ATURAL

CDRRDSIQN OF COPPER BY N
SOLUTIONS

WATER AND MLUTE AQUEOUS

on of M. S.

Thesis fer flu Dog!
WERSETY

MECHiGAN STATE UN
fiéwsrd May

W58

éahn

"I NWWIIIWIWWHHDWWII 4L

_ 3 1293 10229 6336_ -

 

 

 

MSU

LIBRARIES
n

 

 

 

RETURNING MATERIALS:
Place in book drop to
remove this checkout from
your record. FINES will
be charged if book is
returned after the date
stamped below.

 

 

(\l‘ {mu
W

I
a.“ - , fl???
2 4”,: a,
a. “j "

w:

 

 

 

CORROSION OF COPPER BY NATURAL WATER
AND DILUTE AQUEOUS SOLUTIONS

by
JOHN EDWARD MAY

A THESIS

Submitted to the College of Engineering of
Michigan State University of Agriculture
and Applied Science in partial fulfillment
of the requirements for the degree of

MASTER OF SCIENCE

Department of Chemical Engineering
1958

Tho "1
Sum for
in help ext
Hum 13

The vri
uho has toll
valuable sue

 

I- 4. ” ?
J 02“ J ’f.
j} j 4/ (/fa’

ACKNOWLEDGEMENT

The writer wishes to express thanks to Dr. C. Fred
Gurnham for guidance and assistance in this investigation.
The help extended by Dr. Malvern F. Obrecht and Dr. Clyde
C. Dewitt is also greatly appreciated.

The writer is also indebted to Dr. Laurence L. Quill

who has followed the progress of this study and offered many
valuable suggestions.

11

hm: John
Born: Remus
Academic Car

191.0-
19h5-
19L8-
1955-
398ms Helc

Bachc
Hichi

IS'ib'stx'ial 1

1950.

1952.

 

 

VITA

Name: Jehn Edward May

Born: Remus, Michigan; February 11, 1926

Academic Career:

19h0-hh
19h5-h8
l9h8-SO
1955-58
Degrees Held:

Bachelor
Michigan

Mecosta High School
lecosta, lichigan

Central Michigan College
flaunt Pleasant, Michigan

Michigan State University
East Lansing, Iichigan

Michigan State University
East Lansing, Hichigan

of Science, 1950
State University

Industrial Experience:

1950-52

1952-sh

Product Engineer
Dow Corning Corporation
Midland , Michigan

United States Army
Chemical Corps

iii

£4

rf

CORROSION OF COPPER BY NATURAL WATER
AND DILUTE AQUEOUS SOLUTTONS

by
JOHN EDWARD MAY

AN ABSTRACT

Submitted to the College of Engineering of
Michigan State University of Agriculture

and Applied Science in partial fullfillment
of the requirements for the degree of

MASTER OF SCIENCE

Department of Chemical Engineering
1958

Approved:

 

iv

Co
:iluto
:crrosi
to cor
mica
mluat
hereas

53
mm
:alor 1:
ninth
:msi

”PM

ABSTRACT

Copper has excellent corrosion resistance to water and
dilute aqueous solutions but is not completely immune to
corrosion in an aqueous environment. In this investigation
the corrosiveness of dilute aqueous solutions containing
chemical impurities commonly found in natural water was
evaluated in order to determine to what extent these impurities
increase the corrosion rate.

By use of a technique developed for determining the
quantity of corrosion products in surface films and a
colorimetric method for evaluating the copper content of

solutions it was possible to measure the distribution of
I corrosion products and the total extent of corrosion of
capper specimens corroding in water and dilute solutions.

Clean cepper specimens immersed in water were found to
corrode at a.high rate during the first few hours. The
initial corrosion rate at 60°C in water saturated with
atmospheric gases was 5.6 mg. per sq. cm. per day. The
corrosion rate decreased rapidly as surface films formed
on the metal surface indicating that the corrosion resistance
of cOpper to aqueous solutions is due largely to the protection
provided by surface films of corrosion products.

Thermodynamic calculations were carried ou1;which show

that oxygen is necessary for the corrosion of cOpper in water.

V

Ionosion t!
the results
low oxygen
were predom
Soluti
sodium cart
high-puri t3
ppm and am
the rate 0:
affectivelj
Water
tcrmsive
0? corrosi
liter and
tontrati or
In tests 1
m 20 Pm
in the W81
”mad the
313301,” ‘1
In 3.
source d 1‘
113301” d

M

«.3 'ate

Corrosion tests with oxygen-free water confirm.this conclusion.
The results of corrosion tests with copper in water having a
low oxygen content show that the initial corrosion products
were predominantly of the cuprous form.

Solutions of LOO ppm sodium.bicarbonate and 50 ppm
sodium carbonate were found to be no more corrosive than
high-purity'water. Sodium sulfate at a concentration of ho
ppm.and sodium.chloride at a concentration of 10 ppm increased
the rate of film development but fully developed films
effectively inhibited corrosion in these solutions.

Water containing 70 ppm carbon dioxide was much more
corrosive than pure water and dissolved a larger quantity
of corrosion products. In static corrosion tests with pure
water and carbon dioxide~free solutions the capper con-
centration of the aqueous phase was usually less than 0.5 ppm.
In tests with neutral solutions containing sodium.bicarbonate
and 20 ppm free carbon dioxide, 2 to h ppm of capper dissolved
in the water phase. Solubility tests with cuprous oxides
showed that neutral solutions containing free carbon dioxide
dissolved more than 1 ppm.of cuprous oxide.

In staticcorrosion tests, natural water from a local
source did not appear to be excessively corrosive but it
dissolved higher concentrations of copper than pure water.
{fine water might be corrosive under turbulent conditions

because of its ability to dissolve protective surface films.

vi

 

IERODUCTION

Basic Prir
in Aqueous

Thermodyne
Systems .

Factors W}
EKEERIMEN TA]
JSRRCSION .

Materials

Analytical

leaaureme:

Static Co:
or Soluti

Descrip M

Human

Corrosion
Eater . .

fin Effec
“953 of w

COI'I‘OS ion

TABLE OF CONTENTS
PAGE

InmowCTION O O O O O O O I O O O I O O O O O O O O 1

Basic Principles of Corrosion Reactions Occurring
in AunOUB SOlution' e e e e e e e e e e e e e e e 3

Thermodynamdc Relationships of the Copper-Water
sys tom, 0 O O O 0 O O O O O O O O O O O O O O O O O 8

Factors Which Influence Corrosion Reactions . . . . 13
EXPERIMENTAL METHODS USED IN EVALUATING COPPER
CORROSION . . . . . . . . . . . . . . . . . . . . . . 21
Katerials and Reagents . . . . . . . . . . . . . . 22
Analytical Procedures . . . . . . . . . . . . . . . 2h
Measurement of Corrosion Products in surface Films 25

Static Corrosion Test Used to Evaluate Corrosiveness
Of SOlutionl e e o o e e e e e e e e e e e e e e O 27

Description of the Flow System Corrosion Test . . . 29

—

MPER IMNTAIJ RESULTS 0 O O O O O O O O O O O O C O O 31

Corrosion Characteristics of Capper in High-Purity
Water....................... 31

The Effects of Chemical Impurities on the Corrosive-
ness of Water . . . . . . . . . . . . . . . . . . . ho.

Corrosion of Copper in Slightly Acidic Water . . . hB

Corrosion by Water Containing Partially Neutralized
carbon Diox1d° O O O O 0 O O O O O O O O O O O O O 53

Corrosion CharacteTistics_of Copper in Natural Water 63

The Results of Flow System Corrosion Tests . . . . 67

vii

DISCY‘ SE
SW1!
31301

The I
Corr<

Sugg«

ERAFHS

DISCUSSION AND CONCLUSIONS

Summary of Significant Conclusions .

Discussion of Conclusions

The Influence of Chemical Impurities

Corrosiveness of Natural Water .

Suggestions for Future Work

GRAPHS AND TABULATION OF DATA .

Graphic Presentation of Data

,Tabulation of Data
LITERATURE CITED . . .

viii

PAGE

0 O O O O O O O 72

116
127

TABLE
I.
II.

III.

LIST OF TABLES
PAGE
Impurity Limits of HighpPurity Copper . . . . . 22

Solubility of Cuzo in Water Containing 600 ppm
NaHCOBand3lpmeree cog . . . . . . . . . . 61

.Concentration of Cepper in Solution After Two

Days in Contact with Powdered Cu20 or CuO . . . 63

Dissolved Solids and Cases Present in Natural
Water Used in Running Corrosion Tests . . . . . 6h

ix

41'

 

Lme
u atte
math
and to

U! A

'v‘,“ "S‘

INTRODUCTION

Cepper and copper alloys are widely used in the fabri-
cation of water tubing for potable water systems. These
materials usually give excellent service but they are not
immune to corrosion by natural waters. In this investigation
an attempt has been made to evaluate the cOpper corroding
capacity of dilute aqueous solutions and high-purity water,
and to gain information about the fundamental nature of the
corrosion process. The effects of dissolved substances
commonly found in natural water have been studied by evalu-
ating the corrosion characteristics of high-purity water
containing one or more of these substances. Relationships
between the corrosion rate and the concentration of specific
impurities were used to explain some sepects of the corrosion
process and to identify soluble materials capable of caus-
ing high corrosion rates in potable water systems.

Copper has a negative standard oxidation potential
(e.m.f.) and is considered to be a semi-noble metal. 'It
has excellent corrosion resistance to most natural waters;
however, water from a localized geographic area is sometimes
very corrosive to copper water tubing (8). In these specific
instances it appears likely that the corrosion rate is
increased by the presence of some impurity in the water.

The literature contains no data proving the acceleration of

copper c
here is
usual

Oh
if any
could i
extent
alkali
gases,
ewes

tion 1

3f tw

iMes

2

copper corrosion by components in natural water. However,
there is some evidence that points to the existence of
corrosion accelerating materials in certain waters.

One of the objectives of this study was to determine
if any of the chemical constituents found in natural waters
could increase the corrosiveness of water to a measurable
extent. Because most natural waters are neutral or slightly
alkaline and have low concentrations of dissolved solids and
gases, and because they are heated toonly moderatetemper-
atures in use, most of the laboratory work in this investiga-
tion was performed with neutral water containing snall
quantities. of dissolved substances, and all experimental
work was carried out at moderate temperamres.

Although corrosion reactions are no different than
other chemical reactions, they usually involve interactions
of two or more phases and are therefore more difficult to
investigate than reactions occurring in homogeneous systems.
The concentration of metal ions in equilibrium with a metal
surface is often very small and difficult to determine
accurately. Reaction products tend‘to deposit in thin films
on the surface where they~ cannot be measured with exactness.
Kinetic studies of such reactions are difficult because
diffusion through liquid and solid phases often controls
the corrosion rate. Thus, the over-all corrosion rate may

be mCh slower than the kinetics of the chemical reaction

"“161 1 nd ice to .

Corrosion is a universal problem and in economic
magnitude ranks as one of the most important problems with
which present-day technology is confronted. In spite of the
large amount of time and effort devoted to investigation of'
corrosion processes, the fundamental nature of the many
types of corrosion is far from clear. Much.work still
remains to be done before even a few of the commonly en-

countered corrosion reactions are completely understood.

Basic Principles of Corrosion Reactions Occurring in Aqueous
Silution. If

All metals, even those having noble characteristics,
have a solution potential. When metals are in contact with
a solvent, surface atoms tend to leave the metal lattice
and to enter the liquid phase as ions (16). The metallic
ions inlthe solution take on a positive charge while the
nmtaJ.‘becomes negatively charged. The process eventually
reaches an equilibrium, governed by the free energy relation-
ship for the particular reaction. Polar solvents tend to
solvate dissolved ions; the entropy of the system is in-
creased by ion solvation and the reaction free energy becomes
more negative. Non-polar solvents show little tendency to
solvate ions and are much less effective‘in dissolving metal
ions. It is doubtful if any metal atoms would dissolve in
a Infilixi phase in atomic form; physical erosion of metal
surfaces at moderate fluid velocities seems highly improbable.

 

lets
electron:
:‘he pote
is reach
sstsntia
sslving

equatlo

Ct

....

Metal ions, upon entering a solution, leave excess
electrons with the metal surface from which they migrate.
The potential of the metal increases until an equilibrium
is reached between the metal and its ions in solution. The
potential for the ion-electron half-reaction, capper‘ dis-

solving as‘a cuprous ion, can be calculated by the Nernst

equation (17).

Cu‘ 4» e' n = numbers of electrons

Cu

[11
ll

E0 - 2;222 106 cu‘
n acu

In electrOlytic corrosion, two half-reactions combine
to form the complete reaction. Electrons, accumulating
within the metal must be removed by a simultaneously occurr-
ing their-reaction, capable of consuming electrons, in order
to complete the reaction. The two half-reactions do not
have to take place in the same area; it is possible to have
them take place at separate cathodic and anodic areas.

SPhe electrolytic process is not the only means by
'hICh corrosion of a metal may take place. If an oxidizing
agent :is present in the solution the metal surface may be
snufoxunly attacked.‘ There is no need of postulating the
accumulation of electrons within the metal and the build-up

°f a Potential, because the electrons involved inthis reaction

are utilized in forming the chemical bonds of the corrosion

produ c 1: compounds 0

1*"
91...-

In
mastic

This t3

-

 

 

In the corrosion of a metal surface by dry air the
reaction must take place by a direct surface attack (13).
This type of reaction would also be expected with metal
corroding in an aqueous solution when an oxidizing agent is
present. The compound formed at the interface could dissolve
in the solution, indeed the solubility relationship between
it andrits ions in solution would have to be satisfied.

Thus, metal ions may be dissolved in a solution by a
mechanism which is fundamentally different from the electro-
lytic reaction.

All corrosion reactions require the presence of an
oxidizing component in the system. However, almost all
waters contain dissolved oxygen and often are saturated with
it. Water saturated with atmospheric oxygen at 25°C contains
8.14. ppm (parts per million) of dissolved oxygen. The
activity or fugacity of the oxygen is equal to the partial
pressure of oxygen in air with which it is in equilibrium.
Oxygen in the atmosphere has a partial pressure of 0.20
atmospheres. At this concentration its chemical activity
is high enough to enable it to enter into many oxidation
reactions. Solutions in equilibrium with atmospheric oxygen
would likewise be capable of oxidation due the presence of
dissolved oxygen. It cannot be assumed that a metal surface
immersed in such a solution would be protected from the

Otidizing environment of the atmosphere.

kw "‘

 

 

 

When a metal is corroded or oxidized the electrons
lost by the metal must be consumed by some other process
occurring in the system. In pure water free from any
oxidizing agent the only process capable of removing electrons
is the reduction of hydrogen ions to hydrogen gas. This

process can be represented by the following half-reaction.
+ - =
ZR + 2 e H2

The potential of this ion-electron half-reaction is
used as a zero point for the Standard Oxidation Potential
Series. However, in neutral water with a hydrogen ion
concentration‘of l x 10"7 equivalents per liter the potential
or this half-reaction is -0.L|.lh. volts instead of zero. In
spite of its negative standard potential in neutral water,
this half-reaction is in evidence in the corrosion of all
the more active metals. ,

Considering all the factors involved, there are four
possible processes by which corrosion may occur in a system
consistirg of a metal and pure water in equilibrium with the
atmosphere:

1. Electrolytic attack involving hydrogen ion
reduction.

2. Direct surface attack involving hydrogen ion
reduction. ,

3 . Electrolytic attack with oxygen reduction.

it. -Direct surface attack with oxygen reduction.

The
”direct
resctim
plate cl
letal a
in dire
banded
so she
in the

ninth

2N ov
it sep
trons
order
In thi
takes
10:13 I
511st

W

“Kim

 

 

. The distinction between "electrolytic attack" and
"direct surface attack" is based on fundamentally different
reaction mechanisms. Direct surface attack involves a com-
plete chemical reaction taking place in a localized area.
Metal atoms lose electrons to hydrogen ions or oxygen atoms
in direct contact, becoming either ionized or chemically
bonded ‘to hydroxyl ions or to oxygen atoms with which electrons
are shared. Compounds, formed in this process, may‘dissolve
in the solution or remain deposited on the surface if in- '
soluble.

In electrolytic corrosion it frequently happens that
the over-all reaction involves two half-reactions that occur
at separate and distinct cathodic and anodic areas. Elec- '
trons flow from one area to the other through the metal in
order to complete the electron circuit between the areas.
In this type of corrosion the reaction involving reduction
takes place at some distance from the “site where the metal
ions are formed. Solvated metal ions, upon entering the
solution, may travel some distance before combining with
anions to form insoluble products. Insoluble compounds,
formed at a distance, settling on the metal surface are not
likely to be of value in protecting the underlying metal.
Adherent insoluble products formed at the surface are much
more apt to impart subsequent resistance to corrosion.

[The difference between "direct surface attack" and

"electrolytic attack" is the occurrence of the half-reactions

210.5213

at a lo:
occurri
cathodi
they c

betuee

'5’"? ‘
“no.

how“

 

8

at a localized site or at separate sites. If, in a reaction
occurring by the electrolytic mechanism, the anodic and
cathodic areas are imagined to move closer together until
they coincide, it would then be impossible to distinguish
between the two types of attack.

If a metal surface were homogeneous, direct oxidation
would occur rather than electrolytic corrosion. Most surfaces,
however, are nonhomogeneous due to the presence of crystal
defects, grain boundaries, or localized stresses. Adsorbed
gases and solids, which may be present, offer more protection
to one area than another. These factors contribute to the
creation of areas of unequal potential necessary for
electrolytic corrosion. Probably both types of attack occur
sinnfltaneously with one type predominating over the other i

as surface conditions change.
ThermogLnamic Relationships of the Copper-Water System

Corrosion processes, like all other processes involving
chenrical reactions, are subject to thermodynamic treatment.
Corrosion reactions are more often than not extremely com-
Plicated. The liquid phase may contain several ionic species
and dissolved gases. In addition to the metal and water a

third phase is usually present as a surface film'consisting

' 01' pure or mixed metallic salts. Physical equilibria and

s°1\l‘l>:f.2l.ity relationships as well as chemical equilibria‘ will
simultaneously tend to be satisfied for all phases involved (31)-

syste‘
taine
in pL‘

mm
1 act

hair
I

redu

i. I, -

 

While complete thermodynamic treatment of such a
system is difficult, much.valuable information can be ob-
tained by considering a simplified system. A metal immersed
in pure degassed water should represent the simplest possible
corrosion system. For copper metal in such a system the
only films present would be oxides or hydroxides of cupric
or cuprous COpper. In a neutral system of this type the
hydrogen and hydroxyl ion concentrations are also fixed. The
reduction of hydrogen ions to hydrogen gas is the only half-
reaction possible to supplement the oxidation of copper."
Combining the two half-reactions the over-all reaction can

be obtained.

 

 

H+ s' e' = i HZ E° = 0.000
Cu = Cu* t' 9-, 3° = -0.522
H+ + Cu = 0u* _+ s H2 Egexf -0.522

Using the standard potentials (30) for the reactions

involved it can be seen that the potential for the oxidation
of copper to the cuprous form is negative. In this simplified
system the concentration of cuprous ion in equilibrium with
the metal surface may be calculated. The Nernst equation

'hen applied to the over-all reaction takes the following

form :

Ecol], = Eocell "' 9;?22 106 (Cu+) (112%
‘ (3} ) (Cu)

gas as
lents
cupro1

fasten

I n 4,...

ADIr-l

10

Using the activity of metallic cepper as unity, hydrogen
gas as one atmOsphere and hydrogen ion as l x 10"7 equiva-
lents liter, this eXpression can be solved for the equilibrium,

cuprous ion activity or concentration by equating the cell .

potential to zero.

0 = -0.522 - 0.059 log Lent; (1Y1?
' (1 x 10'7) (1)

Solving this equation gives a value of l.h.1 x 10"16
equivalents per liter for the cuprous ion concentration.
This is an extremely minute concentration and indicates that
the reaction proceeds to a very limited extent.

A system consisting of copper and pure water, if
open to the atmosphere would also contain dissolved oxygen.
At room temperature the oxygen concentration would be approx-
imately 8J4 ppm and the possibility exists for corrosion to

occur with dissolved oxygen as the active oxidizing agent.

This reaction can also be considered as the combination of

two _ half-reactions:

 

 

2 Cu = 2 out + 2 e‘ E0 = -0.522
it 02 + 2 11* + 2 e" = 1120 E° = +1.22;
i 02 +' 2 at + 2 Cu = 320 + 2 out 3391i +0.70?

‘The standard cell potential for this reaction is
positxive, indications that the free energy of the reaction
13 negative. It may be noted that hydrogen ions are involved

in this reaction, but not as the active oxidizing agent.

EN.

5
.1:

cent

'31

'r1

wk‘
a

C02:

 

11
_This might account for reports in the literature indicating
that the corrosion rate is dependent on both oxygen and
hydrogen ion concentration (22). ‘

Calculations of the cell potential for this reaction

are further refined by considering the concentrations of
the ionic species present in the system. The maximum con-
centration of cuprous ion is limited by the solubility

product relationship for cuprous hydroxide (23):

Cu* + 03‘ = CuOH Ks = 1 x 10‘1"
(Cu*) (03‘) =_ K8 = 1 x 10'1“
(out) = 1 x 10"7

In a neutral system_yith fixed hydroxyl ion concen-
‘trationthe cuprous ion concentration cannot exceed 1 x 10'7
equivalent per liter. The cuprous ion concentration can,
of course, be less than this value, but if it exceeds this
concentration, cuprous hydroxide will form as a third phase.

The activity of the oxygen in the system is equal to
its activity in the atmosphere and is approximately 0.20
atmosphere.

. Based on these values the actual cell potential can be

calculated.

.- I

Buy

 

 

12

E = 38.11- .ogg 105 (320) (eu+>2
(02)" (11*)2 (Cu)

 

E = 0.707 0 022 log (11 (1 x 10'7)2'g-
( .20)% (1 x 10-7f§.(1)
E = 0.707 - 0.0u1 = 0.666

This reaction has a positive cell potential and there-
fore a negative free energy "changes If this reaction pro-
ceeds beyond the solubility of cuprous hydroxide a precipitate
would be expected to form. The film built upon the surface
will, if the film is protective, limit the extent. of the
reaction. Otherwise, with a continual supply of oxygen,
the reaction will continue until all the copper is consumed.

. These calculations are for copper undergoing oxidation
to the cuprous form. Thermodynamicslly, capper would. be
~ expected to change to cuprous rather than the cupric state
in corrosion reactions. Similar calculations based on
oxidation to. the cupric state, give Values not greatly
different from theresults of the proceeding calculations
(2 x 10'26 for hydrogen. ion concentration, and 0.6111 for
cell potential with oxygen reduction). There are good indi-
cations in many corrosion processes that the cuprous compound
5-3 I'<>J:'n1ed initially and further oxidized to the cupric form

"hen sufficient oxygen is present (27).

Jun
TE‘AJM

a
'3. C1

1101

In!
tr!

IL,

:1-
"t.

\\.. " "”- .—
rev-9"” ‘

 

 

13
Although.these calculations have been made on an

extremely simple system the results are helpful in analyzing
actual systems. One point that is particularly significant
is the indication that corrosion involving hydrogen ion
reduction does not occur to any significant extent. Corrosion
of copper by aqueous solutions must be the result of oxida-
tion by dissolved oxygen. Some indication of the corrosion
characteristics of actual systems are obtained by consider-
ing the specific effects caused by altering the concentration
or adding impurities to the simplified system. In this
way it is possible to predict the effects of increased
velocity or the presence of individual ionic species in

potable water systems.
Factors which Influence Corrosion Reactions

Thermodynamic relationships set limits on corrosion
systems but they provide no information about the kinetics
of"the reaction or the rate of the over-all process. Factors
suck: as the rate of oxygen diffusion, the amount of protec-
tior: provided by surface films, the stability of such films,
and. the complexing of ionic species to form.soluble compounds
can aand.often do influence or control the corrosion process.
Factors present in the system that influence the rate of
forma tion and the extent of pitting, while not clearly under-
8"Odd at the present time, should have a tremendous effect

on the performance of equipment and water tubing in commercial

Water 8 ya tems .

x Y

i Ir

'49:-

 

ions 3

(3‘
a)

srfec

{a1\

1);"

"0n
mt.

can:

.
«‘6'

‘5‘;

111

If components that are capable of complexing with metal
ions are added to water, the corrosiveness of the water
may be increased. The potential of an unprotected metal

surface is influenced by the activity of its ionic species

in solution.

- 0 £2
E - E "' Deg 108 81

A reduction in ionic activity increases the metal
electrode half-cell potential, thereby increasing the
driving force of the corrosion reaction. The extreme
corrosiveness toward cOpper of solutions containing ammonia
is attributed to the complexing of copper ions with ammonia
(3h). Halogen ions can complex with copper ions and may
increase the corrosion rate if present in sufficient con-
centration (3).

Films normally capable of protecting a surface may
lose their effectiveness if complexing compounds are present
in solution. Surface films will attempt to maintain physical
equilibrium.with.the liquid phase. In a static system an
equilibrium would be established in which the instability
constant of coordination compounds and the solubility pro-
duct: of the surface film were both satisfied. The quantity
0f metal in solution could be quite high even when the
ionitz concentration is at a low level.

,IIn a flowing system, the ability of the solution to

remove a large quantity of metal in solution provides a

 

Irfi‘ -

 

large 5
2f tha

a high

CCPPJS

"5““
tied Us

””3.
~11“ .1:

. a
51?.8

a
M. C
Sana.

3~nn
IVie

i“ n‘n

“.5“

1‘“~*

can“

V : ‘
'2114

‘V w“.
3‘s ‘1

a“
'u

, 15

large driving force for film dissolution. Since any portion
of the film depleted tends to be renewed by metal corrosion,
a high rate of film dissolution results in a high over-all
corrosion rate.

In systems where oxygen is the active oxidizing agent,
corrosion can'be limited by the rate at which oxygen
diffuses to the surface. This will probably be true if no
surface film is present and the chemical reaction at the
surface is rapid. Results reported by Brown, Roetheli, and
Forrest (7) indicate that with pure water having a fairly
high.surface velocity the initial corrosion rate for
aluminum, zinc, and iron is limited by oxygen diffusion,
while with copper, nickel, and tin the kinetics of the
surface reaction is the limiting factor.

In corrosion by neutral waters, the protection pro- .
vided by surface films of corrosion products is probably
the most important factor limiting over-all corrosion rates.
It may be true that even perfect films do not provide com-
‘plete protection, yet even poorly developed films covering
part of the surface are capable of slowing down the over-all
corrosion reaction.

In a great many corrosion.reactions films compoSed
0f corrosion products form at the surface. The passivity
01' iron in nitric acid solutions is such an example (20).
The iron film is invisible, but it has been isolated and

its existence definitely proven. Copper too, can be made

 

he

Chi

16
passive and corrosion resistant by careful surface
treatment (19).

Retention of part of the corrosion products as an
adsorbed film.on the metal surface makes quantitative
measurement of the corrosion reaction very difficult. For
this reason weight loss measurements are of limited value
in corrosion studies (h).

A film covering a surface completely does not nec-
-essarily provide complete protection; metal ions and
electrons can diffuse through films of oxides or metallic
salts. Hydroxyl and chloride ions can penetrate surface
films and form.corrosion products at the interface between
the metal and film (38). A corrosion process occurring
under such circumstances would be expected to be slow;
further, the rate should be controlled by the thickness of
the protective film (25). If this is true, the rate
equation for the corrosion reaction takes the following

mathematical form:

rate = %% =1§ X = film thickness

Integrating this expression gives:

2 -
x - 2K (t2 - t1)
{Huis, the relationship between film thickness and time is

Parabolic in nature .

17

Some indication of the protective nature of surface
films is obtained by considering the relationship between
the atomic volume of the metal, and the molecular volume
of its oxide or salt. If the molecular volume of the
corrosion product is less than the atomic volume of the
metal a surface layer of oxide or salt molecules is not able
to protect the underlying metal. A porous film is formed
that is not very effective in preventing further corrosion.
Metals having oxide volumes less than the atom volumes in-
clude sodium, potassium, magnesium, and calcium.

Other metals such as aluminum, lead, iron, and tin
form oxides having bulk volumes greater than their metal

atoms and are capable of completely covering the metal
surface with an oxide film (2h). Copper has a ratio of
oxide volume to metal volume of 1.71 and would be expected
to form oxide films having protective properties.

The ability of surface films to play such a decisive
role in the over-all corrosion process accounts for many of
the irregularities observed in the corrosion of metals in
actusJ.service. Film structure and stability can be influenced

by small variations in solution concentration. Films may
Possess a crystal structure or at least an orderly arrange-
ment <of molecules. Two or more crystal structures may be
PossiJole, one more effective in providing surface protection

than another. Campbell (8) contends that the presence of

?F_

18

extremely small quantities of surface active agents could
alter the film structure and inhibit corrosion.

Metals that form oxide or salt films with bulk volumes
large enough to completely cover the surface may nevertheless
develop defective films. Oxides with large bulk volumes
may crumble due to lateral pressure caused by crowding of

the surface layer of molecules. Cracks developing in the

film expose underlying metal. Surface defects may also

appear at the crystal boundaries in films with a crystal-like
structure. .

The initial film formed may not be stable and may later
change in structure and composition to a more stable form.
Oxide films, for example, may lose or gain water of hydration.
Structural changes may create new defects in what was or-
iginally a film capable of providing complete protection.
Some metals such as copper and iron are capable of existing
in two relatively stable valence states. Films of oxides
in the lower valence state may oxidize to a higher form.

Chemical changes in the film change the structure and

physical properties. A higher oxide might be less capable

01' providing surface protection than the original lower.
oxide. In this way protective films are converted into
non-protective films. 7

Many times, particularly in flowing systems, the
Selubility of the surface films may be extremely important.

Even highly protective films need to attain a certain

19
critical thickness before they are effective in retarding
corrosion. With continuous movement of water overvthe
.surface any portion of the film dissolving in the stream
would, of course, be replaced by additional corrosion of
the metal. Film solubility and the degree of saturation of
flowing streams determine the driving ferce for film dis-
solution. Increased velocity would increase the diffusion
rate and increase the rate of film depletion. Complexing,
or the formation of coordinationcompounds containing metal
ions, is a means by which the metal-holding capacity of the
water is increased. The formation of any non-ionized
soluble compound has the same effect. Reduction of pH also
increases film.solubility. Any of these factors operating

to increase film solubility increase the corrosion rate.

. In electrolytic corrosion reactions metal ions leave
the surface and enter the solution as hydrated ions. It is
the writer's Opinion that insoluble products formed at a
distance from the surface would not be expected to have
.protective value because they are not formed in intimate con-
tactzwith the metal surface. However insoluble corrosion
products formedat a~ distance and having shar'ableT electrons
may,- on contact with a clean.metal surface, cling tightly
forming a relatively continuous local film. To have
Protective value corrosion products must be adsorbed on
the surface and remain firmly attached to it. Insoluble

compounds formed in solution and precipitated as a loose

.M

"Is-‘1

rq'hl H
rlflydl

I
can.

20
porous covering may actually cause pit corrosion by screening
the surface and setting up concentration cells due to
differential oxygen concentration. This could explain.why
pits, once they develop are not self healing and protective
coatings do not form on anodic areas. I I

Corrosion reactions in actual systems are usually
complicated and can be influenced by a large number of
factors.‘ In any one system not all the possible factors
are significant, one or two factors may control the overball
corrosion rate. However, in attempting to analyze a system
it is necessary to consider all possible variables and
determine which are important. The significance of some
factors can be evaluated in the laboratory while the effects
of others are difficult to predict without extensive {field
test. The complexity of corrosion processes is evident in
the fact that at present the mechanisms of many of the I
corrosion reactions occurring in large scale industrial

;processes are not understood.

EXPERIMENTAL METHODS USED IN EVALUATING COPPER CORROSION

In this investigation techniques were develOped for
determining the quantity of corrosion products formed when

c0pper strips are immersed in water or dilute aqueous

solutions. Since part of the corrosion products dissolve

in the aqueous phase and part of them remain attached to
metal as surface films, it was necessary to find methods of
determining the capper content of dilute solutions and to
measure the amount of copper in films of corrosion products.
Most of the experimental work was done with high-purity

cepper and high-purity water or water containing known

quantities of dissolved solids and gases. Some work was

done with natural water and commercial grade copper in order
to get an idea of the relative amount of corrosion occurring
in systems similar to those found in potable water systems.
Natural water contains a large number of impurities and it

is difficult to determine the effect of individual constitu-

ents on the corrosion process.
An attempt was made to design experiments that would

provide information about the fundamental nature of the
corunasion process and show the effects specific components

in the water phase have on the quantity and distribution

015' corrosion products. Information from the experimental

'OPk should be of value in determining the effect of

1mpurities on the corrosiveness of water. A detailed study

22
of highly pure, relatively simple corrosion systems should
provide basic information of value in attempting to describe
the over-all corrosion reactions occurring in more complex

systems.
Materials and Reagents

Nearly all the experimental work was done with high-
purity cepper obtained from the Central Research Laboratory,
American Smelting and Refining Company, Barber, New Jersey.
This high-purity copper was prepared in laboratory equipment
and was cast into rod form.in a nitrogen atmosphere. The
metal has a minimum purity of 99.999 per cent copper. The
impurity limits of this high-purity copper are listed in
Table I, showing the maximum amount of specific impurities

present.

Table I. Impurity Limits of High-Purity Copper.
(Per cent by weight).

Element Definite Probable
Fe 0.00007 0.00005
Sb 0.0001 0.0001
Pb 0.0001 0.00005
Sn. 0.0001 0.00005
Ni 0.0001 0.0001
.31 0.00001 0.00001
.Ag 0.00003 0.000001
.As 0.0002 0.0001
Cr' 0.00005 0.00001
Si 0.00001 0.000001
Te 0.0002 0.0001
Se 0.0001
S 0 .0001

02

0.000027

n are

“H”

a.“
‘Vvhl

'1
11“
(I)

23

The high~purity copper was manufactured by an elaborate
technique in order to produce a metal almost completely
free from.impurities commonly found in commercial grade
copper. The methods of production and specifications are
described in a publication of the American Institute of
Mining and Metallurgical Engineers (36). High-purity copper
rods of 3/8 inch diameter were cold-rolled into sheets
having a thickness of 30 mils for use in this investigation.
These sheets were cut into smaller strips that were used
in running corrosion tests.

O'The water used in preparing solutions used in corrosion
tests was also highly pure. Tap water distilled in a stain-
less steel still are further purified by passing it through
a Bernstead.ion exchange demineralizer. No effort was
made to remove dissolved gases. The purified water (referred
to .. high-purity or highly pure water) had a total dissolved
solids concentration of less than 0.05 ppm and a specific
conductivity of less than 1 x 10"6 mho. The purified water
was stored in Pyrex flasks. f

Oxygen-free water was prepared by refluxing high-
Purity water for several hours in order to expel the dis-
solved oxygen. This water was cooled by passing it through
a heat-exchange coil and was used immediately in corrosion
tests in order to prevent oxygen absorption from the

atmosphere .

21;

Chemicals added to high-purity water to produce dilute
solutions of known concentration were of chemically pure or
reagent grade. Solutions containing carbon dioxide were
prepared by bubbling commercial grade carbon dioxide through
high-purity water until the desired concentration was
reached. In preparing some test solutions it was necessary
to blanket the solution with nitrogen in order to prevent
oxygen absorption. Nitrogen used for this purpose was a

water-pumped commercial grade.

Analytical Procedureg_

The cOpper concentration of aqueous solutions was
determined by a colorimetric method develOped by Hall
Laboratories, Inc. (18). This method can be used to make
rapid and accurate measurements of copper concentration in
the range 0.05 to 2.5 ppm. Fifty ml. of solution is made
slightly acidic with concentrated hydrochloric acid; 5 ml.
of ammonium citrate reagent is added to increase the pH
to about 9.2, and 1.0 ml. of 0.1% aqueous solution of
sodium diethyldithiocarbamate is then added. The sodium
diethyldithiocarbamate complex with copper producing a
Yellowabrown.coloration in the solution. The cOpper complex
18 extracted with carbon tetrachloride, and its concentration
is measured with a Coleman Universal SpectrOphotometer at
MAO uni. The instrument reading is converted to a value for
the COpper concentration of the solution by use of a calibra-

tion curve .

25
Concentrations of oxygen and carbon dioxide in water
solutions were determined by methods described in Standard
Methods 225 the Examination.g£'flgtgg, Sewage, and Industrial
Wastes, 10th edition (1). The modified Winkler method was
used for oxygen determinations. The concentratibn of free
carbon dioxide was measured by titration with sodium.car-
bonate using phenolphthalein as an indicator. Measurements

of pH were made with a Beckman glass electrode pH meter.
Measurement of Corrosion Products in Surface Films

Corrosion of capper strips in contact with aqueous
solutions results in the formation of visible surface films
of corrosion products. The amount of copper dissolved in
the aqueous phase is often not large due to the low
solubility of copper oxides and salts in water. Most of
the corroded cepper is contained in the surface film. In
order to determine the total amount of corrosion products
it is necessary to make a quantitative measurement of the
copper.contained in the surface film.

Surface deposits of copper corrosion products are
rather loosely bound to the underlying metal. A technique
was developed by which it was“ possible to remove films
01' corroded copper and to measure the amount of copper they
contained. ' It is possible to remove most films on copper
“1131 a soft rubber eraser. Corrosion products remain

attached to the eraser crumbs, and the copper is extracted

1;_.=___

 

w
”4

26

from the crumbs by leaching with concetrated hydrochloric
acid. Distilled.water was added to dilute the COpper concen-
tration to a range where it could be measured colorimetrically.

Although this technique may seem somewhat crude it is
probably as accurate as any other method of determining
the copper content of surface films. It is possible to
remove some metal from a clean copper strip with an eraser
but errors due to metal wear are believed to be no greater
than ten per cent of the copper contained in the film.

Visual inspection of the metal surface indicates when all
surface deposits have been removed. The clean copper surface
has a bright shining appearance while even very thin films
cause the surface to have a dull cast. It is quite easy to
tell when the surface is clean and it is not necessary to
wear away bare metal by continued erasing.

An Eberhard Pink Pearl eraser was found to be satis-
factOry for removing surface films. Some erasers contain
abrasive materials that scratch the metal surface but the
Eberhard eraser did not appear to cause any surface scratches.
Newly formed films are quite easy to remove; films that are
more than a week old are more difficult. An electric eraser
was used in removing some of the more adherent films but it
was not used any more than was necessary. This method of
determining the copper contained in films of corrosion pro-

ducts may not-be extremely accurate but it seems to‘ give

27
reproducible results and is accurate enough to indicate the
general trend of the film build up in corrosion reactions.

Static Corrosion Test Used to Evaluate Corrosiveness of
SOlutions ‘

A static corrosion test was used to evaluate the extent
to which copper was corroded.when immersed in aqueous so-
lutions. Glass-stopped Pyrex bottles having a capacity of'
300 ml. were filled with the solution being tested. Strips
of copper 2 3/h by 3/8 inches, having a total area (both
sides) of 13.3 square centimeters were immersed in the test
solution. The strips were supported in a diagonal position
in the bottles. and touched the container walls only at the
tips. The bottles containing the solution and capper strips
were placed in a constant temperature bath where the temper-
ature was controlled to $0.500.) Tests could be run with
the bottles open in order to maintain an equilibrium oxygen
concentration with the atmosphere or in a closed system by
steppering the bottles. All tests were run in a covered
temperature bath in orderto keep the test specimens from
contact with light.

In running a test to determine the corrosiveness of
a solution, several test bottles were prepared and filled
Vith the same solution. Copper test strips were prepared
and one strip placed in each bottle. The test strips were
cleaned by washing in carbon tetrachloride, then placed in
a solutlon of 5% nitric acid in absolute methyl alcohol

28

for two minutes. They were then rinsed in distilled water
and immediately placed in the test bottles. During the
cleaning operation the strips were handled with forceps.

The progress of the corrosion reaction was followed
by removing test bottles at various times and determining
the quantity of corrosion products in the solution and in
the surface film. In some tests corrosion products were
precipitated from the solution and collected on the bottom
of the test bottle. The amount of precipitated corrosion
products was measured by adding concentrated hydrochloric
acid to dissolve the precipitate, then diluting with distilled
water to a concentration range where the copper content
could be measured colorimetrically. With these methods it
was possible to determine the extent of the corrosion re-
action and the distribution of corrosion products at various
time intervals.

Static corrosion tests do not give any indication
of the effects of velocity on the corrosion process but
considerable information can be obtained about the chemical
processes occurring in the corrosion reaction. With closed-
systenltests it is possible to find out whether or not con-
stituents such as oxygen are consumed in the corrosion process.
An approximation of the rate of the over-all corrosion re-
action can be made from examining a plot of the amount of

corrosion products as a function of time. The quantity and

29
distribution of corrosion products within the system provide

considerable information about the nature of the corrosion

process.
Description of the Flow System Corrosion Test

Some test work was done in which.corrosion was studied
under flow conditions. In these tests the metal strip was
placed in a glass tube 2.0 cm. in diameter and 18 cm. in
length. Solution was pumped in the bottom of the cylinder
and overflowed through an outlet at the top. The solution
was pumped with a diaphragm pump having a plastic head.
Plastic lines were used to connect the equipment, so no
metal came in contact with the solution. The flow rate in
these tests was 110 ml. per hour. At this low flow rate
velocity effects were not significant but the flow was high
enough that constituents consumed in the corrosion reaction
could be supplied by incoming solution and all concentrations
maintained at a constant level. Soluble corrosion products
were also removed in the flowing stream.

Flow system corrosion tests were run at a temperature
of 60°C. The test chamber was covered to prevent exposure
to light. Solutions used in the tests were collected and
their copper content measured. At the end of the test the
filnlcan the copper strip was removed and its COpper content
determined. In this way it was possible to evaluate the

amount. of corrosion occurring during the test. The flow

30
test was used to check solutions tested under static con-
ditions in order to determine whether or not the results of

the two tests are comparable.

A]!

 

EXPERIMENTAL RESULTS
Corrosion Characteristics of Copper in High-Purity Water

Corrosion g; highgpurigy copper.--Static corrosion
tests were used to evaluate the corrosion characteristics
of copper in high-purity water. Considerable care was taken
to obtain data from which a fairly accurate curve showing
the amount of corrosion versus time could be constructed.
The corrosion characteristics of high-purity copper in
highly pure water can be used as a standard to which the
corrosion characteristics of water containing known quan-
tities of chemical impurities can be compared. In this way
(munges in corrosiveness caused by dissolved gases and solids
can be evaluated.

Several individual tests were carried out in order to
obtain sufficient data to plot a curve showing the extent
of corrosion as a function of time. Each point on this
curve represents a test with a single copper specimen. All
copper specimens were prepared at the same time and placed
in individual bottles filled with water. These bottles
were then placed in the constant temperature bath. Although
variations in surface conditions of the metal strips may
hams caused some variation in the results obtained from
similar'tests, it is felt that in general, these results

are representative because they were obtained from several

32
different coPper strips rather than from a single piece of
metal. Test bottles were removed at various times and the
quantity of corrosion products in solution and in the surface
film were determined. In this way it was possible to
measure the extent of corrosion at successive periods of
time and to follow the course of the corrosion process.

The quantity of corrosion products in solution and the
total amount of corrosion,both expressed in milligrams per
square centimeter, for a series of static corrosion tests
with high-purity water and high-purity copper is shown in
Graph la, page 98. (All data is presented in graphs on
pages 97 to 115 and in tabular form on pages 116 to 126.)

The total amount of cOpper corroded includes the corrosion
products in the surface film and the copper dissolved in
the water phase. In these tests the bottles were left open
to the atmosphere in order to allow the water to maintain a
constant oxygen concentration by establishing physical
equilibrium.with oxygen in the air. At 60°C, water in
equilibrium with the atmosphere contains h.5 ppm.dissolved
oxygen and less than 0.2 ppm dissolved carbon dioxide.‘ The
pH is about 6.5. During corrosion tests the pH increased
slightly to about 6.8 but the oxygen concentration of the
water remained constant.

The parabolic shape of the curve indicates that films
of corrosion products provided some degree of protection

for the underlying metal. Initial corrosion rates were

 

a: f «c a I‘l- e c a 1 C bl: +0 1h f e “A v A A int P ha {P -v
Vol iv v} \4.u alt o

 

33

very high; within one hour the film contained about 0.10
mg. per sq. cm. of corroded copper. After the initial
period of rapid film build-up the growth rate appeared to
decrease due to the protection provided by the film. The
film continued to increase in thickness during the entire
periodof the test. Apparently, even well developed films
do not provide complete protection. However, in the latter
part of the test corrosion occurred at a very low rate.

Careful measurement of the initial slope of an expanded
plot of the data in Graph la indicates that the initial
corrosion rate is approximately 5.6 mg. per sq. cm. per day;
this corresponds to a corrosion rate of 90 mils per year
at 60°C with an oxygen concentration of [p5 ppm. Although
this is a high, corrosion rate, it is obvious that the rate
of corrosion drOps to a much lower level as soon as the
metal becomes protected by the film of corrosion products.

No attempt was nnde to identify any of the corrosion
products. With copper corroding in highly pure water the
only possible corrosion products must be some form of
cuprous or cupric oxide or hydroxide. Films formed during
the first few days were light brown with a red tint, and
gradually changed to a darker brown. In a few days
irregularly shaped black spots were noted on the surface.
These continued to increase in size until, after about a

week, the entire surface was covered with a black film. The

31+

initial light brown film may be cuprous oxide or hydroxide
which slowly oxidizes to black cupric oxide.

‘Copper immersed in oxygenated water would be expected
to continue to corrode until the water became saturated
with copper oxide and the film of corrosion products became
thick enough to inhibit corrosion.' With copper corroding
in a flowing system any, film depletion due to corrosion,
products dissolving in the flowing stream would tend to be
replaced by additional corrosion. Because corrosion seems
to be largely controlled by the surface film, factors that
affect film stability and solubility are probably very
significant in determining the over-all corrosion rate.

The copper concentration of the water in these tests
is shown in Graph lb, page 98. The copper content of the
water phase in similar tests with four metal strips in
each test bottle and with a single metal strip that had
been conditioned by allowing a film of corrosion products
to develop under controlled conditions are also shown in
this graph. The strip. having the film initially present
had been allowed to corrode for a two-week period in high-
purity water. The copper concentration of the aqueous
phase remained very low during all these tests. The products
of the corrosion reaction appear to be only slightly soluble
in the water phase.

The presence of film on the metal at the start of the

test seems to have little effect on the amount of copper

35
that goes into solution. Increasing the metal-to-water
ratio increases the copper content to some extent but this
factor is much less than the four-fold increase in the
surface area of the metal. Slight traces of precipitated
cOpper compounds were detected in the bottles in all tests
that had been in progress for more than a few days.‘

The copper concentration attained by the solution is
seen.to be less than 0.5 ppm and is not greatly changed by
the initial presence or absence of a film.or by a change in
the surface area of the metal. This concentration may re-
jpresent the solubility of the corrosion products in the
*water and would therefore be governed only by factors that

affect the physical solubility of copper oxides.

Corrosion 23 commercial 55323 copper.--Strips cut from
commercial grade copper water tubing were also tested with
llighppurity water in open-system static corrosion test. This
lnaterial was a phOSphorus-deoxidized copper containing a
lninimum of 99.9% cOpper plus silver. Typical values for
“the amounts of other impurities in this cOpper are: 0.0175%
jphosphorous, 0.0023% iron, 0.0015% antimony, and 0.0015%
nickel. .

From the results of these tests, shown in Graph 2a,
:page 99, commercial capper appears to be less corrosion
resistant than pure cOpper. The total amount of c0pper,

4corroded during the test period was greater. The amount of

36
capper in solution was very small and the larger amount of
corrosion was reflected in the greater quantity of corrosion _
products in the surface film. In these tests with highly
pure water only a small portion of the corrosion products
'were dissolved in the water phase; most of the corroded
capper was found in the surface film.

The curve in Graph 2a is parabolic in nature. Corrosion
continues to occur all through the test period even though
well developed films are formed on the surface. Films
formed in highly pure water are rather difficult to remove
after they are more than a week old. The cepper content of
the films in the six- and twelve-day tests cannot be measured
‘with.sufficient accuracy to determine definitely the in-
crease in corrosion during the last week of the test. It
is apparent that the corrosion rate in this period is much
less than during the first few days of the test and would
continue to decrease as the film.thickness increased.

Graph 2b, page 99, shows the copper concentration of
the waters in these tests and in similar tests with strips
of commercial copper that had been conditioned and had a

:film.present at the start of the test. The capper concen-

tration remained low during the test. The solution in con
tact with the cleaned strips reached a higher copper
concentration than in tests with high-purity copper during

the early part of the test period.

37

In both series of tests the water phase reached a
maximum copper concentration during the early days of the
test-period and then decreased slightly. Small quantities
of precipitate, roughly equivalent to the amount of capper
that would be precipitated by the observed decrease in
solution concentration, were detected in the test bottles
removed dm'ing the last week of the test period. These
results indicate that the water phase may have been in an
unstable state with respect to dissolved corrosion products,
and small quantities of cepper containing compounds were
precipitated from solution after the tests were in progress
a few days.

In open-system tests run with high-purity water the
small amount of carbon dioxide present makes the water
slightly acidic. In order to determine whether or not the
acidic condition of the water phase influences the capper
concentration of the solution, static corrosion tests were
run with water made slightly basic by addition of approxi-
mate1y 1.5 ppm of sodium hydroxide. This water had an
initial pH of 7.5 and remained slightly basic during the
test. The copper pick-up in these tests, shown in Graph
3, page 100, indicated no marked reduction in copper solu-
bility resulting from the higher pH. The capper concentration
of approximately neutral water in contact with corroding

copper metal seem to show no marked change with small

variation in pH.

38

Corrosion g_i_' capper in 3:95.93 having ‘3 limited oxygen
content.--Unprotected copper is seen to corrode rapidly in
oxygenated water. Most workers in the field of cerrosion
agree that dissolved oxygen is consumed in the corrosion of
cOpper by water (7, 22, 3?), but there is some question
whether or not oxygen is absolutely necessary for the
corrosion reaction to occur. A closed system static corrosion
test was run with high-purity cepper in order to determine
if oxygen-free water is corrosive.

Water for this test was prepared by boiling in a Pyrex
flask equipped with reflux condenser. The water was withdrawn
from the flask, passed through a cooling coil, and allowed
to flow into the test bottle containing a copper specimen.

It is not possible to rancve all oxygen by boiling; the oxygen
content could be reduced to about 0.2.ppm with difficulty.
Ten ppm of sodium sulfite were added to the water to remove
the last trace 'of dissolved oxygen and any oxygen absorbed
during the filling operation. Water prepared in this manner
shows no indication of oxygen when tested by the modified
Winkler method.

Graph 15., page 100, shows that very little corrosion
occurred in a closed system static test with this water. It
is evident that corrosion is reduced to a very low level by
the removal of oxygen. The water contained 0.10 ppm cepper
after 13 days; the metal. strip was still bright and showed

no visual evidence of corrosion. The surface was erased

39
lightly to remove for analysis any film.that may have been

present. Although.the extent of corrosion in this test
amounted to 0.006 mg. per sq. cm. this may have been caused
‘by a small amount of oxygen initially absorbed on the copper
strip or to traces of oxidized film not removed in the clean-
ing operation.

While these results do not conclusively prove that no
corrosion occurs in oxygen-free water they do show that it
is reduced to an extremely low level and that the corrosion
rate is essentially zero.

If oxygen is necessary for corrosion, copper metal
corroding in water with a limited oxygen content should
continue to corrode only until the oxygen is consumed.

This assumption was checked by running corrosion tests in '
sealed bottles with water containing l.h8 ppm.oxygen. The
results of these tests, shown in Graph 5, page 100, show

the total extent of corrosion to be less than in tests with
inigh purity copper in which the oxygen content remained con-
stant. The quantity of corrosion appears to approach a
definite limit as the test period continues.

Calculation can be made showing the stoichiometric
relationship between the amount of oxygen consumed and the
quantity of copper corroded. The 300 ml. of water having
an oxygen content of l.h8 ppm contained 5.55 x 10"5 equiva-
lents of oxygen. This amount of oxygen is capable of

oxidizing 3.5h'mg. of copper to the cuprous form; dividing

this quanti
shows that
mg. per sq.
amount (0.:
Oxyge
these test
the oxyger
these rem
sample 31
test, the
proceeds.
The
they Shoe
cuprous
00101‘ be
these re
reasons
are cup

pT‘BSQnt

no

this quantity of cepper by the strip area (13.3 sq. cm.)
shows that the corrosion products should amount to 0.266
mg. per sq. cm. Graph 5 shows that almost this exact
amount (0.27h mg. per sq. cm.) of copper was corroded.

.Oxygen determinations were also made on the water in
these test bottles by removing 100 ml. samples and measuring
the oxygen content by the modified Winkler method. Although
these results may not be very accurate because of small
sample size and oxygen absorption from the air during the
test, that do show a regular decrease in oxygen as corrosion
proceeds.

The most significant thing about these results is that
they show the corrosion products to be predominantly of the
cuprous form. In these tests the strips were dark brown in
color but none of the deep black coloration developed. From
these results and visual observation of other tests it seems
reasonable to conclude that the initial corrosion products
are cuprous oxide or hydroxide. If sufficient oxygen is
present, these may slowly oxidize further to cupric oxide.

ghe Effects of Chemical Impurities on the Corrosiveness of
ater

Water containing sodium bicarbonate and sodium carbonate.
--Both calcium bicarbonate and calciumcarbonate are often
;present in natural waters. In the zeolite process for

water softening, these calcium compounds are converted

k1
into the corresponding sodium salts. Since it is not de-
finitely known what effects these compounds have on the
corrosiveness of water it seems worthwhile to attempt to
evaluate their effect on the corrosion process. The corrosion
characteristics of bicarbonate- and carbOnate-containing
waters were evaluated by testing solutions containing known
quantities of sodium.bicarbonate and sodium.carbonate in
open system static corrosion tests with strips of high-purity
copper. '

The results of open system static corrosion tests at
60°C with high-purity copper and water containing hOO ppm
sodium bicarbonate are shown in Graph 6a, page 101. These
data indicate that this solution corrodes copper to about
the same extent as highly'pure water. The curve is quite
similar to the curve in Graph la. Film development occurs
at about the same rate and the fully developed film contains
nearly the same amount of corroded copper. The film seems
to provide very gpod protection; corrosion is reduced to a
very low level in the second week of the test period.

Graph 6b, page 101, shows the capper concentration of
the solution in this series of tests. The solution reached
a maximum.concentration of 0.h0 ppm copper within 2h hours.
The copper concentration then decreased until at the end of
‘the test period the solution contained only 0.09 ppm copper.
.Altheugh.this solution is definitely basic, having a pH of
8.1;. it seems to be capable of dissolving small quantities of

capper.

he

Due to thermal decomposition at temperatures near the
boiling point, solutions containing sodium bicarbonate tend
to lose carbon dioxide and form sodium carbonate if cpen to
the atmosphere. It is possible that this may have occurred
to some extend in these open-system tests. Tests with
solutions containing h00 ppm sodium bicarbonate were also
run under closed-system conditions. In the closed system
tests the pH remained at about 8.3 throughout the test in-
dicating that sodium bicarbonate decomposition did not occur
to any significant extent. The copper content of the
solution in the closed system test was essentially the same
at corresponding periods of time as in the open system tests.

These results indicate that corrosion products are only
slightly soluble in solutions containing sodium bicarbonate.

The c0pper strips in the open-system tests did not
appear to be much different than strips corroded in high-
purity water. In the first few days of the test the surface
appeared brown or dark straw colored; the color gradually
changed to a very dark brown and became black in about a
week. No indication of pitting was observed. The films
seemed to cover all of the surface unifbrmly. Scratches and
Asurface blemishes were covered as well as the smooth portions
of the surface.

Water solutions containing 50 ppm.sodium carbonate
corrode highppurity copper to the extent shown in Graph 7a,

page 102. Here again a fairly rapid film build-up occurs,

to
after which film growth is reduced to a low level because
of protection afforded by surface deposits of corrosion
products. The rate of corrosion in the first few hours was
slightly lower than in tests with highly pure water, which
may indicate a lower rate of attack on unprotected COpper..
The solution had a pH of 9.8; the higher pH might have re-
duced the rate of surface attack. Once fully developed,
the films seem capable of inhibiting the corrosion reaction
and preventing further attack.

Within three hours the solution attained a copper con-
centration of 0.12 ppm; after one week the concentration
had dropped to 0.05 ppm. From the values of the solution
concentration plotted in Graph 7b, page 102, it is seen that
the concentration remained low throughout the test. However,
even in this definitely alkaline solution with a pH of 9.8
capper was still detectable.

No evidence of pitting was noted in these tests. The
films first developed a light brown color, then turned

darker brown and still later black. The black color pre-
sumably was due to cupric oxide formation.

In general these tests show that the presence of sodium
bicarbonate and sodium carbonate does not increase the
corrosiveness of water. The total extent of corrosion is
8lightlyless thanfiwith highly pure water. Surface films
deve10p at about the same rate and contain slightly less

corroded copper when fully develOped. Film formed in water

containing
more prote
After the
have been
The
the surfa
slope of
the com
while 1:
rate thrq
01‘ greet
become 1
Unless 1;
thick er
Egg
Since 81
natural
DPocegs
lith hi
mphet
isticS
differs

of 12831

Mt

containing 50 ppm sodium: carbonate appear to be even
more protective than films developed in high-purity water.
After the first week of the test period corrosion seems to
have been almost completely stepped.

The accuracy of measuring the quantity ofcopper in
the surface films is not sufficient to fix precisely the
slope of the corrosion curve in the plateau region. However,
the corrosion rate in this region appears to be very low
while in high-purity water corrosion occurs at a significant
rate throughout the duration of the test. This may not be
of great significance shnce all films would be expected to
become increasingly protective as their thickness increased.
Unless pitting occurs, all films should eventually become
thick enough to protect the surface under static conditions.

.rWater containing sodium sulphate and sodium chloride.--

Since sodium sulplmte is another impurity often found in

natural water, tests were run to evaluate its effect on the

process of corrosion. Open-system static corrosion tests

with high-purity cepper and water containing 1+0 ppm sodium
stzlphate having a pH of 6.5 showed the corrosion character-
is tics of sulphate-containing solutions to be somewhat
different than highly pure water. The data from this series
of tests, plotted in Graphs 8a and 8b, page 103, show that
surface films effectively reduced corrosion after the initial
film build up. However, the rate of corrosion in the first

for hours of the test was considerably greater than with

hS

highly pure water, indicating that the solution is more
corrosive to unprotected cepper or that partially deve10ped
films are less protective.

Nearly all the corrosion occurred in the first six
hours of the test. After this initial period of rapid film
growth the film contained about 0.30 mg. per sq. cm. of
copper. The film than seemed to be sufficiently developed
to be capable of reducing corrosion and a relatively small
amount of corrosion occurred in specimens under test during
the remainder of the test period. The flatness of the
plateau portion of the curve indicates that, once formed,
the film is effective in inhibiting corrosion.

The color of the film changed from an initial dark
brown to black within six days. Some evidence of pitting
was noted during the second week of the test period.

As-shown in Graph 8b, the solution concentration in-
creased to 0.hh ppm within six hours and then drapped to
about 0.30 ppm and remained at approximately this level for
'the remainder of the test period. Small quantities of pre-
cripitate were detected in test bottles removed during the

second week of the test period.

Sodium chloride, probably more than any other chemical
impurity found in natural waters, is suspected of stimulating.
corrosion when present in aqueous solutions. Solutions
containing 10 ppm sodium chloride with a pH of 6.7 were

tested in cpen system static corrosion tests at 60°C. The

results of
Graph 98,
ions, SOIL
build-up.
sq. on. 01
increased
appears tc
is some v:
specimens
than a re-
on. of co:
In t]
aPpesred .
becom da
other {‘11
t“”936 1363
The
tests 18
Seen to v
higher tt
c”Maths

5
Cf}

‘centra
tendGRCy
with 0th.
(31‘s.;

30mm c}

no

' results of these tests with high-purity cepper are shown in
Graph 9a, page 10h. As with the solutions containing sulphate _
ions, solutions containing sodium chloride cause rapid film ‘
build-up. _Within three hours the film contained 0.25 mg. per
sq. cm. of corroded cepper and within six hours this had I
increasedto over 0.35 mg. per sq. cm. Corrosion then
appears to have been reduced to a low level. While there
is some variation in the film thickness of individual
specimens, the films of all specimens under test for more
than a few hours contained between 0.3 and 0.1+ mg. per sq.
cm. of corroded cepper.

In the first few days of the test the surface films
appeared dark brown with a purple cast; they continued to
become darker but the deep black color observed with some
other films never developed.’ Fitting was not evident in
these tests.

The copper content of the solution in this series of
tests is shown in Graph 9b., page 1015,. The concentration is
seen to vary in the different test bottles but is definitely

higher than in tests with high-purity water. The solution
concentration was as high as 0.60 ppm in one test. Copper
concentration varied considerably in these tests but this
tendency was also evident in static corrosion tests run
with other solutions.

Graph 10a, page 105, shows solutions containing 100 ppm

sodium chloride to be much more corrosive than solutions

cents:
was n'
after

films

salt

late

_ 1+7
containing 10 ppm. The total quantity of corrosion products
was much greater and films continued to increase in thickness
after the initial rapid period of growth. Apparently thicker
films are required to prevent corrosion in solutions with
a high chloride ion content. '

Strips under test for 1h days showed definite evidence
of pitting. The pit areas were covered with thin films
and no bare metal was exposed. Pit areas were surrounded
by visible rings of corrosion products. All pits were
very shallow and did not appear to penetrate into the
metal to any great extent. Fitting may have been responsible
for the increase in corrosion occurring during the second
week of the test period.

The solution concentration, shown in Graph 10b, page
105,was as high as 0.98 ppm on the seventh day. Although
‘there is considerable variation in copper concentration of
'the solutions tested, all are seen to be high when compared
with tests using pure water. Since the sodium chloride
s<>lutions had a pH of 6.8, about the same as high-purity

water, the higher copper concentrations must be due to some
titlier factor. Copper is known to complex with chloride ions
(3) and complexing may account for the higher cepper con-
centrations. 1

Solutions containing sulphate and chloride ions are
8Ben to be more corrosive to partially protected cepper than

hiShapurity water. The presence of 100 ppm sodium chloride

in we
also
pitt:

do it

com
to '
flu
the
ano
cor:
tes

car

h8
in water definitely increases the amount of corrosion. It
also increases the copper pickeup of water and causes
pitting. These facts suggest that both of these materials

do increase the corrosiveness of water.
Corrosion of Copper in Slightly Acidic Water

EEEEE containing dissolved carbon dioxide.--Carbon
dioxide is known to increase the capacity of water to dissolve
capper and is thought to increase its corrosiveness (15. 37).
The dissolution of surface films of corrosion products
would expose the unprotected.mmtal and result in a high
over-all corrosion rate. It isnot clear whether or not the
corrosiveness of dissolved carbon dioxide is due entirely
to its acidic prOperties or to other factors that may in-
fluence corrosion. The total quantity of corrosion products,
the amount of corroded copper in the surface film, the
amount in solution and the amount contained in precipitated
compounds for a series of closed-system static corrosion
tests with high-purity copper and water containing 70 ppm
carbon dioxide are shown in graph 11a, page 106. The total
quantity of corrosion products includes all copper converted
chemically and is the sum of the cepper in the solution, in
the film and in the precipitate.

The water had an initial pH of h.5 and contained 5.h
ppm oxygen. The pH increased during the test to a value of

55e5. Because oxygen is consumed in the corrosion reaction

.L‘Li.

its c
was 11
of co
the t

const

the

00p;

at t

Col]

w is
its concentration decreased rapidly; almost all the oxygen
was used up in the first few days of the test. The extent
of corrosion would undoubtedly have been much greater if
the tests had been run with the oxygen concentration held
constant. 8

These tests show water containing dissolved carbon
dioxide to be much more corrosive than similar water with
no carbon dioxide present. Corrosion occurred at a rapid
rate until all the oxygen was consumed. Then the system
remained relatively static until in the third week of the
test period the copper content of the film increased and the
copper concentration of the solution decreased. Very little
corrosion occurred after the first three days but this may
have been caused by lack of oxygen. The increase in film
thickness in the last few days of the test may have been
caused by insoluble copper compoundscrystallizing out of
solution and depositing on the metal surface. Although
the solution concentration definitely decreased, very little
copper was precipitated from.solution. It is possible that,
if crystal nuclei were present on the metal surface, cOpper
compounds would be crystallized out of solution and deposited
at these sites rather than.precipitated from solution to
collect in the bottom of the test bottles.

As Graph llb, page 106, indicates, the capper concen-
‘tration of the solution increased rapidly, then remained

at a high level for some time, and finally decreased quite

50
suddenly to a much.lower level. The sudden drOp in solution
concentration isdifficult to explain; the solution may be
in an unstable condition because of the slow formation of
insoluble compounds that crystallize out of solution quite
rapidly once crystthzation is initiated.

It is apparent that the presence of carbon dioxide
increases the corrosion rate and the ability of the solution
to dissolve copper. Although films were formed, the limited
supply of oxygen rather than the protective nature of the
film probably controlled the extent of corrosion. The films
themselves would be expected to be soluble in this solution.
In these tests, in contrast to tests where carbon dioxide
was absent, the amount of copper in solution is seen to be
a s ignificant portion of the total quantity of corrosion
products. AlthOugh other factors may be partially respon-
sible,the high copper concentration of the solution could
be explained by the increased solubility of corrosion
products due to the acidic nature of the solution.
~ It isobvious that the reactions occurring in this
— system under closed-system conditions are complicated in

.nature. While the solution attained a high copper concen-
tration early in the test, further change must also have
‘ occurred to cause a decrease in concentration at a later
time... Cupric carbonate compounds are complex and varied
ill structure (32). It is possible that changes within the

system are 'due to the slow conversion of soluble cuprous

51
or cupric compounds to insoluble copper carbonate
compounds.

The high cOpper pick-up of the solution would be ex-
pected to contribute significantly to corrosion under flow
conditions. If surface films were depleted by being dissolved
in the flowing stream, additional corrosion would occur as

f a result of the metal being incompletely protected. Under
such conditions the over-all corrosion rate might be con-
trolled by the rate of film dissolution. ‘

Since oxygen is necessary for corrosion of copper by
neutral water, the question arises whether or not carbon
dioxide-containing solutions are corrosive in the absence
of oxygen. The results of tests with water containing 20
ppm carbon dioxide having a pH of 5.5, show that this so-
lution is not corrosive if oxygen is not present. The results
of these tests shown in Graph 12, page 10?, indicate that a
small amount of corrosion occurred but this Could have been
caused by traces of oxygen that were not removed from the
system.e The c0pper strips in these tests remained bright
and shiny without showing any visual evidence of corrosion.

The water used in these tests was boiled to expel
oxygen; carbon dioxide was then bubbled through the water
while under a nitrogen atmosphere. Ten ppm of sedium
sulphite was added to remove the last traces of oxygen.
Oxygen could not be detected in the water by the modified

Winkler method.

52
From thermodynamic considerations, carbon dioxide-
containing solutions would be expected to be corrosive only
if Oxygen were present. These results seem to confirm this
conclusion. The hydrogen ion concentration is not high
enough to allow hydrogen ion reduction and carbon dioxide
would not be expected to oxidize oOpper.

Corrosion by,oxygen-free dilute acetic acid solution.--

 

Closed system.corrosion tests were run with high-purity
copper and water containing 5 ml. per liter of 99.7% glacial
acetic acid and from.which oxygen had been removed. Five
ppm sodium sulphite was added to remove all traces of oxygen.
In these tests an attempt was made to determine whether
under more highly acidic conditions, corrosion would occur
through hydrogen ion reduction. The results of these tests,
presented in Graph 13, page 107, show that considerable
corrosion did occur even though no oxygen was present. This
solution had an initial pH of 2.9 and, although the pH
increased to 3.6 during the test, the hydrogen ion concen-
tration remained at a high level at all times. Since acetate
ions are thought to complex with cOpper or at least to form
soluble compounds (35) which reduce the activity of the
copper ions, complexing may also have been a factor in pro-
moting the corrosion reaction.

In these tests most of the corrosion products dissolved
1J1 the water phase and uniform surface deposits did not

forun. During the third week of the test period narrow

S3
streaks of surface deposits were noted. Closer examination
revealed that the metal under these deposits was etched.
Apparently corrosion at this stage did not take place uni-
formly on the surface but electrochemical attack occurred
with the formation of definite anodic areas.

The corrosion rate is seen to be relatively constant
in these tests. This would be expected in the absence of
surface films. From the corrosion data presented in Graph
13 the corrosion rate is estimated to be 0.026 mg. per sq.
cm. per day. This is much lower than the initial corrosion
rate of high-purity oxygenated water. Corrosion involving
hydrogen ion reduction occurring under these rather acidic
conditions is seen to be very slow compared to corrosion

with oxygen reduction under neutral conditions.

Corrosion by;Water ContaininggPartial:yNeutralized Carbon
_fiioxide

Slightly acidic ggtgg.containingԤggg carbon dioxide.--
The corrosiveness of aqueous solutions containing carbon
dioxide may be due to the high acidity of the solution or
possibly to some other specific effects of dissolved carbon
dioxide. Tests were run with solutions containing carbon
dioxide buffered with sodium bicarbonate but still having
considerable free carbon dioxide present. Solutions of
sodium bicarbonate are basic with a pH of about 8.3. If

carbon 'dioxide is added to these solutions it is possible

5h

to produce a neutral solution containing significant amounts
of undissociated carbonic acid. In such a neutral solution
both cuprous and cupric 'oxide would be expected to have very
low solubilities.

Graph 114a, page 108, shows the results of closed-systems
static corrosion tests with high-purity c0pper using water
containing 1+8 ppm freecarbon dioxide and 14.00 ppm sodium
bicarbonate with 6.0 ppm oxygen present. The solution had
a pH of 6.7. This water appears to be quite corrosive;
corrosion continued at a significant rate throughout the
test while the oxygen concentration continually decreased.
The cOpper concentration of the solution, shown in Graph
lhb, page 108, exceeded 2.5 ppm by the third day, then de-
creased to some extent but. remained at a fairly high level
throughout the test period. The copper cencentration was
definitely much higher than in tests with high-purity
water at the same pH. The decrease in copper concentration
may indicate an unstable condition within the solution.
Small quantities of. cOpper-containing compounds were pre-
oipitated from the solution in the latter part of the test.
Blue-green deposits, suspected of being some form of basic
Carbonate such as malachite or azurite, had formed on’ the

metal strips by the fourth day of test. These deposits
continued to increase in thickness [and size as the tests
con tinued. Examination revealed. the presence of a layer

or tightly bound black film beneath these deposits. When

55
the films were removed from the surface the areas under
the green deposits appeared to be etched. These observa-
tions indicate that the green surface deposits may be capable
of causing pit corrosion. .1
In another set of similar tests, water containing
carbon dioxide partially neutralized with sodium hydroxide
was also found to be excessively corrosive. Graph 15a,
page MIL shows the results of tests with water containing
55 ppm carbon dioxide to which 60 ppm sodium hydroxide was
added. This water had a pH of 6.7 and was found to contain
35 ppm free carbon dioxide when titrated with sodium car-
bonate. Although this solution contained considerable free
carbon dioxide the bicarbonate ion concentration would not
be very great since only part of the carbon dioxide was
neutralized.
The corrosion characteristics of this solution are
in many ways quite similar to the solution containing
carbon dioxide and sodium.bicarbonate. Corrosion continued
at a rather high rate throughout the test while oxygen
ciecreased to a concentration of 0.h ppm. Copper concentration
cxf the solution increased rapidly as shown in Graph 15b,
vpage 109, The concentration exceeded 10 ppm at the end of
the first week and decreased to 2.0 by the end of‘ the second
“week. The variation in cOpper concentration must be due to
an unstable condition within the solution. It appears that

some process occurs which decreases the solubility of copper

56
after the solution reaches its maximum concentration. How-
ever, very little precipitate was found. Visible blueegreen
deposits were noted on the metal strips on the ninth day of
the test period. These deposits increased in size as the
test continued. The copper lost from solution may have de-
posited on the metal strip instead of precipitating from
solution.

The deposits, thought to be basic carbonates, may be
slow in forming and a supersaturated condition may exist
until the insoluble compounds crystallize out of solution.)
The lower concentration of bicarbonate ion needed to form
the basic carbonates may account for the longer time re-
quired to form the c0pper carbonate deposits, as compared
with solutions containing a higher bicarbonate ion con-
centration.

Slightly alkaline water containing £323 carbon
dioxide.--If a small quantity of carbon dioxide is added
to water containing a substantial amount of sodium bicarbonate
:it is possible to produce a neutral or slightly basic solu-
1:ion containing free carbon dioxide. In such a water the

cxxides and hydroxides of both cuprous and cupric copper

would be eXpected to only sparingly soluble. Values for
the solubility of cupric oxide in neutral water reported
1J3 'bhe literature range from 0.06 to 1.83 ppm andcuprous

oxide is thought to be even more insoluble (29, 37).

57

Graphs 16a and 16b , page 110, show that such a water
is not excessively corrosive but the copper content of the
solution in static corrosion tests is quite high. This
water contained 300 ppm sodium bicarbonate, and 20 ppm free
carbon dioxide and had a pH of 7.2. With respect to bicar-
bonate ion and carbon dioxide content, this water is similar
to many waters supplied by deep wells from limestone forma-
tion. In these tests corrosion is seen to occur at a rapid
rate initially and then is reduced to a low level. Very
little corrosion occurred after the first few days. The
copper concentration of the solution increased to about
2.5 ppm.and remained near that level.

The quantity of precipitate formed was small and no
surface deposits were visible on the metal strips. Under
these conditions the films of corrosion products appear
to be capable of inhibiting the corrosion reaction. The
small decrease in oxygen concentration also indicates that
only a moderate amount of corrosion occurred. It is possible
‘that surface deposits or increased quantities of precipitate
would have formed if the tests had been run for a longer

Imeriod of time. However, this solution is definitely less
corrosive than solutions having a higher free carbon dioxide!
concentration and a lower pH.

.Because the oxygen concentration decreased only a

moderate amount the solution was well supplied with oxygen

throughout the test period. The tests were run at a nearly

58
constant oxygen concentration and the corrosion reaction
was not seriously retarded because of insufficient oxygen.
Although this solution is only moderately corrosive
under static conditions the copper concentration of the
solution is seen to be high. Under flow conditions corrosion
might be high due to dissolved copper being carried away
in the moving stream. I
Graphs 17a and 17b, pageifll, show the results of closed-
system corrosion tests with water containing 55 ppm carbon
dioxide and 9h ppm sodium hydroxide. This solution had a
pH of 7.1 and contained 1h ppm free carbon dioxide. Its
corrosion characteristics are nearly identical to those of
the neutral solution contain sodium.bicarbonate and carbon
dioxide shown in Graphs 16a and 16b, page 110. The oxygen
concentration drOpped moderately during the initial corrosion
period and then remained nearly constant. The copper con-
centration of the solution increased to over 3.0 ppm.and
:remained at nearly this level. The total amount of copper
(corroded exceeded 3.0 mg. per sq. cm. in the early part of
the test but did not increase much above this level. The
quantity of precipitate formed was small and no copper car-
bonate compounds were observed on the surface of the metal
strips.
The same solution used in running the tests shown in
Graphs 16a and 16b was also used in running corrosion tests

With strips of commercial cOpper. These strips were

S9

phosphorus-deoxidized COpper similar to the copper used in
tests with high-purity water presented in Graphs 2a and 2b.
The solution contained 20 ppm free carbon dioxide and 300
ppm sodium bicarbonate. The pH was 7.2. Results of these'
tests are shown in Graph 18a and 18b, page 112. Although
in tests with highly pure water, commercial copper appears
to be more susceptible to corrosion than high-purity cOpper,
their corrosion characteristics in this solution seem to be
very nearly identical. With commercial copper, cOpper
pick-up ofthe solution is slightly lower and the total
amount of corrosion products are not quite as great. No
marked difference is apparent between the two grades of
copper in these tests.a
It is significant that high concentrations of dissolved
copper are found only in tests with solutions containing
free carbon dioxide. In all tests with carbon dioxide-free
water, the copper concentration never exceeded 1.0 ppm at
any time during any of the tests. Even with slightly basic
solutions containing free carbon dioxide, copper concentrations
are seen to be in range of 2 to )4, ppm. It seems logical to
(conclude that free carbon dioxide in solution increases the
tendency 'to dissolve copper corrosion products and enables
tile solution to hold several parts per million of copper in
solution. This may be only a temporary effect; solutions
With high copper content show some tendency to decrease in

cOncentration over a period of time and to precipitate

60

cOpper-containing compounds. These are indications that‘
two opposing processes are operating simultaneously, one
causing the dissolution of copper and another, a slower
process, forming insoluble products that are precipitated
from solution.

'pIn static corrosion tests with slightly basic solutions
containing carbon dioxide, only a moderate amount of corrosion
occurred whereas with similar slightly acidic solutions
having a pH of 6.7 corrosion was considerably higher. .The
corrosiveness of such solutions seems to be altered signifi-
cantly by small changes in the pH of the solution.

2h; solubility of copper oxides in 33.392 containing _f_r_e_g
carbon dioxide.--If solutions containing carbon dioxide
tend to dissolve copper corrosion products it should be
possible to show this effect with copper oxides. Solubility
tests were run with powdered cuprous oxide in an aqueous
solution containing 600 ppm sodium bicarbonate and 31 ppm
free carbon dioxide. This solution contained h.l ppm
oxygen.and had a pH of 7.2. These tests were run in closed
taottles of 60°C. Upon being removed from the temperature
bath part of each test solution was filtered twice through
No. [Q filter paper and then tested for copper content. An
oxygen determination was run with a 100 m1. sample.
The copper and oxygen content of this solution at
var ious times is shown in Table II. The copper content is

Soon to exceed 1 ppm after one day and then to decrease as

61
the duration of test is increased. The oxygen concentration
dropped rapidly indicating that oxygen was being consumed by

some process occurring within the system.

Table II._ Solubility of Cu20 in Water Containing600
ppm NaHCO3 and 31 ppm Free 002.

C0 er Ox on
Time (days) Concentration (ppm). 00ncentration (ppm)
0 "' I 14.01

1 1.1h

2 0.6h . 0.h
h 0.25 tgo.l
6' 0.06 410.1

These results are somewhat puzzling. It is evident
that this solution is able to dissolve appreciable amounts
of cuprous oxide. However, the solubilizing effect is only ~
temporary; the copper concentration increases to a maximum
value and then decreases. The decrease in oxygen concentra-
‘tion is probably due to oxidation of cuprous ions to the
(rupricsform.x This would be expected as cuprous COpper is
tuiought to be unstable in oxygenated aqueous solutions (6).
The decrease in copper content may be in some way connected

with the decrease in oxygen concentration but it is also

Possible that the two phenomena are independent and

unrelated.

62

Additional solubility studies were made to evaluate
the solubility of copper oxides in high-purity water and to
determine if the presence of bicarbonate ions can increase
the copper solubility of aqueous solutions. Solubility
tests were run to compare the solubility of both cupric and
cuprous oxide in highly pure water, in a solution of 600‘
ppmsodium bicarbonate, and in the solution containing 600
ppm.aodium.bicarbonate with 31 ppm free carbon dioxide.
These tests were run in sealed bottles at 60°C for a period
of two days. The solutions were then filtered twice and
tested for copper content.

Table III lists the copper concentration of the solu-
tions tested. Because of the short time the oxide powders
were in contact with the solutions and the possibility of
yfine solid materials passing through the filter paper, these
Values may not represent the true solubility of cuprous and-
cupric oxides in the respective solutions. However, these
values give an indication of the oxide solubilities; since
all were Obtained.by the same method they should give an
-indication of the relative solubilities. The results show
the copper contents of the aqueous phases in all the tests
‘were low except in the test with cuprous oxide in a solution

containing both free carbon dioxide and sodium bicarbonate.

4.63

Table III. Concentration of Copper in Solution after
Two Days in Contact with Powdered Cu20

 

 

or Cu0.
Solution pg EE%§%%%I%§%QS’ £2%§%%d%%%%2
600 ppm HaHCO3 with
31 ppm Free 002 7.2 0.6L; ppm 0.09 ppm
600 ppm NaHC03 8.L 0,0 " 0‘07 u
High-Purity Water 6.7 0.07 " 0.05 n

An approximately neutral solution containing free carbon
dioxide and sodium bicarbonate dissolved 0.6h ppm copper
when in contact with cuprous oxide although cuprous oxide
has very low solubility in pure water and in the solutions
containing sodium bicarbonate. However, the data in Table
II indicate this solubilizing tendency is only temporary
and the copper concentration will decrease again in a short
time. Nevertheless, this phenomenon can be used to explain
the high copper concentration of the aqueous phase, Observed
in corrosion tests with neutral solutions. In each of I
these instances the solution contains free carbon dioxide

in addition to bicarbonate ions.
Corrosion Characteristics_of Copper in Natural Water

Natural waters vary widely in composition but nearly
always contain substantial quantities of a few specific
schenrical impurities. Dissolved solids commonly present in

natlural water include calcium carbonate, calcium bicarbonate,

61+

calcium sulphate, sodium chloride, and sodium silicate.
Oxygen, nitrogen, and carbon dioxide are the dissolved gases
usually present. The effects these dissolved solids and
gases, either collectively or individually, have on the
corrosiveness of water are not definitely known at the
present time.

Static corrosion tests were run with samples of natural
water in order to compare its corrosiveness with highly
pure water and dilute aqueous solutions containing known
impurities. The water used in these tests was a zeolite-
softened water with high bicarbonate content. It contained
moderate amounts of dissolved carbon dioxide and had a pH
of 7.3. Table IV lists the amounts of dissolvedeolids and
gases present. This water was taken from a potable water
system in which numerous corrosion failures of copper water
tubing had occurred. It was suspected or being excessively

corrosive to cOpper water tubing.

Table IV. Dissolved Solids and Gases Present in Natural
Water Used in Running Corrosion Tests.

 

 

Component Concentration‘flppm)
Free carbon dioxide 15
Dissolved oxygen 6.5
Iron 0.1
Hardness 22
Alkalinity 328
Bicarbonate ion h00
Chloride ion 10

Total dissolved solids 506

65

The results of these static corrosion tests with high-
purity cOpper and natural water at 30° and 60°C are shown
in Graphs 19a, 19b, 20a, and 20b, pages 113 and 11h. These
tests were run under closed system conditions by sealing
all bottles.

From the results of these tests, this water does not
appear to be excessively corrosive. As in many of the
corrosion tests, surface films developed quite rapidly but
did not increase in thickness after the initial film build-
up. The copper concentration reached a fairly high level
and then decreased slightly. The copper content of the ,
water in the test at 30°C is definitely higher than in the
test at 60°C. It is somewhat surprising that the COpper
picksup of the water was quite high as this water was
slightly alkaline and the cOpper corrosion products would
be expected to be only sparingly soluble at this pH.

The copper strip removed at the termination of the 60°C
test showed some evidence of pitting. No well develOped
pits were present but small circular areas were visible.
which had thinner films than the rest of the surface.' These
may have been anbdic areas which would eventually develop
into pits.

Very little precipitate formed during the first 20
days of the test period although significant amounts of
Precipitate formed toward the end. The slight amount of

corrwesion occurring in the latter part of the test period

66

did not cause any increase in film thickness or solution
concentration but is reflected in increased quantities of
precipitate formed within the system. It is interesting to
note that the sum of the copper in solution and in the pre-
cipitate continued to increase throughout both tests even
though the amount of corrosion products in the film did

not increase after the first few days. The total quantity
of corrosion products increased rapidly at first and then
continued to increase at a very low rate.

This [water contained a small amount of free carbon
dioxide and there seems to be some significance in the close
similarity between the behavior of the system in the first
20 days of the 60°C test and the corrosion characteristics
of copper in tests run with the solution containing LOO
ppm sodium bicarbonate and 20 ppm free carbon dioxide shown
in Graphs 16a and 16b, page 110. The copper content of the
solution is seen to be about the same in both series of
tests. It may be that the presence of free carbon dioxide
in the natural water is responsible for high copper pick-up
indicated by these tests.

The total extent of corrosion in these tests is not
excessive and these results give no indication that this
water is excessively corrosive. However, under flow con-
ditions the capacity of the water to hold as much as two
or tfllree parts per million or copper in solution may indicate

a tendency for the water to dissolve protective surface

467
.films. If these films were depleted corosion would be ex-
pected to increase since the metal would be only partially

protected.
The Results of Flow System Corrosion Tests

A test was deveIOped in which strips of high-purity
copper similar to those used in the static corrosion tests
weretested under low velocity flow conditions.. The flow
rate in these tests was 110 ml. per hour. At this low
flow rate, velocity effects are not significant but the
corroding strip is continually supplied with fresh solution
and soluble corrosion.products are removed. One of the
objectives of these tests was to determine if solutions shown
to be capable of holding high concentrations of copper in
solution under static conditions would also pick up excessive
amounts of cOpper under flow conditions. If this were true
the corresion rate with such solutions would be high because
in a flowing system a large quantity of water comes in con-
tact with the metal.

Solutions having the desired concentration of dissolved
gases were prepared by mixing degassed, highly pure water
with water having a high.oxygen or carbon dioxide content.
Solutions were stored in a large glass jar and blanketed

with .a mixture of nitrogen, oxygen, and carbon dioxide

mixed in the prOportions necessary to maintain physical

68

equilibrium with the dissolved gases in the aqueous solution.
As the solution.was removed from the storage container the
vacant space was occupied by a mixed gas having the correct
cemposition required to be in physical equilibrium with the
dissolved gases in the water phase. In this way it was
possible to hold the concentration of dissolved gases con-
stant during the test.

Graph 21, page 115, shows the amount of corrosion pro-
ducts dissolved in the water phase and contained in surface
films in corrosion tests with high-purity water and several
dilute aqueous solutions. These tests were all of three
days duration and approximately 7900 m1. of solution was
used in each test. It appears that the extent of film
develOpment is no greater, and with some solutions is less,
than in three-days static corrosion tests with similar
solutions. In the test with high-purity water the film
formed had a higher c0pper content than in tests with many
of the other solutions but this was also true under static
conditions. The film on the strip in the test with the
solution containing 100 ppm sodium chloride is seen to
contain the most corroded copper. In this test there was
definite evidence of pitting; this may be partially re-
Sponsible for the greater amount of corrosion.

It is interesting to note that in the flow test with
highly pure water the copper pick-up of the solution is

0.28 ppm. Since this is about the same as the copper

69
concentration of high-purity water in static corrosion tests,
this value may represent the solubility of corrosion products
in pure water at this temperature. The results of this
test indicate that highly pure water is quite corrosive.

The amount of corrosion products in the film is greater than
in tests with most of the other dilute solutions, and appre-
ciable quantities of copper are removed in the flowing
stream. No doubt the slightly acidic condition of the water
caused by absorption of small quantities of carbon dioxide
from the atmosphere contributes to its corrosiveness.

Both of the solutions containing free carbon dioxide
might be expected, based on the results of static corrosion
tests, to dissolve large amounts of cOpper in flow tests.
However, in these tests neither of these solutions dissolved
excessive quantities of copper.‘ The solution containing
carbon dioxide partially neutralized with sodium hydroxide
diasolved O.h1 ppm in the flow test.‘ This may indicate a
tendency for the solution to dissolve the film of corrosion
products. The results of static corrosion tests, shown in
Graph 17b, indicate that a solution quite similar to this
is capable of dissolving about 3.5 ppm of copper. It is
evident that solutions capable of holding large amOunts of
copper in solution in the static test did not dissolve
excessive amounts of copper in the flow system corrosion
tests. The hold-up time in the flow system test is only

about one-half hour; this may not be long enough for the

70
solution to dissolve all the c0pper it is capable of holding
in solution.

The solution containing hOO ppm sodium bicarbonate
picked up a significant amount of copper in these tests.
It seems surprising that at the higher pH of this
solution copper still appears to be slightly soluble in the
aqueous phase. These results indicate that even under
moderately basic conditions water coming in contact with
copper will dissolve detectable amounts of cepper. It is
possible that some of the copper entered the solution in
the ionic form as the results of electrolytic corrosion
rather than by being disolved from the surface film.

i The solution containing 100 ppm.sodium chloride is
seen to cause the greatest amount of corrosion. Since most
of the corrosion products are found in the film, this solution
Inay not'be excessively corrosive if the surface film eventually
reaches a limiting thickness and provides good protection

fflor the metal. The copper content of the solution is not

excessively high. There seems to be little indication that
copper is complexim with chloride ions. The copper pick-up
is definitely lower than in static tests with this solution.
,It is remarkable what in these tests on solutions with
a vvixie range of composition and pH the quantity of COpper
dissolved in the water phase does not vary greatly. In

all idastances the capper pick-up is between 0.19 and O.hl

ppm. It‘seems evident that water will pick up copper to

71
about this extent even when in contact with metallic cepper
for a short period of time. It also seems probable that
in water distribution systems constructed of cOpper water
tubing the copper pick-up of the water will not be much less
'than 0.20 ppm unless the hold up time in the system is very

short.

DISCUSSION AND CONCLUSIONS

Summary of Significant Conclusions

Data obtained in the laboratory indicate that the
copper corrosion process is complex. This is particularly
true in corrosion by natural waters that may contain several
inorganic salts and significant amounts of gases dissolved
from the atmosphere. An effort was made to keep the number
of variables in individual tests to a minimum so that
definite conclusions could be drawn from the results. While
it was not possible to simplify all tests to the extent that
only a single variable was involved, it is possible, by
analyzing the data, to determine the effect individual
chemical impurities have on the corrosion process.

Information obtained from tests with simplified systems
loan be used to predict the behavior of more complex systems
:1nvolving several components, such as natural water contain-
:Lng numerous chemical impurities. The data should be of
considerablevalue in that they provide basic information
needed to gain an understanding of the nature and causes ‘of
corrosion in potable water systems.

From careful examination of the data it seems evident

tkuii: the following statements are justified with.respect
_ to corrosion of copper by water and dilute aqueous solutions.

1. Dissolved oxygen is necessary for corrosion, and
is the active oxidizing agent in the corrosion of

copper by water.

73

2. The initial corrosion rate of c0pper in high-purity
water at 60°C is 5.6 mg. per sq. cm. per day.

3. The initial reaction products in corrosion of
copper by high-purity water are predominantly of
the cuprous form.

h. The cOpper concentration of pure water in contact
with corroding copper is influenced by the period
of time in contact and the metal-to-water ratio,
but will be less than 0.5 ppm after the system
comes to equilibrium.

5. Films of corrosion products formed on COpper are
effective in inhibiting corrosion.

6. The presence of chloride or sulphate ions in con-
centrations as low as 100 ppm increase the
corrosiveness of water and causes pitting.

7. Bicarbonate and carbonate ions, in concentrations
commonly found in natural water, do not accelerate
corrosion.

8. ‘Free carbon dioxide increases the concentration of
dissolved copper in both acidic and slightly alkaline

water and increases the rate of corrosion under
acidic conditions.

Discussion of Conclusions

‘222.£2i2.2£ dissolved oxygen in corrosion.--Thermodynamic
calculations (page 10) indicate that oxygen-free water will
not corrode c0pper. The results of a static corrosion test
with oxygen-free water shown in Graph 11,, page 100, confirm
this conclusion. The data presented in Graph 12, page 107,
1x130 show that corrosion did not occur to any appreciable

extent even in a slightly acidic solution of carbon dioxide.
The small amount of corrosion occurring in these tests may
well have been due to traces of oxygen that had not been

excluded from the system.

7h

These results show that if all oxygen.is removed from
water the extent of corrosion is insignificant. Corrosion
of copper in the presence of oxygen is thought to involve a
step in which oxygen reduces hydrogen formed in the corrosion
reaction and prevents cathodic polarization (ll). However,
thermodynamic calculations show that hydrogen reduction
cannot occur in neutral water. FUrther, as pointed out by
Bengough and Hudson (5), there is no reason why, ifoxygen
oxidizes hydrogen, it should not also be able to attack
copper directly.

It may be true that in highly acidic solution, in which
the high rate of corrosion would quickly consume the avail-
able dissolved oxygen, corrosion will occur through hydrogen
ion reduction. It is the author's opinion, however, that,
in neutral aqueous solutions both electrolytic corrosion and
direct surface attack occur only through oxidation by
dissolved oxygen.

:22 initial ggtg_gf_copper corrosion.--The corrosion
of capper specimens immersed in water proceeds at a high
rate initially but the rate steadily declines as the film
thickness increases. Examination of the data presented in

C1raph.la, page 98, and tabulated in Table 1., indicates

tfllat in openpsystem static corrosion tests the initial
ccnrrosion rate is 5.6 mg. per sq. cm. per day. The rate
decreases to much lower levels in a short time; the corrosion

I'a to is estimated to be 1.0 mg. per sq. cm. per day after

75
gone hour and O. 25 mg. per sq. cm. per day after three hours.
'These rates are for corrosion under static conditions in
water saturated with atmospheric gases at 600 C. While the

initial rate is high, corresponding to a corrosion rate of
90 mils per year, it rapidly falls to a much lower level
within a few hours. ‘These data emphasize the facts that
unprotected copper‘has very poor corrosion resistance to
oxygenated water and that much of the inertness of copper
in a corrosive environment is due to the protective nature
of surface film.

Initial corrosion rates of copper in water are reported
by Brown, Roetheli, and Forrest (7) and in aqueous solutions
cfi'hydrochloric acid and potassium chloride by Hill(21).
Brown, Roetheli and Forrest report a value of 0.005h cubic
centimeters per square decimater per minute per average
cubic centimeter of oxygen.per liter as the initial rate of
oxygen comsumption for c0pper corrosion at room temperature
in a dynamic system. Assuming the fermation of cuprous
corrosion products, this is equivalent to a corrosion rate
of 2.75 mg. per sq. cm. per day. While this is less than
the value of 5.6 mg. per sq. cm. per day determined in this
study the two rates were measured at different temperatures

ain under different dynamic conditions. These rates were
determined by completely different methods and yet are seen
fit) tee of the same order of magnitude. The rather close agree-

ment between these values seems to provide additional evidence

76
in support of the conclusion that the initial corrosion rate
of c0pper in oxygenated water is very high even at moderate
temperatures.

The state 23 gxidation of copper gorrosion product§.--

 

 

Since copper exists in both the cuprous and the cupric

form it seems reasonable to assume that either might be
produced in the corrosion reaction. The results presented
in Graph 5, page 103,indicate that in these closed-system
tests corrosion stopped when the limited supply of oxygen
was consumed. Calculations show that the quantity of oxygen
consumed was sufficient to oxidize the corrosion products
only to the cuprous form. It seems reasonable to conclude
that the initial corrosion products are of the cuprous form;
if sufficient oxygen is present these may be oxidized in

a second oxidation step to the cupric form. This would be
expected to happen quite rapidly with dissolved cuprous
products (27, 37).

If the conversion of cuprous corrosion products to the
cupric form occurs at an appreciable rate it is possible to
make some interesting and perhaps signifitant speculations
concerning the corrosion process. Films of cuprous oxide
oxidizing to the cupric form would change in structure and
composition. Since cupric oxide is thought to be more
soluble than cuprous oxide (37), oxidation might increase
the rate of film dissolution. Structural changes in the

film may introduce cracks and imperfections that would lead

to pitting.

77

In pit-type corrosion, cuprous ions entering the solution
as the result of electrolytic corrosion would diffuse away
from the anodic area and would be further oxidized to the
cupric form. This would reduce the oxygen concentration in
the pit area and increase the cell activity by creating a
differential oxygen concentration between the anodic and
cathodic areas. The reduced oxygen concentration would also
diminish direct surface attack in the pit area and help
prevent stifling of pit activity by surface films of
corrosion products. The ability of co,per to exist in two
states of oxidation increases the complexity of the corrosion
process. It is intere ting to note that pitting commonly
occurs in the corrosion of copper and iron, both of which
.are capable of existing in two fairly stable states of
oxidation.

The copper concentration of water in contact with

 

corroding copper.--In all static corrosion tests with water

 

that did not contain free carbon dioxide the copper con-
centration of the aqueous phase never exceeded 1.0 ppm. In
most tests the concentration was found to be less than 0.5
ppm after the test had been in progress for a few days. In
tests with commercial grade copper and high-purity water the
concentration increased to nearly one ppm but then decreased
to a lower level as shown in Graph 2b, page 99. In tests
with solutions containing sodium chloride the solution con-

centration fluctuated but remained high throughout the test

78
period. This may have been due to complexing of c0pper with
chloride ions.

There seems to be a tendency in all tests for the cepper
concentration to increase to a maximum and then to decrease
during the remainder of the test period. In nearly all the
tests small amounts of precipitated copper compounds were
detected after the tests had been in progress a few days.
There seems to be definite evidence that the solution I
initially dissolves more copper than it can hold in solution,
thus becoming unstable with respect to copper corrosion
products. Copper compounds are then precipitated from solution.

There are several possible eXplanations for this
phenomenon. Capper ions might enter the solutiOn in the
lower state of oxidation and be oxidized to cupric ions. How—
ever, most cupric compounds are more soluble than the corres-
ponding cuprous compounds and the solubility would be
expected to increase rather than to decrease by this trans-
formation. Another possibility is that the hydroxide is
first formed in solution and is slowly converted to less
soluble oxides. A third possible eXplanation is that capper
oxides dissolving in solution increase the alkalinity and
reduce the solubility of corrosion products. However, in
all the tests in which the concentration of dissolved copper
remained low the pH of the solution never increased more
than three or four tenths of a pH unit. The solubility of

corrosion products does not appear to be influenced much by

79
small changes in.alkalinity and it is doubtful if this factor
is significant.

The fact should be kept in mind that, in electrolytic
corrosion, metal atoms are thought to lose electrons and to
migrate from the anodic area under the influence of an
electrical potential. The concentration of ionic species
in solution influences the electrical potential, but the
degree of saturation with reSpect to anions present in
solution would not affect the electrolytic reaction. Thus
it might be possible, in corrosion occurring by the electro-
lytic mechanism, for the water phase to become supersaturated
if precipitation did not occur as soon as the copper ion
concentration increased to the concentration at which the
solubility of the least soluble copper compound in solution
was'exceeded.

In cepper corrosion with high-purity water the water
phase reached a higher copper concentration in tests with
commercial copper than with high-purity cepper (Graphs lb
and 2b, pages 98 and 99). If the unstable condition of the
solution is Caused by copper ions entering the solution by
the electrolytic process, the impurities present in commercial
grade capper would be expected to increase the amount of
electrolytic corrosion and to cause a higher degree of super-
saturation in the solution.

The data presented in Graphs lb, 2b, and 3, pages 98,
99, and 100, show that the cepper pick-up of the solution is

80
not greatly altered by the presence or absence of a film on
copper at the start of the test, by changes in the ratio of
metal surface to water volume, or by slight changes in pH.
Graph lb, page 98, shows that a four-fold increase in surface
area increases the copper concentration to some extent but
the increased concentration is considerably less than four
times as great. In one-and-one-fourth-inCh copper water
tubing the metal to water ratio is 28 times as great as in
static corrosion tests run with a single c0pper strip. It
seems likely that water standing in such a tube for a few
days would pick up considerably more copper than would be
picked up in static corrosion tests with a single c0pper
specimen. The presence of a well develOped film on the
metal surface, as would be found with water tubing in service,
reduces the maximum concentration attained (Graph 2b, page
99) but has littleinfluence on the cepper concentration the
water will reach in longer periods of time.

The results of corrosion tests with slightly alkaline
water, shown in Graph 3, page 100, show that the copper pick-up
of the solution is not significantly altered by a small in-
crease in pH. With commercial grade copper the maximum in
the concentration curve is suppressed Slightly while with
high-purity c0pper the copper concentration is seen to be
slightly greater than in tOStS‘With water that is slightly
acidic due to carbon dioxide absorption. It is evident

that small changes in pH do not greatly reduce the concentration

 

81
of corrosion.products in solution. In corrosion teats with
sodium.carb0nate solution having a pH of 9.8 (Graph 7a, page
102) detectable amounts of copper were still present in
solution.

It is interesting to note that, in corrosion tests with
non-carbon dioxide-containing solutions varying in pH from
6.5 to 9.8, the maximum copper pick-up never exceeded 1.0
ppm, and after a period of about one week the solution con-
centrations were found to be between 0.1 and 0.5 ppm. It
is rather difficult to explain the amount of copper dissolved
in the flow tests when compared with the c0pper pick-up in
static tests. In static corrosion tests with high-purity
copper, using 300 ml. of highly pure water, the cOpper con-
centration never'exceeded O.h ppm. In a three day flow test
using 7900 ml. of water the copper pick-up was 0.28 ppm.‘ In
the flow test the quantity of copper dissolved in the water
phase was approximately 18 times-as great as under static
conditions. The amount of cOpper dissolved was also quite
large in all flow tests. Even the solution of too ppm
sodium bicarbonate picked up 0.18 ppm of cOpper and the
amount of capper dissolved exceeded 0.1 mg. per sq. cm. in
the three-day test.

In general, it seems as ifwater and dilute aqueous
solutions dissolve corrosion products to the extent of 0.2
to O.h ppm even whenin contact with corroding copper for

a short period of time. It may well be that the solubility

9 82
of corrosion products is in this range, and the copper con-
‘centration attained in both static and dynamic tests represents
the solubility of the corrosion products.

The solubility of corrosion products does not appear to
change greatly in the pH range of 6.5 to 8.3. This may be,
due to the amphoteric properties of capper oxides. The
solubility of cupric oxides is known to decrease as the pH
increases and reaches a minimum solubility after which the
solubility increases as the acidic form of the oxides become
soluble under highly basic conditions. McDowell and Johnston
(29) report that the minimum solubility occurs under slightly
alkaline conditions at a base strength of less than 0.01 N
potassium hydroxide. It is possible that under neutral or
slightly basic conditions the cOpper oxides are near their
minimum.solubility and are in the transition region.where
both their basic and acidic properties are apparent., Under
Ithese conditions the solubilities might not change greatly
with small changes in pH of the solution.

ihe large quantity of cOpper dissolved under flow con-
ditions as compared with static conditions may result from
the increased diffusion gradient of the solution. McAulay
and Spooner (28) maintain that with metal corroding in dilute
solutions the concentration of ions in the water phase at the
metal-water interface way be independent of the ionic con-
centration in the bulk solution. For bulk solution con-

centrations below the equilibrium surface concentration the

83
potential of the metal is constant and does not obey the
Nernst relationship. For cadmium this equilibrium concen-
tration is approximately 1 x 10'5 moles per liter; COpper
would also be expected to be of this order of magnitude.

In a static system the bulk ionic concentration would
tend to increase to the equilibrium surface concentration
quite rapidly but might tend to increase beyond the equilibrium
concentration at a slower rate. In a flowing system the
concentration of the water phase would also increase to the
equilibrium surface concentration rapidly and might be able
to dissolve this quantity of c0pper even though the hold-up
time is short. Diffusion of copper ions away from the metal
surface definitely is important in the corrosion process.
Evans (1h) reports that a copper specimen under flow :oh-
ditions will become anodic with respect to a Specimen in
contact with similar solution under static conditions.

There appear to be many factors that can influence the
cepper content of aqueous solution in contact with corroding
copper. After a sufficient period of time under static cen-
ditions the c0pper concentration is probably determined by
the solubility of the corrosion products in the aqueous
phase. The solution concentration appears to increase to
some maximum value, due to electrolytic corrosion or to
formation of unstable soluble compounds, and then to decrease.
The value of the maximum concentration in dilute non-carbon

dioxide-containing solutions is influenced by the metal-to-water

8h
ratio, the surface condition of the metal, the degree of
purity of the metal, and variations in pH. While the copper
concentration remains small and prdbably would never be much
above 1.0 ppm, it is impossible to say precisely what the
concentration will be at any time without a rather complete
knowledge of the system. This is also true with respect to
the amount of copper that will be dissolved under flow con-
ditions.

The protective nature pf surface £ilm§.--In static
corrosion tests with solutions that do not contain carbon
dioxide and with alkaline solutions containing carbon
dioxide, films developed rapidly during the first few days
of test and were effective in reducing corrosion. Since
in all these tests the bulk oxygen concentration remained
high, it is evident that the reduced rate of corrosion re-
sulted from formation of protectivesurface films. Copper
corroding in dry air at 209°C has been shown to form pro-
tective oxide films. The relationship between film thickness
and time was found in most instances to be parabolic in
nature (10). Deviation from the parabolic relationship in
the formation of films at higher temperature is attributed
to structural changes occurring in the oxide film during
the corrosion process (26). Films formed on cepper corroding
in aqueous solutions appear to be tightly bound and continuous.
Since cOpper forms protective films in dry air it is not sur-

prising that films formed in aqueous media are alSo protective.

85

In most of the static corrosion tests the relationship
between film thickness and time is nearly parabolic. This
relationship would be expscted in corrosion processes in
which diffusion through the film is the controlling factor
in the corrosion reaction (25). Solutions containing chloride
or sulphate ions cause rapid film development until the film
reaches a thickness corresponding to a cepper content of
about 0.3 mg. per sq. cm. after which corrosion falls to a
much lower level. In tests with natural water and slightly
alkaline solutions containing sodium bicarbonate and free
carbon dioxide, films developed rapidly in the first row;
days, then film growth decreased to a very low level.7 The
films maintained a nearly constant thickness for the remainder
of the test period. ‘

Although in tests with different solutions the rate of
film‘builddup'varies considerably, in all tests the growth
rate of the films decreased as the tests continued. Since in
most static corrosion tests substantially all of the corrosion
products remained in the surface films, corrosion was reduced
to a much lower level after surface films developed. It is
apparent that the degree of protection afforded by films of
corrosion products, rather than the rate of chemical reaction
of unprotected metal, controls the over-all rate of corrosion.

Films deveIOped on high-purity cepper in contact with
highly pure water are shown in Graph 1, page 98, to contain

0.31 mg. per sq. cm. of corroded copper after two days.

86

Assuming the film to be cuprous oxide having a density of

6.0 grams per cubic centimeter, calculations show the film

to have a thickness of 6,h50 angstrom units. >Since none of

the corrosion products were identified it is impossible to

assign definite thicknesses to any of the films but most of

them.probably range from 5,000 to 10,000 angstrom units.~
While there are probably adequate explanations for the

variations in shape of the corrosion curves, the most signifi-

cant conclusion drawn from these corrosion tests is that

films are capable of inhibiting corrosion. However, it

cannOt be assumed that films which are initially protective

will continue to inhibit corrosion indefinitely. If pitting

later deveIOps the pretective preperties or the film would

be destroyed. Thus, two factors are of primary importance

in determining the long range corrosion resistance of metals

in corrosive solution: the protection afforded by surface

films, and the extent of pit activity which may later deve10p.
The effects 2; chloride 32g sulphate iggg.--In static

corrosion tests cepper in contact with solutions containing

chloride or sulphate ions develOpsfilms containing 0.3

mg. per sq. of corroded copper within three to six hours.

With the exception of the solution containing 70 ppm of free

carbon dioxide all other solution did not corrode copper to

this extent in less than two days time. These results indicate

that chloride and sulphate ions in solution accelerate the

rate of corrosion of unprotected or partially protected copper.

87

Graphs 8a and 91a, pages 103 and 101;, show that although
solutiom of 10 ppm sodium chloride and to ppm sodium sulphate
cause rapid film development, the films, once formed, are
quite effective in reducing further corrosion. COpper corrod-
ing in 100 ppm sodium chloride solution, however, is seen
to corrode considerably in the second week of the test period.
In these tests pitting was evident and this may be an example
of film failure due to pit activity.

The increased corrosiveness, indicated by rapid film
development, may be due to the high penetrating power of
chloride and sulphate ions. During most of the time films
are deve10ping, the metal must be covered by films varying
from a few hundred to several thousand angstrom units in
thickness. Evans maintains that both chloride and sulphate
ions have considerable penetrating power and thatit is
difficult to make metals passive in the presence of chlorides
(12). This observation may offer an explanation for the
corrosiveness of solutions containing chloride and sulfate
ions that is evident during the early stages of film develop-
ment.

In addition to the high rates of initial corrosion
observed with (solutions of chloride and sulphate ions, these
ions also appear to cause pitting which might result in
failure of fully deve10ped films. There is reason to suspect
that significant quantities of these ions in natural water
Wohldincrease its corrosiveness and reduce the life of copper

water tubing.

88

1kg effect 9}; bicarbonate g_n_d_ carbonate Logan-Copper
corroding in solutions of sodium.bicarbonate and sodium
carbonate is shown, in Graphs 6a and 7a, pages 101 and 102,
to develOp films quite slowly and to produce films contain-
ing about 0.35 mg. per sq. cm. of capper in about one week.
These curves show that corrosion is reduced by film deve10p-
ment as would be expected in corrosion reactions controlled
by diffusion of reactant through protective films. There
is little indication that these ions would accelerate. the
corrosion process. Solutions of sodium carbonate appear to
be less corrosive than high-purity water. This may be due
to the higher pH of the solution, which.reduces the corrosion
rate and decreases the solubility of corrosion products.

It is also significant to note that the data in Table
III, page 63, show both cuprous and cupric oxide to have low
solubilities in a solution containing 600 ppm of sodium bicar-
bonate. Bicarbonate ions apparently are not responsible for
the high copper pick-up in static corrosion tests with nearly
neutral aqueous solutions containing sodium bicarbonate and
free carbon dioxide.

While there is little indication that sodium bicarbonate
or sodium carbonate increases the corrosiveness of water
there is some evidence to show that these compounds have
beneficial effects. These substances are mildly alkaline
in aqueous solutions and their buffering capacity tends to

keep natural water containing a variety of dissolved substances

. 89
in the neutral range. Dissolved carbon dioxide causes water
to be acidic and would be much more aggressive when present
in natural water if it were not buffered by carbonates.

The effect 2£|££gg carbon dioxide.--Dissolved carbon
dioxide produces an acidic condition in aqueous solutions and,
as shown by Graph lla, page l06, the corrosiveness of the‘
water is greatly increased. The cepper concentration of»
the solution rapidly reached a high.level and corrosion‘
continued at a high rate until the available oxygen was con-
sumed. Although films deve10ped quite rapidly they were of
limited value in inhibiting corrosion because the corrosion
products were quite soluble in the water phase.

Both the increased.corrosiveness and the high copper
solubility of carbon dioxide-containing solutions can be
‘attributed to the acidic nature of the solution. However,
the results presented in Graphs lb. and 15. pages 108 and 109
show solutions containing free carbon dioxide to be excessively
corrosive at a pH of 6.7. . I

~ Slightly alkaline solutions containing carbon dioxide
buffered with sodium bicarbonate or partially neutralized
.with sodium hydroxide do not appear to be excessively
corrosive. Such “solutions are capable of dissolving much
largmsquantities of corrosion products than non-carbon
dioxide-containingsolutions.. It seems evident that the
presence of small quantities of free carbon dioxide in natural

water is responsible for the high copper pick-up of such

9O
waters. It is difficult to see how free carbon dioxide
could increase the copper solubility of solutions but it is
possible the undissociated carbon dioxide complexes with
copper ions (35). The data in Table III, page 63, indicate
that the solubilizing effect of carbon dioxide is specific
for cuprous ions and does not increase the solubility of
cupric compounds.

Whether or not the ability of water to hold as much.as
three or four parts per million of cOpper in solution would
seriously increase the corrosiveness under conditions found
in potable water systems is debatable. It seems that water
capable of holding several parts per million of cOpper in
solution would dissolve considerable copper under flow con-
ditions. However, the rate of dissolution would influence
the quantity of copper dissolved.

From the data presented in Graphs 11b and 15b, pages
106 and.lO9, it is evident that solutions containing large
amounts of dissolved copper lose their capacity to hold
capper in solution and decrease in copper concentration after
a short period of time. The copper concentration increases
to a maximum and then rapidly decreases. Riley (35) reports
that aqueous solutions of sodium acetate, sodium malonate,
and sodium succinate, in which strips of copper are corroding,
attain a maximum copper concentration and then decrease in
copper content. He attributes this maximum in the con-

centration curve to the formation of unstable copper

91
coordination compounds which are later converted to insoluble
basic carbonates. In this investigation precipitated cepper
compounds were found in the test bottles and surface deposits
were noted on the metal specimens in corrosion tests with
solutions containing dissolved carbon dioxide.

It is evident that insoluble compounds form.and complex
changes occur in such systems. The nature of these changes
probably cannot be fully described without additional
laboratory investigation.

The high copper pick-up of natural waters containing
free carbon dioxide may constitute a health hazard. It is
apparent from.the results of this investigation that if such
water was allowed to stand in copper water tubing for a few
days it would.dissolve several parts per million of copper.
Drinking water standards for interstate carriers adopted by
the U. 8. Public Health Service limit the capper content of
potable water to 3.0 ppm (2). water with a high carbon
dioxide content, if allowed to stand in capper water lines
for’a few days, might dissolve enough cepper to render it
unsafe for drinking.

The Influence of Chemical Impurities on the Corrosiveness of
WEEEFEI_WEter ‘

In this investigation the corrosion process was followed

during the period of initial film deve10pment and for a two

or three week period. The results of these short term tests

92
cannot be expected to predict the extent of corrosion that
may occur over a period of several years. They should,
however, give some indication of the corrosiveness of water
containing many of the impurities commonly found in natural
water.

Velocity effects were not studied and no attempt was
made to determine the factors responsible for pitting. Evi-
dence of pitting was noted on copper specimens corroding in
solutions containing sulphate or chloride ions and in
solutions containing free carbon dioxide. While films of
corrosion products were shown to be protective, these films
would lose their effectiveness if extensive pitting occurred.
In water containing these materials, films that effectively
reduced corrosion during the period of initial film develop-
ment might lose their effectiveness if tested for longer
periods of time.

The high copper pick-up of solutions containing free
carbon dioxide might cause higher rates of corrosion under
flow conditions. If films of corrosion products were dis-
solved the thickness of the films might be reduced to such
an extent that they would no longer provide adequate pro-
tection. Water having a high cepper solubility could also
cause increased pitting activity. The potential of anodic
areas is influenced by the concentration of copper ions at
the metal-water interface. Diffusion gradients would be
higher with water having high copper-holding capacity. Rapid

93

diffusion of copper ions from the surface would increase the
potential of the metal and also reduce the chance of pits
being stifeled or deactivated by the formation of insoluble
corrosion products. These effects would be expected to be
more pronounced in dynamic systems than under static conditions.
Some conclusions can be drawn with regard to possible
methods of reducing corrosion in potable water systems. Since
fundamentally oxygen is the active oxidizing agent in water
and is responsible for all corrosion, if oxygen were completely
removed all corrosion would be eliminated. Carbon dioxide
is also undesirable because it accelerates corrosion in the
presence of oxygen. Deaeration may be a practical means of
reducing the corrosiveness of natural water, as both oxygen
and carbon dioxide could be effectively removed by this
method. It is not known to what extent oxygen and carbon
dioxide would have to be removed in order to prevent corrosion,
but if both of these components were reduced to a concentra-
tion of less than one part per million corrosion would
probably be reduced to a very low level. Commercial deaerators
are available that could accomplish this economically (33).
Carbon dioxide can also be removed by' neutralization
with sodium.carbonate or sodium hydroxide. The corrosiveness
of carbon dioxide-containing solutions appear to be con-
siderably'reduced by'a small increase in.pH of the water.
Comparison of the results of corrosion tests presented in

Graph lha, page 108, with data in Graph 16a, page 110, shows

91;

that a small increase in the alkalinity of the solution is
effective in reducing corrosion. Partial or complete
neutralization of the free carbon dioxide in.water would
reduce its corrosiveness and might be an economical method
of reducing corrosion.

While solutions containing only 10 ppm of sodium
chloride do not appear to be excessively corrosive there is
some evidence to show that solutions containing higher
chloride concentrations may be moderately corrosive. Chloride
ions also appear to accelerate corrosion of partially pro-
tected copper and.may increase pit activity. There is little
that can be done to reduce the chloride content of natural
water. However, zeolite water softeners are often regenerated
with sodium.chloride and highmconcentrations of sodium.chloride
enter the water systems if the units are not flushed preperly.
In order to keep corrosion at a low level it seems worthwhile
to make sure that, after regeneration, all softening units

are thoroughly flushed before being placed in service.

Suggestions for Future Work

The results of this investigation demonstrate the signifi-
cant role played by films of corrosion products in reducing
corrosion and controlling the over-all corrosion reaction.
Although the initial corrosion rate of unprotected capper
is high, the fermation of protective films rapidly reduces

the corrosion rate. The corrosion rate of unprotected cepper

95

is not indicative of the extent of corrosion that may occur
under service conditions. In attempting to learn what factors
are responsible for corrosion, primary emphasis should be
placed on the study of film.fbrmation, pitting, and rate of
diffusion of products and reactants to and from the metal-
water interface.

Since pitting could destroy the effectiveness of surface
films, a study of the extent of pitting activity on copper in
water and dilute solutions is necessary in order to predict
long range corrosion tendencies. Static or low velocity flow
tests may be satisfactory for studying pit corrosion, because
pitting does occur under these test conditions. Suitable
means would have to be found to evaluate the extent and
intensity of pitting. A more satisfactory means needs to
be developed for removing surface films. Films that are
more than one or two weeks old are difficult to remove by
erasing. It may be possible to dissolve surface deposits
with dilute oxygen-free acid without corroding the underlying
metal.

An investigation of corrosion under dynamic cenditions
would undoubtedly be worthwhile. Although the chemical re-
actions occurring in the corrosion process would probably not
be altered by increased fluid velocity, the diffusion rate
would be increased and the over-all corrosion rate might be
increased. Film erosion might also occur under turbulent

conditions. The corrosiveness of solutions containing free

96

carbon dioxide, which are capable of dissolving several parts
per million of copper, should be investigated under dynamic
conditions. Under dynamic flow conditions the cepper pick-up
of the solution may be much higher than in the low velocity
flow tests used in this investigation.

A study of the solubility of corrosion products would
provide some worthwhile information. Film stability, and
hence the protection afforded the corroding metal, depends
primarily on solubility of the corrosion products. The
solubility of copper carbonates appears to be complex and
may play a significant role in the corrosion process.

In attempting to evaluate the corrosiveness of natural
water it is probably best to work with dilute solutions con-
taining only one or two dissolved substances. In this way
the effects of individual chemical impurities can be evaluated,
whereas with more complex systems it would be difficult to
determine which effects are caused by specific components.

It also seems important to work with dilute solutions in
which the concentrations of dissolved solids or gases are
in the same range as in natural water.

Basically, corrosion reactions are chemical reactions
and, although they are more complex than reactions occurring
in homogeneous systems, through.careful investigation it
should be possible to gain an understanding of the mechanisms

involved in the corrosion process.

GRAPHS AND TABULATION OF DATA

Grgphic Presentation of Datg

 

The results of all corrosion tests are presented in
graphic form on pages 97 to 115. Most of these graphs show
the distribution of corrosion products as well as the total
amount of copper corroded. In many series of tests the
copper concentration of the water phase is plotted in a
separate graph in order to show the copper content which

the solution attained in static corrosion tests.

422£%2522 czzeuecunaao {2zgz./’<;aof)

cat/65417764770” 4/ 23/1

98
GQAPH io wears/0y 0; Cu 14/ file}! Rae/TV H20

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I I '
l
I /h
. . N I ’,,J/’
- _ _ -i__I----____.L.___._s_.___..;_- - _ ,W/‘(i _____‘ .__-,-_._._._
04 % L/ 3 I
' I I i ‘.
g - i i a 9
9 =: , (I . .
i ;//’ ! 'no77uL czuazmzanaaq
I : 5 1 i .‘
ass—m- cum—ea“. -_ *
, I ‘ i s
. ' i 1 ' i
I E Argos caeaoeczamzuzz>ancwy
' ; i [60.20.59 /4/ .504 077041
i : .
ZZZ—o/6~—-. f I AfloV%-,6ozyrv/czuyzue
. ? g swan/.5x57274 —v-c27°c:
é §L~—~---r—~ 1 "~~~‘
' l i I
I f i z
i . 4 7
0.1 L i is
l ‘ ! i
: . i ‘
l ' I
1 1 i
f ? S;nsazovvzw/
% . i .
' l
0 '+-_-: ________ e-- j—‘-———r-—-—-——.._.J,
0 £2 41 6 <8 /67 A2 A4

77/45 [A] .DA Y5

6 PAPH 1b coeeoa/ou 0; Cu 1W .1/16/1’ Mary 4% 0
l, 0

 

 

 

 

 

 

 

 

 

 

 

r .r r r l
0, 8 . 41.579qu 4 JPIC/ME'A/S f4 » see/As) .
06._____- "*1 czar/MI JPJC/HEAJ ; _-____.._--_,.I__-,
(24 f.'” I fig“ 1. "m. ¢Vfliigxofl
new; in". ,_.- 1L; “.31;de ~+
0 0 caA/p/ T/OA/CO VIC/M!”

4

6 6 /0 AZ /4
77/75 [xv £14 7.5

. cap/0459 caeeooso (fly. / cm 3)

award/734274;” /’(/ 29.4214.

99

 

I

0.6%-”; I. fél’AA— [oEEOs/a

 

 

 

\
M» 4‘

 

 

 

 

 

 

 

 

 

 

 

 

; /
I /
0'4 -.... --. .-. .-_ -._.I: . I .. MIL / . I. I . , I "_._ I “I -- I _-- ~—+~-~-~--~-——-—--—
E I 5 a
i ' t I
f l _
9 —— - —-ro 77M. coppéze
,I 5013420050 Au/A y
!
0.2-Mm _._._3 —- m - - - - -— COP/056 //{/ gag/Way
I CONMEQC/AL cop/05,9
/' £10541 5K5/25l-4 .—- 60° c;
- l i I
g f I
I : j—fioz—aIf/au ?
0.0 ”'1‘1’4’“ ”“ +—-—---—-— -- ——l— -_ -—¢
0 2. 4 6 5 M /2

GEAPH 2a f/ME [U MYS
6063 05/0” 0;: cop/0:79 AV A0619 name/7V Wit/fie

 

I

a) _.

- CA [AA/£0 chose/M I

     
 

 

 

 

 

a 9 2 4 6 <9 /0 /2 /4
GRAPH 2b 729/5 11! £44565
[0,8,6 0.5/0A/ 0F 6010/9529 AV fl/éé’ «me/7y W472?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

% 100
Q .
§ E a6 0054/ 6YST£M -— 60°C
k \Z 44 /\.‘~L\ ‘Nh;-_3— ~- -- A 4.1-. -' --
§ \ /' ."\..—_—._. T_M—.——__.
3 0.2 - 1116,92 Baggy maze- -- 601144525144. ware
S S
\i a 4 6 8 /0 /Z
GRAPH 3 TIME [JV DA Y5
g coeeos/OA/ or coppae xv 506/! 72. Y lLKAA/l/l I’M/£56
v “S !
§ u 6206“ § 544.5527 sysrm- 4
(Q \ QM - -, _. wégjwzyfiapnge
‘11 $0.02 - mrflg caeeagw -440_°C_- .- - 4-
g Q — i — ‘ [7 -* -f " W"
E 2 4 6 8 /o xa
GRAPH 4- TIME IA! DID/.5
caeeas /m/ a; coppze M/ atyms'd/ -Fe££ W4 7‘59
'~\
. ' g T \
‘ W‘” R
\ ’ i i . Q
\ // fi__---_--___p-_ ___.--._? - -_ - L Q
Q 02L:\————/ - 7 ~—+’ -— —--- ' 5‘“50 5Y575/4 ‘A/Ifl S
'q / i . may - PUB/7V cap/351*?» 5
I ;
Q I 6006 . R
o N I -__-_-__.---_- m4- _ . Q
31 3 . # : M
Q . ~ ‘5 ,1 g g
V \ at , A .. .. p.4- 4 +_-.-_--_ 45 Q
5‘“ Q I \ ! §
% \ x at:
\) £649.!” 6&1/562/734 flax/W z‘
4 ~ 7 - i \g Q
2 J 4 5 6

GRAPH 5 77MB W DAYS

caee as lav 0F cap/2.529 //V WA 756 W/T/V 444/7229
0! Véix/

(cease caeecwzo (47;. / 0,92)

cox/6:” raeAf/au /A/ «42/4.

 

 

 

 

 

 

 

 

 

 

a4 : +. , , W —
: / ‘;
I f 4
i / 1 7'0 774;; cae edixox/
/ 3 i i
1 : ‘ .
/ ' 3
as» — » ;/—————--~—~i--———---——-—+—~ 4»- ~—————»——-——-——.--
/
/ " I
f . i i
; I ; ‘ "—-
; i g ; 7? ML cap/>22? came»
; ' [—---- ”my
0.2 - -;______-_-!T_----_.--~;_-... i mange w5aar/ou
’ F I ~ ‘ am” 5K5)?”
' Max-pug”? cap/we
60° C
I '“W - . V _—._i—__ |
f , 4 I
0.] r «-..- L 7;, 4 44 .. __ 4 ._ __ ._4.____
.' .
l
{ ' I
‘ 5040f/(Jx/ ,
_...-,_ ........ ..I- .....
2 4 6 5 la /2
GWPH 60 77/45 I” DA Y5

(06605/01/ 0F

coy/’52 AU mppm [4/0 #605 504 U776?”

 

 

 

 

 

 

05-. —~ —— mi
06. . _-- __
a4»--- _ g 4
622 - _ ’ _-- ----
;f 1 T 4* L : ޴
2 4 6 8 /0 x2
GRAPH 6 b TIME [A] DAMS

caeeos/cu/ 0F (029052? M/ 460,40”? #0 #603 50407700

caecoaev (My. / any-7)

COP/0528

Cal/667073634 7701/ 44/ ADM

Q
0A

9
N

aa
06
0.4

02 --

102

 

 

 

. i
! ,_..; ram; w££0510u

.... .---_._- -_ .-4 -_-_.-.

—
‘ T

y ._ _._ -_-_.._.,-_--

I

i ’27? 7741. can-425.9 weeooso AMI/IV

 

 

Izcolqoéie w 531.0770”
4 § 0,050 5xsrm

é fl/é/x V— Page: cop/660
:L 6o .6

-..... .. -..-..“ __._ ._ A._.-_—-.. ..h. .. _. ," ... .-4._ .... _., -.-..—

jazwt/w/

 

 

 

 

 

 

 

 

 

 

_-. ...... -LC___-____-__--_
2 4 6 <3 I0 /2
GRAPH- 7C1 TIME Ml DA Y5
60660.5 /0/¢/ or €010,052 /A/ .50me A/ae 603 50407701!
1
fl ; 9
2 4 6 8 l0 [2
GPADH 7b TIME W DAYS

(fleewloxv or 609056 /A/ .50me A/az 605 5040270”

cage 0.050 //2:? . / 602-9

6090253

cat/66¢! far/o” /// PPM

0.6 .- ----

I‘ I
l
MW : ~- J /-—«— - ’ ~ ————:
z / i .
a / f ! .
i / ’ f ,

. ‘ 3 i E

’ l , l' '

i a l :

0.3. ' '____---- r -,--_-_--_--__ _ ---... . L --- J, 4:-

E —— - - far/u. cop/arse 60££OM0 «My
I t 1 ..... cap/0&6 /4/ sou/270”

0./

 

 

 

 

 

 

 

i I '
xg’ I
1 / 4,.- 444,-,
i :
i
I

025v .9)!er
III“ -pfl£/7Y 60M“
0.2-4!-----_,4 4?-- _._--_.T._----__ far/2- 0" c

._ ’—-- .

,/t/

 

.4

/

’

 

 

 

 

 

 

 

 

 

% ; ‘ E

5 1 ' i
' i

t

? : ; 5051277041
f“:’“ -—-+---—.--*——i-’-'—-:~-—--;—-——-4

2 4 6 8 /0 la
GRAD-H 6 a 771%: //U .2045

caeeos/m/ or comm All 40,90,» fl/dg 504 Jazz/770M

 

 

 

 

 

 

 

 

 

 

as.-- .- _- - _ f- - _
616 i ‘ 4 . ,- ___-_-L _ i ;
0.4-. . 1., + _ ' _. __
0.sz . f 1_‘T
. i J ! _
2 4 6 8 l0 /2
GRAPH 6 b 7724;: nu p.473

caeeosxav a: cap/068 M/ 4030»: Ab; 504 Wflf/dfl

(04056 caeeaogp / x77; / cm 2]

WCf/(lfie‘l 7701/ /l/ 33M

 

0.4

as

 

 

>
-4_-,---_-----_-._--..._ ._ - .- -__- 4 .- 4-,..- -_ .- --L4-‘--__.- ---_-_._- ._-._.~--.-4 -- . .,0 _.

 

 

 

 

i __ _ __- r0744 (OP/066 Caeeeoa'b AIM/AV
0-2 ; . i _ - COP/058 M/ 50/. 077041
I ‘ 1
i oar” 5Y57£M
fl/é/V- Ame/7V cwpze
E 60 0c:
0014—4 4 - - 4-". __L:4
I
I . 9 :
‘,A-\4,__,A_ ....... ;+-——..a—_-' ____ _______§______J
. 2 4 6 8 /0 /2
GQADH 901 TIME 11! DA Y5

caeeas/ou 0F co/OPEQ M/ /0,o,am 1/0 C/ .504 arm/U

 

 

 

 

T ,
0.2 I g , L
0.6- ; . f.
a4 ; - ; I _
0.2 r M; +1-4“ _.__..L-..- -_-. ..T
2 4 6 8 /0 /2.
GRAPH- Qb 77MB ll/ MKS

caee 0.5/04/ 0;- COP/94576 //l/ /0 ppm. A/a (/ 50w7704/

(aucgflreaflad MI 3/2/44.

(0,0958 609800460 [”257 / 0293)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

/.0 235
, , - ,/
0.8..---- - .- - . -- --_ -- -- - _. - 4 4- .- -4-.-/--__.-___ -._.-4
5 ’/- FOI'AL geeas/OAK /
____4__________
/ i 9 _\ . /
/ z \\/
064/ . f -- 4 ....... ”.4
I ‘
Q44 _-.-_- - 71—70744 coppse meeoow AWAY
* —* - -- (09052 M/ sou/now
f Gad—Al 5Y57£M
a f fl/él/«PMQ/rr COO/3&8
60" c
0.2, _. __ __ it -- -- —- - __ ---_----.-;_-_
l
. ..... 4k sonar/o” j
I£~~~+ ————— \ -.-_-a___-.+———-Jf
6 c9 /0 /2
GQAPH Ioa TIME IN DAYs
caeeas/o/v ox: coppae /A/ /00p,007. l/a' C/ 60LUT/OU
0.8_ - - - _--____---.
as -_--. - _ -..___--__ --
0.4 __-__ - _-_ ....- _ i 4 1
02 4 _- __ 4 _ _;- _ - _ _. ---- __ ___.4'-_ _ _.
6 8 /0 /2
GRAPH IOb TIME. 11v DAYS

caee 05/04/ a; Coppge M/ /00p,0m- A/a (/ 504 077W

cawxwue cowaeov£z>(épg/Qyné}

aaoczx/ne47wau Ab’/%QA¢

106

 

 

c205504mcma~7 : /3
a7 -fimaynaw97ycaaaae;__w. _-_ --4:Kfflmn;m
60°C : -——r“"/I/ ‘
-— ~ -- --.-----—;-7-—v—rf’~~4 7 : w f
0.6 - . . ; f 7017M -: -- .--i---- / ‘-
,/f/ . ' / f
06- / ' 4-4-“4-_,-4 5
, - /
a4 : . raflnu4 T . 4
, , Jr.- 4. .- -/ i
/ \ fif” 5 1 '- i
05 ' ” "“ “ ””-;;;£Z;EE€§@%§ww-~~. 444 - »5
“{r/ k-‘E""""‘ ‘ \' ,
/

 

 

 

 

 

2 4 6 8 IO /2 /4 /6 /6
GRAPH- 11 c1. TIME IN DAYS
(aeeos mu or coppae //V 70,9009. (0.? 50407/01/

 

 

 

 

 

2 4 6 c3 /0 /2 l4 /6 k9
GQADH 11 b TIME IN DAYS
cae£v6«20'6V‘60fiwzxaav'zagwwmczk-flmwmmu/

0! YM/ cat/CED 7364 770d /4/ «0.2/4.

107

 

 

 

 

 

 

Q
Q I I = I
$32} as ' ‘ ! ?
Q S , I ((0.550 sysrA’M»
8 \ 02 , fl/d/r - flue/fr 50/0252 ‘
55 g ; 60°C . ..IL
61/ .
Q 2
Q \' f I
e) L : “ - - 9 § 4
2 4 6 a /0 /2 x4 /6 /a
GRAPH 12. TIME [A] DAYS
COéedé/ou or CO p.052 M/ 010/6641
re 55: 5040770” or zap/2m (03
' 6205410 5Y5 rm

a 6 f . -- fl/éfi wade/fr cappae
2 60 °C ‘ . i

-l

(ac/05:9 caeeoagzp (by/m 2y

 

 

 

“A

 

2 4 6 6 /O A? 14 I6 /8 30
62.41944 15 77M: [11 DAYS
(@2805/01/ 07‘ cop/059 /// axyééxl F855
Ema/Z ACEWC 4.: ID 50407704/

108

 

   

 

 

 

 

 

(2:25.50 5V572‘M

23 H/(rH' PUE/WVCO/%?é’£ - .- - -. - _;- -- -- /h
Q 5005‘ , /0//- 6. 7 //
S 0.7 . . , 757,224. I, 3 _/4*

I . 1_/

\. .. , -/
g 0.6 - //../ 6
\ - l / -

db F/LM 3
f“
{I
3: a4 4
$1
Q 0.5 5
Q
Kl
« 0.2 '- 1 ' ‘- 2
9x a 1 ‘ 9656/9177! - /
\ A... I __4 . /
8 i V’/ ?—"——‘—_ ———————————— J__..I_::+

 

2 4 6 6 /0 /2 /4 /6 /<9 20
GRAPH 14a TIME IN DA Y5
Coeeas/o/u or cap/Gae M/ 5040770/1/ 604/ M/A/M/é
4&0 ppm. Mal/C03 Ado 42 ppm. F222! (:02

 

 

 

 

 

2463/0/2/4/6/620

GQADJA 14 b 77/45 IN DA Y6

(Degas/0v or (@4022? /A/ 502 «7704/ (DA/ffl/U/w
1100/30/27. A/a//64; 41/2 45 ,a/am. #3022? 6'03

 

(04/ch 7647/02/ /,¢/ 22/4.

a1 V6 270 (at/er/QA 7/01/ //(/ RAM

ca/ A5€ (042,4 @5292? {xv r. /c/2,)-2/"

(Ob/CW 7237477534/ //(/

d8

0. 7

db

35

0.4..

0.35 ..

0.2

0.]

 

 

 

 

 

 

 

 

(40550 sys/25M i" _._---7¢-_ v
Mrs/x - Fae/I‘V (ac/0549 é ; , /
60°C , P#-— 6.7 g #5 _J___.
L - - ...--,__ .v-_ .- V-A .; .../...._....#.H w 6
° 5
-4 4
5
2
1.7.5:... .éwm#£>\; 4 /
I / - PA”! ((10/7le 1\ ‘
s5 ' _ i ‘ \"
. h; 3 ' J 4 a; L A i - 4—3.
2 4 6 8 lo /2 /4 /6
GRAPH 15d TIME WDAYS

COBeOé/OA/ 0F cop/0:2 //U 5WI7fl/U (OA/fA/"

vV/zVo-‘fi 60,0/0/79. /Va' 06’, 000’ 35/30/77. 7,675! 6%

 

 

 

 

 

2 4 6 49 IO /2 /4 /6
GRAPH '5 b ZYME [ll DALE/5
Co&€09/m/ a; (aw/£19 //t/ 504 of/axz/ (Qt/[4%

vV/A/é 6930/22 fl/o’aéj 44/0 557,007. fez: (02

(ONCE/U 7:64 7704/ //1/ RR xv.

OIVJvE/t/

109

fly. /fl7c;'y

W02“- £17

cappge

(waft/2764774910 //(/ IQ/Pfly.

 

110

 

I i.

CA 05 5/) 57.5 72-74
fi/(rfi 'Fae/fr cap/959

 

 
 

6006: -~ ,0/7’ 72 jfifrof/M
/ I ‘ \ ‘ >
0.5 ,.,/ f . \T
. ,/ :
5"“ “ .....w/ A: F/AM
§ / \ - \ + 6
/., . ___5
‘ .4
[OXVétfl—F‘
ff _. $- 5
. , i
/ i i -_P.-.2
‘ 3 5 ! OAVT/OA/ :
« ' W i ,..- ..9——- -85 _____ ..0
j 4’ h”;— ' , z" 1 fleé'qP/zxrg’ I

 

/* i i
1:...»— --"*~—~i’ A"), i ; ‘.
2 4 6 c9 /0 /2 l4 /6
GQAQH 160! TIME IN DAYS
mews/0M 0F 60242059 //1/ 50407767” (04/ M/A/M/é
5040/0/79 ”GA/€03 50/729 20/90/99 7855' 602

 

I
L

L.-.

 

 

 

4L

2 4 6 6 /0 /2 /4 /6
GQAD-H 16b TIME [1V DAYS
(0,6905/04/ at“ (ac/350 /4/ 5544/7741/ (04/254W/4/&
500,45”; A/aA/(Qg w/r/I 20/2/00). F-Té’é’é’ (63,3

 

//t/ 20/9/‘4

OXYGé'Z/ C (JA/C EA/ 77847” /&A/

 

 

 

 

 

 

 

 

 

 

 

 

 

 

F——'_"‘"' ' *
620.560 57.57274 3
HIGH- Pz/e If)” (op/0.52 707:4 z.
’W\ _ 60°C _ pfifi 7'] . _ l _ ' -__4_.._.+
9 a5 ‘ , — «M - ”J”—
C 2’" ‘
\ / w I
Q ' - 1
& wrap/L; ; "1.5
\J 1 .’_______.. -:_ . 1 . . . ...—.....— . ...—-... . .4}
Q = 29"" “’fi i - f
. ' ‘ ‘ 3 —-v-‘5
31
g
Q) -0»? Yéé/V "r __w__4
U ; ‘ “r45
<0 5041/7767”: . _
.N _______,____ _;_ —-—qr-~2
\ ._*__idn :
K i
g [“Pefd/P/ffl 72" ______ 1
2 4 6 a /0 /2 /4 /6 /¢9
GEARU 170: TIME W OAKS
(63660.5 /0/U a; 60,67,653 //t/ .554 arm/v C(Uf4/fl/(Vé
75 ppm. A/o’ flK/ ”077/ 14 ppm. mega (02
*\ 5 x I , , §
S
Q: 2
S =
g 1 s z '
l .
K 1
§ ; 4 - 1 ~ A
‘3 2 4 6 a /o /z /4 /6 /5
§ 612APH 17 b T/Mé’ W @4145

(086050” a; cap/«>56 M/ 5&4 af/flx/ CaA/r/I/xv/z/é
75 lop/)9. 4/0694 W/l'fl /4 ,0/5/5'9. Fjéé'é' (Ca

111

OZYGAT/t/ CfiA/C’éf/U7A/47/fl/t/ /A/ 924%

(0,405.8 608809.52? (my. / 0772/

Cort/65.0 7.64 T/OA/ Ml RAM

Q
m

Q
AN

.Q
\

 

640.551.? .5 Y57'E/V
CWMAIEC/AL cap/0&8
60 O C I p f/ ' z 2

O
I
l
.

 

 

 

 

 

GRAPH 16 a

a /o /2 /4 /6
TIME IN DA Y6

60.6605/04/ or 600.0538 /A/ .504 af/OA/ COMfA/A/M/é
JOOppm. Mai/€03 , WNW 20,49,007. F7665 (02

 

 

 

 

 

 

2 4 6
GQADLJ. 15b

6 /0 /2 M /6
TIME. [N M Y6

coezas/ac/ or (0.0.0.58 //(/ 504a 7700 (er/A/M/é
JOO pp; 27. flof/Caj , W/ 77/ 20 ppm. 7724575 602

112

OZYéE/t/ COflCQf/T’é‘AT/QA/ //0 293/'7.

113

GQAPH 19 o. (0,820.5/0A/ 0r cap/case //t/ #470,941 540

 

0.4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

52232-5 bysnsM ; a i/
”a? // 6H Fae/7y Japafz ! ' ‘ 3 /-
S @3116 _- 1 1 I: #707341, -'———/ Z ‘
\ ; ; !’/*;__)_‘___:: ... j ‘
.0’3.._._.-". --.—.4. v/ I .___L_.~_ ..-—--.“--- i -1 ....___+__‘ ...
g‘ 5 ; i L
I i. ‘ 1 . i i '
8 3 , i ~*'L""*.""‘J"¥~-~;. ‘ !
No.2» —-~-§ ff +  ————-—-—~- 5— — -——+.—— ~ :_;.\_..___,.,_:+_;__,__$
w i ‘ . a i ‘ ‘ '
Q 1‘
U .
I .
Q .
8‘3
(Lax
C)
K)
/ E ‘ . 1. l # g
0.0 ‘ *‘fi—‘f ‘ ‘ i' ‘ L ' 1 i 1 L l .
0 4 8 /2 /6 20 24 25 58 56 40 44
77MB [Al DAYS
2% 612A P44 19 b (ye/2 05/0” or cap/95.2 [4/ Mari/2.4L H30
Q4 7 E #2....”
a 1 /:
§5’ ' / F .
s. V a w .
e / + .-.—wi- m.-
g I 7. // * ~: ¢ I z
‘3 ' g 1 I
so 1 ‘ ' A A - -. A L _; L
no 4 8 /2 /6 .20 .24 28 52 56 4o 44
TIME Ml DAMS"

 

 

 

650551" .5 Y5 72.74
”/6/9 - Fae/WV 50/19/056

 

 

    

 

 

 

GQADH 20 d

R\ 600C

§ 0.5 , .1 .1, .

g x ,Jfl ...», “.....-

\S I _____._- M"

Q 0,2 I A . . *F/LM

//’~---- - / *

Q I . +—— __.___.__._____, ____ _“_. ~45

$2 /

\3 I/

Q2 a1 //

§ '/

a / _, ’ —-O~ .- * .. E [’;S0£U7/0”
,/ ,.. ~~~~~~~ ———————
‘ , . ’ I ; ,2, peECAO/fflfé’
W M . e ' _

4 5 /2 /6 20 24 26’ :52 .56 40 44

T/I‘ZE {Al DAYé
(aéuazsmMU'aW‘ccwuz5€’Aw’44WQA94L u%¢ZEZ?

 

 

 

 

 

Q:

R;
$5
$21
k

g 1.
‘6

§

(112.4914 20 b

/6 20 24 26 52 .36
TIME lN DA Y5
(oeeos/at/ or (ac/>532 ax ,t/A 70244 WA 752

4 a /2 40 44

 

115

 

GQAPH 2|
FZQW 55/5727! 60.5206/0/V 225575
0 70.0 ' A/
.5 égp/ pg Urgg (OQéOé/OA/ (”79/5/27 2)

_ -_ 4
I
l
l

H/6// flaw/V Wflffie 6 5 ii 7 F

9
I

400,0/303. A/aflw; , 6. 9 _ 4.8

gap/om. F1355 6&3

60,0,0/27. [0.2 f/flflppfl). 7 0 6. 0
1/0 UA’

/4ppo7. 22255 (42

i
i
” ’ I
I
l

Mppw. M #64:;

M0 ppm. 4/0 C/

44 - .-. ....a .. .- ... -o~:-

. +

8.5 5.5 I

 

 

.g

 

—‘ I

cop/9:76 Mow/x? 0. 28 ppm,

 

 

 

 

 

 

 

 

_ (Of/Tile” //(/ 1574M

:22]

 

(av/556 /4/ 5040/70”

- i j C (JP/751’ I/cxop a zap/m.
: : .
4} ~ ;
; cap/Dre papa/Q 04/ ppm.
5 (010/956 P/dr m0 4/, /7pp/y.

 

 

Tabulation of Data

116

The data presented in the graphs are tabulated on

pages 116 to 126.

In the heading of each table the con-

ditions under which the data was obtained are summarized.

These tables list all the data obtained in this investigation

and also describe the test conditions under which the data

was acquired.

ponding graphs in which the data are plotted.

Table 1.

A11 tables are numbered the same as corres-

Blts of

open—system static corrosion tests at 60 C with

 

 

Four
§pecimens

 

 

genes;
oncen-

tration

(22m)

0.u1
0:h3
0-h5
0:33
0.hh

Data presented in Graphs la and lb.
high-purity cepper and highly pure water.
Single Cleaned Specimen
Copper Co er Total
Concen- En Corro-
tratIon Solution sIon
(22m) (mgzcmz) (mg/cmzl
- 0.0%2
- - 0.0 k
- - 0.09
0.15 0.003 0.13h
0.20 0.00h 0.150
0.20 0.00 -
0.26 0.00 0.20
0.25 0.006 0.23
0.38 0.009 0.309
0.33 0.007 0.377
0.25 0.006 0.038

Conditioned

gpecimen

Co or
Concen-

tration

 

117

Table 2. Data presented in Graphs 2a and 2b. Results of
Open-system static corrosion tests at 60°C with
commercial grade c0pper and highly pure water.

 

 

 

Cleaned Specimen Conditioned Specimen
Copper 00 er Total
Concen- En Corro- Co er
Tine tration Solution sion ConcenEration
$29.18.). 121-22). Lusaka-12.1 Les/.2129. 3.22311
1 0.51 0.011 0.18u o.%s
2 - - O O. O
3 0.88 0.020 0.2 7 -
6 0.6a 0.01h 0.1 S 0.55
10 - - - 0-h0
12 0.h5 0.010 0.582 -
12+ - - - 0.32

Table 3. Data presented in graph 3. Results of open-system
static corrosion tests at 60°C with high-purity
and commercial grade copper in slightly alkaline

 

 

 

 

water.
Commercial Copper High Puritz Copper
Copper Co or
Time Concentration Concentration
(dais) (ppm! Sppm)
1 0.13 0.10
3 0.h9 . 0.h0
7 0.32 0.37
1h .29 0.35

118

Table A. Data presented in Graph h. Results of closed-system
static corrosion test with high-purity copper and
oxygen-free water at 60°C.

 

 

Time Total Corrosion
da s {EEZcm2)
13 0.006

Table 5. Data presented in Graph 5. Results of closed-system
static corrosion tests at 60°C with high-purity
copper and.water containing 1.h8 ppm dissolved

 

 

 

oxygen 0
Total Ox en
Time Corrosion Concen ration
Sdazsz (m. on2 (ppm!

:%
.2

.0

ONO)

WHO
0
0
3’1
;:;r
CDOCDIv-J

 

Table 6. Data presented in Graphs 6a and 6b.

119

Results of

open-system static corrosion tests at 60°C with
high-purity copper and water containing hOO ppm

sodium‘bicarbonate.

Table 7. Data presented in Graphs 7a and 7b.

Copper
Conca1tration

E

COO 0000
PM WWI-J
SUN OJON

er in

Copp
gngcmzz

So u 135

0.00
0.00

0.009
0.009

0.005
0.003
0.002

Total

Corros

SEEK 01122

Results of

open-system static corrosion tests at 60°C with
high-purity copper and water containing 50 ppm
sodium carbonate.

Time

Co r
ConcenEration

O 12

.13
13

O

O
0

OOO
OOH
O‘U'lU'l

12 '

Copper 13
(mg/cm?)

0 u ion

0.003
0.003
0.003
0.003

0.001
0.002
0.002

Total

Corrosion

1.254922;

0.095
0.137
0.180
0.291

0.317
0.298
00292

on

120

Table 8. Data presented in Graphs 8a and 8b. Results of
open-system static corrosion tests at 60°C with
high-purity copper and water containing L0 ppm
sodium sulfate .

 

 

 

 

Copper Co or in Total
Time Concentration SoiutiEB Corrosion
(ppm) (Echmz) (mg/cnzl
3 hours 0.06 0.001 0.20
" 0. 0.010 0.31
12 " 0.3 0.009 0.291
at " 0.33 0.007 0.317
7 days 0.38 0.009 0.Lh
1L " 0.23 0.005 0.51E

Table 9. Data presented in Graphs 9a and 9b. Results of
cpen-system static corrosion tests at 60°C with
high-purity c0pper and water containing 10 ppm
sodium chloride.

 

 

 

 

Coppgr Co or in Total
Time Concentration SoiutiEfi Corrosion
(22m) ( cm? (mg/cm?)
3 hours 0.07 0.001 0.2h8
6 " 0.2h 0.005 0.372
2h " 0.55 0.012 0.3u7
2 days 0.31 0.007 0.321
3 " 0.%3 0.010 0.391
7 " 0. o 0.013 0.320
10 " 0.27 0.006 0.38u

121

 

 

 

Table 10. Data presented in Graphs 10a and 10b. Results of
open-system static corrosion tests at 60°C with
high-purity copper and water containing 100 ppm
sodium chloride .

COpper Co or ip Total
Time Concentration So u ion Corrosipp
12220 M 12312231

6 hours 0.20 0.005 0.230

at " 0.63 0.015 0.5 6

2 days 0.3 0.008 0.702

7 " 0.98 0.023 0.676

8 " 0.27 0.006 0.600

1L " 0.63 0.01h 0.998
Table 11. Data presented in Graphs 11a and 11b. Results of

closed-system.static corrosion tests at 60°C with
high-purity copper and water containing 70 ppm
dissolved carbon dioxide.

 

 

 

Co or 00 er Co or COpper Total 0x en
Concen- ‘25 Precipi- n Corros- Concen-
Time EFEEIER Solution tated FITm. $23 EFEEIoh
$22111) Smgz cm2) 0112/ cm?) (mg/cm?!) (25161322 (ppm)
0 hours - - - - 5.h
6 " 2. 0.050 - 0.218 0.273 -
12 " 3. 0.083 - 0.288 0.331 -
28 " 8.0 0.183 0.001 0.20u 0.388 -
3 da s 11. 0.267 0.002 0.3 0 0.5 0.96
X 12. 0.288 0.008 0.3 0.62% 0.53
13 " 1 .0 0-29h 0.005 0.z68 0.667 0.00
18 " 3.5 0.102 0.003 0. 50 0.755 0.05

 

 

122

Table 12. Data presented in Graph 12. Results of closed-system
static corrosion tests at 60°C with high-purity

c0pper in oxygen-free water containing 20 ppm
dissolved carbon dioxide.

 

 

 

Total
Time 005F3§ion
(daze) Lma/ c1212)
8 0.013
11; 0.011

Table 13. Data presented in Graph 13. Results of closed-system
static corrosion tests at 60°C with high-purity

copper in oxygen-free water containing 5 ml. per
liter of acetic acid.

 

 

 

Total
Time Qprrosion
(dais) (mg/c1112)
0.18
1?; 0.1169
21 0 . 53

Table 10.

Time
0 hours
H

12 N
2,." N

3 days
N

7
10 fl

15 n
22 H

Tablo 15 e

123

 

 

 

 

 

 

 

 

Data presented in Graphs 10a and 10b. Results of
closed-system static corrosion tests with high-
purity copper in water containing 000 ppm sodium
bicarbonate and 08 ppm free carbon dioxide having
a pH of 6.7.
Copper Copper COpper Copper Total Oxygen
Concen- n Precipi- in Corros- Concen-
tration Solfition tated Riim _pp tration
(ppm! (mggcmzz (mglcmzz ng/cmg) Sgchmz) Sppmz
- - - - - 6.0
0.88 0.020 - 0.102 0.162 -
1.08 0.020 - 0.105 0.169 -
1.33 0.030 - 0.1 1 0.191 -
2.56 0.0 8 " 0.1 0 0.188 39E2
2.10 0.0 8 0.000 0.3 7 0.019 2. o
1.90 0.003 0.050 0.338 0.031 2.10
1.27 0.029 0.035 0.530 0.590 0.6
1.38 0.031 0.037 0.717 0.785 0.0
Data presented in Graphs 15a and 15b. Results of
closed-system static corrosion tests with high-
purity copper in water containing 60 ppm sodium
hydroxide and 35 ppm free carbon dioxide having a
pH of 6.7.
C0pper Copper Copper Copper Total Oxygen
Concen- n Prec p - n Corros- Concen-
tration Solfition tated 0110 on tration
(ppm: (mg/0mg) (mg/0mg) (mglcmzl Lag/9mg) (ppm:
- - - - - 6.8
1.12 0.025 - 0.197 0.222 "'
3.60 0.082 - 0.215 0.297 -
7.60 0.172 0.005 0.222 0.399 3.9
10.90 0.206 0.006 0.268 0.520 -
. 2.00 0.005 0.003 0.767 0.815 0.5

 

 

 

120

Table 16. Data presented in Graphs 16a and 16b. Results of
closed-system static corrosion tests with high,
purity copper in water containing 300 ppm sodium
bicarbonate and 20 ppm free carbon dioxide having
a pH of 7.2.

Copper 00pper Co or Co or Total Oxygen
Concen- n Prec p - in Corros- oncen—
Time Eration Solfition tated Fiim ion tration
hays) (ppm! . (mglcmzz (mg/cm?) (ngcm?! (pglcmzz (ppm!
0 "’ "’ " " "' 14.06“-
2 1.22 0.028 - 0.233 0.261 3.52
5 1.68 0.038 0.005 0.222 0.265 3.20
10 2.60 0.059 0.021 0.260 0.300 3.20
18 2.38 0.050 0.028 0.225 0.337 3.10
Table 17. Data presented in Graphs 17a and 17b. Results of

closed-system static corrosion tests with.high-
purity c0pper in water containing 90 ppm sodium.
hydroxide and 10 ppm.free carbon dioxide having a
pH of 7.1.

Go r
Concen-

Time tration

(days)

0 -
f; :-

9 3320
20 3.90

68‘
2

00pper

Solfiiion —tated

(ppm) (pgécmzz (pglcmzz (among) (ngcmzz (ppm!

0:037
0.071
0.071

0.080

 

0.009
0.012

0.009

pgfifiiii- 'EQEEEE

0110

0:208
0.217
0.226

0.228

Total

Corros-

on

0:205
0.297
0.309

0.317

Oxygen

Concen-
tration

5.2
3.80
3.12
3.72
3.70

125

 

 

 

 

 

 

 

 

Table 18. Data presented in Graphs 18a and 18b. Results of
closed-system static corrosion tests with commer-
cial grade copper in water containing 300 ppm
sodium bicarbonate and 20 ppm free carbon dioxide
having a pH of 7.2.

COpper Copper COppgr er Total Ogygen
Concen- in Precip - tin: Corros- Concen-
Time tration Solfition "tated ion tration
(days) (ppm! (mg/cmzj (mg/cle (mgjic;z (gchmZ ) (ppm!
0 - - - - ' - 0. 61.
2 1. 21 0. 027 - 0.190 0.217 .80
5 1.78 O. 000 0.012 0.201 0.25 38
10 1.80 0. 002 0.015 0.197 0.25 3.00
18 2.20 0.051 0.023 0.207 0.281 3.10

Table 19. Data presented in Graphs 19a and 19b. Results of
closed-system static corrosion tests at 30°C with
high-purity copper and natural water.

Copppr Copper Copper :ppe r Total
Concen- n Prec p - Corros-
Time tration Solfition tated ion
(days) (ppm) (mggcmz) (m c1112) ”(Z/0:21 (ma/cw?)
7 .60 0.082 0.002 0.212 0.296
19 .00 .0 0 0.007 0.225 0.320
5 2.65 0.0 O 0.067 0.200 0.327

126

Table 20. Data presented in Graphs 20a and 20b. Resglts of

closed-system static corrosion tests at 60 C with
high-purity cOpper and natural water.
Copper Total

Concen- 991293 £35111 5”. 29133 cm-

n
Time tration Solfiiiog a e Fiim ion

 

 

 

 

 

 

__I._(da a) 1229). 01512221 125401-31 12312911122 (mg/ma)
7 1.90 0.002 0.005 0.190 0.201
19 2.50 0.0E6 0.007 0.180 0.203
S 1080 000 1 - C .-
06 2.00 0.005 0.058 0.166 0.269
Table 21. Data presented in Graph 21. Resu ts of three-day
flow system corrosion tests at 60 C with various
solutions.
Copper Copper Total
n in Corros-
FITm Soififiog on
Solutipn (mg/99%; (mglcmzz (mg/0mg)
High Purity Water 0.308 0.166 0.510
000 ppm NaHCOB,
20 ppm free co2 0-132 0-167 0.299
d 100 m NaOH
Ea fig: ggéeago pp , 0.173 0.222 0.395
2
000 ppm NaHCO3 0.139 0.112 0.251
100 ppm NaCl 0.655 0.113 0.768

127
LITERATURE CITED

1. American Public Health Association, "Standard Methods
for the Examination of Water, Sewage and Industrial
Wastes," 10th Edition, New York, 1955.

2. Babbitt, H. E., and J. J. Doland, "Water Supply Engineer-
ing," McGraw-Hill Book Co., New York, 1955, p. 377.

3. Bailar, J. C., "The Chemistry of Coordination Compounds,"
Reinhold Publishing Corp., New York, 1956, p. 11,

0. Bengough, G. D., and O. F. Hudson, J. Inst. Metals, 21,
37-252 (1919); specifically p- 00-

5. 0p. cit., p. 110.

6. Bengough, 0. D., R. M. Jones, and R. Pirret, J. Inst.
Metals, 22, 65-158 (1920); specifically p. 97.

7. Brown, R. R., B. E. Roetheli and H. O. Forrest, Ind. Eng.
0116111.. 2 p 350-2 (1951).

8. Campbell, H. 3., J. Inst. Metals, 11, 305-56 (1950).
9. Campbell, H. 3., J. Appl. Chem., London, 0, 633-07 (1950).
10. Dunn, J. 3.. Proc. Roy. Soc. London A., 111, 210-19 (1920).

ll. Eliassen, R. E., and J. 0. Lamb, J. Am. Water Works
Assoc., 05, 1281-90 (1953).

12. Evans, U. R., "Metallic Corrosion Passivity and Protection,"
Edward Arnold and 00., London, 1937, p. 70.

)3. GP. 6115., pp. 95-80
10. 0p. cit., p. 232.
15. Op. cit., p. 209.

16. Glasstone, 3., "An Introduction to Electrochemist ," 0.
Van Nostrand Co., Inc., New York, 1902, p. lr0.

17. Glasstone, 8., "Elements of Physical Chemistr ," D. Van
Nostrand 00., Inc., New York, 1906, p. 3.

18. Hall Laboratories, Inc., "Water Booklet," 2nd Edition,
Pittsburg.

19.

20.
21.
22.

23.

25.
26.

27.
28.

29.
30.
31.
' 32.
33.
30.
35.

36.

128

Hedges, E. 8., "Protective Films on Metals," D. Van
Nostrand Co., Inc., New York, 1937, pp. 62-5.

0p. Cito’ pp. 173-90
Hill, G. R., J. Electrochem. Soc.,‘lgg, 305-50 (1953).

Hill, C. R., and J. R. Weeks, J. Electrochem. Soc., 103,
203-8 (1956).

Hogness, T. R., and W. C. Johnson, "Quantitative Analysis
and Chemical Equilibrium," Henry Holt and Co., New
York, 1907, p. 500.

Jost, W., "Diffusion in Solids, Liquids Gases," Academic
Press, Inc., New York, 1952, pp. 3 0-1.

Op. cit., pp. 301-3.
Lustman, B., Trans. Electrochem. Soc.,‘Ql, 359-75 (1902).
May, R., J. Inst. Metals, Qg, 65-70 (1953).

McAulay, A. L., and E. C. R. Spooner, Proc. Roy. Soc.
London A, 138, 090-501 (1932).

McDowell, L. A. and H. L. Johnston, J. Am. Chem. Soc.,
58. 2009-10 (1936).

Moeller, T., "Inorganic Chemistry," John Wiley and Sons,
New York, 1952, p. 286.

Paurbaix, M. J. N., "Thermodynamics of Dilute Aqueous
Solutions," Edward Arnold and Co., London, 1909,
p. 12.

Pickering, 3. V., J. Chem. Soc., 25, 1009-29 (1909).

Powell, S. T., Water and Sewage Works, 9;, No. 3, 93-7
(1906).

Reed, J. H., "Theoretical Qualitative Analysis," McGraw-
Hill Book Co., Inc., 1938, p. 112.

Riely, H. L., Proc. Roy. Soc. London A, 103, 399-010
(1930).

Smart, J. 3., Jr., A. A. Smith, and A. J. Phillips,
"Preparation and Some Preperties of High-Purity
Copper," Am. Inst. Mining Met. Eng., Tech. Publ.

No. 1289; 1901.

129
37. Tronstad L., and R. Veimo, J. Inst. Metals,‘é§, 17-32
(1900)

38. Wagner, 0., American Society for Metals, Cleveland,
"Atomic Movements," 153-73; 1951.

 

 

f1

 

vb.

.3 3::

Y

01000