter-
granular attack [6]. C o m m o n nickel alloys are listed in Table 7.
TABLE 7--Typical nickel alloys.
Designation Common Name Main Constituents
Nickel 200 Commercial Nickel 99.50 Ni-0.25 Mn

"Duranickel" 301 Age Hardenable 99.50 Ni-0.20 Mn
Nickel 211 Manganese-Nickel 95.00 Ni-4.75 Mn
Nickel 213 Cast Nickel 95.00 Ni-1.6 Si-l.5 C
Hastelloy D 82 Ni-9 Si-3 Cu
Monel 400 66 Ni-31.5 Cu-1.4 Fe
Inconel 600 76 Ni-16 Cr-7,2 Fe
HasteUoy F 47 Ni-22 Cr-17 Fe-6 Mo
Ilium 98 55 Ni-28 Cr-8.50 Mo-5.5 Cu
Ni-O-Nel 825 42 Ni-22 Cr-30 Fe-3 Mo
Titanium
Titanium and its alloys exhibit excellent corrosion resistance in many
atmospheres and waters. Generally, titanium is resistant to stress corrosion,
erosion-corrosion, galvanic corrosion, and oxidation in many environ-
ments, including seawater [6].
This corrosion resistance is a result of the stable oxide film normally
present on the titanium surface. The diffusion of the oxygen into the metal
at high temperatures may cause embrittlement and corrosion of the metal.
Titanium stress corrodes in a limited number of environments such as
nitrogen tetroxide, hydrobromic acid, red fuming nitric acid, and chloride
salts at high temperatures (260 C and above). Table 8 shows some common
titanium alloys.
T A B L E 8--Titanium and its alloys.
Major Alloying Elements, 7o
Designation Fe a Pdb Alb Snb Vb
Ti 35 A 0.12 . . . . . . . . . . . . .
Ti 50 A 0.20 . . . . . . . . . . . . .
Ti 65 A 0.25 . . . . . . . . . . . . .
Ti 75 A 0.30 . . . . . . . . .
Ti-0.20 Pd 0.25 01 i5 ...
Ti-5 A1-2.5 Sn 0.50 .... ,~-6 2-3
Ti-6 A1-4V 0.25 .... 5.76-6.75 ... 3.5-4.5
a Maximum.
b Nominal.
Z/no
Zinc is used in corrosion engineering largely as a structural material or as
a coating material on steel or other metals. Here the protection is primarily
galvanic in nature, with the zinc being anodic to iron, nickel, lead, tin, and
copper. With aluminum, zinc can be either cathodic or anodic. In the case
of magnesium, zinc is cathodic. Zinc is also commonly used as an anode
material for cathodic protection of ships, pipelines, structures and so on.
The corrosion of zinc usually is uniform in nature--deep pitting is rare.
The rate of attack on zinc in natural waters is increased by the presence of
oxygen, carbon dioxide, aeration, high temperatures, and agitation.
Between pH values of 7 and 12.5 corrosion of zinc is relatively low. Soft
water is more corrosive than hard water and a lack of oxygen can initiate
pitting. Some commercial zinc alloys are listed in Table 9 [6].
Beryllium
Beryllium has good resistance to water, particularly when the water is
aerated and has a velocity of 5 to 8 fps. At temperatures of 260 C (500 F) or
TABLE 9--Common zinc alloys.
Impurity Content, 7o max.
ASTM Grade Pb Fe Cd Total
Special High Grade 0.006 0.005 0.004 0.010

High Grade 0.07 0.02 0.07 0.10
Intermediate 0.20 0.03 0.50 0.50
Brass Special 0.60 0.03 0.50 1.0
Selected 0.80 0.04 0.75 1.25
Prime Western 1.60 0.08 . . . . . . . .
more and a water velocity of 27 fps, intergranular attack may develop.

Pitting occurs in the presence of chloride ions [6].
Refractory Metals
Such metals as tungsten, tantalum, and molybdenum are used because of
high structural strength at high temperatures. These metals are resistant to
corrosion in many media at low temperatures and do not oxidize until a
temperature of 300 C (572 F) or more is reached.
Discussion of Common Corrosion Tests

Laboratory Tests
Cabinet Tests
1. Neutral Salt Spray (ASTM B 117). Specimens are placed in a sealed
cabinet having a 5 percent salt fog at 35 C. Now used for evaluation of
painted, plated, and anodized parts.
2. Acidified Salt Spray (ASTM B 287). The pH of 3.2 is obtained by
acidifying the fog solution of B 117 with acetic acid. For evaluation of
organic and inorganic coatings, plated metal and anodized specimens.
Also used for exfoliation testing of aluminum.
3. CASS Test (ASTM B 368). A 5-percent salt fog maintained at 49 C with
a pH of 3.2 and made more aggressive by the addition of copper to the
test solution. Primarily used for the testing of decorative copper-nickel-
chromium or nickel-chromium coatings on zinc-base die castings,
anodized aluminum, and painted metals.
4. Corrodkote Test (ASTM B 380). A slurry containing corrosive salts is
applied to test specimens and allowed to dry. Specimens are then
placed in a high humidity cabinet at 38 C. Used for decorative chromium
plating and stainless steels.
5. Kesternick Test (German Standard DIN50018). Hot moist SO2, followed
by ambient air, is circulated at controlled temperature and humidity.
Used for plated coatings and stainless steel fittings.
6. Humidity Tests. Controlled high humidity (70 to 100 percent) at tem-

peratures from ambient to 49 C. Used for organic and metallic coatings.
7. S W A A T Test (Reynolds Metals Company STP AC7). Cyclic acidified
3.5 percent salt spray. Used for exfoliation testing of aluminum.
8. Cargill Test. A test simulating road conditions and involving inter-
mittent melting and cycling temperatures after application of a slurry
on specimens.
Beaker Tests
1. Kape Test. An immersion test having a sodium sulfate-sulfuric acid
solution at 95 C. Used for testing the sealing of anodized aluminum.
2. Dip and Dry Test. Alternate immersion of specimens in a corrosive
solution. Solution may be 3 percent sodium chloride or salt solutions
having sulfur compounds. Used for stainless steel materials.
3. Ferric Chloride Spot Test. A drop of FeCla-NaCI-HC1 solution is placed
on the cleaned metal surface. Used for stainless steels to indicate
chromium depletion.
Electrochemical Tests
1. FACT Test (Ford Anodized Aluminum Corrosion Test). A technique
where a d-c current passed through an acid salt solution on an anodized
aluminum surface. The integral of the time-voltage breakdown curve is
the FACT number.
2. AZTAC Test (Alcoa Impedance Test for Anodic Coatings). This test
measures the a-c impedance of an anodic aluminum coat at a spot
wetted with chloride solution.
3. EC Test (General Motors Electrolytic Corrosion Test). A potentiostatic
test for chromium-nickel plating systems. The specimen is held at a
potential of +0.3 V versus a saturated calomel electrode for 1 min,
followed by free corrosion for 2 min.
4. Electrographic Printing Test. A paper sensitized with indicators for a
certain metal is pressed on the specimen surface (anode) and a potential
1.5 to 6.0 V is applied. Colored spots on the paper indicate the presence
of discontinuities in the coating.
5. Polarization Resistance Test (Linear polarization technique). Used for
many systems, including pitting sensitivity of stainless steels.
Fteld Tests
Field tests are those tests carried out in natural environments of the types
in which specimens are likely to be exposed. Field tests may be made in air,
water, or soil [12].
Atmospheric specimens are most often of the panel type (for example,
4 by 6 in., 4 by 8 in., or 4 by 12 in.) and are supported at the edges by
porcelain or plastic knobs. The usual exposure rack faces south and panels
are exposed at 30 or 45 deg f r o m the horizontal. Test racks should not be

sheltered but should have a bold exposure free o f local unusual air currents
or corrosive effects. Since the time of initial exposure establishes the first
surface film, it is i m p o r t a n t to m a k e exposures on an annual basis at
approximately the same time. Even in waters this can be i m p o r t a n t due to
temperature effects.
Seawater exposures m a y be tidal (at the tidal zone), full immersion, or
partially buried in the ocean bottom. Soil tests require extra care due to
variations in soils and their air and water contents.
Service Tests
Service tests are tests involving actual components or assemblies in the
field rather than panels. Presumably a better judgement as to the corrosion
resistance of the test specimen is obtained in this fashion. However, m o r e
care m a y be required in planning and executing these tests than is needed
for other methods [18].
APPENDIX
Corrosion Tests
Laboratory Corrosion Tests
B 117 Salt Spray (Fog) Testing
B 136 Test for Resistance of Anodically Coated Aluminum to Staining by
Dyes
B 137 Test for Weight of Coating on Anodically Coated Aluminum
B 244 Measuring Thickness of Anodic Coatings on Aluminum with Eddy
Current Instruments
B 287 Acetic Acid Salt Spray (Fog) Testing
B 356 Specification for Zirconium and Zirconium Alloy Forgings and Ex-
trusions for Nuclear Applications
B 368 Copper-Accelerated Acetic Acid Salt Spray (Fog) Testing (CASS Test)
B 380 Corrosion Testing of Decorative Chromium Plating by the Corrodkote
Procedure
B 449 Recommended Practice for Chromate Treatments on Aluminum
C 464 Corrosion Effects of Thermal Insulating Cement on Base Metal
C 486 Test for Spalling Resistance of Porcelain Enameled Aluminum
D 69 Specification for Friction Tape for General Use for Electrical Purposes
D 130 Test for Detection of Copper Corrosion from Petroleum Products by the
Copper Strip Tarnish Test
D 235 Specification for Petroleum Spirits (Mineral Spirits)
D 801 Dipentene, Sampling and Testing
D 807 Corrosivity Test of Industrial Water (USBM Embrittlement Detector
Method)
D 849 Test for Copper Corrosion of Industrial Aromatic Hydrocarbons
D 930 Total Immersion Corrosion Test of Water-Soluble Aluminum Cleaners
D 1141 Specification for Substitute Ocean Water (not a corrosion test but con-
tains directions for preparing a corrodent)
D 1261 Test for Effect of Grease on Copper

D 1275 Test for Corrosive Sulfur in Electrical Insulating Oils
D 1280 Total Immersion Corrosion Test for Soak Tank Metal Cleaners
D 1374 Aerated Total Immersion Test tbr Metal Cleaners
D 1384 Corrosion Test for Engine Antifreezes in Glassware
D 1567 Testing Detergent Cleaners for Evaluation of Corrosive Effects on
Certain Porcelain Enamels
D 1611 Test for Corrosion P, oduced by Leather in Contact with Metal
D 1616 Test for Copper Corrosion by Mineral Spirits (Copper Strip Test)
D 1654 Evaluation of Painted or Coated Specimens Subjected to Corrosive
Environments
D 1735 Water Fog Testing of Organic Coatings
D 1743 Test for Rust Preventive Properties of Lubricating Greases
D 1748 Test for Rust Protection by Metal Preservatives in the Humidity Cabinet
D 1838 Test for Copper Strip Corrosion by Liquefied Petroleum (LP) Gases
D 1838 Test for Copper Strip Corrosion by Liquefied Petroleum (LP) Gases
D 2043 Test for Silver Tarnishing by Paper
D 2251 Test for Metal Corrosion by Halogenated Organic Solvents and Their
Admixtures
D 2570 Simulated Service Corrosion Testing by Engine Antifreezes
F 64 Test for Corrosive and Adhesive Effects of Gasket Materials on Metal
Surfaces
G 1 Recommended Practice for Preparing, Cleaning and Evaluating Cor-
rosion Test Specimens
G 2 Recommended Practice for Aqueous Corrosion Testing of Samples of
Zirconium and Zirconium Alloys
G 4 Recommended Practice for Conducting Plant Corrosion Tests
G 28 Detecting Susceptibility to Intergranular Attack in Wrought Nickel-
Rich, Chromium-Bearing Alloys
Recommended Practice for Laboratory Immersion Corrosion Testing of Metals
Recommended Practice for Recording Data from Atmospheric Corrosion Tests of
Metallic-coated Steel Specimens
Recommended Practice for the 3.5 percent Sodium Chloride Solution Alternate
Immersion Stress-Corrosion Test (1)
Method of Test for Exfoliation Corrosion Susceptibility in 7000 Series Copper
Containing Aluminum Alloys
Federal Test Methods
Method 812 Synthetic Sea Water Spray Test
Method 822 Intergranular Corrosion Test for Aluminum Alloys
MIL-STD-171 Finishing of Metal and Wood Surfaces (for aluminum)
MIL-STD-186 Protective Finishing Systems for Rockets, Guided Missiles,
Support Equipment and Related Materials
MIL-STD-193 Painting Procedures, Tactical Vehicles (Tracked and Wheeled)
MIL-STD-194 Painting and Finishing Systems for Fire Control Instruments
MIL-STD-276 Impregnation of Porous Nonferrous Metal Castings
MIL-T-152 Treatment, Moisture- and Fungus-Resistant, of Communica-
tion, Electronic, and Associated Electrical Equipment
MIL-V-173 Varnish, Moisture- and Fungus-Resistant (for the treatment of
Communications. Electronic, and Associated Electrical Equip-
ment)
MIL-F-495 Finish, Chemical, Black, for Copper Alloys
MIL-M-3171 Magnesium Alloy; Processes for Corrosion, Protection of

MIL-C-5541 Chemical Films and Chemical Film Materials for Aluminum
and Aluminum Alloys
MIL-S-7124 Sealing Compound, Pressure Cabin
MIL-F-7179 Finishes and Coatings; General Specifications for Protection of
Aircraft and Aircraft Parts
MIL-P-8116 Putty, Zinc Chromate, General Purpose
MIL-P-8585 Zinc Chromate Primer (for aluminum)
MIL-A-8625 Anodic Coatings for Aluminum and Aluminum Alloys
MIL-C-8837 Coating Cadmium (Vacuum Deposited)
MIL-T-10727 Tin Plating; Electrodeposited or Hot Dipped for Ferrous and
Nonferrous Metals
MIL-S-11031 Sealing Compound, Noncuring, Polysulfide Base
MIL-C-11796 Corrosion Preventive, Petrolatum, Hot Application
MIL-L-13762 Lead Alloy Coating, Hot Dip (for Iron and Steel Parts)
MIL-L-13808 Lead Plating (Electrodeposited)
MIL-I-13857 Impregnation of Metal Castings
MIL-F-13924 Coating, Oxide, Black, for Ferrous Metals
MIL-F-14072 Finishes, for Ground Signal Equipment
MIL-P-14458 Paint, Rubber, Red Fuming Nitric Acid Resistant
MIL-P-14538 Plating, Black Chromium (Electrodeposited)
MIL-C-14550 Copper Plating (Electrodeposited)
MIL-Z-17871 Zinc Coating (Hot Dip Galvanizing)
MIL-P-23408 Plating Tin-Cadmium Electrodeposited
MIL-C-26074 Coating, Nickel-Phosphorus, Electroless Nickel, Requirements
for
MIL-A-40147 Aluminum Coating (Hot Dip) for Ferrous Parts
MIL-M-45202 Magnesium Alloys, Anodic Treatment of
MIL-G-45204 Gold Plating (Electrodeposited)
MIL-P-45209 Palladium Plating (Electrodeposited)
MIL-A-46063 Aluminum Alloy Heat-treatable Armor Plate
MIL-C-46079 Coating, Epoxy, Baking Type for Magnesium Castings
MIL-M-46080 Magnesium Castings, Process for Anodic Cleaning and Surface
Sealing of Chromium
MIL-A-46118A Aluminum Alloy Armor Plate and Forgings, 2219
MIL-C-60536 Hard Coat Anodize (for aluminum)
MIL-C-60539 Anodic Coatings for Aluminum
QQ-C-32o Chromium Plating (Electrodeposited)
QQ-P-416 Plating, Cadmium (Electrodeposited)
QQ-N-290 Nickel Plating (Electrodeposited)
QQ-P-35 Passivation Treatments for Austenitic, Ferritic, and Martensitic
Corrosion-Resisting Steel (Fastening Devices)
QQ-S-365 Silver Plating, Electrodeposited: General Requirements for
QQ-Z-325 Zinc Coating, Electrodeposited, Requirements tor
TT-C-520 Coating Underbody (for Motor Vehicles)
AMS 2468 Hard Coating Treatment of Aluminum Alloys
National Association of Corrosion Engineers (NACE)
TM-01-69 Laboratory Corrosion Testing of Metals for the Process Industries
Corrosion Tests Not Issued by Any Standards Organization
FACT Test (Ford Anodized Aluminum Corrosion Test) Electrochemical Test in
Acid Salt Solution for Evaluating Anodizing
SWAAT Test--Standard Method for Exf01iation Testing of Aluminum Alloys--
Reynolds Metals M R D - - S T P AC 7; Acid Seawater Spray 30 min; 100,~ RH,

90 min
Ferric Sulfate--Sulfuric Acid Test for Ni-Cr-Mo Alloys. Quality Control and
Acceptance Test for Satisfactory Heat Treatment
Kesternich Test (German Standard D I N 50018) Hot Moist SOs followed by
Ambient Conditions for Testing Metallic Protective Coatings
Kape Test--Immersion in Acidified Sodium Sulfite for Testing the Sealing of
Anodized Aluminum
References
[1] Ailor, W. H., Ed., Handbook on Corrosion Testing and Evaluation, Wiley, New York,
1971.
[2] Metal Finishes Manual, 2nd ed., National Association of Architects Metal Manu-
facturers, Chicago, 1969.
[3] Metals Handbook, Vol 2, Heat Treating, Cleaning and Finishing, American Society
for Metals, Metals Park, Ohio, 1964
[4] Payne, H. F., Organic Coating Technology, Wiley, New York, 1961.
[5] "Weather Resistance of Porcelain Enamels; Effect of Exposure Site and other Variables
after Seven Years," Building Science Series No. 4, National Bureau of Standards,
Washington, D.C.
[6] "Corrosion and Corrosion Protection of Metals," MIL-HDBK-721(MR), Watertown,
Arsenal, 1965.
[7] Stress Corrosion Testing, A S T M STP 425, American Society for Testing and Material;,
1967.
[8] Metal Corrosion in the Atmosphere, ASTM STP 435, American Society for Testing and
Materials, 1968.
[9] van Horn, K. R., Ed., Aluminum, Fabrication and Finishing, Vol. 3, ASM, 1967.
[10] Standards Handbook--Copper, Brass, Bronze, Data-Specifications, Part 7, 1970.
[11] Stainless Steel Handbook, Allegheny Ludlum Steel Corp., 1956.
[12] "A Study of the Correlation of Laboratory and Field Tests for Automotive Corrosion
Testing," NACE Pub. 3N170, Materials Protection and Performance, Aug. 1970.
[13] Black, H. L. and Lherbier, L. W. in Metal Corrosion in the Atmosphere, ASTM STP
435, American Society for Testing and Materials, 1968.
[14] Bush, G. F., Metal Progress, Vol. 92, Oct. 1965, p. 152.
[15] Wirshing, R. J., "Effect of Deicing Salts on the Corrosion of Automobiles," presented
at the annual meeting of the Highway Research Board, 19 Jan. 1956, Washington, D.C.
[16] Ireland, D. T., 1965-66 Minneapolis-Milwaukee Carguard Field Test, Cargill, lnc,.
26 May 1966.
[17] Saur, R. L. and Basco, R. P., Plating, Vol. 35, 1966, p. 53.
[18] LaQue, F. L. in Relation of Testing and Service Performance, ASTM STP 423, American
Society for Testing and Materials, 1967, p. 61.
STP534-EB/Nov. 1973
Chapter 11

Iron and Steel Industry
H. P. Leckie~
Corrosion testing in the steel producing industry is utilized to (a) provide

a basis for internal quality control standards, (b) establish quality and
performance parameters with respect to the final product utilization and
(c) serve as a comparison standard in the development of new ferrous base
products having improved corrosion resistance. The major markets served
by the steel industry are automotive, construction, containers, rails, elec-
trical, appliance and agriculture and both accelerated and field tests on steel
products reflect the specific requirements in corrosion resistance properties
for these industries. Any discussion of corrosion testing of steels must
necessarily include coatings (both metallic and nonmetallic) since the
shipped tonnages of coated steel products are large and continue to increase
on a percentage basis year by year. Furthermore, although stainless and
heat resisting steels represent less than one percent [1] 2 of steel shipments
in the United States, the more critical applications to which these materials
are subjected, together with the specific forms of corrosion related failures
experienced, warrant their inclusion in any description of the use of corro-
sion standards in the steel industry.
The widespread applications for steels and steel products are so diversi-
fied in scope as to require testing to determine resistance or susceptibility
to generalized corrosion, localized (pitting) corrosion, stress corrosion
cracking, hydrogen stress cracking, oxidation, galvanic corrosion and
corrosion fatigue. On a more restricted basis there may also be the need to
test for such forms of corrosion attack as graphitic, cavitation and fretting
corrosion. For metallic coatings the corrosion resistance offered to the steel
substrate depends primarily on thickness, and corrosion resistance is, there-
fore, often related indirectly to such factors as adhesion, ductility, hardness
and porosity. Similar indirect corrosion testing is also applied to non-
metallic coatings. However, the situation with organic coatings differs from
metallic corrosion testing in two major respects: (1) organic coatings are to
1 Inland Steel Research Laboratories, East Chicago, Indiana.
209
a greater or lesser degree permeable to moisture and (2) organic coatings

are susceptible to degradation due to the action of ultraviolet light. As a
result of either the separate or combined action of moisture and ultraviolet
light, the physical properties of organic coatings may be changed in such a
way (chalking, crazing, blistering, peeling, etc.) as to detract significantly
from the corrosion resistance characteristics of the initially applied film.
Corrosion testing procedures for steels and steel products may be gen-
erally classified according to the three general categories of service, field and
accelerated tests. Service tests refer to those conditions where a material is
evaluated as to its performance as a specific end product in a specific in situ
application. Corrosion service tests may involve the material under test
experiencing the corrosive action of only one relatively uniform environ-
ment (for example, a domestic water tank) or varying environmental inter-
actions (automobile body component). Field tests, on the other hand, are
not normally conducted on finished component parts, but rather serve to
gain general corrosion resistance information in the atmosphere, under-
ground or under immersed conditions. Where field tests are conducted
under immersed conditions, the environment must be and is always defined
(for example, flowing sea water). For atmospheric corrosion field tests the
necessity for describing local atmosphere conditions, while not as critical as
the need for describing immersed conditions, nevertheless is most desirable.
The atmospheric corrosion behavior of steels varies considerably with
environment. Table 1, for example, shows the relative corrosivity of several
atmospheric test site locations for mild steel.
TABLE 1--Relative corrosivity o f atmospheric corrosion test sites. ~
Site 1 Year 2 Years 4 Years 8 Years
State College, Pa. (Rural) 1.0 1.0 1.0 1.0

South Bend, Pa. (Semi-Rural) 1.5 1.5 1.6 1.7
Kure Beach, N.C. (Marine) 2.0 2.5 3.5 5.8
Kearney, N.J. (Industrial) 3.3 2.7 2.5 2.6
- State College, Pa. as unity.
Underground (soil) test conditions are specified with the least frequency,
although the variations in soil characteristics with respect to pH, electrical
resistivity, bacterial activity and composition are sufficient to cause signifi-
cant variations (up to orders of magnitude) in corrosion rate for the same
material. For this reason, the location, together with a general description
of soil type, should always be specified in describing underground field
corrosion test data.
In view of the great diversity of natural environmental conditions, the
extrapolation of accelerated laboratory test data to predict corrosion per-
CORROSION IN THE IRON AND STEEL INDUSTRY 21 1
formance of a given material in service must be approached with great

caution. Accelerated tests, by definition, are designed to represent condi-
tions more severe than those normally encountered in service.
By using a control material of known corrosion resistance performance
under real conditions, it is possible through judicious choice of the acceler-
ated test procedure to gain comparative corrosion rate information on a
relative basis. However, rather than specifying materials for service appli-
cations based on accelerated test data, accelerated tests are more commonly
used in the steel industry either as quality control procedures or as means
for evaluating the relative corrosion resistance of materials in the area of
new product development.
To the best of the writer's knowledge, one of the first published proce-
dures for accelerated testing designed to simulate atmospheric corrosion
behavior was described by Capp [2] in 1914. Most of the now widely
accepted humidity and salt spray test methods derive from variations and
improvements on this early work. The importance of such variables as
relative humidity, temperature, and contaminants on the atmospheric
corrosion behavior of both ferrous and nonferrous metals was first de-
scribed in detail by Vernon [3-6]. An early humidity cabinet designed to
provide continuous condensation on test specimens was developed and
described by Darsey [7]. Various procedures were developed [8-11] which
allowed test specimens within a cabinet to be physically rotated through
various locations having different humidity-temperature conditions, and in
this manner simulate more closely real atmospheric exposure conditions.
Other variations [12-14] alternately changed conditions within the test
cabinet so as to permit condensation to occur during specified portions of
the test cycle. The use of sprays and fogs particularly from salt solutions was
first described in 1937 [15] and modification and upgrading continues to the
present day. A tentative procedure for salt fog testing was first published by
the American Society for Testing and Materials (ASTM) in 1939 followed
by revisions in 1941, 1944, 1949, 1954, 1957, 1961, and 1964. The most
recent revision is designated ASTM B 117-64. Variations on the salt fog
test procedure involving pH control using acetic acid and the additic a of
copper chloride to increase aggressiveness of the fog, carry the designations
ASTM B 287-62 and ASTM B 368-68, respectively. The deficiencies in
attempting to correlate accelerated corrosion tests with atmospheric corro-
sion behavior for steels and coated steels is well recognized and were briefly
described earlier. Some of the deficiencies in accelerated test procedures
have been detailed by Schlossberg [16]. A literature survey by Kuensler and
Shur [17] covering the period 1956 to 1966 covers the accelerated testing of
organic coatings with respect to weathering resistance.
It is now generally accepted that the single factor contributing most to
variations in atmospheric corrosion behavior is the concentration of sulfur
dioxide [18]. Testing in SOs atmospheres, although reported much earlier,
2]2 INDUSTRIALCORROSION STANDARDS AND CONTROL
did not gain wide acceptance until the apparatus developed by Kesternich
[19] was described in 1951. This equipment combined exposure to SO2
together with continuous condensation on the surface of the test specimens.
More recent work [20-30] attests to the use of continuous condensing sys-
tems containing SO2 as providing reasonably close correlation between an
accelerated corrosion test procedure and atmospheric corrosion behavior
in an industrial environment.
For convenienee in presentation, specifics of eorrosion and corrosion
testing in the iron and steel industry will be discussed under various
materials categories.
Plain Carbon and Low-Alloy Steels

The corrosion of carbon and low-alloy steels is primarily governed in
most cases by the combined action of water and oxygen. Under most condi-
tions of natural corrosion, the rate is controlled by the cathodic reaction
(normally oxygen reduction) which explains the relatively minor effect of
small alloying additions on the corrosion rate of iron. In simple terms, the
corrosion of iron may be characterized by the following chemical reactions:
2 Fe + 2 H20 + O2---+ 2 Fe (OH)2 (1)
4 Fe (OH)= + 2 H20 + 02 ~ 4 Fe (OH)a (2)
In practice rust films on iron and steel generally comprise several layers,
representing iron oxides in various states of oxidation. A recent publication
by Evans and Taylor [31] details the chemistry of the atmospheric rusting
of iron.
Under immersed conditions the corrosion of iron may or may not be pri-
marily controlled by the cathodic reduction of oxygen. In acid solutions the
reduction of hydrogen ions at cathodic sites becomes the rate controlling
process. In actual practice a combination of both oxygen and hydrogen ion
reduction takes place. As a guideline, it may be considered that over the pH
range 4 to 10 the corrosion rate of iron is essentially unaffected by pH and
is controlled by the diffusion rate of oxygen to the metal surface. In this
region the corrosion product film is ferrous hydroxide (hydrated ferrous
oxide) producing an effective pH at the iron-ferrous hydroxide interface of
approximately 9.5. At pH < 4 ferrous hydroxide is unstable and the cor-
rosion rate rises rapidly with decreasing pH. At the lower end of the pH
scale, surface coverage by hydrogen approaches unity and the corrosion
rate is controlled by the rate of hydrogen evolution. In highly alkaline solu-
tions (pH > 10) iron is passive and the corrosion rate is correspondingly
reduced. Concentrated alkaline solutions (pH > 14) cause iron to dissolve
as the ferrite (FeO2-) anion, although the kinetics of reaction are so slow as
to produce only a very minor increase in corrosion rate over passive iron.
Salt tends to increase corrosion rate due to a combination of increased
CORROSION IN THE IRON AND STEEL INDUSTRY 213
electrolytic conductivity of the solution and the formation of nonprotective

corrosion products at some location removed from the iron surface [32].
Certain oxidizing ions readily reducible at cathodic sites (ferric, cupric,
mercuric) may also stimulate the overall corrosion rate significantly, due to
their depolarizing action on the cathodic reaction kinetics.
Under conditions of oxygen reduction, as the controlling cathodic
reaction, the effects of coldwork and small amounts of alloying additions
on corrosion rate are minimal. In acid solutions, however, a small but
apparently real effect for coldwork in increasing the corrosion rate of iron
in acid solutions has been reported [33] due to the formation of finely dis-
persed low overvoltage areas of nitride and carbide precipitates. Similarly,
heat treatment to the extent that it may vary the volume fraction of low over-
voltage i r o n carbide may also affect the corrosion rate of iron in acid
solutions.
Although it has previously been stated that the corrosion rate of iron
under conditions of oxygen reduction as the primary cathodic reaction is
relatively unaffected by additions of small amounts of alloying elements,
the composition and protectiveness of the corrosion product film is indeed
affected by steel composition, particularly under atmospheric but also to a
lesser extent under immersed exposure conditions. The ability of copper to
retard corrosion in iron and steel for instance was first reported in 1900 [34].
Howe [35] showed as early as 1901 that a 3-percent nickel steel corrodes at a
substantially lower rate in the atmosphere than unalloyed steels. The effect
of alloying additions on the corrosion resistance of iron and steel has been
studied extensively since that time [36-53]. (The list of references given here
is intended to be representative and is by no means exhaustive.) The earliest
extensive atmospheric exposure tests of various ferrous materials were
started by ASTM in 1916--the final report on this investigation being pub-
lished in 1953 [48]. This study showed that both copper and phosphorus
contributed significantly to enhance corrosion resistance. Greenidge and
Lorig [42] published the results of a three-year atmospheric corrosion
program test on 43 steels which indicated that copper at levels between 0.2
and 0.5 percent markedly improved the corrosion resistance of low-carbon
and low-alloy steels. Higher levels of copper did not appear to increase the
resistance substantially. They concluded that 0.4 percent nickel added to
copper steels is not very effective, but Pilling and Wesley [43] showed that a
nickel addition of 2 percent results in a marked improvement. Sims and
Boulger reported in 1944 [45] that the addition of 2 percent nickel was less
effective than much smaller additions of phosphorus. In 1948, Pilling and
Wesley [46] reported that phosphorus and silicon additions to copper-nickel
steels were beneficial, but that carbon and manganese were unimportant in
an industrial atmosphere. In 1952, Copson [47] reported the results of
9-year exposure of 71 low alloy steels at Bayonne, New Jersey, and Block
Island, Rhode Island. This data confirmed the beneficial effect of fractional
percentages of copper, of small amounts of phosphorus, and of 1 percent or

more nickel. Chromium and copper additions to complex steels were seen
to be more helpful in the industrial atmosphere at Bayonne, whereas
manganese appeared to be more helpful in the marine atmosphere at
Block Island. Hudson and Stanners in 1955 [49] reported that the most
useful alloying elements for the purpose of enhancing atmospheric corro-
sion resistance are chromium, copper, and nickel. They indicated that
aluminum and beryllium might also be of value. Larrabee and Coburn in
1961 [51] reported corrosion data on 270 steels with systematic variations
of copper, nickel, chromium, silicon, and phosphorus after exposure for
15.5 years in an industrial, a semirural, and a marine atmosphere. This
work indicated that although an improvement in corrosion resistance can
be obtained by relatively small additions of these elements singly, the great-
est improvement derives from interactions between specific combinations
of these alloying additives. Wiester and Ternes [52], who reviewed the
development of low alloy steels for atmospheric corrosion resistance, con-
cluded that copper, phosphorus, chromium, nickel, and molybdenum were
the additives sufficiently effective in this regard to warrant singling out.
In summary, over the years at least eleven elements have been reported
as contributing to corrosion resistance in low-alloy steels: aluminum,
antimony, beryllium, chromium, copper, manganese, molybdenum, nickel,
phosphorus, silicon, and titanium. A full factorial experimental design for
this many additives at only one concentration level would require the
preparation and processing of 211 or 2048 alloys, and this would allow for
no variation in preparation or processing. The establishement of nonlinear
effects would require the preparation of a significantly greater number of
experimental compositions. A method of avoiding the complete factorial
experiment in producing compositions for alloy development when the
number of variables to be considered is impractically large has been sug-
gested by Plackett and Burman [54]. This experimental design makes
possible the determination of main effects from data obtained in a minimum
number of experiments. For eleven variables, for example, main effects can
be determined from twelve experime/ats. A Plackett-Burman design can
thus be used as a screening procedure to identify the most promising
additives with respect to corrosion resistance which can subsequently be
examined for interactions in a conventional factorial design. Multiple
regression analysis can then be used to determine a mathematical relation-
ship between atmospheric corrosion rate and alloy additive concentrations
which may be used for predictive purposes.
The most commonly quoted elements for improving the atmospheric
corrosion resistance of low-alloy steels are copper, phosphorus, chromium
nickel, molybdenum and to a lesser extent silicon. From these additives was
developed a series of low-alloy steel compositions commonly known as the
"weathering" steels, having corrosion resistance up to six times that of
University of Washington (University of Washington) pursuant to License Agreement. No further repr
CORROSIONIN THEIRONAND STEELINDUSTRY 215
14
~ CARBON STEEL
"-'12
z
v 8
J ~ COPPER-BEARING
STEEL
;6
z
o4
u
~2 ~ . HIGH-STRENGTH
LOW-ALLOY
WEATHERING STEEL
I I I I
5 10 15 20
TIME, YEARS
F I G . 1--Atmospheric exposure weight gain versus time for weathering steel, miM steel and
copper-bearing steel.
plain carbon steel in many atmospheric test locations. Typical weight-gain

versus time curves for a weathering steel compared to mild steel and copper-
bearing steel are shown in Fig. 1.
Laboratory accelerated tests designed to predict atmospheric corrosion
performance for low-alloy steels have for the most part been ineffective.
Initial tests attempted to correlate atmospheric exposure rate with weight-
loss measurements obtained in 20 percent sulfuric acid [55]. It was recog-
nized early [56], however, that the atmospheric corrosion rate of low-alloy
steels was affected more by the protective qualities of the rust film than
inherent corrosion rate of the bare steel. While it has been observed that
certain accelerated tests involving exposure to sprayed solutions of various
kinds and to moist sulfur dioxide correlate qualitatively with the observed
corrosion rates of selected low-alloy steels in certain atmospheres, there has
been no great success in demonstrating the observed beneficial contribu-
tions of specific alloying elements using accelerated tests.
Pourbaix [57] has recently attempted to predict weathering character-
istics of low alloy steels from their potential/time behavior in an alternating
immersion dry test. It is claimed that those alloys showing a propensity for
the formation of protective "patinable" rust films exhibit a rapid rise in
potential (see Fig. 2) in the noble direction to a value approximately 200 mV
more noble than the "non-patinable" steels which exhibit a much reduced
rate of potential increase. It was also demonstrated in these experiments
that steel surface preparation was an important factor in the rate of forma-
tion of protective rust films. The potential/time data obtained by Pourbaix
would indicate that a protective patina forms more rapidly on pickled
rather than sand-blasted surfaces.
1 I I 1 I I I 1 I I I I 1
~200 PATINABLE
e9 ~ .e
STEELS
144
-r
X
>
+100
E
Z
/
0
NONPATINABLE
~ 0 STEELS
T4~/+/+.'~
I e"
-100 .
z
z
z
m,n
-200
<
<
z
=,- -300
0
0
I=-
u
144
-400
,-4
~M
-500
0 1 2 3 4 5 6 7 8 9 10 11 12 13
TIME, DAYS
FIG. 2--Potential time behaviorfor patinable and nonpatinable steels.
Matsushima and Ueno [58] showed that the protective rust films formed
on low-alloy weathering steels exhibited less "active corrosion sites" than
plain carbon steels exposed for a similar duration in the atmosphere. Active
corrosion sites were identified using an autoradiographic technique in
which samples of steels corroded for various lengths of time in the atmos-
phere were immersed in a sodium sulfate solution containing radioactive

SO4-- and subsequently placed on an X-ray film. These authors explained
the increased corrosion resistance of weathering steels on the reduced
tendency for their rust films to catalyze the conversion of SOs to SO3 and
to the more continuous nature of the rust coating which is reflected in an
increased anodic polarization.
In general, immersed electrochemical test procedures have been singu-
larly unsuccessful in predicting even relative corrosion rates of carbon and
low alloy steels in the atmosphere. This is again due to the fact that im-
mersed exposure conditions do not produce the same physical and chemical
properties of the rust films formed on the same steel compositions under
atmospheric exposure conditions. Electrochemical test procedures have,
on the other hand, shown reasonable correlation with observed corrosion
rates in specific environments under immersed conditions. Thus, Cohen
and Jelinek [59] obtained good correlation between the corrosion rate
measured directly by weight loss and indirectly from the linear polarization
method for the corrosion of mild steel in alkaline lithium bromide solutions.
Linear polarization resistance is obtained from the slope of the linear por-
tion of a polarization curve measured at a potential range close (usually
5 to 10 mV) to the corrosion potential. The method has been demonstrated
as applicable to the measurement of corrosion rate in a series of diverse real
environments [60-65]. Anomalous data may be obtained, however, where
more than one electrochemical oxidation reaction is proceeding simul-
taneously and, in particular, where such reactions do not involve metal
dissolution.
Metallic Coatings for Steel

Metallic coatings are applied to low-carbon steels primarily to effect an
improvement in corrosion resistance, oxidation resistance and to a lesser
extent for aesthetic purposes. Zinc and tin remain by far the most widely
used coating materials, the former being applied by hot dipping and finding
primary utilization as galvanized steel in the automotive industry. Produc-
tion of hot-dip galvanized sheet and strip steel in the United States ex-
ceeded five million tons during the year 1970 [1]. Uses for galvanized steel
other than for automotive applications include major appliances, con-
struction and drainage products. Although the greatest tonnage by far of
zinc-coated steel is produced by hot dipping, significant tonnages of steel
are electroplated with zinc, producing what is commonly referred to as
electrogalvanized steel. The corrosion resistances afforded by the two appli-
cation methods are essentially the same, with total corrosion resistance
being a direct function of the zinc coating weight. Hot-dip zinc coatings are
produced to several standard coating weights, with coating weight control
on continuous galvanizing facilities being achieved through gas impinge-
ment. ASTM Standard A 525 specifies the coating designation numbers
under which zinc coating weights on sheet steel are specified. The most
commonly produced material, G.90, has a total coating weight (both sides)
of 1.25 oz zinc/ft 2 with minimum check limit (triple spot test) of 0.90
oz/ft 2. In like manner, galvanized sheet steel having the designation G.60
corresponds to a minimum coating zinc weight of 0.60 oz/ft 2. In 1971, the
designation G.01 was introduced referring to extremely light hot-dip zinc
coatings close to those produced during continuous electrogalvanizing in
which no minimum coating weight is specified. Galvanized sheet steel is
produced in four basic forms, as follows: (1) regular (or full) spangle--
produced on continuous coating lines such that the surface exhibits the
well-known "flowery" dendritic spangle; (2) minimized spangle--elimina-
tion of the spangle by modifying the nucleating characteristics of the
coating during solidification; (3) iron-zinc alloy--a nonspangled matte
finish suitable for painting, produced by processing the galvanized steel at
sufficiently high temperatures to cause increased alloying; and (4) differ-
ential--galvanized steel having a specified zinc coating weight on one side
and a significantly lighter zinc coating weight on the other side of the
steel strip.
It was previously mentioned that the corrosion protection afforded by
zinc coatings is a direct function of coating weight and this is attested to
by the observed linearity in corrosion rate obtained during the atmospheric
exposure of galvanized steel. Figure 3 shows the corrosion rate expressed
as total weight loss in 4 by 6-inch test panels for commercial galvanized steel
exposed for 6 years at a semi-industrial atmospheric corrosion test site in
Porter County, Indiana. The data shown in Fig. 3 correspond to an annual
loss in thickness of approximately 0.045 mil and thus, for a standard
1.25 oz/ft 2 coating (G.90) having a specified minimum zinc coating thickness
of ~0.7 mil, the first sign of red rust due to complete removal of the zinc
coating might be expected in 15 to 16 years. This calculation, however,
assumes uniform removal without preferential localized attack and applies
to one environmental (semirural) test location only. A similar calculation
for a more industrial atmospheric test site shows a corrosion rate for gal-
vanized steel of 0.07 mil/year resulting in a predicted life for a G.90 coating
of only 10 years. It should be pointed out, however, that nonuniform attack
at coating defects will generally reduce the predicted life based on uniform
attack assumptions.
In spite of the fact that zinc is electrochemically more active than iron, a
comparison plot of atmospheric corrosion rates shows bare steel to corrode
at a much higher rate than a galvanized coating, indicating that the zinc
corrosion products offer a contribution to the retardation of corrosion rate
by slowing the dissolution kinetics. In contrast to the uniform corrosion
rate with time observed for zinc, bare low-carbon and alloy steels exhibit
a pronounced reduction in rate with time. However, even after years the
corrosion rate of bare steel exceeds that of the zinc coating by one order of
1.8 D
1.6-
1.4-
(3
1.2--
O
"~ 1 . 0 - -
~ o.s-
i,,,,-
0.6--
0.4--
0.2
br I I I I I I
0 1 2 3 4 5 6
TIME, YEARS
FIG. 3--Atmospheric exposure weight loss for galvanized steel exposed six years in
Porter County, Indiana.
magnitude. This effect is shown in Fig. 4 for galvanized steel and a copper-
bearing uncoated steel exposed at Porter County, Indiana for 6 and 8 years,
respectively, with exposure for both commencing in October 1963.
Since 1926, ASTM Committee A-5 has coordinated an atmospheric cor-
rosion test program on corrugated galvanized sheets at five different test
locations in the United States [66]. A second program undertaken by Sub-
committee XIV of ASTM Committee A-5 to evaluate the atmospheric
corrosion resistance of galvanized steel produced both by batch and con-
tinuous dipping techniques was initiated in 1960. Test samples were exposed
at State College, Pennsylvania (rural), Newark, New Jersey (industrial),
Kure Beach, North Carolina (marine), Brazos River, Texas (marine), and
220 INDUSTRIALCORROSIONSTANDARDSAND CONTROL
1.4
1.2
1.0
E 0.8
~_BEARING STEEL
~ 0.6
0
0 0.4
u
O'2~GALVANIZED STEEL
I I
0 I 2 3 TIME, 4 YEARS
5 6 7 8 9
FIG. 4--Comparative atmospheric exposure weight loss for galvanized steel and uncoated
copper-bearing steel in Porter County, Indiana.
Point Reyes, California (marine). Hudson and Stanners [67] reported on a

12-year test program for various types of zinc-coated steel exposed to
various British and tropical environments and again concluded that the life
of a zinc coating in a given environment is a direct linear function of
thickness.
In addition to the economic attractiveness of zinc as a coating for extend-
ing the usable life of steel, a further advantage is derived from the sacrificial
action afforded by zinc to the steel substrate at cut edges, scratches and
coating discontinuities in general. Under all conditions of atmospheric
exposure, zinc has been found to be anodic to iron [68]. Protection of the
steel base at such discontinuities is achieved both by the preferential dissolu-
tion of the zinc in the zinc-iron cell and the deposition of zinc corrosion
products which further stifle the reaction. An excellent survey of the
corrosion behavior of galvanized steel is contained in the Zinc Develop-
ment Association publication, Zinc: Its Corrosion Resistance, by Slunder
and Boyd [69].
Under immersed conditions zinc coatings continue to afford sacrificial
protection to the base steel substrate, although ready removal of non-
adherent zinc corrosion products results in a continued high dissolution
rate of the zinc coating. Furthermore, it was first reported by Schikorr [70]
that in various hot aqueous solutions a reversal in polarity between zinc and
iron occurs such that the zinc may become cathodic to the steel. It should be
emphasized that both temperature and solution composition are contribut-
ing factors to the potential reversal. This phenomenon has been the subject
of intensive study due to the widespread use of hot-dip galvanized steel in

domestic hot water systems.
Salt spray testing (ASTM B 117) is widely used as a quality control
criterion in evaluating the corrosion resistance of zinc coatings. A common
acceptance criterion for standard 11~ oz/ft 2 (G.90) galvanized sheet steel in
certain segments of the automotive industry is 240 h exposure to salt spray
without the occurrence of red rust. For lighter coating weight, hot-dip or
electrogalvanized steel and depending on the application of the fabricated
component, salt spray life requirements may be specified at times con-
siderably less than 240 h. The Preece test (ASTM A 239) is widely used as a
control in establishing uniformity of zinc coatings on steel. Test panels are
subjected to immersion in a copper sulfate solution for periods of 1 min
and the total number of dips required to dissolve the zinc coating and
deposit an adherent layer of copper over a specified area of the steel substrate
is determined. The test is used primarily to determine the thinnest portions
of the coating and finds widespread utilization for quality control in such
critical materials applications as electrical raceways. The Kesternich
SOs-humidity test also provides a means for accelerated testing of galva-
nized coatings, and in addition may be used to demonstrate the effective
sacrificial protection afforded by zinc at scribes and cut edges.
Hot-dip aluminized steel is produced on continuous coating lines by a
limited number of steel companies in the United States. Hot-dip aluminum
coatings are generally used where oxidation resistance is a requirement at
temperatures to ~1250 deg F in such applications as oven construction,
heat shields and automotive exhaust system components. Silicon ( ~ 1 0
percent) is normally added to the molten aluminum bath in order to reduce
the thickness of the brittle iron-aluminum alloy layer which forms at the
steel-coating interface. Silicon also reduces the viscosity of the bath and
results~'in a lighter coating than that produced from the "pure" aluminum
bath. Silicon-free hot-dip aluminum coatings are also applied to steel for
atmospheric corrosion resistance, although the utilization of the straight
aluminum coating is significantly less than for the aluminum-silicon alloy
coating. The aluminum-silicon alloy coating carries the A S T M designation
Type I and is specified as containing silicon in the range of 5 to 11 percent.
Further designation according to coating weight recognizes two classes,
T1 40 (regular) and T1 25 (light) having minimum coating weights (triple
spot test both sides) of 0.40 and 0.25 oz/ft 2, respectively, according to
ASTM standard A 463-69.
Although having good atmospheric corrosion resistance, the Type I
aluminum coating is not generally recommended for bold exposure applica-
tions for aesthetic reasons, in that the silicon in the coating tends to impart
a dark grey-black stain on exposure to many environments. At the present
time, there is no ASTM designation for the unalloyed aluminum coating
which is used in outdoor atmospheric exposure applications such as farm
silos and metal building r o o f decks. In general, for equivalent coating thick-
nesses the atmospheric corrosion rate of aluminized coatings are approxi-
mately one-third those of galvanized coatings. However, the sensitivity of
aluminum coatings to the nature of environment is significantly greater
than for galvanized coatings and probably relates both to the active-passive
behavior of aluminum and the polarity reversals with steel which occur in a
number of environments. For example, in a 3-year test conducted at a semi-
industrial atmospheric corrosion test site in Porter County, Indiana,
aluminized steel showed a pronounced increased corrosion resistance over
galvanized steel (0.016 mpy versus 0.049 mpy), while the same materials
tested in the highly industrial environment (HC1 fumes) of a chemical plant
in the Gulf area of Texas exhibited essentially reverse behavior.
Under immersed conditions, aluminum generally exhibits a potential
noble to steel resulting in accelerated attack on the steel base at discon-
tinuities in the coating. In environments containing ions conducive to the
breakdown of passivity (CI-, deaerated SO4-) the potential of the alumi-
num coating becomes more active than that of iron and under these
conditions cathodically protects the steel substrate by sacrificial action.
For a holiday-free aluminum coating, extremely long salt fog life, in the
range of 500 to 1500 hours, is obtained prior to the incidence of red rust. At
the present time there is no ASTM test procedure for measuring uniformity
and thickness of aluminized coatings, although some consideration has
been given to a test similar to the Preece test for galvanized coatings. The
test consists of repeated immersion in a solution containing copper and
fluoride ions in the presence of sulfuric acid. As with the Preece test, the
number of dips required to expose the steel substrate and deposit a layer of
adherent copper is determined.
A test developed for determining coverage of aluminum coatings involves
immersion in a 35-percent nitric acid solution at room temperature and
measuring the quantity of hydrogen produced by reaction of the acid with
exposed areas of bare steel.
Tin-coated steel (tin plate) represents one of the major products of the
steel industry accounting for approximately 5.7 million tons of steel shipped
during the year 1970 [1]. Prior to 1937 all tin plate produced was manu-
factured by a hot-dipping process similar to that presently used for the
continuous zinc and aluminum coating of steel. Today, essentially all tin
plate is produced on continuous electrodeposition lines and the product is
generally referred to as electrolytic tin plate.
The electrochemical sensitivity of tin-steel couples to the composition of
the environment is very critical and tin may act either as a noble or a sacri-
ficial coating. On exposure to the atmosphere and most aerated solutions,
tin is noble with respect to steel and would tend to promote corrosion of the
steel substrate at discontinuities in the coating. However, in the absence of
air (food container applications) tin is almost always anodic to steel. The
reversal in polarity is further promoted due to the complexing of stannous

ions by many food products which results in a shift in the equilibrium
potential for tin in the active direction due to the decreased activity of
stannous ions. Furthermore, although dissolution of tin occurs in food
product and beverage environments, the lack of toxicity of tin salts provides
tin coatings with the required properties for can linings. In certain food
product environments the dissolved tin salts may themselves act as inhibi-
tors, thus decreasing the subsequent dissolution rate of the tin coating.
It was established many years ago that the composition of the base steel
has a significant effect on the corrosion behavior of tin plate [71-73].
Phosphorus and silicon and in certain cases copper have been shown to be
detrimental with respect to corrosion resistance, and for this reason these
elements are normally specified to a minimum level. Corrosion of food and
beverage cans manufactured from tin plate may occur by reaction with the
atmosphere prior to filling the container, by chemical reaction with the
alloy coatings for certain specific corrosion resistance applications. Such
coatings normally contain 10 to 25 percent tin and the coated steel product
is known as terne plate. Terne plate has a cost advantage over tin plate and
in addition is more readily drawn, stamped and soldered. The primary dis-
advantages of the coating are unattractive appearance and unsuitability for
contact with foodstuffs and beverages due to the toxicity of lead.
Although other metal coatings (nickel, cadmium, chromium) are applied
to steel for improved corrosion resistance, only insignificant quantities are
produced as primary steel industry products and, therefore, will not be
considered here.
Organic Coatings for Steel

In 1971, approximately 1.25 million tons of precoated painted steel strip
were consumed in North America, and according to statistics supplied by
the National Coil Coaters Association production since 1962 has increased
at an average annual rate of ~ 18 percent. Although large quantities of pre-
fabricated structural and sheet steel are painted in the field, the present
section will be restricted to a discussion of precoated strip since in many
cases this is a basic product produced within the steel industry.
The continuous painting of steel (galvanized steel, aluminum) strip is
referred to as coil coating. At the present time, most coil coated steel finds
application in the construction market primarily for pre-engineered steel
buildings. This market is followed in size by those of container and packag-
ing products. It is anticipated that the major growth areas for coil-coated
steel in the near future will be in the automotive and appliance industries.
On a modern continuous coil coating line the steel strip is cleaned and
pretreated followed by the application of a primer and top coat each of
which is cured at a specified temperature range. A typical coil coating
finishing system for galvanized steel building panels might consist of the
c
0
0
0
TABLE 2--Some comparisons of chemical and physical properties of selected generic-organic coating systems. Z
~>
Silicone Z
Fluorocarbon Polyesters Polyesters Acrylics Vinyls Alkyds
Acid Resistance Excellent Excellent Good to Excellent Good to Excellent Excellent Excellent
Alkali Resistance Excellent Good to Excellent Good Good to Excellent Excellent Fair to Good Z
Solvent Resistance Excellent Good to Excellent Good to Excellent Excellent Poor to Fair Good
Erosion Excellent Excellent Good Fair to Good Fair to Good Fair to Good r~
O
Chalk Excellent Excellent Good Fair to Good Fair to Good Fair to Good z
Fade Excellent Excellent Good Fair to Good Fair to Good Fair to Good
O
r-
General Corrosion Resistance Excellent Good to Excellent Good Good Good Good
following sequence: (1) zinc phosphate conversion coating, (2) 0.2-mil

epoxy primer with appropriate corrosion inhibitors added, and (3) 0.8-mil,
silicone modified polyester top coat. Many coil coating lines also have
facilities for applying organic coatings as laminates; the most common sys-
tems applied in this fashion are polyethylene, polyvinyl chloride and various
copolymers of polyethylene with acetic acid and vinyl acetate. Generic
organic systems commonly applied to cold-roUed and galvanized sheet strip
include fluorocarbons, silicone polyesters, polyesters, acrylics, vinyls,
urethanes, and alkyds. Specific mechanical and chemical properties require-
ments often require combining two or more of the above in one formula-
tion. Table 2 outlines the relative properties of various organic systems.
Accelerated corrosion testing of precoated steel fails into the two general
categories of accelerated chemical resistance testing and accelerated
weathering testing. The former varies according to end use and includes, for
example, salt fog testing for pre-primed automotive stock and SO~-humidity
tests for industrial building siding. Water, whether as a vapor, liquid, or
solution, is the most universal chemical involved in the degradation or
organic coatings [74] and several standard tests are in existence designed to
evaluate resistance to water vapor and liquid in the atmosphere [74].
ASTM standard E 96 is the general technique for measuring water vapor
permeability through organic films in sheet form, while ASTM Standard
D 1653 is used to determine water vapor permeability of organic coatings.
The most widely used test for measurement of resistance to water condensa-
tion utilizes the Cleveland Condensation Tester (ASTM Standard D 2247)
in which coated test panels are subjected to continuous water vapor con-
densation at elevated temperatures. Coil coatings are normally evaluated
at 60 deg C for periods ranging from 6 to 240 h.
Spot tests for determining resistance to specific environments include
such specifications as ASTM Standard D 1303 (Household Chemicals) and
ASTM Standard D 1540 (Transportation Industry), the latter including the
effect of such chemicals as anti-freeze, lubricating oils, hydraulic fluids and
polishing creams and waxes.
For outdoor exposed applications the only truly reliable test for organic
coated steels is exposure under actual conditions of use. Guidelines for
conducting exterior exposure tests to determine the service life of organic
finishes are presented in the National Coil Coaters Association (NCCA)
Technical Bulletin No. 111. ASTM Standards D 609 and D 823 describe the
configurations and preparation of organic-coated test panels for both out-
door weathering and accelerated testing procedures.
Procedures designed to simulate natural outdoor weathering have been
available for many years. In general, the devices used employ a high intensity
ultraviolet light source and some form of either constant or cyclic tempera-
ture control. More recent devices incorporate humidity control, water
sprays, and additives designed to simulate atmospheric pollution (for
example, SO2, CO, NO~) [75]. Accelerated weathering tests of this type
enjoy a mixed reputation as to the correlation obtained with natural outdoor
weathering [76,77]. To a large extent the degree to which artificial weather-
ing machines are useful in predicting actual service performance depends to
a large extent on how closely the spectrum of ultraviolet light source
approximates that of the solar spectrum. Various complex light-dark,
wet-dry and contaminant-no contaminant combinations of cyles can now be
programmed into the control systems of modern artificial weathering
machines. ASTM Recommended Practices E 42 and E 239 cover operation
of carbon arc and water cooled xenon arc-type weathering machines,
respectively. In general, however, weathering machines are not to be recom-
mended for comparing organic coatings based on different polymers (for
example, acrylic versus polyester). General descriptions of these various
testing procedures, together with more detailed descriptions of paint
evaluation methods not covered by ASTM standards, are provided in the
extensive Gardner-Sward Paint Testing Manual [78].
Infrared analysis is gaining in acceptance as a quality control tool in the
identification and "fingerprinting" of paint formulations and is now being
routinely used as a first-step check in the determination of deviation from
formulation in paint failure analysis. The advent of Fourier Transform
Spectroscopy (FTS) permits not only the establishment of more rapid and
accurate infrared spectra, but by interfacing with a computer data bank
containing standard spectra will allow an almost instantaneous determina-
tion of quantitative deviations from a given formulation.
Evaporative Rate Analysis (ERA) continues to gain wider utilization as a
means of investigating such parameters as surface cleanliness, degree of cure
or cross-linking of adhesives and organic coatings, modifier migration and
film forming. ERA involves measurement of the rate of evaporation or
desorption of a minute amount of radioactive high boiling point material
(for instance, tetrabromoethane C14) which is deposited on the surface
being investigated. The rate of desorption is a reverse function of the
"activity" of the surface.
The degree of corrosion protection offered by organic coatings is often
more affected by the metal surface preparation and pretreatment and by
primer composition rather than by top coat composition [79]. This is
particularly true for coil coatings where the film thickness (typically H1
mil) offers little protection from water vapor and oxygen permeation. Paint
pretreatments such as phosphates and chromates are designed not only to
improve adhesion but to passivate the steel surface leading to a high degree
of underfilm corrosion protection. Further protection is provided by inhibi-
tive pigmentation such as zinc or strontium chromates within the primer
system.
In general, there are no widely employed corrosion standards for pre-
painted steel. The relative performance of various generic coatings is
reasonably well known, and the appropriate standards for any particular
application are established between paint supplier and the coil crater.
Typically, a 10-year warranted coating for exterior building panel applica-
tions would be required to meet the corrosion and durability criteria
outlined in Table 3.
TABLE 3--Typical requirements to meet a lO-year coating warranty for exterior building
panel applications.
Accelerated Tests
Salt Spray (ASTM B 117-64) A. 1000hours--No blistering or loss of adhesion on score

line when tested with No. 600 Scotch tape.
B. 1500hours--No more than 20 % of the area may con-
tain blisters: none larger than ASTM D 714-56 #6.
No loss of adhesion further than 1A in. from score
line when tested with No. 600 Scotch tape.
Humidity (ASTM D 1735-62) A. 500 hours--May show only slight softening and no
blistering.
B. 1000hours--Slight softening with no more than 10%
ASTM D 714-56, # 8 blisters.
Water Immersion (77 deg F- A. 500 hours--Shall show no marked color change after
Distilled Water) a 24-h recovery period.
Accelerated Weathering Atlas XW-R "Dew Cycle" weatherometer--300 light
hours (600 total hours). No adhesion loss or spotting
(other than normal water spotting) will be acceptable.
Slight fading and no chalking as tested with No. 600
Scotch tape.
Weathering South Florida--45 degrees South. After 5 years the
painted surface shall show no evidence of checking,
cracking, blistering, or loss of adhesion. There shall be no
more than slight chalking (#9 ASTM D 659-65) and
slight color fade (5 NBS units).
Stainless Steels
Although constituting only a very smaU proportion of total steel ship-
ments, the corrosion test procedures involved in establishing utilization
feasibility and design limitations for the stainless steels are probably more
numerous and extensive than those for all coated and uncoated low-alloy
steels combined. The reason for this is basically twofold and results from
(1) the increased aggressiveness of those environments to which the stainless
steels are exposed and (2) the localized nature of corrosion attack which can
lead to catastrophic failure. These localized forms of corrosion behavior,
although found on non-passive low-alloy steels, tend to be more pronounced
for the passive stainless steels and include intergranular corrosion, pitting
corrosion, crevice corrosion and stress-corrosion cracking. Furthermore,
susceptibility to hydrogen cracking becomes greater with the increase in
tensile strength associated with many of the heat treated higher alloy steels.
The increased susceptibility of the stainless steels to these localized forms

of corrosion attack derives in simplest terms from the nature of the small
anode-large cathode galvanic cell produced on breakdown or rupture of the
passive film often enhanced by local compositional differences. Although
the presence or absence of a passive film is a function of environment com-
position [80], the stainless steels are characterized by the presence of a
passive film over a broad range of environment pH, temperature and ionic
species. Stainless steels are normally recognized as iron-base alloys contain-
ing a minimum of 12 percent chromium, although this level has been shown
to be modified slightly due to the addition of other alloying elements. At the
12 percent chromium level, the critical current for passivity in neutral and
slightly acid solutions is sufficiently small as to permit spontaneous passiva-
tion due to local cell current action. The pH at which spontaneous passivity
is induced in aerated solutions is a function of alloy composition and values
for the common stainless steels are shown in Table 4 [81].
Many factors can contribute to local attack on metal surfaces and are
covered in the excellent review paper by Payer and Staehle [82]. These
authors discuss localized corrosion processes in terms of homogeneous
(dissolution at solute segregates, grain boundaries) and heterogeneous
(second phases) phenomena.
Improper heat treatment of both ferritic and austenitic stainless steels
may cause compositional changes at grain boundary areas resulting from
either solute segregation or precipitation. A schematic representation of
these conditions is shown in Fig. 5, taken from the work of Aust, Armijo
and Westbrook [83]. A general review of solute redistribution at grain
boundaries has been published by Westbrook [84]. The heat treatment
necessary to produce solute distribution at grain boundaries sufficient to
cause accelerated attack is known as sensitization, and the time-temperature
requirements necessary to induce such a condition differ considerably for
the ferritic and austenitic stainless steels. The sensitizing temperature range
for austel~itic steels is in the range 750 to 1550 deg F, such that slow cooling
or prolonged heating operations in this temperature range produce suscepti-
bility to intergranular attack. Sensitization of asutenitic steels results in
diffusion of carbon to the grain boundaries and precipitation of chromium
carbides, resulting in a chromium depleted zone at some finite distance
from the boundary. It has been clearly established that degree of suscepti-
bility to this form of failure is strongly influenced by carbon content to the
extent that austenitic stainless steels containing carbon in the range below
0.02 percent are relatively immune to this form of attack [85]. At tempera-
tures higher than the sensitizing range, the mobility of carbon is sufficiently
great to cause a uniform distribution throughout the alloy, while at lower
temperatures the diffusion rate is not sufficient to cause major migration to
grain boundaries (within reasonable time limits).
(a) SEGREGATED (b) C O N T I N U O U S PRECIPITATE
(c) D I S C O N T I N U O U S PRECIPITATE (d) HOMOGENEOUS
FIG. 5--Schematic representation of grain boundary segregation effects.
In addition to minimizing intergranular corrosion by reduction in carbon

level, significant reductions in susceptibility may be achieved by stabilizing
the carbon with titanium or columbium. In certain cases, heat treatment at
temperatures in the range 1900 to 2000 deg F followed by quenching may
be used to dissolve chromium carbides, and in this manner cause desensiti-
zation of a susceptible alloy. Other metallurgical effects (for example grain
growth) may preclude such heat treatments, however.
The most commonly used corrosion test for determining intergranular
corrosion susceptibility due to compositional variations in grain-boundary
areas is a modification of that first described by Huey [86] in 1930 and
involves exposure to a boiling solution of 65 percent nitric acid. Data is
normally reported in inches penetration/month for each of five 48-h
successive tests. A nonspeeific ASTM standard method for total immersion
corrosion testing of stainless steels is designated A 279-63, while ASTM
A 262-70 describes procedures for detecting susceptibility to intergranular
attack in stainless steels. Within the scope of the latter designation is in-
eluded oxalic acid, ferric sulfate-sulfuric acid, nitric acid, nitric-hydrofluoric
acid (for molybdenum-bearing austenitic steels) and copper-copper sulfate
sulfuric acid environments. Intergranular corrosion due to chromium
carbide precipitation is detected by all five test environments, whereas high
corrosion rates due to sigma phase precipitation in wrought chromium-

nickel-molybdenum and titanium or columbium stabilized steels are
observed only in nitric acid plus ferric sulfate-sulfuric solutions, respec-
tively. A less sensitive test for detecting intergranular corrosion than that
described in A 262 is contained in ASTM A 393-63, this being an acidified
copper sulfate test for use on severely sensitized alloys. This latter test is
based on that initially described by Strauss et al [87].
Nonsensitized (including stabilized) grades of austenitic stainless steel
have been found to show intergranular corrosion in highly oxidizing media
which has been attributed to grain boundary segregation of phosphorus and
silicon [88,89]. Intergranular corrosion of ferritic stainless steels has also
been reported [90,91], although the conditions under which sensitivity
occurs and the degree of susceptibility are quite different from those for the
austenitic grades [92].
The localized corrosion of stainless steels by pitting in chloride-containing
solutions including sea water has been recognized for many years [93]. It has
been established that pitting also occurs in bromide solutions. Susceptibility
is reduced with molybdenum additions, and in this respect AISI Type 316
stainless steel is less susceptible than Type 304. Nickel additions also reduce
susceptibility, although less significantly than molybdenum and the auste-
nitic grades are, therefore, generally less susceptible than the martensitic or
ferritie stainless steels. Pitting is readily induced in chloride solutions con-
taining oxidizing cations which cause depolarization of the cathodic reduc-
tion process. In this respect, ferric chloride has been used for many years to
evaluate the resistance of stainless steels to localized pitting attack and a
description of the test procedure was first described by Smith [94]. It should
be noted, however, that the test tends to be erratic (probably greatly in-
fluenced by surface imperfections) and thus is used only to qualitatively
determine gross differences in pitting susceptibility between materials.
Brennert [95] established the existence of a "breakthrough" potential at
which pitting initiates, which was investigated in some greater detail by
Mahla and Nielson [96]. More recent studies [97-101] have attested to the
validity of the "critical potential" (Vc) as a parameter in quantitatively
comparing the susceptibility of stainless steels to the initiation of pitting
corrosion. Thus, for passive Type 304 (18 Cr-8 Ni) stainless steel, the corro-
sion potential must be more noble than 0.26V (versus standard hydrogen
electrode) in 0.1 N NaC1 at 25 deg C in order to induce pitting [102]. The
effect of chloride ion concentration and pH on the eritical pitting potential
for Type 304 stainless steel is shown in Fig. 6 [100]. In addition to providing
a rapid method for evaluating the tendency for localized pitting attack,
measurements of critical pitting potentials in conjunction with statistical
regression analysis of alloying element effects show promise as a tool in the
development of pit resistant stainless steels.
Although the factors contributing to intergranular corrosion failure are
CORROSION I N THE I R O N AND STEEL I N D U S T R Y 231
t
,.n
0
z
1.0
0.9
xl
d 0.8
u'l
0.7
I--
0.6
0.5
u 0.4
0.3
0.2
Z
0.1
0 -'-c" i o o --e,
O. 0 I
IMMUNITY
U
I
IX
--0.1
--0.2
0 2 4
1
6 8 10 12 14
pH
FIG. 6---Critical pitting potential (V~)versus p H and chloride ion concentration for AISI
type 304 stainless steel.
now reasonably well understood, considerable controversy continues to

exist in the area of transgranular stress-induced cracking of stainless alloys
in certain specific environments. Tensile stresses are a prime requirement
for this mode of failure and time-to-failure is a function of applied stress.
In general, there is no clear-cut threshold stress below which failure does
not occur for even extended testing times.
Ferritic nickel-free stainless steels are essentially immune to stress corro-
sion cracking both in the chloride and hydroxyl media which cause rapid
failure in stressed austenitic materials. On the other hand, austenitic stain-
less steels containing greater than 45 percent nickel are again immune to
stress-corrosion cracking [103] in chloride solutions. Transgranular stress-
corrosion cracking of austenitic stainless steels generally requires high tem-
perature and the most commonly utilized test media for determination of
cracking susceptibility is a boiling (154 deg C) solution of 42 percent
magnesium chloride. ASTM Subcommittee G01.06 has established a
recommended practice for the boiling magnesium chloride test which was
first described by Schiel [104] in a study of the cracking susceptibility of
various stainless steels. A detailed review of available information on the
TABLE 4--Critical pH values for spontaneous development of passivity for stainless steels
in sulfate and chloride solutions.
Critical pH
AISI Type Sulfate Chloride
304 1.4 1.4

430 4.0 2.3
410 5.0 2.6
stress-corrosion cracking of iron-nickel-chromium alloys has been pub-

lished by Latanision and Staehle [105]. A similar review of the stress-
corrosion cracking behavior of high-strength steels is provided in the same
publication by Phelps [106].
It should be noted that at the present time there is a lack of standardized
stress-corrosion test methods for both stainless and high-strength alloy
steels in general. Continued activity by ASTM Subcommittee G01.06 is
directed towards the development of recommended practices in the stand-
ardization of test specimen configuration, method of stressing, precracking
and laboratory test environments.
Conclusion
The scope of the present subject has precluded more than a superficial
survey of corrosion control and standard test procedures for ferrous-base
products. At some time in the future it may be appropriate to subdivide the
subject matter into separate areas of review, and in this manner effect a
more in-depth evaluation into corrosion testing and quality control test
procedures for the various ferrous-base product categories.
References
[1] "Annual Statistical Report," American Iron and Steel Institute, 1970.
[2] Capp, J. A., Proceedings, American Society for Testing and Materials, Vol. 14,
Part 2, 1914, p. 474.
[3] Vernon, W. H., Transactions of the Faraday Society, Vol. 23, 1927, p. 113.
[6] Vernon, W. H., Transactions of the Electrochemical Society, Vol. 64, 1933, p. 31.
[7] Dassey, V. M., Industrial and Engineering Chemistry, Vol. 30, 1938, p. 1147.
[8] Swinden, T. and Stevenson, W. W., Journal of the Iron and Steel Institute, Vol. 142,
No. 2, 1940, p. 165.
[9] Pray, H. and Gregg, J., Proceedings, American Society for Testing and Materials,
Vol, 41, 1941, p. 758.
[10] Battelle Memorial Institute, Proceedings, American Society for Testing and Ma-
terials, Vol. 41, 1941.
[11] Opinsky, A. J., Thomson, R. F., and Boegehold, A. L., ASTM Bulletin 47, American
Society for Testing and Materials, 1953.
[12] Burns, R. M., Industrial and Engineering Chemistry, Analytical Edition, Vol. 17,
1945, p. 299.
[13] Campton, K. G., Transactions of the Electrochemical Society, Vol. 91, 1947, p. 705.
[14] Cooke, C. D. and Merritt, C., Materials and Methods, Vol. 25, No. 5, 1947, p. 77.
[15] Dix, E. H. and Bowman, J. J., presented at ASTM Symposium held in Chicago on
2 March 1937.
[16] Schlossberg, L., Metal Finishing, 1964, p. 93.
[17] Kuentsler, H. G. and Shur, E. G., Journal of Paint Technology, Vol. 40, 1968, p. 48A.
[18] Uhlig, H. H. in Corrosion and Corrosion Control, Wiley, New York, 1965, p. 144.
[19] Kesternich, W., Stahl und Eisen, Vol. 71, 1951, p. 587.
[20] Barton, K. and Beranek, E., Chemicke Listy, Vol. 50, 1956, p. 1388.
[21] Beranek, E., Barton, K., Smreck, K., and Sekerka, J., Chemicke Listy, Vol. 50, 1,956,
p. 1563.
[22] Edwards, J., Transactions of the Institute of Metal Finishing, Vol. 35, 1958, p. 55.
[23] Preston, R. and Stroud, V. E., Journal of the Institute of Petroleum, Vol. 36, 1950,
p. 457.
[24] Stankiewicz, S., Przemysl Chemiczny, Vol. 11, 1955, p. 151.
[25] Biestek, T. (Institut Mechaniki Precyzyjej, Warsaw); paper presented at 1st European
Corrosion Congress, Paris, 1956.
[26] Kutzelnigg, A., Werstoffe und Korrosion, Vol. 7, 1956, p. 65.
[27] Kutzelnigg, A., Werstoffe und Korrosion, Vol. 8, 1957, p. 492.
[28] Kutzelnigg, A., Werstoffe und Korrosion, Vol. 9, 1958, July.
[29] Kutzelnigg, A., Werstoffe undKorrosion, Vol. 14, 1963, p. 181.
[30] Kesternich, W., Werkstoffe und Korrusion, Vol. 16, 1965, p. 193.
[31] Evans, U. R. and Taylor, C. A. J., Corrosion Science, Vcl. 12, 1972, p. 227.
[32] Uhlig, H. H. in Corrosion and Corrosion Control, Wiley, New York, 1971.
[33] Fouroulis, Z. and Uhlig, H. H., Journal of the Electrochemical Society, Vol. 3, 1964,
p. 522.
[34] Williams, F. H., Iron Age, Vol. 66, 1900, p. 16.
[35] Howe, H. M., Congress International des Methodes d'Essia des Materiaux de Con-
struction, Vol. 2, Part I, 1901, p. 229.
[36] Buck, D. M., Journal of Industrial and Engineering Chemistry, Vol. 5, 1913, p. 447.
[37] Buck, D. M. and Handy, J. O., Journal oflndustrial Engineering Chemistry, Vol. 8,
1916, p. 209.
[38] Richardson, E. A. and L. T., Transactions of the American Electrochemical Society,
Vol. 30, 1916, p. 379.
[39] Kalmus, H. T. and Blake, K. B., Journal of Industrial and Engineering Chemistry,
Vol. 9, 1917, p. 123.
[40] Richardson, E. A. and L. T., Transactions of the American Electrochemical Society,
Vol. 38, 1929, p. 221.
[41] Vernon, W. H. J., Transactions of the American Electrochemical Society, Vol. 64,
1933, p. 31.
[42] Greenidge, C. T. and Lorig, C. H., Iron Age, Vol. 145, 1940, p. 21.
[43] Pilling, N. B. and Wesley, W. A., Proceedings, American Society for Testing and
Materials, Vol. 40, 1940, p. 643.
[44] Taylerson, E. S., Journal of the Iron and Steel Institute, Vol. 143, 1941, p. 287.
[45] Sims, C. E. and Boulger, F. W., Report to the Inland Steel Company from Battelle
Memorial Institute, 15 April 1944.
[46] Pilling, N. B. and Wesley, W. A., Proceedings, American Society for Testing and
Materials, Vol. 48, 1948, p. 610.
[47] Copson, H. R., Proceedings, American Society for Testing and Materials, Vol. 52,
1952, p. 1005.
[48] Proceedings, American Society for Testing and Materials, Vol. 53, 1953, p. 110.
[49] Hudson, J. C. and Stanners, J. F., Journal of the Iron and Steel Institute, Vol. 180,
1955, p. 271.
[50] Vernon, W. H. J., Journal of the Iron and Steel Institute, Vol. 196, 1960, p. 334.
[51] Larrabee, C. P. and Coburn, S. K. in 1st International Congress on Metallic Corrosion,
Butterworths, London, 1961, p. 276.
[52] Wiester, H. J. and Ternes, H., Stahl und Eisen, Vol. 87, 1967, p. 746.
[53] Kaluza, F. and Michalik, J., Przgald Spawalnictwa, Vol. 19, 1967, p. 261.
[54] Plackett, R. L. and Burman, J. P., Biometrika, Vol. 33, 1946, p. 305.
[55] Report of Committee V on the Corrosion of Iron and Steel, Proceedings, American
Society for Testing and Materials, Vol. 7, 1907, p. 209.
[56] Copson, H. R., Proceedings, American Society for Testing and Materials, Vol. 45,
1945, p. 554.
[57] Pourbaix, M., CEBELCOR Report RT 160, 1969 (in French).
[58] Matsushima, I. and Ueno, T., Corrosion Science, Vol. 2, 1971, p. 129.
[59] Cohen, A. and Jellinek, R., Corrosion, Vol. 22, 1966, p. 39.
[60] Legault, R. A. and Walker, M. S., Corrosion, Vol. 19, 1963, p. 222.
[61] Wilde, B. E., Corrosion, Vol. 23, 1967, p. 379.
[62] Walker, M. S. and France, W. D., Jr., Materials Protection, Vol. 8, 1969, p. 47.
[63] Jones, D. A., Corrosion Science, Vol. 8, 1968, p. 19.
[64] Evans, S. and Koehler, E. L., Journal of the Electrochemical Society, Vol. 108, 1961,
p. 5O9.
[65] Meany, J. J., Materials Protection, Vol. 8, 1969, p. 27.
[66] Report of Subcommittee XIV, Committee A-5, Proceedings, American Society for
Testing and Materials, Vol. 44, 1944, p. 92.
[67] Hudson, J. C. and Stanners, J. F., Journal of the Iron and Steel Institute, Vol. 175
1953, p. 381.
[68] Report of Subcommittee VIII of Committee B-3, Proceedings, American Society for
Testing and Materials, Vol. 284, 1928, p. 1.
[69] Stunder, C. J. and Boyd, W. K., "Zinc: Its Corrosion Resistance," Zinc Development
Assoc., London, 1971.
[70] Schikorr, G., Transactions of the Electrochemical Society, Vol. 76, 1939, p. 247.
[71] National Canners Assoc., "Canned Food Containers: A Study With Special Ref-
erence to the Steel Base on Resistance to Perforation," Research Lab Bulletin, No.
22L, 1923.
[72] Caulfield, K. W., Kerr, R., and Angles, R. M., Journal of the Society of Chemical
Industry, Vol. 66, 1947, p. 5.
[73] Price, J. W. and Hoare, W. E., Tin Research Institute Publication TTE, 1949.
[74] Official Digest, Federation of Societies for Paint Technology ODFPA, Vol. 37, 1965,
p. 626.
[75] Nass, L. I., Plastics Technology, Oct. 1971, p. 91.
[76] Hoffmann, E., Journal of Paint Technology, Vol. 43, 1971, p. 107.
[77] Mitton, P. B. and Richards, D. P., Journal of Paint Tecnnology, Vol. 43, 1971, p. 107.
[78] Sward, G. G , Ed., Gardner-Sward Paint Manual, ASTM STP 500, American Society
for Testing and Materials, 1972.
[79] Tropp, F. E., Journal of Paint Technology, Vol. 39, 1967, p. 225.
[80] Uhlig, H. H., Journal oJ the Electrochemical Society, Vol 108, 1961, p. 327.
[81] Leckie, H. P., Corrosion, Vol. 24, 1968, p. 70.
[82] Payer, J. H. and Staehle, R. W., Proceedings of the 1st International Conference on
Corrosion Fatigue, NACE, Houston, Texas, 1972, in press.
[83] Aust, K. T., Armijo, J. S., and Westbrook, J. H., Transactions of the American
Society for Metals, Vol. 59, 1966, p. 544.
[84] Westbrook, J. H., Metal Review, Vol. 9, 1964, p. 415.
[85] Bain, E., Aborn, R., and Rutherford, J., Transactions of the American Society for
Metals, Vol. 21, 1933, p. 481.
[86] Huey, W. R., Transactions of the American Society of Steel Treating, Vol. 18, 1930,
p. 1126.
[87] Strauss, B., Schottky, H., and Hinnuber, J., Z. Anorganische und Allegemeine Chemie,
Vol. 188, 1930, p. 309.
[88] Armijo, J. S., Corrosion, Vol. 24, 1968, p. 24.
[89] Duncan, R., Armijo, J., and Pickett, A., Materials Protection, Vol. 8, 1969, p. 37.
[90] Lula, R., Lena, A., and Kiefer, G., Transactions of the American Society of Metals,
Vol. 46, 1954, p. 197.
[91] Hockmann, J., Revue de Metallurgie, Vol. 48, 1951, p. 734.
[92] Uhlig, H. H. in Corrosion and Corrosion Control, Wiley, New York, 1971, p. 307.
[93] Uhlig, H. H. in Corrosion Handbook, Wiley, New York, 1948, p. 165.
[94] Smith, H. A., MetalProgress, Vol. 33, 1938, p. 596.
[95] Brement, S., Journal of the Iron and Steel Institute, Vol. 135, 1937, p. 101.
[96] Mahla, E. M. and Nielson, N. A., Transactions oJ the Electrochemical Society, Vol. 89,
1946, p. 167.
[97] Kolotyrkin, Y. M., Corrosion, Vol. 19, 1963, p. 261.
[98] Pourbaix, M. et al, Corrosion Science, Vol. 3, 1963, p. 239.
[99] Leckie, H. P. and Uhlig, H. H., Journal of the Electrochemical Society, Vol. 113, 1966,
p. 1262.
[100] Leckie, H. P., Journal of the Electrochemical Society, Vol. 117, 1970, p. 1152.
[101] Wilde, B. E., Corrosion, Vol. 28, 1972, p. 283.
[102] Horvath, J. and Uhlig, H. H., Journal of the Electrochemical Society, Vol. 115, 1968,
p. 791.
[103] Copson, H. in Physical Metallurgy of Stress-Corrosion Fracture, T. Rhodin, Ed.,
Interscience, New York, 1959, p. 247.
[104] Schiel, M. A. in Symposium on Stress-Corrosion Cracking of Metals ASTM-AIME,
1944, p. 395.
[105] Latanision, R. M. and Staehle, R. W., Proceedings of the Conference "Fundamental
Aspects of Stress-Corrosion Cracking," NACE, Houston, Texas, 1969, p. 214.
[106] Phelps, E. H., Proceedings of the Conference "Fundamental Aspects of Stress-
Corrosion Cracking," NACE, Houston, Texas, 1969, p. 398.
STP534-EB/Nov. 1973
APPENDIX A-1
Tablulated List of Current Corrosion Standards, Test Methods, and Recom-
mended Practices Issued by the American Society for Testing and Materials
(ASTM) and the National Association of Corrosion Engineers (NACE)
American Society For Testing and Materials, 1916 Race Street, Philadelphia, Pa.,
19103.
Designation Title
A 262-68(3) 1 Recommended Practices for Detecting Susceptibility to
Intergranular Attack in Stainless Steels
A 279-63(3) Total Immersion Corrosion Test of Stainless Steels
A296--68(2) Specification for Corrosion Resistant Iron-Chromium,
Iron-Chromium-Nickel, and Nickel Base Alloy Castings for
General Application
A 380-57(3) Recommended Practice for Descaling And Cleaning Stain-
less Steel Surfaces
A 393-63(3) Recommended Practice for Conducting Acidified Copper
Sulfate Test for Intergranular Attack in Austenitic Stainless
Steel
B 117-64(7, 21, 31) Salt Spray (Fog) Testing
B 287-62(7, 21, 31) Acetic Acid-Salt Spray (Fog) Testing
B 368-68(7, 21) Copper-Accelerated Acetic Acid-Salt Spray (Fog) Testing
(CASS Test)
B 380-65(7) Corrosion Testing ot Decorative Chromium Plating by the
Corrodkote Procedure
B 537-70(7) Recommended Practice for Rating of Electroplated Panels
Subjected to Atmospheric Exposure
B 538-70(7) Method of FACT (Ford Anodized Aluminum Corrosion
Test) Testing
C 464-64(14) Test for Corrosion Effect of Thermal Insulating Cements on
Base Metal
C 621-68(13) Test for Static Corrosion of Refractories by Molten Glass
C 622-68(13) Simulated Service Test for Corrosion Resistance of Re-
fractories to Molten Glass
D 69-67(28) Specification for Friction Tape for General Use for Elec-
trical Purposes
D 130-68(17) Test for Detection of Copper from Petroleum Products, by
the Copper Strip Tarnish Test
D 484-71(17) Specification for Hydrocarbon Drycleaning Solvents
D 665-60(17) Test for Rust-Preventing Characteristics of Steam-Turbine
Oil in the Presence of Water
D 801-57(20) Dipentene, Sampling and Testing
D 807-52(23) Corrosivity Test of Industrial Water (United States Bureau
of Mines Embrittlement Detector Method)
1Numbers in parentheses indicate the part number of the Book of Standards in which the
standard appears, as of 15 July 1971. Standards are also available separately. The number
after the dash is the year of adoption or of latest revision.
236
Copyright9 1973by ASTMInternational www.astm.org
APPENDIXES 2 3 7
Designation Title
D 849-47(20) Test for Copper Corrosion of Industrial Aromatic Hydro-
carbons
D 930-67(22) Total Immersion Corrosion Test of Water-Soluble Alumi-
num Cleaners
D 1141-52(23) Specification for Substitute Ocean Water
D 1261-55(17) Test for Effect of Grease on Copper
D 1275-67(18, 29) Test for Corrosive Sulfur in Electrical Insulating Oils
D 1280-67(22) Total Immersion Corrosion Test for Soak Tank Metal
Cleaners
D 1374-57(22) Aerated Total Immersion Corrosion Test for Metal Cleaners
D 1384-70(22) Corrosion Test for Engine Antifreezes in Glassware
D 1567-62(22) Testing Detergent Cleaners for Evaluation of Corrosive
Effects on Certain Porcelain Enamels
D 1611-60(15) Test for Corrosion Produced by Leather in Contact with
Metal
D 1616-60(20) Test for Copper Corrosion by Mineral Spirits (Copper
Strip Test)
D 1654-61(21) Evaluation of Painted or Coated Specimens Subjected to
Corrosive Environments
D 1735-62(21) Water Fog Testing of Organic Coatings
D 1743-64(17) Test for Rust Preventive Properties of Lubricating Greases
D 1838-64(18, 19) Test for Copper Strip Corrosion by Liquified Petroleum
(LP) Gases
D 2043-69(15) Test for Silver Tarnishing by Paper
D 2059-63(25) Test for Resistance of Zippers to Salt Fog
D 2251-67(22) Test for Metal Corrosion by Halogenated Organic Solvents
and Their Admixtures
D 2570-70(22) Simulated Service Corrosion Testing of Engine Antifreezes
D 2649-70(18) Determining Corrosion Characteristics of Dry Solid Film
Lubricants
D 2688-70(23) Test for Corrosivity of Water in the Absence of Heat
Transfer (Weight Loss Methods)
D 2776-69T(23) Tests for Corrosivity of Water in the Absence of Heat
Transfer (Electrical Methods), Tentative
D 2803-70(21) Test for Filiform Corrosion Resistance of Organic Coatings
on Metal
D 2809-69T(22) Test for Cavitation-Erosion Corrosion Characteristics of
Aluminum Automotive Water Pumps with Coolants,
Tentative
G1-72(31) Recommended Practice for Preparing, Cleaning, and
Evaluating Corrosion Test Specimens
G2-67(7, 31) Recommended Practice for Aqueous Corrosion Testing of
Samples of Zirconium and Zirconium Alloys
G3-68(31) Recommended Practice for Conventions Applicable to
Electrochemical Measurements in Corrosion Testing
G4-68(3, 31) Recommended Practice for Conducting Plant Corrosion
Tests
G5-72(31) Recommended Practice for a Standard Reference Method
for Making Potentiostatic and Potentiodynamic Anodic
Polarization Measurements
G 7-69T(30) Recommended Practice for Atmospheric Exposure Testing
of Nonmetallic Materials, Tentative
'~38 INDUSTRIAL CORROSION STANDARDS AND CONTROL
Designation Title
G 9-69T(21, 30) Test for Water Penetration into Pipeline Coatings, Tentative
G 11-69T(21, 30) Test for Effects of Outdoor Weathering on Pipeline Coatings,
Tentative
G 15-71(31) Definitions of Terms Relating to Corrosion and Corrosion
Testing
G 16-71(31) Recommended Practice for Applying Statistics to Analysis
of Corrosion Data
G 28-72 Method of Detecting Susceptibility to Intergranular Attack
in Wrought Nickel-Rich Chromium Bearing Alloys
G 30-72 Recommended Practice for Making and Using U-Bend
Stress Corrosion Test Specimens
G 31-72 Recommended Practice for Laboratory Immersion Cor-
rosion Testing of Metals
G 33-72 Recommended for Recording Data from Atmospheric
Corrosion Tests of Metallic Coated Steel Specimens
G 34-72 Standard Method of Test for Exfoliation Corrosion Sus-
ceptibility in 7XXX series Copper-Containing Aluminum
Alloys (Exco Test)
G 35-73 Recommended Practice for Determining the Susceptibility
of Stainless Steel and Related Ni-Cr-Fe Alloys to Stress
Corrosion Cracking in Polythionic Acids
G 36-73 Recommended Practice for Performing Stress Corrosion
Cracking Tests in a Boiling Magnesium Chloride Solution
G 37-73 Recommended Practice for the Use of Mattsson's Solution
of pH 7.2 to Evaluate the Stress Corrosion Susceptibility of
Cu-Zn Alloys
National Association of Corrosion Engineers, 2400 West Loop South, Houston,
Texas, 77027.
TM-01-69 ~ Laboratory Corrosion Testing of Metals for the Process
Industries
RP-01-69 Control of External Corrosion on Underground or Sub-
merged Piping Systems
TM-01-70 Visual Standard for Surfaces of New Steel Airblast Cleaned
with Sand Abrasive
RP-01-70 Protection of Austenitic Stainless Steel in Refineries Against
Stress Corrosion Cracking by the Use of Neutralizing Solu-
tions During Shut Down
TM-02-70 Method of Conducting Controlled Velocity Laboratory
Corrosion Tests
TM-01-71 Autoclave Corrosion Testing of Metals in High-Tempera-
ture Water
RP-01-71 Method for Lining of Lease Production Tanks with Coal
Tar Epoxy
TM-01-72 Antirust Properties of Petroleum Products Pipeline Cargoes
RP-01-72 Surface Preparation of Steel and Other Hard Materials by
Water Blasting Prior to Coating or Recoating
RP-02-72 Direct Calculation of Economic Appraisals of Economic
Control Measures
The last two digits indicate the year of adoption. TM denotes a test method and RP a
recommended practice.
APPENDIXES 2 3 9
Designation Title
RP-03-72 Methods for Lining Lease Production Tanks with Coal Tar
Epoxy
RP-04-72 Methods and Controls to Prevent In-Service Cracking of
Carbon Steel (P-l) Welds in Corrosive Petroleum Refining
Environments
RP-05-72 Design, Installation, Operation, and Maintenance of Im-
pressed Current Deep Ground Beds
TM-01-73 Methods for Determining Water Quality for Subsurface
Injection Using Membrane Fitters
RP-01-73 Collection and Identification of Corrosion Products
RP-02-73 Handling and Proper Usage of Inhibited Oil Field Acids
STP534-EB/Nov. 1973
A P P E N D I X A-2
Selected Tabulation of British, French, and German Standards Concerned
with Corrosion Testing Methods and the Evaluation of the Corrosion
Resistance of Materials and Products
British Standards: Issuing Agency--British Standards Institution

Designation Title and Description
B.S. 135 Specifications for Benzines and Benzoles
The corrosive sulfur content is specified in terms of the discoloration of a freshly
prepared copper strip exposed in a reflux condenser.
B.S. 245 Specifications for White Spirit
Similar to B.S. 135
B.S. 441 Rosin-Cored Solder Wire "Activated" and "Non-Activated"
(Non-corrosive)
The corrosive action of flux residue is assessed in terms of the discoloration and
possible pitting of a copper sheet exposed to the flux at 35 C for 48 h.
B.S. 489 Specification for Steam Turbine Oils
Specifies corrosivity, rust preventing characteristics, and oxidation behavior in
terms of ASTM Standard Methods D 130, D 665, D 943, and D 974.
B.S. 1133 British Standard Packaging Code. Section 6. Temporary
Prevention of Corrosion
Salt, humidity, and hydrogen bromide exposure tests are used to evaluate the
effectiveness of corrosion inhibiting coatings and solutions.
B.S. 1224 Specification for Electroplated Coatings of Nickel and
Chromium
CASS, Corrodkote, and acetic acid salt spray tests (similar to ASTM B 368, B 380,
and B 287) are used to evaluate corrosion resistance.
B.S. 1263 Hypodermic Syringes for Use in Medical and Surgical
Practice
Autoclaving in steam, boiling in distilled water, and boiling in 0.9 percent sodium
chloride solution consecutively for 30 rain each, are used to evaluate corrosion
resistance.
B.S. 1344 Part 2A Vitreous Enamels-Group A, Kitchen Equipment
Disks of filter paper saturated with 100 g/1 of citric acid are placed onto the surface
and the deterioration observed after 20 min at 20 C.
B.S. 1391 Performance Tests for Protection of Light-Gauge Steel and
Wrought Iron Against Corrosion
Corrosion tests axe described involving either daily exposure to a sea water spray or
continuous exposure to vapor condensation above a heated solution of sulfur
dioxide. These tests are aimed at evaluating both metallic and paint coatings.
B.S. 1615 Anodic Oxidation Coatings for Aluminum
An acetic salt spray test (similar to A S T M B 287) and the sulfur dioxide test of
B.S. 1391 are used to evaluate corrosion resistance and effectiveness of scaling.
B.S. 1706 Specification for Electroplated Coatings of Cadmium and
Zinc on Iron and Steel
240
APPENDIXES 241

The effectiveness of passivation is measured by means of a 95-percent relative
humidity exposure at 55 C for 16 h, followed by cooling to 30 C and holding for 1 h.
B.S. 1872 Specification for Electroplated Coatings of Tin
Exposure to a controlled moist sulfur dioxide atmosphere is used to determine
coating discontinuities.
B.S. 1916 Hypodermic Syringes for Insulin Injection
The same procedure as in B.S. 1263 is used to evaluate corrosion resistance.
B.S. 2011 Basic Climatic and Durability Tests for Components for
Radio and Allied Electronic Equipment
A 2-h synthetic sea water spray at 20 C followed by storage ai 35 C and 90-95 per-
cent relative humidity is used. Humidity tests at 55 C and 95 percent humidity with
2 deg C temperature fluctuations ibur times an hour are also included
B.S. 2056 Rust, Acid, and Heat Resisting Steel Wire for Springs
Susceptibility to intergranular corrosion is evaluated by a sensitizing heat treatment
followed by exposure to a solution of copper sulfate and sulfuric acid.
B.S. 2983 Hypodermic Dental Needles
A 5-h exposure to 10 percent citric acid solution at room temperature followed by
boiling in distilled water for 30 min is used.
B.S. 3116 Specification for Automatic Fire Alarm Systems in Buildings
Part I. Heat-Sensitive (Point) Detectors
A 16-day exposure to condensing sulfur dioxide is used, similar to that detailed in
B.S. 1391.
B.S. 3597 Specification for Electroplated Coatings of 65/35 Tin-
Nickel Alloy
A 24-h exposure to a controlled sulfur dioxide atmosphere at room temperature is
used to evaluate the presence of discontinuities.
B.S. 3745 Method for the Evaluation of Results of Accelerated Cor-
rosion Tests on Metallic Coatings
A detailed procedure for the counting and evaluation of corrosion sites observed
after acetic acid salt spray, Corrodkote, and CASS tests is described.
B.S. 4601 Specification for Electroplated Coatings of Nickel Plus
Chromium on Plastic Materials
CASS and acetic acid salt spray tests are used.
B.S. 4292 Specification for Electroplated Coatings of Gold and Gold
Alloy
Exposure to sulfur dioxide followed by exposure to hydrogen sulfide is used for
coatings greater than 5 t,m thick while hydrogen sulfide exposure alone is used for
thinner coatings.
B.S. 4758 Specification for Electroplated Coatings of Nickel for
Engineering Purposes
Exposure to sodium chloride and gelatine-soaked filter papers for 10 min followed
by dipping into a solution of potassium ferricyanide is used to evaluate coating
porosity.
French Standards: Issuing Agency--L'Association Francaise De Normalisation

(AFNOR)
NF X 41-002 Essai au brouillard salin
Gives specifications for both 5 and 20 percent salt spray testing at 35 C and 85-90
percent relative humidity.

NF A 91-020 RevStements M6talliques Clichds-t~talons Pour Essais de
Corrosion
Provides color photographs showing the difference in behavior of an anodic metal
plating (zinc) and a cathodic metal plating (nickel) on steel during salt spray testing.
NF A 05-159 D6termination de la R6sistance a la Corrosion Inter-
granulaire des Aciers Inoxydables Aust6nitiques
Describes the determination of intergranular corrosion susceptibility in austenitic
stainless steels using the Monypenriy-Strauss Test (immersion in a solution of
sulfuric acid and copper sulfate).
NF A 91-021 M6thode d'I~valuation des R6sultats des Essais de Corrosion,
Applicable aux D6pots t~lectrolytiques Cathodiques
Provides a detailed rating procedure and classification system for evaluating (in
conjunction with NF A 91-020) the performance of cathodic metal electroplates in
accelerated corrosion tests.
NF A 05-160 D&ermination de la Resistance h la Corrosion Inter-
granulaire des Aciers Inoxydables Austenitiques Essai de
Corrosion en Milieu Nitrique
Describes the determination of intergranular Corrosion susceptibility in austenitic
stainless steels by means of the Huey Test (nitric acid exposure).
German Standards: Issuing Agency--Fachnormenausschuss Materialpriifung im

Deutschen Normenausschuss
Designation Title
DIN 1548 Zinkfiberzi~ge runder Stahldr~thte
DIN 2444 Entwurf, Zinkiiberzi~ge aut Stahlrohren; Technische Liefer-
bedingungen for Feuerverzinkung in handelsiiblicher Qualit~it
DIN 8565 Rostschutz von Stahlbauwerken durch Metallspritzen
DIN 20578 Zinki~berzi~ge fi~r F~rderwagen; Feuerverzinkung der
K~sten
DIN 50010 Werkstoff-, Bauelemente- und Ger~,teprfifung; Klima-
beanspruchung, Allgemeines, Begriffe
DIN 50016 Werkstoff-, Bauelemente- und Ger~tepri~fung; Bean-
spruchung im Feucht-Wechselklima
DIN 50017 Werkstoff-, Bauelemente- und Ger~iteprtifung; Bean-
spruchung in Schwitzwasser-Klimaten
DIN 50018 Werkstoff-, Bauelemente- und Ger~teprtifung; Bean-
spruchung im Schwitzwasser-Wechsel-klima mit schwefel-
dioxydhaltiger Atmosphere
DIN 50021 Vornorm Korrosionsprtifungen; Sprtihnebelpri~fungen mit
verschiedenen Natriumchloridl6sungen
DIN 50900 Korrosion der Metalle; Begriffe
DIN 50901 Korrosionsgroben bei ebenm~,ssigem Angriff; Begriffe,
Formelzeichen, Einheiten
DIN 50902 Entwurf Korrosionsschutz; Behandlung yon Metallober-
fl~chen, Begriffe
DIN 50903 Metallische (Jberztige; Poren, EinschRisse, Blasen und
Risse, Begriffe
DIN 50905, Korrosionsversuche; Richtlinien fiir die Durchfiihrung und
Auswertung
APPENDIXES 243
Designation Title
DIN50906 KorrosionspriJfung in kochenden FiJssigkeiten
(Kochversuch)
DIN50907 auf Meerklima- u. Meerwasserbest~idigkeit, ffir Leicht-
metalle
DIN50908 Priifung yon Leichtmetallen; Spannungskorrosionsversuche
DIN50910 Einflussgr6ssen und Messverfahren bei der Korrosion im
Erdboden in Gegenwart yon elekrisehen Erdstr6men
DIN50911 Priifung von Kupferlegierungen; Quecksilbernitratversuch
DIN50914 Priifung nichtrostender St~ihle auf Best~ndigkeit gegen
interkristaline Korrosion; Kupfersul fat-Schwefels~iure-
Verfahren
DIN50930 Vornorm Korrosion der Metalle; Beurteilung des korrosion-
schemischen Verhaltens kalter W~isser gegenfiber unver-
zinkten und verzinkten Eisenwerkstoffen, Richtlinien
DIN50932 Priffung metallischer Uberziige; Bestimmung der Dicke von
Zinktiberziigen auf Stahl durch 6rtliches anodisches Abl6sen
DIN50933 Entwurf, Prtifung metalliseher Uberztige; Messung der
Dicke von Oberztigen auf Stahl mittels Feinzeigers
DIN50938 Entwurf, Korrosionsschutz; Briinieren yon Eisenwerk-
stoffen
DIN50940 Prfifung von chemischen Entrostungsmitteln und Spar-
beizzus~itzen (Inhibitoren) ffir Stahl und Eisen: Labora-
toriumsversuche
DIN50941 Korrosionsschutz; Chromatieren yon galvanischen Zink-
und Cadmiumfiberzi]gen
DIN50942 Entwurf, Korrosionsschutz; Phosphatieren von Stahlteilen
DIN50943 Priifung von anorganischen nichtmetallischen Uberziigen
auf Aluminum und Aluminiumlegierungen; mikroskopishe
Messung der Schichtdicke
DIN 50944 Priifung von anorganischen nichtmetallischen Oberziigen
auf Reinaluminum und Aluminumlegierungen; Bestimmung
des Fl~ichengewichtes von Aluminiumoxidschienten durch
chemisches Abl6sen
DIN 50945 - - ; Zerst6rungstreie Messung der Dicke transparenter
Oxidschichten nach dem Differenzverfahren mit dem Mikro-
skop
DIN 50946 - - ; Prfifung der Gtite der Verdichtung anodisch erzeugter
Oxidschichten im Anf~irbeversuch
DIN 50947 - - ; Prtifung anodisch erzeugter Oxidschichten im Korro-
sionsversuch (Dauertauchversueh)
DIN 50948 Prtifung yon anorganisehen nichtmetallischen Deckschichten
auf Reinaluminum und Aluminiumlegierungen; zerst6rungs-
freie Messung der Schichtdicke yon transparenten Oxid-
schichten naeh dem Lichtschnittverfahren
DIN 50949 Priifung von anorganischen nichtmetallischen Oberztigen
auf Reinaluminum and Aluminumlegierungen; Zerst6rungs-
freie Priifung yon anodisch erzetigten Oxidschichten durch
Messung des Scheinleitwertes
DIN 50950 Prfifung galvanischer Lrberztige; mikroskopische Messung
der Schichtdicke
DIN 50951 Entwurf, PriJfung g.alvanischer 0berztige; Messung der
Dicke galvanischer Uberziage nach dem Strahlverfahren
Designation Title
DIN50952 Priifung metallischer Crberz~ige; Bestimmung des Fl~ichen-
gewichtes von Zinkfiberzfigen auf Stahl durch chemisches-
Abl6sen des Uberztiges, gravimetrisches Verfahren
DIN50953 Priifung galvanischer Uberziige; Bestimmung der Dicke von
diinnen ChromiiberziJgen nach dem TiJpfelverfahren
DIN50954 Priifung metallischer UberziJge; Bestimmg. des mittleren
Fl~ichengewichtes von Zinniiberziigen auf Stahl durch chem.
Abl6sen des Lrberziiges
DIN50955 Entwurf, PriJfung metallischer Uberziige; Messung der
Dicke galvanishcer Lrberziige, coulometrisches Verfahren
DIN50957 Priifung galvanishcer B~ider; Galvanisierungspriifung mit
der Hull-Zelle, aUgemeine Grunds~t.ze
DIN50958 Entwurf, Priifung galvanischer Uberziige; Korrosions-
prtifung von verchromten Gegenst~inden nach dem modifi-
zierten Corrodkote-Verfahren
DIN50960 Korrosionsschutz; galvanische Oberztige, Kurzzeichen,
Schichtdicken, allgemeine Richtlinien
Bbl. Vornorm, Galvanische UberziJge auf Stahl; allgemeine
Hinweise zur Anwendung ais Schutz gegen atmosph~irische
Korrosion in Mittel- und Westeuropa
DIN 50961 Korrosionsschutz; galvanische Zinkiiberztige auf Stahl
DIN 50962 Korrosionsschutz; galvanische KadminiumiJberziige auf
Stahl
DIN 5O963 Korrosionsschutz; galvanische Nickel- und Nickel-Chrom-
Uberztige auf Stahl
DIN 50964 Korrosionsschutz; galvanische Kupfer-Nickel-Chrom-
(]berziige auf Zink und Zinklegierungen
DIN 50965 Korrosionsschutz; galvanische Zinn- und Kupfer-Zinn-
UberziJge auf Stahl, Ku.p.fer und Kupferlegierungen
DIN 50967 Entwurf, Galvanische UberziJge, Nickel-Chrom-Uberziige
auf Stahl, Ku.p.fer und Zinkwerkstoffen sowie Kupfer-
Nickel-Chrom-Uberztige auf Stahl und Zinkwerkstoffen
DIN 50971 Entwurf, BI.1, Elektrolytisch erzeugte Lrberztige; Chemi-
kalien ftir cyanidische B~ider, Anforderungen
DIN 50972 Entwurf, BI.1, Elektrolytisch erzeugte Ctberziige; Kupfer-
sulfat fiir galvanische Bhder, Anforderung.en
DIN 50973 Entwurf, BI.1, Elektrolytisch erzeugte Uberzfige, S/iuren
ftir galvanische B~ider, Anforderungen
DIN 50975 Korrosionsschutz; Zinktiberziage durch Feuerverzinken,
Richtlinien
DIN 50976 Entwurf, Anforderungen an Zinktiberztige auf Gegenstanden
aus Eisenwerkstoffen, die als Fertigteile feuerverzinkt werden
DIN 50980 Entwurf, Prtifung metallischer Crberziige; Auswertung yon
Korrosionsprtifungen
DIN 51213 Vornorm Priifung metaltischer/JberziJge auf Dr~ihten
STP534-EB/Nov. 1973
APPENDIX B
Selected ASTM Standards Referred to Frequently in Book
Designation: A 279 - 63 American Nat,onal Standard G81 9 - 1 9 7 0

American National Standards Institute
Standard Method of
TOTAL IMMERSION CORROSION T E S T OF
STAINLESS STEELS 1
This Standard is issued under the fixed designation A 279: the number immediately following the designation indicates the
year of original adoption or, in the case of revision, the year of last revision. A number in parentheses indicates the year of
last reapproval.
1. S c o p e for control of t e m p e r a t u r e and aeration will be

1.1 This method describes procedures for essentially the same with all types of appara-
m a k i n g total immersion corrosion tests on tus. The principal differences will be with
stainless steels. N o one procedure is rigorously respect to the means of providing a control of
described since the most desirable procedure to the velocity.
follow in any specific case will depend on the 2.2 Velocity."
p a r t i c u l a r aim desired. 2.2.1 Ordinarily, velocity will be fixed at
1.2 When the total i m m e r s i o n test is to some value which shall be held uniform over
serve as a control test for d e t e r m i n i n g whether the whole surface of the specimen, especially
successive lots of the same material differ when changes in mechanical properties are to
significantly in some property from each other, be used as a measure of corrosion; however, for
the test conditions should be arbitrarily se- p a r t i c u l a r purposes i t may be desired to vary
lected and closely controlled so that any varia- the velocity from,point to point on a specimen.
tion in results can safely be attributed to Any device for m o v i n g a specimen through a
v a r i a t i o n s in the m a t e r i a l being tested. solution, or a solution past a specimen, as
1.3 When the test is used to assist in the through a tube, will be satisfactory provided
choice of material for a specific use, the test that relative motion can be held constant and
conditions should simulate the conditions of when desired, substantially uniform over the
service as closely as practical. Where the whole surface of the specimen (Note I). It
t e m p e r a t u r e and composition of the solution, s h o u l d b e recognized that at very high rates of
aeration, and similar factors vary widely under motion the effects of skin friction will reduce
service conditions, rigorous control of these the true velocity below the apparent velocity
factors is not necessary, provided all of the w'ithout, h o w e v e r , i n t e r f e r i n g with the
c o m p e t i n g materials are subjected to the same reproducibility of tests m a d e with the same
exposure conditions. apparatus.
1.4 In designing any total immersion test, NOTE 1: Example--Specimens may be moved in
consideration should be given to the various a vertical, circulating path, 2 or specimens may be
factors discussed in this method since these
This method is under the jurisdiction of ASTM Com-
factors have been found to be of importance in mittee A-I on Steel, Stainless Steel and Related Alloys, and
affecting the results obtained. is the direct responsibility of Subcommittee A01.14 on
Methods of Corrosion Testing.
Current edition accepted Sept. 30, 1963. Originally is-
2. A p p a r a t u s sued 1944. Replaces A 279 - 44 T.
2.1 Any a p p a r a t u s capable of providing the For a description of an apparatus to move specimens in
a vertical, circular path see Fraser, O. B. J., Ackerman D.
p r o p e r control of the i m p o r t a n t factors: aera- D., and Sands, J. W., "'Controllable Var ab es in the
tion, temperature, and velocity, may be used to Quantitative Study of the Submerged Corrosion of Metals,"
Industrial and Engineering Chemistry, Vol 19, 1927, pp.
achieve the required degree of reproducibility 332-338; also Searle. H. E., and LaQue, F. L., "Corrosion
in a total immersion corrosion test. Methods Testing Methods," Proceedings, Am. Soc. Testing Mats,.
Vol 35, Part II. 1935. p. 249.
245
A 279
mounted on a carrier attached to'a rotating spindle blowing air through it using an Alundum
or mounted on the spindle itself in a way that will thimble ~ or a sintered glass diffusion disk of
ensure substantially uniform velocity over the princi-
pal surfaces of the specimens when this is desired, a It medium porosity to break the air stream up
is also possible to mount a specimen on a rotating into small bubbles. Such air bubbles should be
disk or spindle, so that the surfaces of the specimens introduced at the base of a glass chimney over
move through the solution at velocities that vary
with the distance from the center of rotation. each aerator so as to prevent the impingement
Obviously, this procedure will not measure the effect of the stream of air bubbles on the test
of a single velocity but rather the combined effect of specimen.
variable velocities. So long as the other test condi-
tions are kept the same, such variable velocity tests 2.4.2 The rate of air flow required to main-
may also be expected to give reproducible results. tain air saturation will depend on the volume of
the testing solution, the area of the test speci-
2.2.2 The test velocity should approximate
men, and its rate of corrosion. The volume of
that expected in the proposed service use of the
air should be measured and controlled as
alloys or metals being tested. It may be
accurately as possible, preferably within
impossible to control the motion of boiling
• by the use of a flowmeter such as a
liquids, especially when a reflux condenser is
calibrated differential manometer, a rotame-
used to prevent rapid loss of some constituent
ter, or other suitable device. The volume of air
of the testing solution. However, the velocity
per litre of testing solution should be at least 20
induced by boiling and aeration together may
cm 3/min when the recommended solution vol-
be sufficient to give satisfactory check results,
ume to specimen area (4 litres/dm 2) is main-
even though it will not suffice to duplicate
tained. I f the indicated rate of corrosion should
service conditions involving high velocity as
exceed 300 m g / d m 2 . d a y , it may be necessary
well as high temperature.
to increase the rate of air flow or to employ
2.2.3 Where velocity appears to be the con-
some extraordinary means of supplying the
trolling factor, tests should be made at differ-
oxygen required to maintain saturation. The
ent velocities, keeping the other conditions
air should be purified by passing it through
constant, in ,some cases it may be proper to
some porous packing material, such as wool or
omit any kind of mechanical stirring; however,
excelsior, to remove suspended solids, and then
it should be recognized that zero velocity is
through a solution of sodium hydroxide (ap-
difficult to maintain and that stagnant tests
proximately 3%) to remove carbon dioxide and
shall be subjected to exceptionally careful
sulfur compounds, and finally through a water
control to achieve a p r o p e r degree of
wash bottle which also serves to humidify the
reproducibility.
air and avoid crystallization of salts in the
2.3 Temperature Control--The tempera-
pores of the aerator.
ture of the corroding solution should be con-
2.4.3 When it is desired to maintain the
trolled within • F (I C). For control testing
dissolved oxygen concentration at a value
at room temperature, it is suggested that the
lower than the point of saturation with air, this
solution be maintained at 95 • 2 F (35 • I C)
should be accomplished by altering the compo-
which, being slightly above most room temper-
sition of the saturating gas (as by the addition
atures, is easy to maintain by heat input. When
of nitrogen), rather than by altering the rate of
a water bath is used to maintain the proper
flow of the gas. Similarly, where it is desired to
temperature, the level of the water in the bath
have zero aeration, the solution should be kept
should be the same or slightly above the level
saturated with, and under, an atmosphere of an
of the solution in the test jars. The water bath
inert gas, such as oxygen-free nitrogen. Merely
should be large enough to permit free circula-
eliminating aeration will not ensure an air-free
tion of the water around the test jars.
solution, nor can reproducible results be ex-
2.4 Aeration:
pected from such attempts to achieve un-
2.4.1 The degree of aeration should be sub-
jected to close control. It is not possible to do
this by depending on diffusion from the surface s For a description of methods for attaching specimensto
of the solution to maintain uniform conditions a rotating spindle, see Journal, Am. Soc. Naval Eng., Vol
55, No. 1, February 1943, pp 64-65.
even in a well-agitated solution. To achieve air 9 4Thimble RA-98 of the Norton Co.. or equivalent, is
saturation, the solution should be aerated by satisfactory.
APPENDIXES 247
A 279
aerated solutions. As a general rule, corrosion be required corresponding in composition with

of stainless steels is retarded rather than ac- the vapors from the solution, so as to maintain
celerated by aeration. Conditions favoring the original volume within •
oxygen exclusion, therefore, favor corrosion 3.4 The volume of the test solution should
and represent just as adverse testing conditions be large enough to avoid any appreciable
for stainless steels as fully aerated solutions do changes in its corrosiveness either through the
for non-ferrous metals, and ordinary irons and exhaustion of corrosive constituents, or the
steels. accumulation of corrosion products or other
2.5 Specimen Supports--Supports for the contaminants that might affect further corro-
specimens will vary with the apparatus used, sion.
but should be designed so as to insulate 3.5 A recommended ratio between the vol-
specimens from each other, and from any ume of the test solution and the area of the
metallic container or supporting device used specimen is 250 ml/in. 2 of specimen area (4
with the apparatus. The supporting device and litres/dm ~).
container should not be affected by the corrod- 3.6 Whatever volume of test solution is
ing agent to an extent that might cause con- used, possible effects of corrosion on the
tamination of the testing solution so as to concentration of corrosive constituents should
change its corrosiveness. The shape and form be determined by analysis, and, when required,
of the specimen support should be such as to appropriate action should be taken by replac-
avoid, as much as possible, any interference ing the exhausted constituents or providing a
with free contact of the specimen with the fresh solution.
corroding solution. Where it is desired to set up 3.7 When the object of the test is to deter-
conditions favoring contact corrosion, "de- mine the effect of a metal or alloy on the
posit attack," or other forms of concentration- characteristics of the test solution (for exam-
cell action, the means by which these types of ple, the effects of metals on dyes), it is
attack are favored should be such as to ensure desirable to reproduce the ratio of solution
exact reproducibility from specimen to speci- volume to exposed metal area that exists in
men and test to test. practice. It is also necessary to take into
account the actual time of contact of the metal
3. Test Solution with the solution. If all of these factors cannot
3.1 Test solutions should be made up accu- be reproduced directly in the laboratory test,
rately, using reagents conforming to the speci- then it will be necessary to make proper
fications of the Committee on Analytical Rea- allowances as by reducing the time of contact
gents of the American Chemical Society, dis- to compensate for necessary decreases in the
solved in distilled water except in special cases. ratio of volume to area. Any necessary distor-
such as naturally occurring solutions, or those tion of the testing conditions must be taken
taken directly from some plant process. into account when interpreting the results.
3.2 The composition of any test solution
should be controlled to the fullest extent possi- 4. Test Specimens and Their Preparation
ble and, in reporting results, it should be 4.1 The size and shape of specimen will vary
described as completely and as accurately as with the purpose of the test, the nature of the
possible. Chemical content should be reported materials to be tested, and the testing appara-
either as weight percent of the solution, grams tus to be used. The size may also be limited by
per litre, or in terms of normality. the necessity of preserving a proper ratio
3.3 The composition of the test solution between the area of the specimen and the
should be checked by analysis at the end of the volume of the testing solution when the latter
test to determine the extent of any changes in must be limited. In general, an effort should be
composition, such as might result from evapo- made to have the ratio of surface to mass large
ration favored by aeration. Evaporation losses, and that of edge area to total area small.
if any, ~hould be made up by means of a 4.2 When quantitative determinations of
constant level device, or by frequent additions changes in tensile properties are to be used as
of distilled water, or other components as may the principal measure of corrosion, then ter~-
A 279
sion test specimens or a piece from which such lint. The dried specimens should then be
specimens may be cut after exposure shall be weighed with an'accuracy of • g.
used. 5 In such cases, also, a set of similar 4.8 Surface passivation sometimes has an
specimens should be preserved in a noncorro- important effect on the resistance of stainless
sive environment for comparison with the steels in certain types of nonoxidizing solu-
exposed corrosion test specimens as to tensile tions. It is, therefore, sometimes desirable to
properties. use pre-passivated specimens in corrosion
4.3 The shape and dimensions of specimens tests. This passivation may be accomplished by
shall be such as to permit weighing on an exposure of the finall3 polished specimens for I
accurate balance and to facilitate accurate h in nitric acid (30 ,height. %) at 60 C. Such
measurement and calculation of the area of passivated specimens represent the most nearly
each specimen. Such measurements of dimen- identical starting conditions possible for a
sions shall be made to the nearest 0.01 in. (0.25 series of stainless steel specimens(Note 2). If u
mm), unless for some special purpose greater passivation treatment is employed, it must be
accuracy is required. recognized that subsequent disturbance of the
4.4 All sheared edges should be trimmed surface by scraping or abrasion may greatly
beyond the shear marks by sawing, machining. affect the results obtained under certain condi-
or filing or grinding, with the final cut to be as tions of exposure.
light as possible so as to minimize hardening NoTe 2--Instead of the-passivation treatment
and distortion of the edges. just described, it may be desirable simply to clean the
4.5 When the test is being made for engi- surface of the specimens chemically bY treatment in
nitric acid (10 weight, %) at 60 C for 30 min.
neering purposes and a special finish is speci-
fied, it may be desirable to make the surface of 4.9 When the proposed application will re-
the test specimen correspond to the surface to quire welded assemblages, welded specimens
be used in service. In general, however, results approximating the thickness to be used should
that are more reproducible may be expected if be included in the test. Such specimens should
a standard surface finish for the test specimens represent the same condition of heat treatment
is used. and finish as contemplated for the service unit.
4.6 It has been shown that more uniform 4.10 To facilitate interpretation of test re-
results may be expected ifa substantial layer of' sults and their duplication by others, the details
n~etal is removed from the specimens to elimi- of the methods of preparation of the specimens
nate variations in condition of the original should be described when reporting the results
metal surface. This may be done either by a of a test.
preliminary chemical treatment Ipickling) or 4.11 The test report should include a de-
b) surfacing with a coarse abrasive paper or scription of the nature and composition of the
cloth, such as No. 50. The thickness of metal specimens (see Section 9). The composition
so removed should be at least 0.003 mm or 2 to preferably should be that actuall 3 determined
3 mg/cm 2 in the case of heavy metals. by analysis of the material from which the
4.7 The final treatment should include re- specimens were cut. If it should not be practi-
surfacing with No. 120 abrasive paper or cloth cal to provide this information, then reference
or equivalent. This resurfacing may be ex- should be made to the ~pproximate or nominal
pected to cause some surfuce work-hardening composition of the material or, as a last resort.
to an extent that will be determined by the the trade name or grade of the material may be
vigor of the surfacing operation, but is not given. The form and metallurgical condition of
ordinarily significant. The resurfaced speci- the specimen, including the nature and se-
mens should then be degreased by scrubbing quence of any hot or cold working, welding.
with clean pumice powder, followed if neces- and heat-treatment should also be described as
sary, by rinsing in water and a suitable solvent. completely as possible.
such as acetone or a mixture of 50% alcohol
and 50~ ether, and drying. The use of towels
for drying may introduce an error through 5See Figs. 7 and 8 of ASTM Methods E 8, Tension
Testing of MetaUic Materials, Annual Book of ASTM
contamination of the specimens with grease or Standards, Part 31.
APPENDIXES 2 4 9
A 279
5. Number of Specimens metals and alloys is described in Appendix X1.

5.1 In general, it is recommended that total This method can be used for stainless alloys if
immersion tests be made in duplicate. For desired.
precise work it may be desirable to test a larger 6.4 Other chemical cleaning methods that
number of specimens and for routine tests a may be used are:
single specimen may be considered sufficient. 6.4.1 Treatment in nitric acid ( 10 weight, %)
In certain types of nonoxidizing solutions at 60 C provided no chlorides are present since
highly variable results may be obtained on chlorides will promote attack of the base
stainless alloys with only slight differences in metal.
surface or exposure conditions. This should be 6.4.2 Treatment in a hot solution of sodium
borne in mind in conducting tests on stainless hydroxide (20%) containing 200 g/litre of zinc
alloys in. for instance, sulfuric acid solutions. dust may be effective in loosening deposits
5.2 Each specimen should preferably be which can then be rubbed off.
tested in a separate container, since testing 6.4.3 Immersion.of the specimens in a hot
several specimens of either the same material solution of ammonium acetate to remove rust.
or of different materials in a single container
may give erratic results. However, under spe- 7. Duration of Test
cial conditions it may be permissible, or even 7.1 The duration of any test will be deter-
desirable, to test more than one specimen in a mined by its nature and purpose. In some cases
single container provided it is recognized that it will be desirable to expose a number o]"
the corrosion products from a specimen show- specimens so~ that certain of them can be
ing a high rate of attack may accelerate removed after definite time intervals so as to
corrosion of another specimen, or specimens. provide a measure of change of corrosion rates
in the same container. with time. Any procedure that requires re-
moval of solid corrosion products between
6. Methods of Cleaning Specimens After Test periods of exposure of the same specimens will
6.1 It is essential that corrosion products be not measure accurately normal changes of
removed from specimens if changes in weight corrosion with time.
are to be used as a measure of corrosion. There 7.2 The higher the rate of corrosion, the
are man 3 satisfactory means of cleaning speci- shorter may be the testing period.
mens after exposure, such as the use of bristle 7.3 Where the object of the test is to predict
brushes with mild abrasives and detergents. corrosion rates over a long period, it is obvi-
treatment with appropriate chemical solutions, ously desirable to run the test for as long as
especially suitable with certain metals and may be practical, provided that the testing
corrosion products, and electrolytic methods. conditions and the corrosive characteristics of
The use of bristle brushes should ordinarily be the solution can be maintained constant over a
limited to heavily corroded specimens. Drastic long test period.
cleaning methods should not be used when the
specimens are small or the amount of weight 8. Interpretation of Results
change expected is slight. For most tests on 8.1 After the corroded specimens have been
stainless alloys scrubbing of the specimen with cleaned, they should be reweighed with the
a rubber stopper under running water has been same accuracy as the original weighing
found adequate. (• g). It will then be possible to calcu-
6.2 Whatever the treatment, its effect in late the loss in weight per unit of area during
removing metal, if any, should be determined the test period. This may be used as the
for each material and the results of weight loss principal measure of corrosion.
determinations should be corrected accord- 8.2 Corrosion rates calculated from the loss
ingly. The method of cleaning should be rein weight data should be reported in milligrams
po rted. per square decimetre per day (24 h), abbrevi-
6.3 An electrolytic cleaning method that has ated mdd. Factors for converting these units to
been found to be useful with a large number of other common corrosion.rate units are given in
A 279
Appendix X2. The expression of corrosion loss 8.9 Under certain conditions the stainless
as a percentage of original weight is usually steels are susceptible to stress-corrosion crack-
valueless. ing. This effect may be studied by the exposure
8.3 Corrosion rates in milligrams per square to the solution of specimens previously stressed
decimetre per day may be expressed in terms of to a known degree by some suitable method.
inches penetration per year (abbreviated ipy)
by the following equation: 9. Report
ipy = mdd • (0.001437/d) 9.1 To the fullest extent that may be possi-
ble, the investigator should follow the recom-
where:
mendations embodied in the ASTM Manual
d = density of the metal, g/cmL
on Quality Control of Materials. e In any event
8.4 It should be remembered always that
the report should include the following infor-
any calculations of corrosion rates, such as
mation:
"'mdd" or "ipy," will be subject to error on
9.1.1 The chemical compositions of the met-
account of nonuniform distribution of corro-
als and alloys tested (see 4.11),
sion and changes of corrosion rates with time.
9.1.2 The exact size, shape, and area of the
In connection with the liatter, it is often
specimen,
desirable to carry out the testing program so as
9.1.3 The forms and metallurgical condi-
to provide data from which curves can be
tions of the specimens,
plotted to ill.ustrate changes in corrosion rates
9.1.4 The treatment used to prepare speci-
with time.
mens for test,
8.5 After reweighing, the specimens should
9.1.5 The number of specimens of each
be ",examined carefully and the average and
material tested, whether each specimen was
maximum depths of pits, if any are present,
tested in a separate container or which speci-
determined by means of a calibrated micro-
mens were tested in the same container.
scope, or by direct measurement with a depth
9.1.6 The chemical composition of the test-
gage or sharp pointed micrometers. If the
ing solution and information as to how and to
number of pits is very large, it should suffice to
what extent the composition was held constant
report the average depth of the ten deepest pits.
or how frequently the solution was replaced,
8.6 The depths of pits should be reported in
9.1.7 The temperature of the testing solu-
thousandths of an inch for the test period. The
tion and the maximum variation in tempera-
size, shape, and distribution of pits should be
ture during the test,
noted. A distinction should be made between
9.1.8 The degree of aeration of the solution
local attack or pitting that occurred under-
in terms of cubic centimetres of air per litre of
neath supporting devices and those pits that
solution per minute and the maximum varia-
developed at the surfaces that had been ex-
tion in this flow, or similar information for any
posed freely to the testing solution.
gas or mixture of gases other than air. The type
8.7 For special purposes it may be desirable
of aerator should also be described,
to subject the specimen to simple bending tests
9.1.9 The v,elocity of relative movement
and microscopical examination to determine
between the test specimens and the solution
whether any embrittlement or intergranular
and a description of how this movement was
attack has occurred. Electrical resistance
effected and controlled,
measurements of specimens of special type
9.1.10 The volume of the testing solution,
may be employed for studying these effects by
9.1.11 The nature of the apparatus used for
comparing with specimens not subjected to
the test,
test.
9.1.12 The duration of the test or of each
8.8 With suitable specimens, it may be
part of it if made in more than one stage,
possible to make quantitative mechanical tests
9.1.13 The method used to clean specimens
comparing the exposed specimens with uncor-
after exposure and the extent of any error
roded specimens reserved for the purpose. By
such means the effects of corrosion may be
observed by measuring changes in mechanical
e Issued as Special Technical Publication 15-C. Decem-
properties. ber 1957.
APPENDIXES 251
A 279
introduced by this t r e a t m e n t , exposure if d e t e r m i n e d , results of microscopi-

9.1.14 T h e actual weight losses of the sev- cal e x a m i n a t i o n or qualitative bond tests, and
eral specimens, depths of pits (plus notes on 9.1.15 Corrosion rates for individual speci-
their size, shape and distribution, as by sketch), m e n s c a l c u l a t e d in m i l l i g r a m s per square dec-
d a t a on mechanical properties before and after i m e t r e per day.
APPENDIXES
XI. M E T H O D FOR ELECTROLYTIC CLEANING OF CORROSION TEST SPECIMENS

AFTER EXPOSURE
XI.1 After scrubbing to remove loosely attached XI.2 After the electrolytic treatment, the speci-
corrosion products, treat the specimens as a cathode men should be scrubbed. The weight losses of
in hot, diluted sulfuric acid under the following specimens 0.5 dm ~ in area treated by the method
conditions: described have been found to be less than 0.0002 mg.
XI.3 Instead of using 2 ml of any proprietary
Test solution sulfuric acid (5 weight, %)
inhibitor, about 0.5 g/litre of such inhibitors as
Inhibitor 2 ml organic inhibitor/litre of diorthotolyl thiourea, quinoline ethiodide, beta-
solution naphthol quinoline may be used.
Anode carbon
XI.4 It should be noted that this electrolytic
Cathode test specimen treatment may result in the redeposition of adherent
Cathode current 20 A / d m ~ metal from reducible corrosion products and thus
density
lower the apparent weight loss. However, general
Temperature 165 F (74 C) experience has indicated that in most cases of
Exposure period 3 min corrosion in liquids the possible errors from this
source are not likely to be serious.
X2. CONVERSION FACTORS
Multiply By To Obtain
Grams per square inch per hour 372 000 milligramsper square deeimetre per
day (mdd)
Grams per square metre per year 0.0274 milhgrams per square decimetre per
day (mdd)
Mdhgrams per square decimetre 0.0003277 ounces per square foot
Milligrams per square decimetre per 0.00000269 grams per square inch per hour
day (mdd)
Milligrams per square decimetre per 0.001437/densit.~of metal in g/cm a penetration
inches per year
day (mddp
Milligrams per square decimetre per 0.0001198/densityof metal in g/cm 3 penetrationinches per month
day (mdd)
Milligrams per square decimetre per 36.5 grams per square metre per year
day (mdd)
Milligrams per square decimetre per 0.00365/densityof metal in g/cm s penetration
centlmetres per year
day (mdd)
Milligrams per square decimetre per 0.00748 pounds per square foot per year
day (mdd)
Ounces per square foot 3052 milligramsper square decimetre
Pounds per square foot per year 133.8 milligramsper square decimetre per
day (mdd)
t Factors for converting milligrams per square decimetre per day to inches penetration per year. for different AISI types
.of stainless steels are given in Table XI.
@ A 279
TABLE Xi Factors for Converting Milligrams per
Square Decimetre per Day to Inches Penetration per Year
for Different Types of Stainless Steel
Multiply Corrosion Rate in

AISI Type No. m g / d m 2 day by indicated
lactor to get in. penetration
per year
410 0.000186
430 0.000186
446 0.000189
302 0,000182
304 0 000182
308 0.000182
309 0.000182
310 0.000182
316 0.000180
317 0.000180
321 0.000182
347 0.000180
By publication of this standard no position is taken with respect to the validity of any patent rights in connection
therewitlh and the A merican Society for Testing and Materials does not undertake to insure anyone utilizing the standard
against liability for infringement of any Letters Patent nor assume any such liabthtv.
STP534-EB/Nov. 1973
•l• ) Designation: B 117-73 American National Standard Z 118.1

American National Standards institute
Endorsed by American
Electroplaters" Society
Endorsed by National
Association of Metal Finishers
Federation of Societies for
Paint Technology Standard No. Ld 18-62
Standard Method of
SALT SPRAY (FOG) TESTING ]
This Standard is issued under the fixed designation B 117; the number immediately following the designation indicates the
last reapproval.
This method has been approved by the Department of Defense to replace method 811.1 of Federal Test Method Standard
No. 151b and for listing in DoD Index of Specifications and Standards. Future proposed revtsions should be coordinated with
the Federal Government through the Army Materials and Mechanics Research Center, Watertown, Mass. 02172.
].scope 3. Test Specimens

1.1 This method sets forth the conditions 3.1 The type and number of test specimens
required in salt spray (fog) testing for specifi- to be used, as well as the criteria for the ev-
c a t i o n p u r p o s e s . S u i t a b l e a p p a r a t u s which aluation of the test results, shall be defined in
may be used to obtain these conditions is de- the specifications covering the material or
scribed in Appendix AI. The method does n o t product being tested or shall he mutually
prescribe the type of test specimen or expo- agreed upon by the purchaser and the seller.
sure periods to be used for a specific product,
nor the interpretation to be given to the re- 4. Preparation of Test Specimens
sults. Comments on the use of the test in re- 4.1 Metallic and metallic-coated specimens
search will be found in Appendix A2. shall be s u i t a b l y cleaned. T h e cleaning
NOTE I--This method is applicable to salt spray method shall be optional depending on the
(fog) testing of ferrous and non-ferrous metals, and is nature of the surface and the contaminants,
also used to test inorganic and organic coatings, etc.,
especially where such tests are the basis for material except that it shall not include the use of abra-
or product specifications. sives other than a paste of pure magnesium
oxide nor of solvents which are corrosive or
will d e p o s i t either c o r r o s i v e or p r o t e c t i v e
2. Apparatus
films. The use of a nitric acid solution for the
2.1 The apparatus required for salt spray chemical cleaning, or passivation, of stainless
(fog) testing consists of a fog chamber, a salt steel specimens is permissible when agreed
solution reservoir, a supply of suitably condi- upon by the purchaser and the seller. Care
tioned compressed air, one or more atomizing shall be taken that specimens are not recon-
nozzles, s p e c i m e n s u p p o r t s , p r o v i s i o n for taminated after cleaning by excessive or
heating the chamber, and necessary means of careless handling.
control. The size and detailed construction of 4.2 Specimens for evaluation of paints and
the apparatus are optional, provided the con- other organic coatings shall be prepared in ac-
ditions obtained meet the requirements of this cordance with applicable specification(s) for
method. the material(s) being tested, or as agreed upon
2.2 Drops of solution which accumulate on by the purchaser and supplier. Otherwise, the
the ceiling or cover of the chamber shall not test specimens shall consist of steel meeting
be permitted to fall on the specimens being the requirements of A S T M Methods D609
tested. for Preparation of Steel Panels for Testing
2.3 Drops of solution which fall from the
specimens shall not be returned to the solution This method is underthe jurisdiction of ASTM Com-
reservoir for respraying. mittee G-I on Corrosionof Metals, and is the direct respon-
sibility of SubcommitteeG01.05 on Laboratory Corrosion
2.4 Material of construction shall be such Tests.
that it will not affect the corrosiveness of the Current edition approved March 29, 1973. Published
June 1973. Originallypublishedas B 117 - 39 T, Last prc-
fog. viouseditionB 117 04.
253
Bl17
Paint, Varnish, Lacquer, and Related Prod- strips are suitable for the support of flat pancls.
Suspension from glass hooks or waxed string may
ucts, ~ and shall be cleaned and prepared for be used as long as the specified position of the speci-
coating in accordance with applicable proce- mens is obtained, If necessary by means of second-
dure of Method D 609. ary support at the bottom of the specimens.
4.3 S p e c i m e n s coated with p a i n t s or
6. Salt Solution
nonmetallic coatings shall not be cleaned or
handled excessively prior to test. 6.1 The salt solution shall be prepared by
4.4 Whenever it is desired to determine the dissolving 5 • I parts by weight of sodium
development of corrosion from an abraded chloride in 95 parts of distilled water or water
area in the paint or organic coating, a scratch containing not more than 200 ppm of total
or scribed line shall be made through the coat- solids. The salt used shall be sodium chloride
ing with a sharp instrument so as to expose substantially free of nickel and copper and con-
the underlying metal before testing. The containing on the dry basis not more than 0.1 per-
ditions of making the scratch shall be as de- cent of sodium iodide and not more than 0.3
fined in ASTM Method D 1654, Evaluation of percent of total impurities. Some salts contain
Painted or Coated Specimens Subjected to additives that may act as corrosion inhibitors;
Corrosive Environments, 2 unless otherwise careful attention should be given to the chemi-
agreed upon between the purchaser and seller. cal content of the salt. By agreement between
4.5 Unless otherwise specified, the cut purchaser and seller, analysis may be required
edges of plated, coated, or duplex materials and limits established for elements or com-
and areas containing identification marks or pounds not specified in the chemical composi-
in contact with the racks or supports shall be tion given above.
protected with a suitable coating stable under 6.2 The pH of the salt solution shall be such
the conditions of the test, such as ceresin wax. that when atomized at 35 C (95 F) the collected
Nor~ 2--Should it be desirable to cut test speci- solution will be in the pH range of 6.5 to 7.2
mens from parts or from preplated, painted, or oth- (Note 4). Before the solution is atomized it
erwise coated steel sheet, the cut edges shall be pro- shall be free of suspended solids (Note 5). The
tected by coating them with paint, wax, tape, or pH measurement shall be made electrometri-
other effective media so that the development of a
galvanic effect between such edges and the adjacent cally at 25 C (77 F) using a glass electrode with
plated or otherwise coated metal surfaces, is a saturated potassium chloride bridge in ac-
prevented. cordance with Method E 70, Test for pH of
5. Position of Specimens During Test Aqueous Solutions with the Glass Electrodea;
or colorimetrically using bromothymol blue as
5.1 The position of the specimens in the indicator, or short range pH paper which reads
salt spray chamber during the test shall be in 0.2 or 0.3 ofa pH unit (Note 6).
such that the following conditions are met:
NoTE4--Temperature affects the pH of a salt
5.1.1 Unless otherwise specified, the speci- solution prepared from water saturated with carbon
mens shall be supported or suspended between dioxide at room temperature and pH adjustment
15 and 30 deg from the vertical and preferably may be made by the following three methods:
parallel to the principal direction of horizontal (I) When the pH of a salt solution is adjusted at
room temperature, and atomized at 35 C (95 F), the
flow of fog through the chamber, based upon pH of the collected solution.will be higher than the
the dominant surface lacing tested. original solution due to the Joss of carbon dioxide at
5.1.2 The specimens shall not contact each the higher temperature. When the pH of the salt
solution is adjusted at room temperature, it is there-
other or any metallic material or any material fore necessary to adjust it below 6.5 so the collected
capable of acting as a wick. solution after atomizing at 35 C (95 F) will meet the
5.1.3 Each specimen shall be so placed as pH limits of 6.5 to 7.2. Take about a 50-ml sample
of the salt solution as prepared at room tempera-
to permit free settling of fog on all specimens. ture, boil gently for 30 s, cool, and determine the
5.1.4 Salt solution from one specimen shall pH. When the pH of the salt solution is adjusted
not drip on any other specimen. to 6.5 to 7.2 by this procedure, the pH of the
atomized and collected solution at 35 C (95 F) will
NOTE3--Suitable materials for the construction come within this range.
or coating of racks and supports are glass, rubber,
plastic, or suitably coated wood. Bare metal shall
not be used. Specimens shall preferably be sup Annual Book of A S T M Standards, Part 21.
ported from the bottom or the side. Slotted wooden J Annual Book of A S T M Standards. Parts 16. 22, 30.
APPENDIXES 255
B 117
(2) Heating the salt solution to boiling and cool- placed w i t h i n the e x p o s u r e zone t h a t no d r o p s
ing to 95 F or maintaining it at 95 F for approxi-
mately 48 h before adjusting the pH produces a so- o f solution f r o m the test s p e c i m e n s or a n y
lution the pH of which does not materially change o t h e r source shall be collected. T h e collectors
when atomized at 35 C (95 F). shall be placed in the p r o x i m i t y o f the test
(3) Heating the water from which the salt solu- s p e c i m e n s , one n e a r e s t to a n y nozzle a n d the
tion is prepared to 35 C (95 F) or above, to expel
carbon dioxide, and adjusting the pH of the salt o t h e r f a r t h e s t f r o m all nozzles. T h e fog shall
solution within the limits of 6.5 to 7.2 produces a be such t h a t for e a c h 80 c m ~ o f h o r i z o n t a l col-
solution the pH of which does not materially lecting a r e a there will be collected in e a c h col-
change when atomized at 35 C (95 F).
NOTE 5--The freshly prepared salt solution may lector f r o m 1.0 to 2.0 ml o f s o l u t i o n per h o u r
be filtered or decanted before it is placed in the res- b a s e d on a n a v e r a g e run of at least 16 h ( N o t e
ervoir, or the end of the tube leading from the solu- 10). T h e s o d i u m c h l o r i d e c o n c e n t r a t i o n o f the
tion to the atomizer may be covered with a double
layer of cheesecloth to prevent plugging of the noz- collected solution shall be 5 4- I w e i g h t per-
zle. cent ( N o t e l l ) . T h e p H o f the collected solu-
NOTE ~ T h e pH can be adjusted by additions of tion shall be 6.5 to 7.2. T h e p H m e a s u r e m e n t
dilute cp hydrochloric acid or cp sodium hydroxide
solutions. shall be m a d e e l e c t r o m e t r i c a l l y o r c o l o r i m e t -
rically using b r o m o t h y m o l blue as the indica-
7. Air S u p p l y
tor.
7.1 T h e c o m p r e s s e d a i r supply to the nozzle NOT~ 10~Suitable collecting devices are glass
or nozzles for a t o m i z i n g the salt s o l u t i o n shall funnels with the stems inserted through stoppers
be free o f oil a n d d i r t ( N o t e 7) a n d m a i n - into graduated cylinders, or crystallizing dishes.
Funnels and dishes with a diameter of l0 cm have
t a i n e d between 69 a n d 172 k N / m ~ (10 a n d 25 an area of about 80 cm 2.
psi) ( N o t e 8). NOT~ l I - - A solution having a specific gravity of
NOTE 7--The air supply may be freed from oil 1.0255 to 1.0400 at 25 C (77 F) will meet the con-
and dirt by passing it throu[gh a water scrubber or at centration requiremenL The concentration may also
least 610 mm (2 ft) of statable cleaning material be determined as follows: Dilute 5 ml of the col-
such as asbestos, sheep's wool, excelsior, slag lected solution to 100 ml with distilled water and
wool, or activated ~ alumina. mix thoroughly: pipet a 10-ml aliquot into an
NOTE 8--Atomizing nozzles may have a "critical evaporating dish or casserole: add 40 ml of distilled
pressure" at which an abnormal increase in the cor- water and 1 ml of 1 percent potassium chromate
rosiveness of the salt fog occurs. If the "critical solution (chloride-free) and titrate with 0.1 N silver
pressure" of a nozzle has not been established with nitrate solution to the first appearance of a perma-
certainty, control of fluctuation in the air pressure nent red coloration. A solution that requires be-
within plus or minus 0.7 kN/m'-' (0.1 psi), by in- tween 3.4 and 5.1 ml of 0.1 N silver nitrate solution
stallation of a suitable pressure regulator valve 5 will meet the concentration requirements.
minimizes the possibility that the nozzle will be
8.3 T h e nozzle or nozzles shall be so di-
operated at its "critical pressure. ''~
rected o r baffled t h a t n o n e o f the s p r a y c a n
8. Conditions in the Salt Spray Chamber i m p i n g e directly on the test s p e c i m e n s .
8.1 Temperature--The e x p o s u r e z o n e o f
the salt s p r a y c h a m b e r shall be m a i n t a i n e d a t 9. Continuity of Test
35 + 1.1 - 1.7 C (95 + 2 - 3 F). T h e 9.1 U n l e s s o t h e r w i s e specified in the speci-
t e m p e r a t u r e within the e x p o s u r e zone o f the fications covering the material or product
closed c a b i n e t shall be r e c o r d e d a t least twice b e i n g tested, the test shall be c o n t i n u o u s for
a d a y a t least 7 h a p a r t (except o n S a t u r d a y s , the d u r a t i o n o f the entire test period. C o n t i n -
S u n d a y s , a n d h o l i d a y s w h e n the salt s p r a y u o u s o p e r a t i o n implies t h a t the c h a m b e r be
test is n o t i n t e r r u p t e d for e x p o s i n g , r e a r r a n g -
ing, o r r e m o v i n g test s p e c i m e n s or to c h e c k
Registered U. S. Patent Office.
a n d replenish the solution in the reservbit'). The Nullmatic pressure regulator (or equivalent) man-
NOTE9--A suitable method to record the tem- ufactured by Moore Products Co., H and Lycoming Sts.,
perature is by a continuous recording device or by a Philadelphia, Pa. 19124, is suitable for this purpose.
It has been observed that periodic fluctuations in air
thermometer which can be read from outside the pressure of 4-3.4 kN/m ~ (0.5 psi) resulted in about a two-
closed cabinet. The recorded temperature must be fold increase in the corrosivity of the fog from a nozzle
obtained with the salt spray chamber closed to which was being operated at an average pressure of 110
avoid a false low reading because of wet-bulb effect kN/m ~ 0 6 psi). Controlling the fluctuations within 4-0.7
when the chamber is open. kN/m z (0.1 psi), however, avoided any increase in the
corrosivity of the salt fog. See Darsey, V. M. and Cava-
8.2 Atomization and Quantity of Fog - A t hugh, W. R., "Apparatus and Factors in Salt Fog Test-
ing," Proceedings,ASTEA, Am. Soc. Testing Mats., Vol.
l e a s t t w o c l e a n f o g c o l l e c t o r s s h a l l be so 48, 1948, p. 153.
Bl17
closed and the spray operating continuously material or product being tested or by agree-
except for the short daily interruptions neces- ment between the purchaser and the seller.
sary to inspect, rearrange, or remove test spe-
cimens; to check and replenish the solution in 13. Records and Reports
the reservoir, and to make necessary record- 13.1 The following i n f o r m a t i o n shall be
ings as described in Section 8. Operations recorded, unless otherwise prescribed in the
shall be so scheduled that these interruptions specifications covering the material or product
are held to a minimum. being tested:
13.1.1 Type of salt and water used in pre-
10. Period of Test
paring the salt solution,
10.1 The period of test shall be as desig- 13.1.2 All readings of temperature within
nated by the specifications covering the mate- the exposure zone of the chamber,
rial or product being tested or as mutually 13.1.3 Daily records of data obtained from
agreed upon by the purchaser and the seller. each fog-collecting device including the fol-
NOTE I2 Recommended exposure periods are lowing:
to be as agreed upon by the purchaser and seller, but 13.1.3.1 Volume of salt solution collected in
exposure periods of multiples of 24 h are suggested.
milliliters per hour per 80 c m 2,
!1. Cleaning of Tested Specimens 13.1.3.2 Concentration or specific gravity
at 35 C (95 F) of solution collected, and
11.1 Unless otherwise specified in the spec-
13.1.3.3 pH of collected solution.
ifications covering the material or product
13.4 Type of specimen and its dimensions,
being tested, specimens shall be treated as fol-
or number or description of part,
lows at the end-of the test:
13.5 Method of cleaning specimens before
11.1.1 The specimens shall be carefully
and after testing,
removed.
13.6 Method of supporting or suspending
11.1.2 Specimens may be gently washed or
article in the salt spray chamber,
dipped in clean running water not warmer
13.7 Description of protection used as re-
than 38 C (100 F) to remove salt deposits
quired in 4.5,
from their surface, and then immediately
13.8 Exposure period,
dried. Drying shall be accomplished with a
13.9 Interruptions in test, cause and length
stream of clean, compressed air.
of time, and
12. Evaluation of Results 13.10 Results of all inspections.
12.1 A careful and immediate examination NOTE 13--1f any of the atomized salt solution
shall be made for the extent of corrosion of which has not contacted the test specimens is re-
turned to the reservoir, it is advisable to record the
the dry test specimens or for other failure as concentration or specific gravity of this solution
required by the specifications covering the also.
APPENDIXES
A 1. CONSTRUCTION OF APPARATUS
AI.I Cabinets tuFe.

AI.I.I Standard salt-spray cabinets are available AI.I.3 Accessories such as a suitable adjustable
from several suppliers, but certain pertinent acces- baffle or central fog tower, automatic level control
sories are required before they will function accord- for the salt reservoir, and automatic level control for
ing to this method and provide consistent control for the air-saturator tower are pertinent parts of the
duplication of results. apparatus.
AI.I.2 The salt spray cabinet consists of the Al.l.4 The cabinet should be of sufficient size to
basic chamber, an air-saturator tower, a salt solu- test adequately the desired number of parts without
tion reservoir, atomizing nozzles, specimen sup- overcrowding. Small cabinets have been found diffi-
ports, provisions for heating the chamber, and suita- cult to control and those of less than 0.43-m~ 05-
ble controls for maintaining the desired tempera- ft:~) capacity should be avoided.
APPENDIXES 257
Bl17
AIA.5 The chamber may be made of inert mate- selected. Nozzles are not necessarily located at one
rials such as plastic, glass, or stone but most prefer- end, but may be placed in the center and can also be
ably s constructed of metal and lined with ~mper- directed vertically up through a suitable tower.
vious plastics, rubber, or epoxy-type materials or
equivalent. AI.4 Air for Atomization
AIA.1 The air used for atomization must be free
AI.2 Temperature Control of g r e a s e , oil, and d i r t before use by p a s s i n g
AI.2.1 The maintenance of temperature within through well-maintained filters. Room air may be
the salt chamber can be accomplished by several compressed, heated, humidified, and washed in a
methods. It is generally desirable to control the water-sealed rotary pump, if the temperature of the
temperature of the surroundings of the salt spray water is suitably controlled. Otherwise cleaned air
chamber and to maintain it as stable as possible. may be introduced into the bottom of a tower filled
This may be accomplished by placing the apparatus with water, through a porous stone or multiple noz-
in a constant-temperature room, but may also be zles. The level of the water must be maintained au-
achieved by surrounding the basic chamber of a tomatically to ensure adequate humidification. A
jacket containing water or air at a controlled ~em- chamber operated according to this method and
pcrature. Appendix will have a relative humidity between 95
AI.2.2 The use of immersion heaters in an inter- and 98 percent. Since salt solutions from 2 to 6 per-
nal salt-solution reservoir or of heaters within the cent will give the same results (though for uniform.
chamber is detrimental where heat losses are appre- ity the limits are set at 4 to 6 percent), it is
ciable, because of solution evaporation and radiant preferable to saturate the air at temperatures well
heat on the specimens. above the chamber temperature as insurance of a
AI.2.3 All piping which contacts the salt solution wet fog. Table A2 shows the temperatures, at differ-
or spray should be of inert materials such as plastic. ent pressures, that are required to offset the cooling
Vent piping should be of sufficient size so that a effect of expansion to atmospheric pressure.
minimum of back pressure exists and should be in- AI.4.2 Experience has shown that most uniform
stalled so that no solution is trapped. The exposed spray chamber atmospheres are obtained by in-
end of the vent pipe should be shielded from ex- creasing the atomizing air temperature sufficiently
treme air currents that may cause fluctuation of to offset heat losses, except those that can be re-
pressure or vacuum in the cabinet. placed otherwise at very low-temperature gradients.
AI.3 Spray Nozzles AI.5 Types of Construction

A I.3.1 Satisfactory nozzles may be made of hard AI.5.1 A modern laboratory cabinet is shown in
rubber, plastic, or other inert materials. The most Fig. AI. Walk-in chambers are not usually con-
commonly used type is made of plastic. Nozzles cal- structed with a sloping ceiling due to their size and
ibrated for air consumption and solution atomized location. Suitably located and directed spray noz-
are available. The operating characteristics of a typ- zles avoid ceiling accumulation arid drip. Nozzles
ical nozzle are given in Table AI. may be located at the ceiling, or 0.91 m (3 It) from
A1.3.2 It can readily be seen that air consump- the floor directed upward at 30 to 60 deg over a
tion is relatively stable at the pressures normally passageway. The number of nozzles depends on
used, but a marked reduction in solution sprayed type and capacity and is related to the area of the
occurs if the level of the solution is allowed to drop test space. A 11 to 19-dm ~ (3 to 5-gal) reservoir
appreciably during the test. Thus, the level of the is required within the chamber, with the level con-
solution in the salt reservoir must be maintained trolled. The major features of a walk-in type cabi-
automatically to ensure uniform fog delivery during net, which differs significantly from the laboratory
the test. ~ type, are illustrated in Fig. A2. Construction of a
AI.3.3 If the nozzle selected does not atomize plastic nozzle, such as is furnished by several sup-
the salt solution into uniform droplets, it will be pliers, is shown in Fig. A3.
necessary to direct the spray at a baffle or wall to
.pick up. the larger drops and prevent them from
~mptngmg on the test specimens. Pending a com- A suitable device for maintaining: the level of liquid in.
plete understanding of air-pressure effects, etc., it is either the saturator tower, or rescrvo,r of test solution may
tmportant that the nozzle selected shall produce the be designed by a local engineering I~(oup. or may br pur-
desired condition when operated at the air pressure chased from manufacturers of test cabinets as an accessory.
A 2 . UsE OF THE SALT SPRAY (FOG) TEST IN RESEARCH
A2.1 The detailed requirements of this method ASTM Salt Spray Tests: Method B 117, Method
are primarily for quality acceptance and should not B 287, Acetic Acid-Salt Spray (Fog) Testing, s and
be construed as the optimum conditions for research Method B 368, for Copper-Accelerated Acetic Acid-
studies. The test has been used to a considerable Salt Spray (Fo~) Testing (CASS), a into useful
extent for the purpose of comparing different mate- tools for many industrial and military production
rials or. finishes with an acceptable standard. The
recent elimination of many cabinet variables and
the improvement in controls have made the three A nnual Book ofA STM Standards, Part 7~
~[~ B 117
and qualification programs. (chromated, phosphated, or anodized), although

A2.2 The test has been used to a considerable final conclusions regarding the validity of test re-
extent for the purpose of comparing different mate- sults related to service experience have not been
rials or finishes. It should be noted that there is sel- reached. Method B 117 is considered to be most
dom a direct relation between salt spray (fog) resis- useful in estimating the relative behavior of closely
tance and resistance to corrosion in other media, related materials in marine atmospheres, since it
because the chemistry of the reactions, including simulates the basic conditions with some accelera-
the formation of films and their protective value, tion due to either wetness or temperature or both.
frequently varies greatly with the precise conditions A2.3 When a test is used for research~ it may
encountered. Informed personnel are aware of the prove advantageous to operate with a different solu-
erratic composition of basic alloys, the possibility of tion composition or concentration or at a different
wide variations in quality and thickness of plated temperature. In all cases, however, it is desirable to
items produced on the same racks at the same time, control the temperature and humidity in the manner
and the consequent need for a mathematical deter- specified, and to make certain that the composition
mination of the number of specimens required to of the settled fog and that of the solution in the
constitute an adequate sample for test purposes, in reservoir are substantially the same. Where differ-
this connection it is well to point out that Method ences develop, it is necessary to control conditions
B l l 7 is not applicable to the study or testing of so that the characteristics of the settled fog meet the
decorative chromium plate (nickel-chromium or specified requirements for the atmosphere.
copper-nickel-chromium) on steel or on zinc-base A2.4 Material specifications should always be
die castings or of cadmium plate on steel. For this written in terms of the standard requirements of the
purpose Methods B 287 and B 368 are available, appropriate salt-spray method, thereby making it
which are also considered by some to be superior possible to test a variety of materials from different
for comparison of chemically-treated aluminum sources in the same equipment.
TABLE AI Operating Characteristics TABLE A2 Temperature and Pressure

of Typical Spray Nozzle Requirements for Operation of
Test at 95 F
Air Flow, Solution
Siphon liters/rain Consumption, ml/h Air Pressure, psi
Height,
in. Air Pressure, psi Air Pressure, psi 12 14 16 18
5 10 15 20 5 10 15 20 Temperature, deg F 114 117 119 121
4 19 26.5 31.5 36 2100 3840 4584 5256 Air Pressure. kNlm'
8 19 26.5 31.5 36 636 2760 3720 4320
12 19 26.5 31.5 36 0 1380 3000 3710 83 96 llO 124
16 19 26.6 31.5 36 0 780 2124 2904 Tempcra'ture, deg C 46 47 48 49
Solution Consumption,
Air Flow, dm'/min cm*/h
Siphon
Height Air Pressure, kN/m ~ Air Pressure, kN/m =
cm
34 69 103 138 34 69 103 138
10 19 26.5 31,5 36 2100 3840 4584 5256
20 19 26.5 31.5 36 636 2760 3720 4320
30 19 26.5 31.5 36 0 1380 3000 3710
40 19 26.6 31.5 36 0 780 2124 2904
APPENDIXES 259
II6~ B 117
13'
!--====:17 ~
2O
20
f
12A
2
2O
I I
0 - - Angle of lid, 90 to 125 deg
I Thermometer and thermostat for controlling heater (Item No. 8) in base
2 - - Automatic water levelling device
3 - - Humidifying tower
4 Automatic temperature regulator for controlling heater (Item No. 5)
5 - - Immersion heater, non-rusting
6 - - Air inlet, multiple openings
7 - - Air tube to spray nozzle
8 Strip heater in base
9 - - Hinged top, hydraulically operated, or counterbalanced
1 0 - - Brackets for rods supporting specimefis, or test table
11 - - Internal reservoir
12 - - Spray nozzle above reservoir, suitably designed, located, and baffled
12A - - Spray nozzle housed in dispersion tower located preferably in center of cabinet
13 Water Seal
1 4 - - Combination drain and exhaust. Exhaust at opposite side of test space from spray nozzle (Item 12), but preferably in
combination with drain, waste trap, and f o r c ~ draft waste pipe (Items 16. 17, and 19).
16 Complete separation between forced draft waste pipe (Item 17) and combination drain and exhuast (Items 14 and 19)
to avoid undesirable suction or back pressure.
17 - - Forced draft waste pipe.
18 Automatic levelling device for reservoir
19 - - Waste trap
20 Air space or water jacket
21 - - T e s t table or rack, well below roof area
FIG. AI Typical Salt Spray Cabinet.
NOTE--The controls are the same in general as for the

aboratory cabinet (Fig. A l L but are sized to care for the
larger cube. The chamber has the following features:
(1) Heavy insulation,
(2) Refrigeration door with drip rail. or pressure door
with drip rail, inward-sloping sill,
(3) Low-temperature auxdiary heater, and
(4) Duck hoards on floor, with floor sloped to comb/na-
3 tion drain and air exhaust.
FIG. A2 Walk-in Chamber, !.5 by 2.4 mm (5 by 8 It) and

Upward in Over-all Size.
~ B 117
Air ~
t
Solution
FIG. A3 Typical Spray Nozzle.
By publication o f this standard no position is taken with respect to the validity o/any patent rights in connection there-
with, and the American Society for Testing and Materials does not undertake to insure anyone utilizing the standard
against liability/or infringement o/any Letters Patent nor assume any such liability.
STP534-EB/Nov. 1973
(~l~ Designation: G 1 - 72
Standard Recommended Practice for

PREPARING, CLEANING, AND EVALUATING
C O R R O S I O N TEST S P E C I M E N S 1
This Standard is issued under the fixed designation G 1; the number immediately followingthe designation indicates the
year of original adoption or. in the case of revision,the year of last revision. A numberin parenthesesindicates the year of
last reapproval.
1. Scope 3.1.2.2 Pickle in an appropriate solution

1.1 This recommended practice gives sug- (in some cases the chemical cleaners de-
gested procedures for preparing bare, solid scribed in Section 5 will suffice) if oxides or
metal specimens for laboratory corrosion tarnish are present.
tests, for removing corrosion products after 3.1,2.3 Abrade with a slurry of an appro-
the test has been completed, and for evalu- priate abrasive or with an abrasive paper (see
ating the corrosion damage that has occurred. Methods A 262 and A 279 and Recom-
Emphasis is placed on procedures related to mended Practice D 1384). The edges as well
the evaluation of corrosion by mass-loss and as the faces of the specimens should be
pitting measurements. abraded to remove burrs.
3.1.2.4 Rinse thoroughly and dry.
2. Applicable Documents 3.2 Metallurgical C o n d i t i o n - - W h e n spec-
2.1 A S T M Standards: imen preparation changes the metallurgical
A 262, Recommended Practice for De- condition of the metal, other methods should
tecting Susceptibility to lntergranular be chosen or the metallurgical condition must
Attack in Stainless Steels.2 be corrected by subsequent treatment. For
A 279, Total Immersion Corrosion Test of example, shearing a specimen to size will cold
Stainless Steels, 2 work and may possibly fracture the edges.
D 1384, Corrosion Test for Engine Anti- Edges should be machined or the specimen
freezes in Glassware. s annealed.
3.3 The clean, dry specimens should be
3. Methods for Preparing Specimens for Test measured and weighed. Dimensions deter-
3.1 Surface Condition: mined to the third significant figure and mass
3.1.1 For laboratory corrosion tests that determined to the fifth significant figure are
simulate exposure to service environments, a suggested.
commercial surface, closely resembling the
4. M e t h o d for E l e c t r o l y t i c Cleaning After
one that would be" used in service, will yield
Testing
the most significant results.
3.1.2 For more searching tests of either the 4.1 Electrolytic cleaning is a satisfactory
metal or the environment, standard surface method for many common metals.
finishes may be preferred. A suitable proce- 4.1.1 The following method is typical; after
dure might be: i This recommendedpractice is under the jurisdiction of
3.1.2.1 Degrease in an organic solvent or ASTM Committee G-I on Corrosion of Metals and is the
direct responsibility of SubcommitteeG01.05 on Labora-
hot alkaline cleaner. tory CorrosionTests.
NOTE l--Hot alkalies and chlorinated solvents Current edition approved May 30, 1972. Published
may attack some metals (for example, aluminum). July 1972. Originally published as G 1 - 67. Last pre-
viousedition G 1 - 67.
NOTE 2--Ultrasonic cleaning may be beneficial 2Annual Book of ASTM Standards, Part 3.
in both pre-test and post-test cleaning procedures. aAnnual Book of ASTM Standards, Part 22.
261
@ G1
scrubbing to r e m o v e loosely a t t a c h e d corro- 5.4 Lead Alloys--Suitable m e t h o d s in-

sion products, electrolyze the specimen as fol- clude:
lOWS: 5.4.1 P r e f e r a b l y , use the e l e c t r o l y t i c
Sulfuric acid (H,SO4, 28 ml cleaning p r o c e d u r e of Section 4.
sp gr 1.84) 5.4.2 Dip in:
Organic inhibitor 2 ml (sr162
Note 3)
Water to make 1 liter Acetic acid (99.5 percent) l0 ml
Temperature 75 C (167 F) Water to make 1 liter
Time 3 rain Temperature boiling
Anode carbon or lead (see Note 4) Time 5 min
Cathode test specimen
Current density 20 A/din ~ 5.4.3 A l t e r n a t i v e l y dip in:
Ammonium acetate 50 g
NOTE 3--Instead of using 0.2 volume percent of Water to m~ke I liter
any proprietary inhibitor, about 0.5 g/liter of such Temperature hot
inhilbitors as diorthotolyl thiourea, quinoline ethio- Time 5 rain
dide, or betanaphthol quinoline may be used.
NOTE 4 - - I f lead anodes are used, lead may de- 5.4.3.1 This r e m o v e s lead oxide ( P b O ) and
posit on the specimen and cause an error in the
lead sulfate ( P b S O , ) .
mass loss. If the specimen is resistant to nitric acid,
the lead may be removed by a flash dip in 1+ 1 ni- 5.5 Zinc--The following m e t h o d s are suit-
tric acid. Except for this possible source of error, able:
lead is preferred as an anode as it gives more effi- 5.5.1 Dip in:
cient corrosion-product removal.
Ammonium hydroxide (NH,OH, 150 ml
4.2 It should be noted t h a t this electrolytic sp gr 0.90)
t r e a t m e n t m a y result in the redeposition of Water to make I liter
Temperature room
metal, such as copper, f r o m reducible corro- Time several minutes
sion products and, thus, lower the a p p a r e n t
m a s s loss. 5.5.2 T h e n dip in:
Chromic acid (CrOa) 50 g
5, M e t h o d s for Chemical Cleaning After Silver nitrate (AgNes) 10 g
Water to make [ liter
Testing
Temperature boiling
NOTE 5: Caution--These methods may be haz- Time 15 to 20 s
ardous to personnel. They should not be carried out
by the uninitiated or without professional supervi- NOTE 6 - - I n making up the chromic acid solu-
sion. tion, it is advisable to dissolve the silver nitrate sep-
arately and add it to the boiling chromic acid to
5.1 Copper and Nickel Alloys--Dip in: prevent excessive crystallization of the silver chro-
Hydrochloric acid (HCI, sp gr 1.19) or 500 ml mate. The chromic acid must be free from sulfate
Sulfuric acid (H~SO,, sp 8r 1.84) 100 ml to avoid attack on the zinc.
Water to make 1 liter
Temperature room 5.5.2 Dip in:
Time I to 3 min Hydriodic acid (HI, sp gr 1.5) 85 ml
5.2 Aluminum Alloys--Dip in:
Temperature room
Chromic acid (CrO,) 20 g Time 15 s
Phosphoric acid (HsPO,, sp gr 1.69) 50 ml
Water to make I liter
5.5.2.1 This procedure'dissolves a little zinc
Temperature 80 C (176 F)
Time 5 to 10 rain and corrections m u s t be m a d e as noted in 6.1.
5.6 Magnesium Alloys--Dip in:
5.2.1 l f a film remains, dip in:
Chromic acid ( e r e , ) 150 g
Nitric acid (HNO,, sp gr 1.42) ...
Silver chromate ( A g a t e , ) 10 g
Time I min
5.2.2 R e p e a t CrO8 dip. Temperature boiling
Time I rain
5.2.2.1 Nitric acid alone m a y be used if
there are no deposits.
5.7 Iron and Steel--Suitable methods are:
5.3 Tin Alloys--Dip in:
5.7.1 T h e hot sodium hydride m e t h o d is
Trisodium phosphate (NajPO,) 150 g excellent for cleaning iron and steel both f r o m
Water to make I liter
Temperature boiling the point of view o f ease of r e m o v a l of corro-
Time 10 min sion products and m i n i m u m a t t a c k on the
APPENDIXES 263
G1
metal. 4 Because o f the h a z a r d involved a n d tion o f the c o r r o s i o n rate. T o c h e c k this, one

the s o m e w h a t m o r e s o p h i s t i c a t e d e q u i p m e n t or m o r e cleaned a n d weighed s p e c i m e n s m a y
required, o t h e r m e t h o d s m a y be p r e f e r r e d . A n be r e c l e a n e d by the s a m e m e t h o d a n d re-
a l t e r n a t i v e choice is electrolytic c l e a n i n g (see weighed. L o s s due to this s e c o n d w e i g h i n g
Section 4). m a y be used as an a p p r o x i m a t e c o r r e c t i o n to
5.7.2 Dip in C l a r k e ' s solution: the first one (see A p p e n d i x A 2 for a m o r e
Hydrochloric acid (HCI, 1 liter exact method).
sp gr 1.19) 6.2 T h e initial t o t a l s u r f a c e a r e a o f the
Antimony trioxide (SbCOa) 20 g specimen ( m a k i n g a l l o w a n c e s f o r the c h a n g e
Stannous chlecide (SnCI 2) 50 g
Temperature room in a r e a d u e to m o u n t i n g h o l e s ) a n d the m a s s
Time up to 25 gin lost d u r i n g the test are d e t e r m i n e d . T h e av-
5.7.2.1 S o l u t i o n s h o u l d be v i g o r o u s l y e r a g e c o r r o s i o n rate m a y t h e n be o b t a i n e d as
stirred o r the s p e c i m e n should be r u b b e d with follows.
a nonabrasive implement of wood or rubber. Corrosion rate = (K • W ) / ( A • T • D)
5.7.3 Dip in: where:
Sulfuric acid (H ~SO,, sp gr 1.84) 100 ml K = a c o n s t a n t (see 6.2.1),
Organic inhibitor 1.5 ml T = t i m e o f e x p o s u r e in h o u r s t o the
Water to make I liter n e a r e s t 0.01 h,
A = a r e a in c m 2 to the n e a r e s t 0.01 c m ~,
5.8 Stainless Steels: W = m a s s loss in g, to n e a r e s t 1 m g , a n d
5.8.1 M e t h o d s in 5.7.1 a r e also a p p l i c a b l e D = density in g / c m 3 (see A p p e n d i x A 1 ) .
5.8.2 Dip in: 6.2.1 M a n y different units are used to ex-
Nitric acid (HNOs, sp gr 1.42) 100 ml press c o r r o s i o n rates. U s i n g the a b o v e units
Water to make I liter for T, A , W, a n d D the c o r r o s i o n r a t e c a n be
Time 20 rain c a l c u l a t e d in a v a r i e t y o f units with the fol-
lowing a p p r o p r i a t e value o f K :
5.8.3 A l t e r n a t i v e l y dip in:
Constant (K) in
Ammonium citrate 150 g Corrosion Rate
Water to make 1 liter Corrosion Rate Units Desired Equation
Temperal,ure 70 C (158 F) mils per year (mpy) 3.45 x l06
Time 10 to 60 min inches per year (ipy) 3.45 x l0 s
5.9 In place o f c h e m i c a l c l e a n i n g , use a inches per month (ipm) 2.87 • l0 s
millimeters per year (mm/y) 8.76 • 10'
b r a s s s c r a p e r or b r a s s bristle b r u s h , or b o t h , micrometers per year (,~m/y) 8.76 x l07
followed by s c r u b b i n g with a wet bristle b r u s h picometers per second (pm/s) 2.78 • 10'
a n d fine s c o u r i n g p o w d e r . H o w e v e r , t h i s grams per square meter per hour 1.00 x 104 x Da
(g/m~.h)
m e t h o d m a y n o t r e m o v e all-the p r o d u c t s 4"rom milligrams per square decimeter 2.40 x l0 s x Dr
pits. per day (mdd)
NOTE 7--Such vigorous mechanical cleaning is micrograms per square meter per 2.78 • 10~ x Da
applicable when mass losses are l a r g e and hence second 0,g/m ~.s)
errors in mass loss will produce only small errors in
corrosion rates. Blank corrections will be ~difficutt "Density is not needed to calculate the corrosion rate in
to apply. these units; the density in the constant K cancels out the
density in the corrosion rate equation.
5.10 In all the f o r e g o i n g m e t h o d s , ~ s p e c i - NOTE 8--1f desired, these constants may also be
m e n s should be rinsed following c l e a n i n g a n d used to convert corrosion rates from one set of
s c r u b b e d lightly with a bristle b r u s h u n d e r units to another. To convert a corrosion rate in
units X to a rate in units Y, multiply by K y / K x for
r u n n i n g water. T h e c l e a n i n g dip m a y be re- example:
p e a t e d as n e c e s s a r y . A f t e r the final rinse, 15 mpy = 15 • (2.78 • 10e)/(3.45 • 10e) p m / s
s p e c i m e n s s h o u l d be dried a n d weighed.
6.3 C o r r o s i o n r a t e s c a l c u l a t e d f r o m m a s s
6. Calculation of Corrosion Rate
6.1 W h a t e v e r c l e a n i n g m e t h o d is used, the 4Technical Information Bulletin SP29-370 "'DuPont
p o s s i b i l i t y o f r e m o v a l o f s o l i d m e t a l is Sodium Hydride Descaling Process Operating Instruc-
tions," available from E. 1. duPont de Nemours & Co.
present; this resuRs in e r r o r in the d e t e r m i n a - (Inc.), Electrochemicals Dept., Wilmington, Del. 19898.
@ 01
losses can be misleading when deterioration is reduced by corrosion. Loss in tensile strength
highly localized, as in pitting or crevice corro- will result if a m e t a s o m a t i c change, such as
sion. If corrosion is in the form of pitting, it p a r t i n g h i s t a k e n place. L o s s in tensile
may be measured with a depth gage or mi- strength and elongation will result from local-
c r o m e t e r calipers with pointed anvils. Micro- ized attack, such as cracking.
scopical methods will determine pit depth by 6.4.3 Electrical Properties--Loss in ap-
focusing from top to b o t t o m of the pit, when parent conductivity will result from cracking
it is viewed from above (using a c a l i b r a t e d or pitting.
focusing k n o b ) or by e x a m i n i n g a section that 6.4.4 Microscopical Examination--Parting,
h a s been mounted and m e t a l l o g r a p h i c a l l y pol- exfoliation, cracking, or intergranular attack
ished. The pitting factor is the ratio of the may be m e a s u r e d by metallographic examina-
d e e p e s t m e t a l p e n e t r a t i o n to the a v e r a g e tion of suitably prepared sections.
m e t a l penetration (as measured by mass loss).
6.4 O t h e r methods of assessing corrosion 7. Report
d a m a g e are: 7.1 The report should include the composi-
6.4.1 Appearance--The d e g r a d a t i o n of tions and sizes of specimens, their metallur-
a p p e a r a n c e by rusting, tarnishing, or oxida- gical conditions, surface preparations, and
tion. post-corrosion cleaning methods, as well as
6.4.2 Mechanical Properties--An a p p a r e n t measures of corrosion d a m a g e such as corro-
loss in tensile strength will result if the cross- sion rates (calculated from mass losses), max-
sectional area of the specimen (measured be- i m u m depths of pitting, or losses in mechan-
fore exposure to the corrosive environment) is ical properties.
APPENDIXES
AI. DENSITIES FOR A VARIETY OF METALS AND ALLOYS

Density Density
Aluminum Alloys g/cm a Ferrous Metals g/cm'
1100, 3004 2.72 Gray cast iron 7.20
1199, 5005, 5357, 6061, 6062, 6070, 6101 2.70 Carbon steel 7.86
2024 2.77 Silicon iron 7.00
2219, 7178 2.81 Low alloy steels 7.85
3003, 7079 2.74 Stainless steels:
5050 2.69 Types 201,202, 302, 304, 304L,321 7.94
5052, 5454 2.68 Types 309, 310, 311,316, 316L, 317, 7.98
5083, 5086, 5154, 5456 2.66 329, 330
7075 2.80 Type 347 8.03
Copper Alloys Type 410 7.70
Copper 8.94 Type 430 7.72
Brasses: Type 446 7.65
Commercial bronze 220 8.80 Type 502 7.82
Red brass 230 8.75 Durimet 20 8.02
Cartridge brass 260 8.52 Carpenter Stainless No. 20 Cb-3 8.05
Muntz metal 280 8.39
Admiralty 442, 443, 444, 445 8.52 Lead
Aluminum brass 687 8.33 Aatimonial 10.80
Bronzes: Chemical I 1.33
Aluminum bronze, 5 percent 608 8.16
Aluminum bronze, 8 percent 612 7.78 Nickel Alloys
Composition M 8.45 Nickel 200 8.89
Composition G 8.77 Monel Alloy 400 8.84
Phosphor bronze, 5 percent 510 8.86 Inconel Alloy 600 8.51
Phosphor bronze, 10 percent 524 8.77 Incoloy Alloy 825 8.14
85 5-5-5 8.80 Illium G 8.31
Silicon bronze 655 8.52 Hastelloy B 9.24
Copper nickels 706, 710, 715 8.94 Hastelloy C 8.93
Nickel silver 752 8.75 Hastelloy G 8.27
APPENDIXES 265
~T~) G 1
Density Density
Other Metals g/cm J Other Metals g/cm a
Magnesium 1.74 Tin 7.30
Molybdenum 10.22 Titanium 4.54
Platinum 21.45 Zinc 7.13
Silver 10.49 Zirconium 6.53
Tantalum 16.60
A2. M E T H O D FOR DETERMINING M A S S L O S S W H E R E CLEANING MAY ATTACK T H E

BASE M E T A L OF T H E S P E C I M E N
A2.1 Repeat the cleaning procedure a number of the intersection of the two lines is the mass loss
times. Weigh after each cleaning and plot the mass caused by removal of corrosion products alone. The
loss against the total time of cleaning or the method is particularly applicable to electrolytic
number of clcanings, see Fig. AI. The ordinate at cleaning, see Section 4.
~WEIGHT OF CORROSION
~ PROOUCTS REMOVED
/ o ~METAL
-- REMOVAl.
--~
kd ~CORROSION PROOUCTS
I /6 REMOVAL
/
CLEANING TIME
FIG. A I Mass Lass Versus Exposure Time for Specimens During Cleaning.
By publication of this standard no position is taken with respect to the validity o/any patent rights in connection there-
against liability for infringement of any Letters Patent nor assume any such liability.
STP534-EB/Nov. 1973
(I~I~) Designation:G 4 - 6 8

CONDUCTING PLANT CORROSION TESTS ~
This Standard is issuedunderthe fixeddesignationG 4; the numberimmediately followingthe designationindicatesthe year
of original adoption or, in the ease of revision,the year of last revision. A numberin parenthesesindicates the year of last
reapproval.
1. Scope 2.1.3 The behavior of certain metals and

1.1 This recommended practice outlines alloys may be profoundly influenced by the
procedures for conducting corrosion tests in presence of dissolved oxygen. It is essential
plant equipment under operating conditions. that the test coupons be placed in locations
It is not intended for atmospheric or under- representative of the degree of aeration nor-
ground corrosion tests but may possibly apply mally encountered in process.
to other tests under natural conditions where 2.1.4 Corrosion products may have undesir-
the procedure has proven satisfactory. Corro- able effects on the product. This possibility is
sion testing by its very nature precludes com- frequently recognized in advance. The extent
plete standardization. This recommended of possible contamination can be estimated
practice, rather than a standardized proce- from the loss in weight of the specimen, with
dure, is presented as a guide so that some of proper application of the expected relation-
the pitfalls of such testing may be avoided. ships among (1) the area of corroding surface,
NoTE--The values stated in U.S. customary (2) the mass of the product handled, and (3)
units are to be regarded as the standard. the duration of contact of a unit of mass of the
product with the corroding surface.
2. Interferences 2.1.5 Corrosion products from the plant
2.1 Tests described herein are probably the equipment used in the test may influence the
best means available for approximating the corrosion of one or more of the test metals.
behavior of metals in service, short of actually For example, when aluminum specimens are
constructing and operating a piece of equip- exposed in copper equipment, corroding cop-
ment. For best results, certain variables must per will exert an adverse effect on the corro-
be considered. These include: sion of the aluminum. Contrariwise, stainless
2.1.1 Metal specimens immersed in a spe- steel specimens can have their corrosion re-
cific hot liquid may not corrode at the same sistance enhanced by the presence of the oxi-
dizing cupric ions.
rate or in the same manner as in equipment
where the metal acts as a heat transfer me- 2.1.6 The accumulation of corrosion prod-
dium in heating or cooling the liquid. In cer- ucts sometimes can have harmful effects. For
tain services, the corrosion of heat-exchanger example, copper, corroding in intermediate
tubes may be quite different than that of the
shell or heads. This restriction also applies to
specimens exposed in gas streams from which aThis recommendedpractice is under the jurisdiction of
the ASTM Committee G-I on Corrosion of Metals. This
water or other corrodents condense on cool standard is the direct responsibilityof Subeommittee GOI. 12
surfaces. Such factors must be considered in on In-plantCorrosionTests.
Current edition effective Sept. 13, 1968. Originally is-
both design and interpretation of plant tests. sued as A 224 in 1939. Replaces A 224 - 46.
2.1.2 Effects caused by high velocity, abra- Revised with the aid of Unit Committee T-5A. Corro-
sion in Chemical Processes,National Associationof Corro-
sive ingredients, etc. (which may be empha- sion Engineersand ASTM Committee A-10on Iron-Chro-
sized in pipe elbows, pumps, etc.), may not be mium, Iron-Chromium-Nickeland Related Alloys which
formerly held jurisdiction over ASTM Recommended
easily reproduced in coupon tests. Practice A 224.
266
APPENDIXES 267
G4
strengths of sulfuric acid will have its corro- uation than a simple weight loss measurement
sion rate increased as the cupric ion accumu- is required to detect this phenomenon.
lates. Phenomena such as this will not be rec- 2.1.7.5 Certain metals and alloys are sub-
ognized from coupon tests in the plant, and ject to a highly localized type of attack called
must be anticipated from general knowledge pitting corrosion. This cannot be evaluated by
and experience, or studied under controlled weight loss. The reporting of nonuniform cor-
(laboratory) conditions. rosion is discussed below. It should be appre-
2.1.7 Coupon corrosion testing is predomi- ciated that pitting is a statistical phenomenon
nantly designed to investigate general corro- and the incidence of pitting can be directly
sion. There are a number of other special related to the area of metal exposed. For
types of corrosion phenomena of which one example, a small coupon is not as prone to
must be aware in the design and interpreta- exhibit pitting as a large one, and it is possi-
tion of coupon tests. ble to miss the phenomenon altogether in the
2.1.7.1 Galvanic corrosion may be investi- corrosion testing of certain alloys, such as the
gated by special devices that couple one cou- AISI Type 300 series stainless steels in chlo-
pon to another in electrical contact, as for ride-contaminated environments.
example, by substituting a spacer made from 2.1.7.6 All metals and alloys are subject to
the more noble metal of the couple in place of stress-corrosion cracking under s o m e cir-
an insulating spacer. The behavior of the cumstances. This cracking attack occurs un-
specimens in this galvanic couple are com- der conditions of tensile stress and it may or
pared with that of insulated specimens exposed m a y not be visible to the naked eye or upon
on the same holder and the galvanic effects casual inspection. A metallographic examina-
noted. It should be observed, however, that tion will confirm this mechanism of attack. It
galvanic corrosion can be greatly affected by is imperative to note that this usually occurs
the area ratios of the respective metals. The with no significant loss in weight of the test
coupling of corrosion coupons then yields only coupon, although certain refractory metals are
qualitative results, as a particular coupon re- an exception to these observations.
flects only the relationship between those two
metals at the particular area ratio involved. 3. Apparatusfor Mounting Specimens
2.1.7.2 Crevice or concentration cell corro- 3.1 Although it is possible to expose speci-
sion m a y occur where the metal surface is mens to corrosive environments in operating
partially blocked from the corroding liquid, as equipment by attaching them to pieces of
under a spacer. At times it is desirable to string, wire, glass, etc., this is usually inade-
know whether a given metal is subject to crev- quate. In general, the m e t h o d of support
ice corrosion in a given environment, whereas should be such as to satisfy the following re-
in other cases the spacers can be designed to quirements:
minimize this effect (see below). An accumu- 3.1.1 Prevent loss of specimens from causes
lation of debris or corrosion products between other than corrosion,
the coupons can produce misleading results in 3.1.2 Eliminate the possibility of galvanic
either accelerating corrosion or protecting the effects resulting from metal-to-metal contact
coupons from corrosion. between specimens or between the vessel and
2.1.7.3 Selective corrosion at the grain the specimen exposed therein,
boundaries (for example, intergranular corro- 3.1.3 Hold specimens firmly in place, and
sion of sensitized austenitic stainless steels) 3.1.4 Provide for protection of specimens
will not be readily observable in weight loss against mechanical damage.
measurements and often requires microscopi- 3.2 While it is possible to clamp a speci-
cal examination o f the coupons after expo- men near its edge, it is not easy to design a
sure. clamping arrangement that will provide the
2.1.7.4 Metasomatic corrosion is a condi- necessary electrical insulation without either
tion in which one constituent is selectively completely or partially shielding a fairly large
removed from an alloy, as in the dezincifica- area of the specimen from free contact with
tion of brass or the graphitization of cast iron. the corroding solution. A better arrangement
Close attention and a more sophisticated eval- is to drill a hole in the specimen and allow the
G4
supporting device to pass through the hole. A acting as a lock nut (Fig. 2). The end plates
satisfactory location for the supporting hole is used as bearing plates should be made larger
at the center of the specimen so that it cannot than the specimens so that they will act as
be lost from the holder unless either the speci- bumpers to keep the specimens from touching
men or the holder is destroyed by corrosion. any flat surfaces with which the holder may
3.3 To ensure effective electrical insulation, come in contact. The end plates need not nec-
it is necessary to provide some insulating essarily be made of insulating material. Addi-
material between the specimen and the metal- tional metal rods may be used to connect the
lic support. This can be accomplished most end plates at points where they will clear the
conveniently by using a tube of some insulat- specimens and can be tightened so as to pro-
ing material that will fit over the metal rod. vide additional reinforcement and rigidity to
Polyethylene, poly(vinyl chloride), saran, ba- the entire assembly. The rods also protect the
kelite, ceramics, or fluorinated plastics are sat- specimens from mechanical injury. Support
isfactory. The hole in the specimen should be and bracing rods from 1/4 in. (6.35 mm) to 3/s
made large enough so that the specimen will in. (9.53 m m ) have been found suitable. All
slide over the insulating tube. More than one metal rods used in the assembly should be
specimen may be exposed on the same holder made of material which is sufficiently corro-
by insulating or separating the s p e c i m e n s sion resistant to ensure the assembly remain-
from each other by means of short tubes of ing intact for the duration of the test. Stain-
insulating material that can be slipped over less steel, Monel Alloy 400, or other suitable
the insulating tube on the supporting rod. The nonferrous metals are commonly employed.
short tube spacers should be large enough in Individual spacers (Fig. 1) may be used, or a
outside diameter to provide firm support for separate insulating tube with insulating wash-
the specimens without covering more than a ers may be employed. All three modes of
small percentage of the total surface. The mounting coupons are illustrated in Fig. 2. Be-
spacing of the specimens in this arrangement cause these spool type racks have the disad-
is determined by the length of the insulating vantage of requiring that the equipment not
spacer. Any insulating material that will with- only be out of service but also be gas-free or
stand the action of the corroding solution may otherwise made suitable for entry, other types
be used. A preferable means of support is the of field corrosion racks (which are preferably
use of individual insulating spacers machined in petrochemical process equipment) have
to the desired shape. Figure 1 delineates the been devised.
dimensions of two types of insulating spacers 3.4.2 The insert rack is fabricated by weld-
designed specifically for field corrosion testing suitable rod or strip to a welding disk that
ing. The first type of specimen is designed to can be held within the bolt circle and flange
minimize concentration cell effects on the face face of a flange in an unused nozzle. A 1.5 to
of the specimen. The second type has a sharp 2-in. (38.1 to 50.8-mm) nozzle is usually con-
shoulder which will tend to lead to crevice cor- venient for this installation (Fig. 3). Such
rosion adjacent to the hole in the coupon. The racks employ a stout member immediately
choice between these two types of spacers will adjacent to the welding disk, for example,
lie with the corrosion engineer, based on the 0.375-in. (9.53-mm) Type 316 stainless steel
type of information he requires. rod with an 0.25-in. (6.35-mm) rod extension
3.4 Although there are many ways of sup- that carries the specimens and spacers. Such
porting corrosion coupons in plant apparatus, racks should also be assembled with a lock
the following basic types of field corrosion nut arrangement. Occasionally, racks of this
racks are described in some detail as a guide: type may be.-required for nozzles which are
3.4.1 A spool rack may be assembled by not "blind" but are employed for the fasten-
threading the ends of a supporting rod, and ing of piping to the equipment in question. In
providing end disks or bearing plates against such cases the disk can be perforated, and the
which nuts on the ends of the rod may be specimens mounted sidewise on the rack, if
turned so as to press the s p e c i m e n s and required, as indicated in Fig. 3.
spacers close together. Two nuts should be 3.4.3 For larger diameter pipes or nozzles,
used at each end of the rod, the second nut a " d u t c h m a n " type rack m a y be employed.
APPENDIXES 2 6 9
G4
Such a rack will consist of a suitable spool influenced by several factors and cannot be
piece with the specimens mounted crosswise rigidly defined. In general, the ratio of surface
on a bar as shown in Fig. 4. Both insert and area to mass should be large so as to favor
dutchman corrosion racks require that the maximum amount of corrosion loss. This can
equipment be out of service, but they may be he accomplished by the use of thin sections.
installed and removed without extraordinary Sufficient thickness, however, should be em-
precautions in gas freeing the equipment. ployed to minimize the possibility of perfora-
3.4.4 The "slip-in" corrosion rack is ideally tion of the specimen during the test exposure.
suited for effecting the entry and removal of The size of the specimen should be as large as
corrosion coupons from operating equipment can be conveniently handled, the limitation
that is in active service. The slip-in rack re- being imposed primarily by the maximum
quires an unused nozzle of suitable size weight (200 g) that can be handled by an ana-
(usually nominal 1 l/2-in, pipe size or greater) lytical balance and, secondarily, by the prob-
and a gate valve. The corrosion rack is then lem of effecting entry into operating equip-
assembled from a short length of pipe or tub- ment.
ing with a suitable flange and a packing-gland 4.2 A convenient size for standard corro-
arrangement made from the bonnet of a 1/2-in. sion coupons is 1.5 in. (38.1 mm) in diameter
(12.7 ram) stainless steel or suitable alloy and 0.125 in. (3.18 mm) in thickness with a
valve. An alloy rod of appropriate length is 0.438-in. ( t l . l - m m ) hole in the center of the
used as the specimen mount as shown in Fig.. round coupon. This size was arrived at as
5. In the "Out" position the specimens are being the maximum size that could easily
mounted on the rod and drawn into the re- effect entry through a nominal 11/2-in. nozzle.
cessed area provided by the pipe or tube sec- However, it is also convenient for larger size
tion. When this is bolted to the gate valve and nozzle entries as well as for laboratory corro-
the valve is opened, the assembly of corrosion sion testing. A convenient standard coupon for
coupons may be slid into the operating equip- spool-type racks is the 2 by 2 by 0.125 in.
ment for exposure. When it is desired to re- (50.8 by 50.8 by 3.18 ram) square, if only a
move the specimens, they are withdrawn into few coupons need be made. A round coupon
the recessed area, the gate valve is closed, and of 2.11 in. (53.5 mm) by 0.125 in. (3.18 ram),
the entire assembly is then physically re- or 2.18 in. (55.5 mm) by 0.062'in. (1.59 ram),
moved from the operating equipment. is sometimes employed. These last measure
3.4.5 The design of corrosion racks for 0.500 dm 2 in area.
plant tests is limited only by the imagination 4.3 Other sizes, shapes, and thicknesses of
and ingenuity of the corrosion engineer. In specimens can be used for special purposes or
specific circumstances, for example, it is pos- to comply with the design of a special type of
sible to convert thermowells into corrosion corrosion rack. When the choice of material
racks by welding a short extension rod on has been reduced to a few in number in pre-
them. Similarly, racks may be designed to liminary tests, special coupons should be em-
clamp onto agitators, thermowells, or other ployed to consider the effect of such factors of
parts of operating equipment. equipment construction and assembly as heat
3.4.6 When the choice of materials of con- treatment, welding, soldering, and cold work-
struction has been narrowed to one or two, it ing or other mechanical stressing.
may be desirable to investigate heat-transfer
effects with a simple bayonet heater of the 5. Preparation of Test Specimens
design shown in Fig. 6. Either a heating or 5.1 The edges of..test specimens should be
cooling medium is circulated through the tube so prepared as to eliminate all cold-worked
side of this testing apparatus, and the effect of metal except that introduced by stamping for
the hot or cold wall upon corrosion is observed identification. Shearing will, in some cases,
by visual observation, pit depth measure- cause considerable attack; and, therefore,
ments, micrometer measurements, etc. specimens having sheared edges should not be
used. The edges should be finished by machin-
4. Test Specimens ing or polishing. The slight amount of cold
4.1 The size and shape of test specimens is working resulting from machining will not
G4
introduce any serious error. percent confidence interval factors t are as fol-
5.2 Usually no specific finish is of interest lows:
except in the sense that uniformity is desira- #1 !
ble when comparing data from different tests. 2 6.4
Furthermore, it may be necessary to remove 3 1.3
4 0.72
dirt or heat-treating scale from the metal sur- 5 0.51
face. It has been found convenient to stand- Confidence interval = ~ • tw
ardize on a 120-grit surface in most cases. A
where:
surface roughness greater than 120 grit should
x = average of n observations,
not be used.
w = range, and
5.3 After the test specimens are cut to size,
t ~ factor.
they should be freed from water breaks by
For example, if four successive tests give cor-
suitable cleaning. In isolated cases, it is desir-
rosion rates of 15, 20, 25, and 20 mils/year,
able to expose specimens with a special sur-
the average (s is 20, and the range is 25 - 15
face treatment. For example, in some applica-
= 10. Then the 95 percent confidence interval
tions, stainless steels may be prepassivated by
is 20 =k 10 (0.72) or 20 -4- 7.2 mils/year (1
a 30-min immersion in 10 to 20 percent nitric
rail/year = 0.0254 mm/year). A special con-
acid at 60 C. In most cases, however, special
sideration can also be applied to evaluate
pretreatments are unnecessary and undesira-
whether a doubtful observation should be dis-
ble.
regarded. This is discussed in detail in the
5.4 The weight of each specimen should be
referenced article.
determined to the nearest 0.1 mg on an ana-
lytical balance. 7. Identification of Test Specimens
6. Number of Test Specimens 7.1 For purposes of identification, a record
should always be made of the relative posi-
6.1 For statistical validity, it is desirable to tions of the test specimens on the holder. If
expose replicate specimens. When corrosion
identification marks are obliterated by corro-
tests are performed in the laboratory under sion, careful handling of the specimens is re-
standard conditions, duplicate specimens will
quired to maintain identity.
suffice for an accuracy of -4-10 percent. How- 7.2 Although it m a y be necessary in special
ever, this reflects the reproducibility of certain
instances to notch the edge of the specimens
standardized tests, and does not necessarily
for identification, it is preferable that they be
hold true for plant corrosion testing. It is pos-
stamped with a cOde number. The stamped
sible, although not probable, to have rather
number has an additional advantage in that,
widely different results on replicate specimens
should a specimen show a preferential attack
exposed on the same rack in a given test.
at the stamped area, a warning is given that
6.2 In multiple exposures, it is probable
the material is susceptible to corrosion when
that there will be considerable variation in the
cold worked. It is also possible in some in-
results from one exposure to the other because
stances to detect stress-corrosion cracking
of changes in operating conditions. Under
emanating from the Stamped areas. With such
such circumstances an evaluation should be
indications, the investigator is forewarned and
based on the statistics of a limited number of
can reject the material from further consider-
observations)
ation or may study fur[her the effects of cold
6.3 For a limited mimber of observations
wo[king or stress upon the corrosion behavior.
(for example, ten o r less) the range w between
Note, however, that although the presence of
m a x i m u m and m i n i m u m values provides
such localized attack is a positive indication,
more definitive values than does the standard
absence of attack is not a guarantee of im-
deviation. In practice, it is usually desired to
munity from attack in operating equipment.
establish a "confidence interval", that is, the
distance on either side of the average in which
one would expect to find the true value 95
percent of the time. This is established by Dean, R. B., and Dixon, W. J., "SimplifiedStatistics
for Small Numbers of Observations," Analytical Chemls-
multiplying the range w by a factor t. The 95 try, ANCHA, Vol 23, No. 4, April, 1951,pp. 636-638.
APPENDIXES 271
G4
8. Installation of Specimen Holder them in proper sequence relative to each other
8.1 The location of the test specimens in so that any specimen may be identified from
the operating equipment will be governed by the original record of its position on the
the information that is desired. This may re- holder. This is important if corrosion has been
quire tests at more than one location in the so severe that identification marks have been
same piece of equipment, such as below the removed.
level of the test liquid, at the level of the liq- 10.2 A record should be made of the ap-
uid, or in the vapor phase. pearance and adhesion of any coatings or
8.2 It is desirable to have the specimen films on the surface of the specimens after
holder securely fixed in place. The preferred washing. It may be desirable to photograph
position of the holder is with the long axis the specimens. Color photographs may be of
horizontal so as to prevent drippage of corro- value. Samples of any products or films result-
sion products from one specimen to another. ing from corrosion may be preserved for fu-
Preferably, the holder should be so placed that ture study.
any flow of liquid will be against the edges of
11. Cleaning ~ and Weighing Test Specimens
the specimens. The same condition of agita-
tion of the liquid should then be encountered 1.1 The surfaces of the test specimens
by all specimens. should be thoroughly cleaned of all corrosion
products. Removal of corrosion products from
9. Duration of Exposure the specimens may not be a simple procedure.
9.1 The duration of exposure may be based No hard and fast rules can be laid down since
on known rates of deterioration of the materi- the cleaning procedure adopted will depend
als in use. More often, it is governed by the on the base material as well as the nature of
convenience with which plant operations may the corrosion products. It will be necessary for
be interrupted to introduce and remove test the investigator to study the problem and de-
specimens. In many tests, some materials may cide upon the most suitable procedure.
show little or no attack while other materials 11.2 It is essential that the base metal be
m a y be completely destroyed. In general, the unattacked either by the cleaning reagent or
duration of the test should be as long as possi- by compounds formed by reaction between
ble commensurate with the resistance of the the cleaning reagent and corrosion products or
materials under test. In special cases, the du- other deposits on the specimen. A preliminary
ration m a y be established in regard to some solvent cleaning may be necessary to remove
specific phase of the operation, as for example organic deposits. The simplest cleaning proce-
to study corrosion in one step of a batch proc- dure is to scrub the specimens with a fiat fiber
ess. Possible changes in the rate of corrosion brush using a mild abrasive soap. Care must
may be studied either by successive exposures be taken that no base metal is removed by
or by the installation of several sets of speci- abrasion. Acid or alkaline solutions of suita-
mens at the same time, which can be removed ble nature and strength may be employed,
one set at a time at different intervals. The contingent upon their being noncorrosive to
m i n i m u m duration is commonly defined by the base metal. For example, a copper flash
the equation: on stainless steel or Hastelloy Alloy C can be
Minimum hours of test = 2000/mils/year. safely removed in concentrated nitric acid. On
the other hand, a copper flash on Monel Alloy
400 or Hastelloy Alloy B should be removed
10. Removal of Specimens from Test with a mixture of peroxide and a m m o n i u m
10.1 The condition and appearance of the hydroxide which will not significantly attack
holder and specimens after removal from the base metal. A solution of 5 percent stan-
e q u i p m e n t should be noted and recorded. nous chloride and 2 percent antimonious chlo-
Specimens should then be carefully washed, ride in concentrated hydrochloric acid may be
either in water or in a suitable solvent, to used to remove rust deposits from steel. This
remove all soluble materials from the surface
of the specimens. In removing the specimens ASTM Recommended Practice G 1, for Preparing,
Cleaning, and EvaluatingCorrosion Test Specimens,which
from the holder, care should be taken to keep appears in Annual Book of ASTM Standards, Part 31.
G4
solution has an added advantage in that the and those occurring on the boldly exposed sur-
final disappearance of the last vestige of ferric face. As previously noted, pitting at or under
oxide is readily apparent to the naked eye. the insulating spacers is an indication of the
For iron, steel, and alloy steels, a hot caustic susceptibility of the material to "concentra-
solution (20 percent) with 200 g of zinc dust tion cell" effects, whereas pitting on the sur-
added per liter is effective for loosening de- face is indicative of the intrinsic pitting tend-
posits which can then be scrubbed off. ency of the environment.
11.3 A cleaning method used by many in- 12.3 In the case of severe pitting of the
vestigators for a variety of materials consists specimen, the weight loss is of little value and
in making the specimen the cathode in a hot the study of the number, size, and distribution
dilute sulfuric acid solution under the follow- of the pits will be of much more importance.
ing conditions: Sometimes a pit-type of corrosion is initiated
11.3.1 Solution--5 weight percent of sul- but is self-healing and stops. A more detailed
furic acid plus 2 ml of an appropriate com- study of pitting is necessary before a definite
mercial inhibitor per liter of solution. conclusion can be reached as to the desirabil-
11.3.2 Anode Carbon--cathode-test speci- ity of rejecting a material because it has a
men. tendency to pit.
11.3.3 Cathode Current Density--20 A~ 12.4 If an alloy is known to be susceptible
dm ~. to localized corrosion on a microscale, such as
11.3.4 "Temperature--165 F (73.9 C). the phenomenon of intergranular corrosion in
11.3.5 Exposure Period--3 min. stainless steel, dezincification in brass, or
11.4 Another method which is sometimes stress corrosion cracking of any kind, the spec-
effective for removal of iron oxides is immer- imen should be bent after the previously out-
sion of the specimen in a hot solution of lined examination is completed, and any
ammonium acetate. cracks which develop on the surface noted.
11.5 After cleaning, the weight of each The results should be compared with those
specimen should be determined to the nearest obtained on similar bend tests on unexposed
0.1 mg on an analytical balance and the loss specimens from the same lot of material.
in weight calculated. The corrosion rate in 12.5 Microscopical examination of the sur-
mpy (mils per year) can be calculated using face and interior of the specimens may be
the following equation: made if deemed necessary. A low power shop-
weight loss, g X 534,000 type binocular microscope is ideal for many of
mpy = these examinations, although a metallo-
metal density, g/eraa X metal area, in.2
graphic examination may be needed.
• hours exposure
12.6 The behavior of the metals in galvanic
The corrosion rate may be translated into couples can be compared with that of insu-
other terms as discussed below.
lated specimens exposed at the same time,
and any galvanic effects, including cathodic
12. Examination of Specimen Surface protection, can be observed. As mentioned
12.1 The specimen should be carefully earlier, such tests are only qualitative as the
examined for type and uniformity of surface magnitude of the galvanic effect will be influ-
attack such as etching, pitting, metasomatic enced by the relative areas of the two metals
attack, tarnish, film, scale, etc. If pitting is comprising the couple. The results will apply
observed, the number, size and distribution, directly only to assemblies where the ratio of
as well as the general shape and uniformity of areas used in making the tests is similar to the
the pits should be noted. The maximum and ratio of areas anticipated in the fabricated
minimum depth of the pits can be measured assembly.
with a calibrated microscope or by the use of
a depth gage. Photographs of the cleaned 13. Report
specimens will serve as an excellent record of 13.1 In reporting results of corrosion tests,
the surface appearance. the conditions of the test should be described
12.2 A distinction should be made between in complete detail with special attention being
pits occurring under the insulating spacers given to the following:
APPENDIXES 273
@ G4
13,1,I Corrosive media and concentration, and changes of corrosion rates with time oc-
13.1.2 Type of equipment in which the test cur. In connection with the latter, it is often
was made, desirable to carry out the testing procedures
13.1.3 Process carried out in the operating so as to provide data from which curves can
equipment, be plotted to illustrate changes in corrosion
13.1.4 Location of specimens in the operat- rates with time.
ing equipment, 13.2 The depth of pits should be reported
13.1.5 T e m p e r a t u r e of corrosive media in 0.001 in. (0.02 ram) for the test period and
(maximum, minimum, and average), not interpolated or extrapolated to thou-
13.1.6 Oxidizing or reducing nature of cor- sandths of an inch per year or any other arbi-
rosive media, trary period. The size, shape and distribution
13.1.7 Amount and nature of aeration and of the pits should be noted.
agitation of corrosive media,
13.1.8 Duration and type of test (if equip- 14. Supplementary Tests
ment was operated intermittently during the 14 1 Supplementary laboratory tests should
tests, the actual hours of operation should be always be made when it is desired to study the
stated as well as the total time of the test), effect of one or more of the variables encoun-
13,1.9 Surface condition Of specimen tered in plant tests. They are particularly de-
(polished,. machined, pickled, 120 grit, etc.), sirable if there is any reason to believe that
and the products of corrosion, or the metal used
13.1.10 Units for expressing corrosion loss. for the equipment in which the test was con-
The unit for expressing corrosion rate should ducted, might have had a controlling influence
be mils penetration per year in cases where on the behavior of any metal in which there is
the corrosion has been substantially uniform further interest.
in distribution over the surface of the speci- 14.2 Special supplementary field tests
men. If this figure is representative, it may be should be made if there is any reason to be-
correlated with the thickness of the equipment lieve that stress corrosion cracking, intergran-
in the evaluation of the probable life. It is ular corrosion, or any other special metallur-
possible to convert this penetration unit into gipal phenomena m a y be anticipated.
other terms such as millimeters per year or
milligrams per square decirmeter per day for
comparison with other d a t a ? Any such ex-
pression will be subject to error to the extent 4Sr162Appendix A2 of ASTM Method A 279, Total
Immersion Corrosion Test of Stainless Steels, which ap-
to which nonuniform distribution of corrosion pears in the Annual Book o f A S T M Standards, Part 3.
1/8"
1/8"
(3.18 ram)
(3.18 ram)
/ / 8))
i8,) 7/16- ) 9)32- 115.9 -)
(11.~ I ) 9/32" (15.9 I )
[ (7. Z~ I )
_ H --
LI "'"
(15.9 ram)
'4
(15.9 ram)
FIG. 1 Tubular ]PlasticSpacers.
(300-6o0 mm)
FIG. 2 Spool Rack.
APPENDIXES
275
~ G4
coupons
~/4,I~od ~ j
3/8' rod
(9.53 ~)
,1
sp~s L i ~
*Disk to fi/
withinbolt
circle / . ~
12" - 18"
(300~460 tam)
[__. i;! "~ I ~'~

....
~----fiat bar
T ~nsulati~ t
*This plate may be ~ ~"~-,~ ~

slotted or otherwise - - / ~ [ ~
nfloO:z~t:spiiTuslieOrneS \[ ~~ - - I ----J----:'
~ '' ~ I
FIG. 3 Insert Racks.
~ G4
--co pon
(150,-300 ,,,,*) (50 ~ )
FIG. 4 Dutchman Racks.
s e r i e s flange
j/~l/#j (to match s a t e
vat re)
3/8" s . s . rod
Bonnet from I/." V (9.52 era)
-,....
rod (use welding TFE -
4| - - l; ,I,? =' 7 \
V r o d and thread to r F l u o r o c a r b o n p /
\ l O/ L4 on end. ) [__ spacer ~/ j I"
(25-75 Bin)
\ I i-ooopoa IY /I .
\ '" .... .'~" L}I
u _ _
6)
4~
Dram V ve L
L ..... i0,, _ _ q-:L "/ Optional -'4"
/'- - (25o.~- ~-~ (600 ram)
HandleV'~
j-- l0 tl
=r
(250 am)
FIG. $ Slip-In CorrosionTest Rick.
x
"4
~4
tt8~ (34
t earn i n l e t
3 / 8 " i n s t r u m e n t tubing
(9.52 tin)
~ slip on f l a n g e
I " c o n d e n s e r tubing
(25 m )
FIG. 6 Hot-Wall Tester.
By publication o f this standant no position is taken with respect to the validity o f any patent rights in connection there.
against liability for infringement o f any Letters Patent nor assume any such liability.
STP534-EB/Nov. 1973
(~I~ Designation:G15-71
Standard Definitions of Terms Relating to

CORROSION AND CORROSION TESTING 1
This Standard is issued under the fixed designation G 15; the number immediately following the designation indicates the
last reapproval.
anode--the electrode of an electrolytic cell at g a l v a n o s t a t i c - - p e r t a i n i n g to an e x p e r i m e n t a l

which oxidation is the principal reactio~a. technique whereby an electrode is main-
(Electrons flow away from the anode in the tained at constant current in an electrolyte.
external circuit. It is usually the electrode inhibitor--a c h e m i c a l substance or combina-
where corrosion occurs and m e t a l ions enter tion of substances, which when present in
solution.) the proper concentration and forms in the
anion--a negatively charged ion. environment, prevents or reduces corrosion.
cathode--the electrode of an electrolytic cell long-line c u r r e n t - - c u r r e n t which flows through
at which reduction is the principal reaction. the earth from an anodic to a cathodic area
(Electrons flow t o w a r d the cathode in the of a c o n t i n u o u s metallic structure. Usually
external circuit.) used only where the areas are separated by
cathodic c o r r o s i o n - - c o r r o s i o n of a m e t a l when considerable distance and where the current
it is a cathode. It usually happens to am- results from concentration-cell action.
p h o t e r i c m e t a l s as a result of a rise in pH at metallizing--See thermal spraying.
the cathode or as a result of the formation open-circuit potential--the p o t e n t i a l of an
of hydrides. electrode m e a s u r e d with respect to a refer-
cation--a positively charged ion. ence electrode or another electrode when no
concentration c e l l - - a n e l e c t r o l y t i c cell, the current flows to or from it.
e m f of which is caused by a difference in overvoltage--the c h a n g e in p o t e n t i a l of an
concentration of some c o m p o n e n t in the electrode from its equilibrium or steady
electrolyte. This difference leads to the for- state value when current is applied.
m a t i o n of discrete cathode and anode re- passivator--a type of inhibitor which appreci-
gions. ably changes the potential of a m e t a l to a
corrosion p o t e n t i a l - - t h e potential of a cor- more noble (positive) value.
roding surface in an electrolyte relative to a potentiostat--an instrument for a u t o m a t i c a l l y
reference electrode measured under open- m a i n t a i n i n g an electrode at a constant po-
circuit conditions. tential or controlled potentials with respect
electrolytic deaning--a process of removing to a suitable reference electrode.
soil, scale, or corrosion products from a redox potential--the potential of a reversible
m e t a l surface by subjecting it as an elec- oxidation-reduction electrode measured
trode to an electric current in an electrolytic with respect to a reference electorde, cor-
bath. rected to the hydrogen electrode, in a given
equilibrium (reversible) potential--the poten- electrolyte.
tial of an electrode in an electrolytic solu- stress-corrosion cracking--a cracking process
tion when the forward rate of a given reac-
tion is exactly equal to the reverse rate. The
equilibrium potential can only be defined ~These definitions are under the jurisdiction of Com-
with respect to a specific electrochemical mittee G-I on Corrosion of Metals and are the direct re-
sponsibility of Subcommittee l I on Nomenclature.
reaction. EffectiveJan. 8, 1971.
279
6,5
requiring the s i m u l t a n e o u s action of a cor- thermal spraying--a g r o u p of p r o c e s s e s

rodent and sustained tensile stress. This ex- wherein finely divided metallic or nonme-
cludes corrosion-reduced sections which fail tallic m a t e r i a l s are deposited in a molten or
by fast fracture. It also excludes intercrys- semimolten condition to form a coating.
talline or transcrystalline corrosion which The coating m a t e r i a l m a y be in t h e form of
can disintegrate an alloy without either ap- powder, c e r a m i c rod, wire, or molten mate-
plied or residual stress. rials.
STP534-EB/Nov. 1973
s(~l ~ Designation: G 16 - 71

APPLYING STATISTICS TO ANALYSIS OF
CORROSION DATA 1
This Standard is issued under the fixed designation G 16; the numberimmediately followingthe designation indicates the
year of original adoption or, in the case of revision, the year of last revision. A numberin parentheses indicates the year of
last reapproval.
1. Scope fleet statistical methods that are most relevant

1.1 Corrosion scientists and engineers are to analysis of corrosion experiments.
making increased use of statistical methods, 1.3 The recommended practice includes the
not only in laboratory programs, but often in following sections:
field failure analysis. Application of statistical Section
methods and interpretation of the results ob- Errors, Their Recognization, and Treatment. 2
tained is rendered difficult by the large Standard Deviation . . . . . . . . . . . . . . . . . . . . . 3
Probability Curves . . . . . . . . . . . . . . . . . . . . . 4
number of complex techniques that are avail- Curve Fitting--Method of Least Squares.. 5
able and a lack of standardization between Estimate of Limits That Include True Value
analytical methods employed by various of Mean (Confidence Limits) . . . . . . . . . . . 6
groups. Statistics as a discipline applies to Comparing Means . . . . . . . . . . . . . . . . . . . . . . 7
Comparison of Data on Probability Carves. 8
nearly all physical, biological, and economic Sample Size . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
sciences; this has led to development of a large Comparison of Effects--Analysis of Vari-
number of methods that are generally appliance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Two-Level Factorial Design . . . . . . . . . . . . . . Il
cable and complex. In contrast, the specific
application of statistics to corrosion problems 2. Errors, Their Recognition, and Treatment
often involves simplification and use of a lim- (1)~
ited number of methods.
2.1 Engineers are frequently faced with the
1.2 The purpose of this practice is to pro-
problem of making measurements in the labo-
vide a set of sample procedures that are in
ratory or in the field that are not completely
current usage in statistical analysis of corro-
accurate. It is common practice to repeat a set
sion experiments. It is recognized that the
of measurements; this repetition allows appli-
procedures selected are but a fraction of
cation of statistics to determine the degree of
methods available and that complete agree-
precision obtained. The basis for this ap-
ment on specific methods selected may not be
proach is that random errors tend to cancel
possible. The examples included are intended
out when a large number of measurements are
to provide a method for planning corrosion
averaged. For example, suppose that a techni-
experiments, analyzing data obtained, and
cian has been asked to prepare a large number
establishing the degree of confidence that can
of nominally ~-in. thick corrosion samples by
be placed in the results of specific experi-
cutting them from a large diameter bar of a
mental or field applications data. Alternative
relatively soft alloy. If the cutting is judged by
methods or improved approaches are con-
stantly being developed that may provide
more complete analysis and understanding of This recommendedpractice is under the jurisdiction of
specific experiments. Accordingly, Subcom- ASTM Committee G-I on Corrosionof Metals and is the
direct responsibilityof SubcommitteeG01.03 on Statistical
mittee G01.03 would welcome the comments Analysisand Planningof Corrosion.
and criticisms of readers so that future revi- Effective Jan. 8, 1971.
2The boldface numbersin parenthesesrefer to the list of
sions of the procedure may be updated to re- referencesat the end of this recommendedpractice.
281
G16
eye, it is obvious that not all of the samples bution curve. However, the results for those
will be the same thickness. However, when all who pressed down hard with their microme-
the samples are measured and an average ters on the soft metal would produce a low
thickness is calculated, it will be found that average and a curve whose maximum was lo-
most values lie close to the numerical average cated to the left of that shown in Fig. 1. Those
or mean. who did not press down firmly would obtain a
2.2 Statistical methods cannot eliminate high average and a curve to the right of that
experimental error, but statistics can provide shown. This type of error is a function of the
an indication of the magnitude of the possible experimental technique and is not a random
errors. Statistical methods are particularly error, it is called bias and is a systematic
useful in establishing the degree of confidence error which cannot be handled by statistical
that can be placed in a given measurement or analysis alone, although statistical methods
in a value calculated from a measurement. can sometimes be used to detect and identify
Statistical analysis is based on the premise bias.
that errors follow a normal distribution pat- 2.2.3 Mistakes--Mistakes either in
tern or some special case. The types of errors carrying out an experiment or in calculations
that occur in experiments arise in measure- are not a characteristic of the population and
ments or in handling of data. Proper care can preclude statistical treatment of data, or
during the experiment and in subsequent cal- lead to erroneous conclusions if included in
culations is essential to minimize unnecessary the analysis. Sometimes mistakes can be iden-
errors and to ensure that all sources of error tified by statistical methods by recognizing
can be properly identified and calculated. that the probability of obtaining a particular
2.2.1 Normal Distribution--In the example result is very low.
above if the number of samples of each meas- 2.2.4 Significant Figures:
ured thickness is plotted against the thickness, 2.2.4.1 Care should be exercised in re-
a curve called a histogram will be obtained. porting results to show the proper number of
Frequently this curve will approximate the significant figures. The location of the decimal
shape shown in Fig. 1. This is the so-called point can be used for this purpose. For ex-
normal distribution curve. It has certain char- ample the number 2700 can be written as 2.7
acteristics. It can be divided into equal-size • l0 s to show two significant digits, that is, an
segments on either side of the mid point which indicated accuracy of :el00. On the other
will include a certain fixed percentage of all hand 2700. indicates an accuracy of • I.
measurements. The first two are located equal 2.2.4.2 In carrying out calculations it is
distances from the mid point on either side to good practice to retain one insignificant digit
include 68.27 percent of the measurements. through the calculation to minimize rounding
These represent one standard deviation (• off errors. This insignificant digit should be
the standard deviation is discussed in the next rounded off in the result. For example the sum
section. Two standard deviations (+2a) on the of 2700. + 7.07 should be reported as 2707.
abeissa will then encompass 95.45 percent of not 2707.07.
the measurements and three standard devia- 2.2.5 Propagation of Error in Calculation:
tions (~: 3 a) will encompass 99.73 percent of the 2.2.5.1 Mathematical operations with ex-
measurements. It should be noted that not all perimental data will cause errors in the data to
experimental error is normally distributed, change in predictable ways. Two types of er-
and it is a good idea to plot histograms to de- rors are frequently discussed: maximum error
termine if the data fit a normal distribution. and probable error. Estimates of maximum
This is only possible when a large number of error can usually be found in descriptions of
points are available, for example, 20 or more. instruments, etc., and generally include sys-
2.2.2 Systematic Error--In the example tematic as well as random error. Probable
above suppose 10 people measured the thick- error refers to the standard deviation due to
ness of each of the corrosion coupons. The random error in systems where it is known or
plots of the resulting data for each person assumed that bias is negligible.
would generally also produce a normal distri- 2.2.5.2 Maximum error calculations can be
APPENDIXES 283
G16
handled by equation: weight. From Eq 5 the maximum error be-

tween the initial and final weights of a cor-
A Q = y~ = ," t aQlaX~ i AX, (I)
reded specimen is 0.4 mg. If the observed dif-
where: ference in weight is only 1 mg, the maximum
Q = calculated quantity of interest, a func- possible error is (0.4/1) • 100 = 40 percent.
tion of n measured variables denoted If the difference is 10 mg, the error is only 4
St, percent.
AQ = maximum error in Q, and 2.2.5.6 For another example, the velocity of
AX~ = maximum error in the independent flowing water is to be measured in a corrosion
variables. test. The water will be bypassed into a con-
This expression assumes that all of the X's are tainer for a given period of time and the
independent variables. If this is not true then amount collected will be weighed. Knowing
all partial deviations with dependent variables the pipe diameter, the average velocity can
must be grouped by independent variables in- then be calculated from the following equa-
side the absolute value bracket and the sign of tion:
the p a r t i a l deviations considered in each Vavz = W / t A p = 4W/TrD2tp (6)
group. where:
2.2.5.3 If standard deviation information is W = weight of water,
available then a different equation should be h = cross-sectional area,
used, namely: D = diameter,
o<Q) = [~, = , , ( O Q / ~ X , ) 2 o ~ ( x , ) ] ~ (2) t = time, and
p = density.
where a (X) represents the standard deviation /xW= • l b - - A n old scale will be used
of X and all the other terms are defined above. which is accurate to • lb. About 100
2.2.5.4 Again we assume that all X's are lb will be collected.
independent of one another. Both Eqs 1 and 2 At = • s --A c c ura c y of watch and ob-
can be simplified in specific cases. For ex- server is estimated to be • s. Total
ample if time will be about 70 s.
Q(x~x~f3...x.) = A(X~)a(X2) . . . . ( X . ) , (3) AD = • in.--Out-of-roundness and cal-
iper errors are expected to be •
then Eq 1 can be simplified to: in. for l-in. diameter pipe (inside di-
ameter).
AQ/Q = a(AXt/X~) + b(AXJX2)
+ 999 + j(AXJX,) (4) Ap = • l b / f t S - - T e m p e r a t u r e measure-
ment is expected to be 60 • 3 F, which
and with simple product functions the percent corresponds to densities limits of 62.38
errors are additive. Another simplification and 62.34 lb/ft ~. Thus, the error is of
occurs in the case when: the order of 0.1 percent. This is an
Q =aX~+bX2+ ..'jX, order of magnitude less than the other
errors and thus the error in this term
Then Eq 1 becomes:
can be neglected.
~Q = aAX~ + bAX2 + ... + jAX, (5) The maximum error then can be calculated by
2.2.5.5 For an apphcation of these calcula- Eq 1.
tions, consider the errors introduced into the
results from limited precision of equipment
such as a balance. Weighing is a common
measurement in conducting corrosion tests. 4•
= AW
The difference between initial and final 70 x 3.14 x 1 x62.3
weights is often used to calculate corrosion 4 x 100x 144
rates. These weighings are usually conducted + At
3.14 X | X 62.3 X (70)2
on a conventional laboratory analytical bal-
ance that is accurate to • mg. In weight 8X 100X 1728
+ AD
loss, L equals the initial weight minus the final 70 • 3.14 • 62.3 X 1
G16
= 0.042AW+0.060At + 100AD 3.2 The standard deviation of a large group

= 0.52 ft/s of n u m b e r s is defined below
Vav ~ = 4 W / r D ' t p
a=~ (7)
4 x 100x 144
= 3.14x I • 70 x 62.3 where:
o = standard deviation,
= 4.2 ft/s
d = x - g (where x = value and g = the
Percent error = (0.522/4.21) >< 100 = 12 per-
cent mean), and
n = total n u m b e r of observations.
2.2.5.7 A simplified form of Eq l can be
The definition also holds for a small group of
used in this case because Eq 6 is of the form
numbers if g is known independently; how-
shown in Eq 3. Rewriting Eq 4 for this case we
ever, if .f is not independently known, then
have:
with a limited n u m b e r of observations, only an
AV/V = I AW/W I + I 2AD/D I + I AT/T'I
estimate of the standard deviation can be
Then by inspection: m a d e which is:
A V / V = (5/100) + (2 (0.03)/I) + (1/70) s = x/-~V(n - 1) (8)
= 0.124 = 12.4 percent
or
and
s = V'n~x 2 - (Zx)'/n(n - I)~ (9)
AV = 0.124 • 4.21 = 0.52 ft/s
Equation 9 is convenient if a desk calculator is
2.2.5.8 A n o t h e r advantage to this simplifi- used. The square of the standard deviation, oa,
cation is that it is dimensionless and so elimi- is calied the variance of the data. The compu-
nates the need for converting units. Note in tations have been carried out in Table 1, and
the example given D is measured in inches but the mean and the estimate of the standard de-
must be converted to feet to be used in the viation are found to be 177.17 and •
equation given. Also note that the greatest mg/dmS-day, respectively.
reduction in error can be m a d e by increasing If d a t a from the 24 samples follow n o r m a l
the accuracy of weighing and of measuring the distribution (well-known bell-shaped curve),
pipe diameter. then
2.2.5.9 Simplification of Eq 2 along the g • s will include 68.27 percent of the results,
lines shown above are also possible. For ex- on the average
ample in the above case if the error figures • 2s will include 95.45 percent of the re-
were standard deviation rather t h a n m a x i m u m suits, on the average
errors the result would become: .f • 3s will include 99.73 percent of the re-
o ( V ) l V = [(0.05)' + (2)2-(0.03) ' + (1/70)211/2 sults, on the average.
= [0.0063]' ~ = 0.0795 or 8.0 percent
Then the standard deviation of V would be 4. Probability Curves (2, 3, 4)
0.34 ft/s. 4.1 A r i t h m e t i c probability paper is so con-
structed that d a t a from a n o r m a l distribution,
3. Standard Deviation (2) when plotted on the paper will be randomly
distributed about a straight line. To plot the
3.1 The 24 values listed under x in Table 1
curve, the data must be arranged in ascending
are weight loss data in m g / d m ~ - d a y tbr a par-
order of value and the cumulative percent of
ticular alloy exposed several m o n t h s to sea
tests must be determined for each observation
water. An over-all description of d a t a may be
from the following equation:
expressed as (1) the mean, ~, the sum of all
values divided by the total n u m b e r of values, P(%) = 100 [(i - 0.375)/(n + 0.25)] (10)
n; (2) the median, the mid value (the average where:
of the 12th. and 13th values since n is even) in i = position of d a t a point in total ranking,
ascending order; often more meaningful than and
the mean when there are one or two values n = total number of d a t a points.
vastly different from the rest, and (3) the The data from Table 1 are used to calculate
standard deviation, a. P(%)'s which are shown in Table 2.
APPENDIXES 2 8 5
G16
4.2 The results are plotted on arithmetic the plotting position for the first specimen is
probability pape# in Fig. 2. It is not necessary 1/(10 + 1) = 0.0909 etc.
to fit a straight line to these data since the line 4.6 The above data when plotted on ex-
can be plotted accurately from the mean and treme probability paper produce the straight
the standard deviation. The mean (~), 177.17, lines shown in Fig. 4, thus indicating an ex-
is plotted at 50 percent on the abscissa and the treme value distribution. By extrapolating the
mean plus the standard deviation (,g + s) plotted lines, one can make certain predic-
077.17 + !0.71) is plotted at 84.13 percent on tions. For example, with the 2-week data it
the abscissa. A straight line through these can be seen that the probability of obtaining a
points establishes the slope of the curve. (The pit 760 #m in depth or less is 0.999 and that
mean minus the standard deviation (~ - s) the probability of obtaining a pit greater than
could also have been used as one of the 760 #m is only one in 1000; whereas the ob-
points.) served deepest pit depth was 580 #m.
4.3 Some data exhibit log-normal distribu-
tion. For example, the time to fracture of
5.1 Curve Fitting--Method of Least Squares
aluminum alloys in stress-corrosion cracking (s)
tests in salt solution follows such a distribu- 5.1 To fit data to a linear plot of the form y
tion. A plot of these data on log-normal proba- = m x + b , it is necessary to solve two equa-
bility paper~ produces a straight line. Endur- tions:
ance times for the stress-corrosion cracking of mZx 2 + bZx = ~xy
an aluminum .alloy and the log-normal proba- mY~x + b n = Z y
bility plot of these data are presented in Table or
rn = ( n X x y - Z x X y ) / [ n Y ~ x ~ - (Zx)~]
3 and Fig. 3, respectively (3). The probability b = ( l / n ) ( Y ~ y - mY, x )
curves are plotted from the log mean (at 50
where:
percent) and the log mean minus or plus the
~x = sum of allx points,
standard deviation of the logs (at 16 percent,
~y = sum of ally points,
84 percent) given in Table 3.
~x 2 = sum of squares o f x points,
4.4 A special form of distribution is ex-
~xy = sum of x points multiplied by y
treme value analysis. This type of distribution
points, and
has been used to analyze maximum pit depths.
n = number of points.
The pit depth distribution on a given number
For a parabola of the form y = a x 2 + b x + c
of corrosion coupons may follow a normal dis-
three equations must be solved:
tribution on each coupon, but the maximum
aZx 4 + bZx 3 + cZx 2 = Zx~y
pit depths on each of the coupons follow a aZx 3 + bZx 2 + c•x = Xxy
special distribution of extreme values. The a X x ~ + b x x + cn = X y
mathematics of extreme values are complex,
5.2 Data for exposure of five replicate speci-
but practical use of the technique has been
mens of Zircaloy-2 to 750 F-1500 psi steam
simplified by the use of extreme probability
are presented in Table 5. it is known that the
paper.
corrosion kinetics of Zircaloy-2 obey two rate
4.5 Aziz (4) has used extreme probability
laws, an initial cubic-to-parabolic ~'ate fol-
paper in the study of the pitting of aluminum
lowed by a linear rate. In 750 F steam the rate
alloys. As an example, consider the maximum
becomes linear after about 42 days. Thus the
pit depths observed for sets of 9 or 10 samples
data in Table 5 comprise the i n i t i a l reaction
exposed to tap water for exposure periods
kinetics which follows a power formula of the
ranging from 2 weeks to 1 year presented in
general type:
Table 4. The data are ranked in order of in-
creasing pit depth. The plotting position for W = kt a
each ranking is determined by R / ( n + 1) where:

where R = rank and n = total number of W = weight gain (the oxide is extremely ad-
specimens. Thus, where 9 specimens were ex- herent),
amined, the plotting position for the first
specimen is 1/(9 + 1) = 0.100, for the second 3Keuffel and Esser No. 359-24 has been found satisfac-
is 2/(9 + 1) = 0.200 etc. For 10 specimens tory.
G16
k = a rate constant, d e p e n d e n t of m e a n s at e a c h e x p o s u r e period.

t = time, a n d ( V a r i a n c e should be i n d e p e n d e n t of m e a n s of
a = dimensionless. function o f the d a t a ( a r i t h m e t i c , l o g a r i t h m s ,
T h e a b o v e e q u a t i o n c a n be expressed in the e x p o n e n t i a l , etc.)) T h e difference, d, f r o m the
l o g a r i t h m i c form: predicted value, y - 5', w a s d e t e r m i n e d for
log 14" = a l o g t + l o g k e a c h e x p o s u r e period. All the aPs were s q u a r e d
a n d s u m m e d a n d s w a s d e t e r m i n e d for t h e en-
A plot of the above l o g a r i t h m s p r o d u c e s a
tire sample.
s t r a i g h t line whose slope is a. T h u s , the l o g a -
rithms of the d a t a c a n be used to fit a c u r v e by s =~ - 2) = V:0.0644/(30 - 2) = 0.048
the least squares m e t h o d f o r a s t r a i g h t line b y
A t the 95 p e r c e n t c o n f i d e n c e interval 2s = (2)
setting: (• = • T h e n the c a l c u l a t i o n s o f
y = log W the lines for 95 p e r c e n t c o n f i d e n c e are as fol-
x = log t
b = log k lows: ( F o r 1- a n d 4 2 - d a y exposures).
m =a l o g w = l o g k + a l o g t • 2s
At l d a y , l o g w = 0.897 + ( 0 . 4 6 9 x 0 ) •
5.3 T h e l o g a r i t h m i c f o r m s o f the d a t a a n d = 0.897 + 0.096 = 0.991
the required s u m m a t i o n s a r e s h o w n in T a b l e 6 = 0.897 - 0.096 - 0.803
( a l t h o u g h only one value is given for x t h e r e is At 42 days, log w = 0.897 + (0.469 • 1.62)
• 0.096
one for each value o f y, t h a t is five for e a c h = 0.897 + 0.760 + 0.096
exposure time). The c a l c u l a t i o n s of m a n d b = 1.753
are s h o w n below: = 0.897 + 0.760 - 0.096
= 1.561
mZxy - ~xZy
m The a n t i l o g s are:
n Y . x ~ _ ('~Yc)~
At I day, 0.991 = 9.79 mg/dm 2
30 • 41.303 - 27.75 • 39.92 0.803 = 6.36 m g / d m ~.
- 0.469 At 42 days, 1.753 = 56.6 mg/dm 2
30 x 35.001 - (27.75)~ 1.561 = 36.4 mg/dm 2
I
b =-- (Zy-mY.x) 5.5 These d e v i a t i o n s d o not include the ef-
n
fect o f v a r i a b l e slope which b e c o m e s increas-
39.92 - (0.469 • 27.75) ingly i m p o r t a n t in regions a w a y f r o m the
0.897 m e a n . T h e v a r i a b l e slope c o u l d be p l o t t e d in
3O
Fig. 5 as t w o lines passing t h r o u g h the d a t a
togk = b=0.897 m e a n with slopes o f m ~ 2s (m).
k = 7.89
a = m = 0.469
6.1 E s t i m a t e o f L i m i t s that Include T r u e
T h u s , the e q u a t i o n of the curve best fitting the Value of M e a n (Confidence Limits) (6)
d a t a is
6.1 W h e n d e a l i n g with a s m a l l n u m b e r o f
W = kt a observations, the e s t i m a t e of the limits t h a t
W = 7.89t ~
include the t r u e value o f the m e a n c a n be ob-
5.4 T h e a b o v e c a r v e is plotted on l o g - l o g tained from:
p a p e r in Fig. 5. T h e c u r v e also s h o w s a plot o f A = • t (s/v;)
the 95 percent (2s) limits o f d a t a . T h e s e were
o b t a i n e d f r o m the e s t i m a t e o f s t a n d a r d devia- where:
tion o f residuals a b o u t the curve, c a l c u l a t e d s = e s t i m a t e o f s t a n d a r d deviation,
f r o m the e q u a t i o n : n = number of observations, and
t = s t u d e n t ' s t f r o m published tables.
s = ~ 2 ) F o r e x a m p l e , the 14-day-75 F s t e a m d a t a for
w h e r e d = y - ~ a n d ~ is the w e i g h t g a i n pre- Z i r c a l o y - 2 a r e c a l c u l a t e d below:
dicted f r o m the a b o v e e q u a t i o n at a given level Five observations: 25.6, 25.5, 24.3, 26.9, and 27.1
mg/dm 2
o f e x p o s u r e t i m e t. T h e l o g a r i t h m s o f d a t a
Mean ~.~1 = 25.9
p r e s e n t e d in T a b l e 6 were used on the a s s u m p - Estimate of standard deviation s = ~ - I)
tion t h a t v a r i a n c e of l o g a r i t h m s o f d a t a is in- = ~ - 1) = 1.15
APPENDIXES 287
G16
The values of t are obtained from Table 7 at 50% 95% 99%

the a p p r o p r i a t e degrees of freedom (n - 1). t = 0.706 2.306 3.355
For 4 degrees of freedom the value of t = A = • (SIVa) = • • •
0.741 at 0.50 probability, 2.776 at 0.95 proba- U p p e r limit 2.2 4.2 5.4
- 1.4 - 2.6
bility, and 4.604 at 0.99 probability. Then cal-
culating the limits that would include the true
The above limits of the differences between the
mean at particular levels of confidence:
two m e a n s do not include zero in the first
Z = • t (siva) column so that the two m e a n s are statistically
Confidence different at the 50 percent confidence level.
L i m i t s , percent Deviation The limits of the differences between the two
50 A-= • • (l.15/X/~) = • means include zero in the last two columns
95 A = • • (I.15/~r = •
99 A = • • (1.15/~r = •
and the two means are not significantly dif-
ferent .at the 95 percent and 99 percent con-
Limits o f M e a n fidence levels.
Lower Upper 7.3 An alternative m e t h o d of c o m p a r i n g
50 25.5 26.3 means has been described by F r e e m a n (7) who
95 24.5 27.3 uses the equation:
90 23.5 28.3
Thus based on the above samples, we could be l=

99 percent confident that the true mean is con- d
tained between 23.5 and 28.3.
x/in~Sx z + n ~ x 2 / n x + n~. - 2)[(l/nx) + ( l / n v)]
7. Comparing Means (5,6,7)
7.1 The m e a n s of two sets of replicate ob-
where:
servations can be c o m p a r e d by determining
= difference between the m e a n s o f x and y,
the estimate of the limits of the difference
nx = n u m b e r of variates of the x's,
between the two means. If the limits include
Sx2 = variance of the x's,
zero, the means are statistically alike; if they
n x = n u m b e r of variates of the y's, and
do not include zero, the m e a n s are different.
sy 2 = variance of the y's
7.2 A s an example, determine whether two
The degrees of freedom are nx + n x - 2.
heat t r e a t m e n t s of Zircaloy-2 produce a signif-
The calculated value of t is c o m p a r e d with
icant difference in corrosion behavior. The
the t a b u l a t e d values of t (see T a b l e 7) for the
heat treatments, corrosion data for 14 days in
appropriate degrees of freedom. If the calcu-
750 F steam, and calculations of an e s t i m a t e
lated value is larger than the value from the
of the standard deviation, s, are presented in
tables, the difference is significant at t h a t con-
Table 8. An s for both m e a s u r e m e n t s is calcu-
fidence level. If the calculated value is smaller,
lated as follows:
the difference is not significant at that confi-
s - ~ x / ( Y , d~2"+ Y~d2*)/l(nz - 1) + (n2 - 1)]
dence level. F o r example, a s s u m e that the cal-
= %/(5.29 + 24.35)/[-(5- I) + (5 - 1)] = 1.92 culated t for two m e a n s obtained from five
The s for both m e a s u r e m e n t s is now multi- samples in each is 2.604. Then in Table 7
plied by V ~ . e x a m i n e the v a l u e s of t at 8 d e g r e e s of
Corrected s = s (V~-) = (1.92)(1.41) = 2.70 freedom (5 + 5 - 2). The calculated value
2.604 is greater than the t a b u l a t e d value of
(s is multiplied by V ~ because differences in-
0.706 at the 50 percent confidence level and
crease s by V~'.) The limits of the difference
greater than 2.306 at the 95 percent confi-
between the means is calculated from
d~,nce level, so that the m e a n s are significantly
a = t (s/~/'~ different at these levels. However, at the 99
Difference between means = 27.3 - 25.9 = 1.4
s = 2.70 percent confidence level the m e a n s are not
n=5 significantly different because the calculated
Degrees of freedom = ( 5 - 1 ) + ( 5 - 1)= 8 value of 2.604 is smaller than the tubular
F r o m Table 7, value of 3.355.
G16
8. Comparison of Data on Probability Curves dm 2. From Section 6 the desired equation is

(s) = • t(slC~)
8.1 Data can be compared by plotting con- where:
fidence limits for each curve on probability A = limits that.include the true value of the
paper and determining whether they overlap. mean (af a particular level of confi-
As an example, the log-probability distribu- dence),
tion of stress-corrosion cracking endurance of t = student's t,
aluminum alloys with and without silver addi- s = estimate of the standard deviation, and
tions are plotted in Fig. 6. Not all specimens n = sample size.
failed. The estimate of standard deviation, s, is Thus,
obtained from the plot as the difference be-
x/n-= tslA or n . = t2s2/A ~
tween the log endurances at 16.2 and 50 per-
cent probability. At the median (50 percent) In the above equation, t is a function of n. For
the 95 percent confidence limits are calculated a first approximation assume that n = 16;
to be: then t = 2.131 (from Table 7).
• t(s/vr~ n = (2.131)~(10)2/(5)2 = 18.2
where: Substituting the value of t corresponding to a
t = student's t (see Table 7), and sample size of 18 and recalculating for n
n = number of failed specimens, would give a more accurate value for n. Ac-
At one standard deviation (• the limits tually, the value of t for 18 samples at the 95
must be expanded by adding • to percent confidence level is not greatly different
• Lines through these points to the from that for 16 samples. Thus, under the
limits at the median establish approximate above conditions it is estimated that a sample
confidence limits. size of 18 would be required to assure that, at
8.2 Since not all specimens have failed, the 95 percent confidence level, the limits of
• is plotted one standard deviation the true value of the mean would not exceed
from the median of specimens that have • m g / d m 2 with an assumed estimate of the
failed. This results in limits being rather standard deviation of 10 m g / d m 2.
broad at high and low probabilities, a con- 9.3 A reasonably small number of speci-
sequence of lack of data in this region. mens can be used if a corrosion experiment is
8.3 Although confidence limits overlap in so designed that replicate specimens are ex-
Fig. 6, the lines from either sample are not posed to an environment and specimens are
included in confidence limits of the other; removed periodically for evaluation (such as
therefore, within the approximate range of descaling). If there are, say, six to eight expo-
probability of 2 to 50 percent, silver addition sure periods, the removal of triplicate speci-
has a significant effect on stress-corrosion mens at each period can furnish statistically
cracking. significant results, by calculating a standard
deviation based on all the data rather than for
9. Sample Size (9) .a single exposnre period. That is, determine
9.1 One of the most frequently asked ques- the mean for each exposure period; determine
tions in corrosion work is "How many sam- the difference, d, from the mean for the speci-
ples should I test for each condition?" The mens at that exposure period; square the d's
and determine the estimate of the standard
statisticians usual answer is "'What are the
deviation by the usual equation:
limits you wish to put on the results?"
9.2 Assume that it is desired to determine s = x / ~ 2 / f n - 1)
the corrosion behavior of a new alloy in a This approach assumes that the variances at
chemical environment and that prior tests with the several exposure times are the same. The s
similar alloys have produced an estimate of determined by this method provides a better
the standard deviation (s) of 10 m g / d m 2. Fur- estimate of the standard deviation for a given
thermore, it is desired, at the 95 percent confi- exposure period than that obtained from the
dence level, that the limits that include the three replicate samples examined at each pe-
true value of the mean do not exceed 5 rag/ riod.
APPENDIXES 2 8 9
481, G16
9.4 On the other hand, extremely large sam- For three-way interactions,
ple sizes are required to obtain significant re- SS = (I/n) [(~tl) 2 + (~t~) ~ + ... +'(]gtz)2]
suits if the evaluation is a go-no-go type, such - [(?n/N) + SS for all main effects and
as pitting versus no pitting or cracking versus interactions]
no cracking. Snedecor has assembled probabili-
For error term,
ties for observation of these types. They are
Zt 2 - l(?n/N) + SS
listed in Table 9. A s an illustration assume for all main effect and interactions]
that ten tubes were selected r a n d o m l y from a
where:
heat exchanger and were e x a m i n e d t h o r o u g h l y
S S = sum of squares,
for stress-corrosion cracking. If cracks were
n = n u m b e r of data within e a c h level
found in only one of the ten tubes, it would be
being compared,
predicted at the 95 percent confidence level
l = sum of d a t a c o m m o n to a given level
that between 0 and 45 percent of the re-
of the m a i n effect,
m a i n d e r of the tubes would contain a stress-
Ix, t~ 9 -. tz = test results c o m m o n to a
corrosion crack. On the other hand, if none of given c o m b i n a t i o n of the levels of
the ten tubes contained a crack that it would the two m a i n effects (two-way inter-
still be predicted at 95 percent confidence that action) or three main effects (three-
between 0 and 31 percent of the r e m a i n i n g way interaction). For example, in the
tubes would contain a crack. It can be seen alloy and test interaction 2;t's is the
from T a b l e 9 that no c r a c k s in 100 tubes sum of 12 data points for a given
would reduce the predicted percentage to 0 to alloy and test and n = 12,
4 for the r e m a i n d e r of the tubes. T = sum o f a l l t h e data, and
10. Comparison of Effects--Analysis of N = total n u m b e r of observations.
10.2 The sum of squares and m e a n s q u a r e
Variance (10)
are determined for each m a i n effect, two-way
10.1 The d a t a presented in Table 10 are the interaction, three-way interaction, and error
results of l a b o r a t o r y i m p i n g e m e n t tests in 3 term. The m e a n square, M S --. S S / D F , where
percent N a C I solution. C o p p e r alloy speci- DF = degrees of freedom. The degrees of
mens 1 by 4 by 0.05 in. were bolted radially to
freedom are:
the periphery of nonmetallic disks. Each disk
Tests: ( 2 - I ) = I
carried four specimens of each of four alloys. Alloys: (4 - 1) = 3
T h e m a x i m u m p e r i p h e r a l v e l o c i t i e s of the Velocities: (3 - 1) = 2
outer edge of the specimens were 20, 25, and Tests-alloys: (2 - l) • (4 - 1) = 3
Tests-velocities: (2 - 1) x (3 - I) = 2
40 ft/s. The test was run for 10 weeks and the Alloys-velocities:(4 - 1) • (3 - 1) = 6
m a x i m u m pit depth was obtained for each Tests-alloys-velocities: (4 - 1) • (3 - 1) •
specimen. The whole test was then repeated. (2 - 1) = 6
Error term: (96 - 1) - (the sum of the DF's of
There were the f o l l o w i n g sources of variation: all main effects and interactions) = 72
4 alloys, 3 velocities, 2 tests, and 4 replicate
specimens. There were the following m a i n ef- 10.3 The m e a n square for each effect is di-
fects: " a m o n g alloys," "'among velocities," vided by the m e a n square of the m o s t signifi-
"between tests"; the following two-way inter- cant interaction containing that effect or the
actions: alloys-velocities, alloys-tests, veloci- error if none of the interactions are signifi-
ties tests; one three-way interaction: alloys- cant. The result is c o m p a r e d with values from
velocities-tests; and an e r r o r t e r m (derived F tables which may be found in m o s t text
from the v a r i a t i o n a m o n g replicate specimens). books on statistics (see T a b l e i 1). The F value
The e q u a t i o n s are: is found by locating the degrees of freedom in
the error term down in the table. If the calcu-
For m a i n effects,
lated value is g r e a t e r than the F value, the ef-
SS = ( I / n ) ~ t ~ - ( ? n / N )
fect is significant. If it is less than the F value,
For two-way interactions, the effect is not significant.
SS = ( l / n ) [(~tt) ~ + (2;t2)2 + -.. + (~tz)~] 10.4 C a l c u l a t i o n of the sum of squares and
- [(?n/N) + SS for each of the main effects] m e a n squares is shown in Table 12. Analysis
4Bib G16
of variance is shown in Table 13. be further subscripted as yA+B+C+, and

YA+B+C+.. The error sum of squares for the
11. Two-Level Factorial Design (8) experiment is:
11.1 The two-level factorial design experi-
ment is an excellent method for determining
which variables have an effect on the outcome.
The significance of each effect can be deter- Each pair of responses must be squared, then
mined by analysis of variance. added, and also added then squared. For ex-
11.2 As many variables as possible that ample, the responses for A + B + C + would be
may be expected to have an effect on the out- treated in the following manner:
come should be included in the original experi- (1.86)~ + (1.95)2 - 1/2 [I.86 + 1.95]3 = 0.004
ment. In order to simplify the following exam-
Each of these figures is then summed, to give
ple, only three variables will be used.
the error sum of squares, which in this ex-
11.3 Assume that the stress-corrosion
ample is 0.0789. The error degree of freedom
cracking endurance of aluminum alloys is
is (2 - 1) (8) = 8. The 2 is the number of
being evaluated on alternate immersion tests
times the response is replicated, and the 8 is
in 3 percent NaCI. Suppose that one alloy
the number of pairs.
contains silver and another does not, and in
11.5 In studying the effects of variables it is
addition, that the effects of cold working and
mathematically easier to work with differences
overaging are to be studiedl The following
between levels rather than with means at each
nomenclature is then assigned:
level. The difference is referred to as a con-
A + Alloy with silver A - Alloy without silver
trast:
B+ Withcold work B - Without cold work
C + With overage C - Without overage 9~ = ( I / N ) [Zy+ - Z y ]
where:
A+ A- = contrast or effect of silver,
B+ B- B+ B- N = number of tests, which is 16,
y+ = any response in the A + columns, and
C+ y_ = any response in the A - columns.
C- and ~ are calculated in a similar manner.
9The interactions A'B, A~'C, 1~, and A~'C use
11.4 This experiment requires eight entirely the same procedure, except the signs for the
different sets of conditions. In order to deter- responses are determined by products of the
mine the within-sample error more accurately, signs for ~ e variables. For example, (A+)
it is wise to replicate each condition. It is thus (BF~) is ~,~.~) and ( A + ) (B+) ( C - ) is
necessary to perform a minimum of 16 sepa- A B C - . For AB, A + B + i s ( + ) , A + B - i s ( - ) ,
rate tests. In this particular example, the out- A - B + is ( - ) , and A - B - is (+). The abso-
come is the log of the endurance of each lute value of each response remains the same.
stress-corrosion specimen. Each effect or contrast has (2 - 1) degrees of
freedom. The 2 is for the levels at each condi-
A+ A- tion.
B+ B- B+ B- 11.6 Each contrast is squared and multi-
plied by the number of tests (16) to obtain the
1.86 2.54 2.01' 3.02 sum of squares. Table 14 shows the values as
C+
1.95 2.43 2.32 2.89 they are used in analysis of variance. F is the
ratio of the sum of squares of the effect to the
1.65 2.32 1.98 2.56 error sum of squares. An F distribution table
C- shows that for 1 degree of freedom for the
1.73 2.25 1.87 2.60
greater sum of squares (numerator) and 8 de-
Each res onse can be identified by its loca- grees of freedom for the lesser sum of squares
tion. For example, Y^+B+c+ has two re- (denominator), the 5 percent and 1 percent
sponses, which are 1.86 and 1.95. They can levels of F are 5.32 and 11.25, respectively.
APPENDIXES 291
@ G16
Thus, in this example, there is less than 1 per- A+ A-
cent probability that the B effect is caused by
random error. On the other hand, the re- B+ B- B+ B-
mainder of the effects are not significant. 1.86 3.02
11'.7 If this were a true problem, it would C+
show that materials without cold work were 1.95 2.89
not as susceptible to stress-corrosion cracking 2.32 1.98
as materials with cold work. The addition of C-
silver and overaging had no significant effect. 2.25 1.87
Note that this is a hypothetical example.
11.8 Each time an additional variable is to 11.11 With the previous method for anal-
be studied, twice as many experiments must be ysis of variance it is found that ABC cannot
performed to complete the two-level factorial be obtained ' because the negative values
~ A
are
missing and that contrasts ~, = BC B =
design. When many variables are involved, the
number of experiments becomes prohibitive. AC, and ~ = AB. In this particular example,
11.9 Fractional replication can be used to an assumption that all the interaction effects
reduce the amount of testing. When this is are unimportant is correct and it is possible to
arrive at the same conclusions that were ob-
done, the amount of information that can be
tained from the full factorial design experi-
obtained from the experiment is also reduced.
ment. In some cases, it may be that the interac-
11.10 The example of the factorial design
tion effects are much greater than the effects
with three variables will be used. However, the
of the main variables, in which case an as-
negative" side of the A'BC contrast will not be
sumption would lead to drastically wrong con-
included.
clusions. It is wise to have some idea about the
effect of interactions before fractional replica-
tion is used.
REFERENCES
(1) Mickley, H. S., Sherwood, T, K., and Reed, C. caloy-2," a thesis presented in partial fulfill-
E., Editor, Applied Mathematics in Chemical ment of the requirements for the degree of
Engineering Second Edition, McGraw-Hill Master of S~:ience, The Ohio State University,
Book Co, New York, N.Y., 1957, pp. 46-99. 1961.
(2) Douglas, D. M., ARMCO Steel Corp. "Are (6) Youden, W. J., Experimentation and Measure-
You Using or Abusing Data'?," paper presented ment, National Science Teachers Assn., Wash-
at the 1965 joint Northeast-North Central Re- ington, D. C., 1962.
gional Meeting National Association of Corro- (7) Freeman, H. A., lndustrial~Statistics, John
sion Engineers, Pittsburgh. Pa., Oct. 4-6, Wiley & Sons, inc., New York, N.Y., 1942.
1965. (8) Haynie, F. H., Vaughan,. D. A., Phalen, D. 1.,
(3) Booth, F. F., and Tucker, G. E. G., "Statistical Boyd, W. K., and Frost, P. D., "A Funda-
Distribution of Endurance in Electrochemical mental Investigation of the Nature of Stress-
Stress-Corrosion Tests," Corrosion, CORRA, Corrosion Cracking in Aluminum Alloys,"
Vol 21, 1965, pp. 173-177. AFML-TR-66-267, June 1966.
(4) Aziz, P. M., "'Application of Statistical Theory (9) Snedecor, G. W., "'Statistical Methods Applied
of Extreme Values to the Analysis of Max- to Experiments in Agriculture and Biology,"
imum Pit Depth Data for Aluminum," Corro- 4th Ed., Iowa State College Press, Ames,
sion, CORRA, Vol 12, 1956, pp. 495-506t. iowa, 1946.
(5) Hanes, H. D., "The Effect of Gas-Pressure (10) Contribution by D. H. Thompson, Anaconda
Bonding on the Corrosion Resistance of Zir- American Brass Co., Waterbury, Conn.
G16
TABLE I Computing Standard Deviation
x d an x d aTM x d an
190 13 169 178 1 1 178 I I

195 18 324 162 15 225 164 13 169
169 8 64 162 15 225 189 12 144
185 8 64 171' 6 36 178 I I
180 3 9 192 15 225 171 6 36
178 I I 172 5 25 172 5 25
170 7 49 195 18 324 156 21 441
179 2 4 181 4 16 185 8 64
= 177.17
a=lx-~l s = ~ ) = ~ = 10.71
TABLE 2 /'(%) = 100[(i - 0.375)/(n + 0.25)] = Cumulative Probability (see Fig. 2)
i P(%) Data i /'(%) Data i P(%) Data
1 2.6 156MDD 9 35.5 172MDD 17 68.5 181MDD

2 6.7 162 10 40 172 18 72.5 185
3 10.8 162 II 44 178 19 77 185
4 15 164 12 48 178 20 81 189
5 19 169 13 52 178 21 85 190
6 23 170 14 56 178 22 89.2 192
7 27 171 15 60 179 23 93.3 195
8 31.5 171 16 64.5 180 n = 24 97.4 195
TABLE 3 Endurances of Aluminum-5 percent Magnesium Stress-Corrosion Specimens Exposed Anodicully in 3 percent
NaC| Solution (see Fig. 3)
I ntensiostatie 66, 70, 72, 73, 75, 75, 76, 77, 80, 80, 82, 82, 82, 88, 89,
40 m A / i n ) 90, 9 I, 9 I, 92, 92, 93, 93, 94, 94, 94, 95, 96, 96, 96, 97,
97, 97, 97, 99, 99, 100, 100, 100, 101, 106, 106, 106, 107, 107, 107,
108, 108, II0, III, 115, 116, 116, 116, 116, 117, 117, 118, 119, 120, 122,
122, 122, 123, 126, 127, 128, 130, 130, 132, 133, 135, 135, 136, 140, 147,
150, 152.
Geometric mean = 103.2
Mean of Iog~o endurance = 2.014
Standard deviation of Iog~o endurance = 0.0844
Potentiostatic 50, 52, 57, 60, 60, 60, 62, 63, 63, 64, 66, 66, 67, 67, 67,
-0.34 V 67, 67, 68, 68, 69. 69. 70, 70. 70, 70, 70. 71, 71, 71, 71,
(S.C.E.) o 72, 72, 72, 72, 72, 72, 72. 73, 74, 74, 74, 74, 75, 75, 75,
76, 76. 76, 76, 76. 76, 77. 77, 77. 78, 78, 78, 78, 78, 78,
80, 80, 80, 80, 80, 81, 81, 81, 82, 82, 82, 83, 83, 83, 83,
84, 84, 85, 85, 85, 85, 85, 86, 86, 86, 86, 86, 87, 88, 88,
89, 90, 90, 92, 92, 92, 92, 92, 93, 93, 94, 94, 95, 95, 97,
97, 97, 98, 98, 99, 99, 99, 99, 99, 100, 100, 100, 102, 105, 105,
108, 112, 112, 115.
Geometric mean = 80.15
Mean of Iog,o endurance = 1.90387
Standard deviation of logto endurance = 0.0697
a Saturated Calomel Electrode.
APPENDIXES 293
G16
TABLE 4 Ordered Maximum Pit Depths Developed on Alcoa 3S-O Coupons Immersed in Kingston Tap Water for the
Time Peiiods Slmwn Together with Their R ~ m k s and Plotting Positions (see Fig. 4)
2 Plotting I Plotting 2 Plotting 4 Plotting 6 Plotting I Plotting

Rank Weeks Position Month Position Months Position Months Position Months Position Year Position
1 330 0,1000 570 0.0909 600 0.1000 620 0.0909 640 0.0909 700 0.0909
2 460 0,2000 620 0.1818 670 0.2000 620 0.1818 650 0.1818 70(" 0.1818
3 500 0.3000 640 0.2727 770 0.3000 670 0.2727 670 0,2727 750 0.2727
4 5(}0 0.4000 640 0.3636 790 0.4000 680 0.3636 700 0.3636 770 '.3636
5 530 0.5000 700 0.4545 790 0.5000 720 0,4545 720 0.4545 700 0.4545
6 540 0.6000 740 0.5454 830 0.6000 780 0.5454 730 0.5454 810 0.5454
7 560 0.7000 780 0.6363 860 0.7000 780 0.6363 750 0.6363 820 0.6363
8 560 0.8000 810 0.7272 930 0.8000 800 0.7272 770 0.7272 830 0.7272
9 580 0.9000 840 0.8181 1030 0.9000 830 0.8181 780 0.8181 830 0.8181
I0 ...... 910 0.9090 ...... 920 0.9090 850 0.9090 930 0.9090
TABLE 5 Weight Gain of Zircaloy-2 in 750 F Steam at

Time Indicated, rag/din 2
Days
l 3 7 14 28 42
9.8 I 1.8 20.3 25.6 34.8 47.2

7.2 I 1.8 19.7 25.5 36.0 49.2
6.6 10.5 19f0 24.3 34.1 47.3
8.5 13,8 22.3 26.9 34.8 48.6
9.9 13.9 22,4 27.1 41.7 52.2
TABLE 6 Least Squares Calculation Zircaloy-2 in 750 F Steam
Log of time (x): 0 0.48 0.85 I, 15 1.45 1,62

Log of weight gain (y): 0.99 1.07 1.31 1,41 1,54 1.67
0.86 1.07 1.29 1,41 1.56 1.69
0.82 1.02 1.28 1,39 1,53 1,67
0.93 1.14 1.35 1.43 1.54 1,69
1.00 1.14 1.35 1,43 1.62 1,72
Zx = 27.75 Z y = 39.92 ~xy = 41,303 Z x z - 35.0015
TABLE 8 Comparing Means--Zircaloy-2 for 14 Days in

TABLE 7 Distribution of t
750 F Steam
Degrees Probability 1450 F WQ 1650 F WQ
of
Freedom 0,50 0.95 0.99 25.6 Mean = 25.9 25.5 Mean = 27.3
25.5 Xan = 5.29 26.8 ~a n = 24.35
I 1.000 12.706 63.657 24.3 s = 1.15 26.8 s = 2.46
2 0.816 4.303 9.925 26.9 27.2
3 0.765 3.182 5.841 27.1 30.5
4 0,741 2.776 4.604 s~,2 for both measurements
5 0.727 2,571 4.032 = X/'{~d, ~ + Y.d2Z)II(n~ - I) + (n~ - l)J
6 0.718 2.447 3.707
8 0.706 2.306 3.355 = "~/(5,29 + 24.35)/[(5 - I) + (5 - I)J = 1.92
15 0.691 2,131 2.947 Correct s = s~. ~(x/'2) = 1,92 (X/2) = 2.70
t
30 0.683 2.042 2.750
99 0.676 1.984 2.626
0.674 1.960 2.576
'0
~a
P._
f't
.-- i
o
i i+ 0
~ ~ ~---
o
Z
o
Z
3"
Z
- - - ~ ~ ....... ~ .... ~ Z ~
~ o ~
o
Z
,-I
n
~ - ~ # ~ R ~ - o .~ . . ~. . -. . .
o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o ~ o o o ~ o o o o o o
o ~ ~ ~ ' o~ ~~ ~~ -~ o~ ~~ ~~ -' o~ 1 .7 . . 6. . 1 ~ 7 - 6o 1~ ~7 -6 ~1 7 6 1 7 6
.m
APPENDIXES 295
TABLE 10 Maximum Depth of Pitting, Mils

Velo- Test I Test 2
city,
ft/s 20 25 40 20 25 40
Alloy A 24 19 33 32 21 37
24 18 35 27 20 44
22 20 31 34 21 42
23 21 34 26 19 43
Alloy B 23 21 31 29 18 40
22 19 36 32 21 37
22 19 30 31 20 36
20 19 33 27 20 38
Alloy C 5 28 21 2 11 6
4 30 19 3 9 8
5 20 24 4 18 7
5 23 18 3 13 6
Alloy D 10 3 7 6 7 11
6 13 14 14 4 12
10 4 8 11 6 12
7 3 9 II 7 10
TABLE 11 Partial Table of the Distribution of F (5 percent) Top (1 percent)-Bottom
DF of DF of Effect MS
Error MS 1 2 3 4 5 6 7 8 9 10
1 161 200 216 225 230 234 237 239 241 242
4052 4999 5403 5625 5764 5859 5928 5981 6022 6056
2 18.51 19.00. 19.16 19.25 19.30 19.33 19.36 19.37 19.38 19.39
98.49 99.00 99.17 99.25 99.30 99.33 99.34 99.36 99.38 99.40
3 10.13 9.55 9.28 9,12 9.01 8.94 8.88 8.84 8.81 8.78
34.12 30.82 29.46 28.71 28.24 27.91 27.67 27.49 27.34 27.23
4 7.71 6.94 6.59 6,39 6.26 6.16 6.09 6.04 6.00 5.96
21.20 18.00 16.69 15.98 15.52 15.21 14.98 14.80 14.66 14.54
5 6.61 5.79 5.41 5,19 5.05 4.95 4.88 4.82 4.78 4.74
16.26 13.27 12.06 11.39 10.97 10.67 10.45 10.27 10.15 10.05
10 4.96 4.10 3.71 3,48 3.33 3.22 3.14 3.07 3.02 2.97
10.04 7.56 6.55 5,99 5.64 5.39 5.21 5.06 4.95 4.85
25 4.24 3.38 2.99 2.76 2.60 2.49 2.41 2.34 2.28 2.24
7.77 5.57 4.68 4.18 3.86 3.63 3.46 3.32 3.21 3.13
50 4.03 3.18 2.79 2.56 2.40 2.29 2.20 2:13 2.07 2.02
7.17 5.06 4.20 3.72 3.41 3.18 3.02 2.88 2.78 2.70
70 3.98 3.13 2.74 2.50 2.35 2.23 2.14 2.07 2.01 1.97
7.01 4.92 4.08 3.60 3.29 3.07 2.91 2.77 2:67 2.59
80 3.96 3.11 2.72 2.48 2.33 2.21 2.12 2.05 1.99 1.95
6.96 4.88 4.04 3.56 3.25 3.04 2.87 2.74 2.64 2.55
100 3.94 3.09 2.70 2.46 , 2.30 2.19 2.10 2.03 1.97 1.92
6.90 4.82 3.98 3.51 3.20 2.99 2.82 2.69 2.59 2.51
~c 3.84 2.99 2.60 2.37 2.21 2.09 2.01 1.94 1.88 1.83
6.64 4.60 3.78 3.32 3.02 2.80 2.64 2.51 2.41 2.32
@ 61e
TABLE 12 Calculations Based on Section l0 and Examples of Table 10
TZ/N = (1811)2/96 = 34163.76041
SS for tests = y4,[(895) 2 + ( 9 1 6 ) 2] - T ~ / n

= 4.59375 DF = 1
MS = 4.59375
S S f o r a l l o y s = y2,[(670) 2 + ( 6 4 4 ) 2 + ( 2 9 2 ) ~ +
( 2 0 5 ) z] - T ~ / N
= 7124.78125 DF = 3
MS = 2374.92708
SS for velocities = ys2[(524) 2 + ( 5 1 5 ) 2 + ( 7 7 2 ) 3]

_ T2/N
= 1329.52084 DF = 2
MS = 664.76042
SS for tests-alloys = ~/1z[(304) 2 + ( 3 6 6 ) 2 + . . . +

( I l l ) 2] - [ T 2 / N + 4 . 5 9 3 7 5 +
7124.78125]
= 811.78125 DF = 3
MS = 270.59375
SS for tests-velocities y~6[(232) 2 + ( 2 8 0 ) 2 + . . . + ( 3 8 9 ) 2]

- [ T 2 / N + 4.59375 + 1329.52084]
= 172.31250 DF = 2
MS = 86.15625
SS for alloys-velocities = Ys[(212) ~ + ( 1 5 9 ) z + ( 2 9 9 ) 2

+ (206) 2 +(157) z + (281) 2
+ (31) 2 + ( 1 5 2 ) 2 + ( I O 9 ) ~ + (75) 2
+ (47) 2 + (83) 2]
- [T2/N + 7124.78125 + 1329.52084]
=
1924.56250 DF = 6
MS = 320.76041
SS for tests-aUoys-velocities = Y41(93) 2 + (78) 2 + . . . + ( 4 5 ) 2]

- [TZ/N + 4.59375 + 7124.78[25
+ 1329.52084 + 811.78123
+ 172.31250 + 1924.56250]
=
129.43750 DF = 6
MS = 21,57291
SS for error = [(24) 2 + (24) * + (22) 2 + . . . + ( 1 0 ) 2]

[T2/N + SS for all main eltects
and interactions]
438.2500 DF = 72
MS = 6.08680
APPENDIXES 297
616
TABLE 13 Analysis of Variance

95 percent Level 99 percent Level
Degrees of Calculated
Freedom M S Ratio F Signi- F Signi-
Value ficant? Value ficant?
Tests 4.59375
1,3 - - = <1 10.13 no 34.12 no
Tests, Alloys 270
Alloys 2374.92708
3,6 - - = 7.45 4.76 yes 9.78 no
Alloys, Velocities 320
Velocities 664,76042
2,6 - - = 2.07 5.14 no 10.92 no
Alloys, Velocities 320
Tests, Alloys 270.59375
3,6 - - = 12.5 4.76 yes 9.78 yes
Tests, Alloys, Velocities 21.57
Tests, Velocities 86.15625
2,6 - - = 4.0 5.14 no 10.92 no
Alloys, Velocities 320.76041
6,6 - - = 14.5 4.28 yes 8.47 yes
6,72 - - = 3.54421 2.23 yes 3.07 yes
Error 6.08680
TABLE 14 Analysis of Variance

Sum Degrees
Effect Contrast of of F
Squares Freedom
-0.1575 0.397 5.03

-0.3275 1.716 21.75
+0.1287 0.265 3.36
+0.0337 0.018 0.23
A'~ -0.025 0.010 0.13
~'~ -0.013 0.0~28 0.04
+0.0188 0.0056 0.07
Error 0.0789
Number of Specimens
of Spe " "
-3s -2s -is O Is 2s 3s

Average
Thickness
FIG. 1 NormalDistribution Curve.
@ G' 8
Sample Mean R = 177.17
Standard Deviation s = 10.71
Corrosion Weight
Loss, MDD
195 - O O
190
O
+ s = 187.88
185
~.]3%
180
= 177.17
175
O0 50%
170 m
Z-s=
166.46
165
O 15.87%
160 B
155
I I I I I [ t [ ~ I I I
2 5 10 20 30 40 50 60 70 80 90 95 98
p (~) - Cumulative percent of Tests
( Cumulative Probability )
FIG. 2 Cumulative Frequency Distribution of 24 Corrosion Tests (see Table 2).
APPENDIXES 299
@ G'~e
Log. (Endurance - Minutes )
2.20
2.15 J
[ntenatostatic~
S o I
\ J /
2.10 /
/
2.05
o
2
2.00 L
f S
1.95
1.90
-Potenttostat| e
1,85 "~
./
1.9o ~//~" v"
1.75 / ~
J
1.70 /
0.10.2 0.5 1 2 5 10 20 30 40 50 60 70 80 90 95 98 99 99.8
99.5 99.9
Cumu[ative Probability
FIG. 3 Endurance of Aluminum--5 Percent Magnesium Alloys Exposed Anodically in 3 Percent NnC] Solution (See
Table 3).
y ffi Reduced Variate

Return Period = T
Frequency ~ (X)
5000
0.9997
8
0.9995
0.9993 Probability Return Period
7 0.999 lOOO_ - 1000
0.9990
0.998
6
0.997
0. 995
5 -0.993
0.990 - 100
4 0.980 1 Year~
I /rl Month
0.970 6 Months/ t { \ /
3 --0.950 2 I / / ~ / a/ - ~-nths
...u -
o o
0.900 onth_s 10
2
1--0.7000"800 ~~ f 5] 8deepest
0 ( )~ ~ - ' /
f'l Mean
o/
0.500 2.00
[ Mode
oo oo
-1 o.loo o X - 1.1o
0.050 //" /
0.010
0.005
-2 0.0010 -- 1.001
-• nasa 1.0001
0.0001 Observed Variate = X
300 400 500 600 700 800 900 1000 1100
Maximum Pit Depth in Microns
FIG. 4 Maximum Pit Depth Data for Alcan 3S-O (AA3003-O) Immersed in Kingston Tap Water for the Time
Periods Shown Plotted Against Their Cumulative Relative Freq~ncies (See Table 4).
APPENDIXES 301
4S[b Gle
Weight Gain, m g / d m 2
100
80
60
40
I
- j f~"//j
J
/
i
20
_
/ J
~~~ / / / ~ - 2 s Limits
i
i0
8
i i W = 7.89t 0"469
i
Zircaloy-2, 750 F Steam
I I I I I I I ] I I
2 4 6 8 10 20 40 60 80 100
Exposure Time, Days
FIG. 5 Zircaloy~ Expo~dto 750F-1500psiSteam.
10,000 II
"r,
1,000
~/
'22 .0ooe
Lira[its
iloo
twith silver = 13
twithout silver
9" ~
ra With Silver
26 of 64 Specimens Failed
O Without S i l v e r
36 o f 6 4 S p e c i m e n s F a i l e d
0.1 I I I Ill i I I I I
0.5 1 2 5 10 20 40 60 80 90 95 98 99.5
Cumulative Probability, %
Confidence limits code:
.g~l Region within 95 percent confidence limits on best fit line for endurances of alloys containing silver
k~ Region within 9.5 percent confidence limits on best fit line for endurances of alloys withom silver
ml Region common to both
FIG. 6 Effect of the Addition of Silver on Stress-Corrosion-Cracking Behavior of 7079 Type Aluminum Alloys.
By publication o f thin standard no position is taken with respect to the validity o f any patent rights in connection there-
agai~tst liability for infringement o f any Letters Patent nor assume any such liability.

STP 534-1973

Uploaded by

Copyright:

Available Formats

STP 534-1973

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

STP 534-1973

Uploaded by

Copyright:

Available Formats

Copyright by ASTM Int'l (all rights reserved); Fri Jan 1 23:09:17 EST 2016

Sponsored by ASTM Committee G-1

ASTM SPECIAL TECHNICAL PUBLICATION 534

List price $16.75

Jt~[~ AMERICAN SOCIETY FOR TESTING AND MATERIALS

Printed in Baltimore, Md.

The Manual of Industrial Corrosion Standards and Control has been

Frontispiece: Photograph of U.S. 35 Highway Bridge, Point Pleasant, W.Va. taken

This manual is a working source book of procedures, equipment, and

Webster [1] 2 defines corrosion as "the action or process of corrosive

Basic Corrosion Principles

--DILUTE HYDROCHLORIC ACID

I- z..- -z.--I I-I

FIG. 3--Schematic drawing o f a metal-ion concentration corrosion cell.

T A B L E 1--Standard electromotive force series (emf) at 25 C [4].

A u +++ + 3e- = Au 41.50 N o b l e ( m o r e c a t hodi c )

TABLE 2--Galvanic series of metals and alloys.

Noble (more cathodic) Platinum

It is also possible for dissolved oxygen to participate directly in the cathodic

or more elements or if the solution environment contained more than one

FIG. 5--Schematic drawing o f an oxygen concentration corrosion cell

increases. In the case of the oxygen reduction reaction, this polarization

I- 2 45j ~ ' J " [ METAL 9

corrodes at a high rate. As the concentration of acid is increased this

Passivity may thus be broadly defined as the decrease in corrosion sus-

would be obtained if Fe, Fe(OH)2 and Fe'(OH)3 were considered. The

Forms of Corrosion Attack

consider the pitting attack of aluminum in an oxygenated solution of

FIG. lO--Schematic drawing illustrating the autocatalytic nature o f pitting attack on

acidic conditions. The accumulation of positive charge in the solution

cussion of cathodic protection as a means of controlling corrosion damage

grain boundary chromium precipitates and hence prevents preferred grain-

FIG. 13--An electronmicrograph showing the preferred formation of MgZn~ precipitates

FIG. 14--An eleetronmierograph showing the selective attack o f MgZn2 precipitates in a

been devised to separate the effects of stress and corrosion in materials

in which corrosion processes and tensile stresses interact will depend

decreasing damage may be due to a lubrication effect. As might be ex-

High Temperature Oxidation

~ M+,,,>20 - ~-02 (b) ,M%<- 2 02 (d)

which combines to give

Thus as in aqueous corrosion, high temperature oxidation consists of an

Methods of Corrosion Prevention and Control

may aggravate corrosive damage. Electrochemical methods too are available

anodic surface treatments. Aluminum may be given a thin adherent

aluminum, for example, anodized coatings are produced by making the

Designing for Minimum Corrosion

port, turbulent flow and gas or air entrapment should be minimized.

Electrochemical Protection Methods

~ J DIRECT CURRENT ! "~-

FIG. 17--Schematic diagram of the cathodic protection of a buried pipe.

cient to achieve complete protection, acceleration of localized corrosive

Corrosion Standards and Control in the

Petroleum refineries use water, steam, hydrogen, acids, and a variety of

1 Research and Development Department, Amoco Oil Company, Whiting, Ind.

extra thick walls as a corrosion allowance, whereas the seating surfaces of

HC~ H!S 0 JET ~UEE

= ORGANIC SULFUR COMPOUNDS AND H2S ABOVE ABOUT 500~

FIG. 1--Major corrodents in a petroleum refinery.

CRUDE NEUTRALIZER .'L

Crude distillation, the first step in refining, is outlined in Fig. 2. Most