Optical studies of dezincification in alpha-brass

MISSING IMAGE

Material Information

Title:
Optical studies of dezincification in alpha-brass
Physical Description:
x, 111 leaves : ill. ; 28 cm.
Language:
English
Creator:
Finnegan, John Edmund, 1946-
Publication Date:

Subjects

Subjects / Keywords:
Copper-zinc alloys -- Corrosion   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Includes bibliographical references (leaves 103-110).
Statement of Responsibility:
by John E. Finnegan.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000097889
notis - AAL3330
oclc - 06630634
System ID:
AA00003874:00001

Full Text
















OPTICAL STUDIES OF
DEZINCIFICATION IN ALPHA-BRASS










BY

JOHN E. FINNEGAN


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY









UNIVERSITY OF FLORIDA


1980















ACKNOWLEDGEMENTS


The author wishes to express his sincere admiration

and appreciation to his advisor, Dr. R. E. Hummel, for

his unselfish devotion to his students and research.

Special thanks are also extended to Dr. E. D. Verink,

Jr., Dr. J. Ambrose, and Dr. G. Schmid for their advice

concerning the electrochemical aspects of this research.

The National Science Foundation provided financial

support for this research, which is gratefully acknow-

ledged.
















TABLE OF CONTENTS


Page

ACKNOWLEDGEMENTS........... ............................. ii

LIST OF TABLES.......................................... v

LIST OF FIGURES................................................ vi

ABSTRACT................................................. ix

CHAPTER
I INTRODUCTION...... ............................. 1

II LITERATURE REVIEW................................ 7
II-1 Foreword .............................. 7
II-2 Mechanisms ............................ 8
II-3 Techniques ............................ 15
II-4 Oxides................................. 25
II-5 Summary ............................... 30

III EXPERIMENTAL PROCEDURE ......................... 32
III-1 Introduction.......................... 32
III-2 Differential Reflectometer........... 32
III-3 Specimen Preparation................. 35
III-4 Solution Preparation................. 40
III-5 Electrochemical Equipment and
Operating Conditions................ 42

IV EXPERIMENTAL RESULTS.......................... 47
IV-1 Preface................................ 47
IV-2 Yellow Brass Dezincification.......... 50
IV-3 Red Brass Dezincification............. 65
IV-4 Discussion............................. 92

V CONCLUSIONS.................................... 98

VI SUGGESTIONS FOR FURTHER WORK.................... 100
VI-1 "Satellite" Structure.................. 100
VI-2 Dezincification Limit................. 100
VI-3 Chloride Solutions.................... 101
VI-4 Potential-pH Diagrams................. 101
VI-5 Kinetics............................... 101













TABLE OF CONTENTS continued.



Page

REFERENCES...... ........................................ 103

BIOGRAPHICAL SKETCH.................................... ... 111















LIST OF TALES


Table Page

3.1 Compositions of Copper-Zinc Alloys Used.......... 38

3.2 Constituents of Electrolyte Used in
Chloride-Free Experiments....................... 41

4.1 Peak Positions for a Copper 20.8 at.% Zinc
Alloy (Average Composition) and a Dezincified
Yellow Brass Alloy (Starting Composition,
21.4 at.% Zn).................................... 62

4.2 Peak Positions for a Copper 6.8 at.% Zinc
Alloy (Average Composition) and a Dezincified
Red Brass Alloy (Starting Composition,
9.2 at.% Zn)..................................... 66

4.3 Calculated Dezincification Zone Thickness....... 97















LIST OF FIGURES


Figure Page

1.1 Trends of dezincification, stress corrosion
cracking, and impingement attack with in-
creasing zinc content in copper-zinc alloys.... 4

2.1 Potential-pH diagram for Cu-30 Zn alloy in
0.1 M chloride solution, at 250C............... 13

2.2 (111) peaks of a mixture of Cu 70-Zn 30
brass alloy and copper fillings. Brass
peak is to the left........... ................. 18

2.3 (111) peaks of a dezincified brass alloy.
A range of brass alloy compositions illus-
trated by broad peak on left.................... 19

2.4 Lattice parameter of alpha-brass as a
function of atomic percent zinc................ 20

2.5 Copper-zinc phase diagram...................... 21

3.1 Schematic of differential reflectometer........ 34

3.2 Photomicrograph of specimen grain size.......... 37

3.3 Schematic of electrochemical cell.............. 43

3.4 Schematic of potentiostat used in dezincifi-
cation experiments.............................. 45

3.5 Block diagram of electrochemical system......... 46

4.1 Potential-pH equilibrium diagram for the
system copper-water at 250C (ionic species
a = 10-6)....................................... 48

4.2 Potential-pH equilibrium diagram for the
system zinc-water at 250C (ionic species
a = 10-6) ..................................... 49












LIST OF FIGURES continued.


Figure Page

4.3 Differential reflectogram of Cu-22.3 Zn/
Cu-19.3 Zn (yellow brass) ........................ 52

4.4 Differential reflectogram of Cu-22 Zn at
1 hour exposure in pH %4 and ESHE = -100 mv.... 54

4.5 Montage of Cu-22 Zn, yellow brass, dezinci-
fication reflectograms......................... 55

4.6 Position of peak "A" versus average zinc
content in copper-zinc alloys.................. 57

4.7 Schematic of zinc concentration gradient in
dezincification zone..................... ... .. 60

4.8 Differential reflectogram of cupric oxide
(CuO).......................................... 64

4.9 Differential reflectogram of Cu-9.2 Zn/
Cu-4.4 Zn (red brass) ........ ...... ......... 68

4.10 Differential reflectogram of Cu-10 Zn at
1 hour exposure in pH %4 and ESHE = -100 mv.... 70

4.11 Montage of Cu-6.8 Zn red brass dezincifi-
cation reflectograms........................... 73

4.12 Differential reflectograms of oxygen con-
centration effects on dezincification be-
havior of red brass, 48 hours exposure
(curve A air saturation, B >300 ppm oxygen,
and C <10 ppm oxygen).............. ............. 75

4.13 Differential reflectogram of Cu-10 Zn at
48 hours exposure with %300 ppm oxygen
contamination ................................ 77

4.14 Differential reflectogram of cuprous oxide
(Cu20)........................................... 79

4.15 Differential reflectograms of oxygen concen-
tration effect on dezincification behavior
of red brass, at 72 hours exposure............. 81


vii












LIST OF FIGURES continued.



Figure Paqe

4.16 Differentail reflectogram of Cu-5 Zn at 10
hours exposure in pH %4 (KHP) solution.
Curve (a) exhibits dezincification, curve
(b) no dezincification............. ............ 84

4.17 Differential reflectogram of cupric
hydroxide Cu(OH)2.. .................. .. 88

4.18 Potential-pH equilibrium diagram for the
system copper-water at 250C, considering
the solid species Cu, Cu20, and Cu(OH)2
(ionic species,a = 10-6) ....................... 90

4.19 Experimental potential-pH diagram for the
system 99.99% pure copper-water, at 250C
(ionic species a = 10-6)..... .................. 91

4.20 Bipolarity of precipitate films favorable
and unfavorable to passivation................. 94


viii












Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment
of the Requirements for the Degree of Doctor of Philosophy



OPTICAL STUDIES OF
DEZINCIFICATION OF ALPHA-BRASS

By

John E. Finnegan
March 1980

Chairman: R. E. Hummel
Major Department: Materials Science and Engineering

Experimental studies were undertaken to investigate

the dezincification behavior of various copper-zinc

alloys. This investigation was conducted in order to

confirm or refute reports in the literature, that red

brasses (copper alloys having a zinc concentration of

less than 15%) are "immune" to dezincification. Dif-

ferential reflectometry was applied in conjunction with

electrochemical methods as a means of identifying and

characterizing possible dealloying of alpha-brasses.

It was established that both red and yellow brasses

undergo dezincification. Additionally, the dealloying

behavior of red and yellow brasses exhibited several

strong similarities; the intensity of the dezincification

spectrum increased with increasing time, the average alloy

composition was not significantly altered with increasing












potentiostating times, and the calculated surface loss

of zinc, as a result of selective leaching, was approxi-

mately 4 at.% for each.

The apparent solid state transport mechanism of

zinc atoms to the solid/solution interface is enhanced

volume diffusion.















CHAPTER I
INTRODUCTION


Alloys of the copper-zinc system exhibit higher

hardness, improved physical properties, and increased

resistance to impingement attack than unalloyed copper.

The extensive use of brasses as condenser tubes, archi-

tectural grillwork, hardware, coinage, plumbing fixtures,

and munitions attests to its good machinability and

forming behavior as well as its cost effectiveness. Un-

fortunately, copper-zinc alloys also have a serious

weakness. Under certain conditions, brass suffers a loss

of zinc, a process that is generally called dealloying or,

specifically, dezincification. Fontana and Greene de-

scribe dezincification as a form of selective leaching

which is defined as "the removal of one element from a

solid alloy by corrosion process" (Ref. 1, p. 67). De-

zincification as defined by Uhlig is "a type of attack

occurring with copper-zinc alloys (e.g. yellow brass in

which zinc corrodes preferentially, leaving a porous

residue of copper and corrosion products. The alloy so

corroded often retains its original shape, and may appear

undamaged except for surface tarnish, but its tensile













strength and especially ductility are seriously reduced"

(Ref. 2, p. 15). The occurrence of dezincification in

copper-zinc alloys is enhanced when brass is exposed to

stagnant acid solutions, high temperatures, and solutions

with high chloride concentrations. There are two proposed

mechanisms for the dezincification behavior of alpha-

brasses. One school suggests that the copper-zinc

alloy first dissolves stoichiometrically, increasing the

concentration of copper and zinc ions in solution. The

copper is then related on the surface, forming a porous

structure. Subsequent dezincification occurs because of

the porous nature of the redeposited copper, permitting

contact between the solution and the brass [2]. The

alternate theory is that the zinc dissolves preferential-

ly, leaving a porous residue of copper. Further dezinci-

fication occurs by volume diffusion of zinc to the alloy/

solution interface [2]. Heidersbach [3] published a

comprehensive literature survey of the dezincification

phenomenon in 1968 which describes the numerous and

varied experimental techniques employed to determine the

controlling mechanismss. Heidersbach and Verink [4]

ultimately clarified the controversy of the controlling

mechanism of dezincification, selective leaching versus

dissolution/replating, by defining the conditions of

potential and pH under which each mechanism operates.












It has been widely reported [1,21 that red brasses

(copper containing less than 15 at.% Zn) are "immune" to

corrosive attack by a dezincification reaction. Uhlig

[2] has characterized the performance of copper-zinc alloys

with respect to their tendency for dezincification, stress

corrosion cracking, and impingement attack as a function of

increasing zinc content (see Fig. 1.1). Fontana and Greene

[1] describe red brass as usually immune to attack by de-

zincification. Brasses in the composition range of 0-15%

zinc do not exhibit the gross features of dezincification

displayed by the much-studied alpha-yellow brasses and

are, therefore, characterized as "immune." Additionally,

previous studies of this phenomenon using conventional

techniques (x-ray diffraction, electron microprobe

analysis, electron diffraction, etc.) have required

relatively long incubation times (>24 hours) to produce

sufficient dealloying for detection.

The purpose of this study is to determine if dezinci-

fication occurs in copper base alloys containing 15 at.%

or less zinc. Because of the inherent sensitivity of the

differential reflectometer, the early stages of dezinci-

fication can be investigated and compositional changes

of less than 0.1 at.% can be detected. Through the use

of this relatively new surface analysis technique, it is




























High

Rate of -
dezincification


30 -


15K-


Immune


High

Sensitivity -
to stress
corrosion
cracking


Low
Immune


Low

Rate of--
impingement
attack

















High


Figure 1.1. Trends of dezincification, stress corrosion
cracking, and impingement attack with in-
creasing zinc content in copper-zinc alloys
(from Ref. 2).


40-













also possible to gain further insight into the kinetics and

sequential behavior, in real time, of the dezincification

attack on brass. Nastasi-Andrews and Hummel [5] have

shown that differential reflectograms (AR/R versus wave-

length of light) of copper-zinc alloys show distinct

peaks which can be ascribed to certain electron transi-

tions. These peaks shift essentially linearly with in-

creasing zinc content so that the surface composition of a

copper-zinc alloy can be determined from these peak

positions. Shanley et al.'s [6] work on the corrosion

of copper identified unique differential reflectograms for

each of the stable copper oxides, CuO and Cu20. These

observations provide the basis for the qualitative and

quantitative interpretation of differential reflectograms

obtained from dealloyed brass specimens.

In this investigation, alpha-brass was potentio-

stated under dealloying conditions. Subsequently, the

differential reflectivity (AR/R) between this specimen

and a specimen having the same original composition (but

in the unexposed state) was measured. The differential

reflectograms thus obtained clearly showed that brasses

containing 22, 10, 7, and 5 at.% Zn undergo dezincifica-

tion. Further analysis of the experimental results re-

vealed the presence of cuprous hydroxide on dealloyed












yellow brass samples. This hydroxide was not observed

on red brass specimens tested under identical conditions.

The following chapter summarizes existing literature

on dezincification in alpha-brasses. Chapter III presents

experimental procedures, details of preparation of samples

and solutions, and methods employed to control the electro-

chemical reactions as well as a description of dif-

ferential reflectometry. In Chapter IV, the experimental

results are described and discussed. Chapter V presents

the conclusions. Chapter VI enumerates suggestions for

further research.















CHAPTER II
LITERATURE REVIEW


II-1 Foreword

Dealloying, which has also been called "selective

leaching" and "parting," can be defined as a corrosion

process whereby one constituent of an alloy is pre-

ferentially removed from the alloy, leaving an altered

residual structure [1,2]. The most common example of

dealloying occurs in the copper-zinc system where the

selective removal of zinc from brass alloys is known as

dezincification.

Dealloying is not, however, restricted to the copper-

zinc system. Other binary copper-base alloys experience

dealloying phenomena (nickel [7-9], aluminum [10-16], and

tin [17]). Sometimes selective corrosion of one element

from an alloy may be beneficial. Enrichment of silicon

observed in the oxide film on stainless steels results

in better passivity and resistance to pitting. In

general, this is an exception rather than the rule and

dealloying normally is considered undesirable in an alloy

system.

Dezincification manifests itself in two general

classifications: (a) uniform or layer type and













(b) localized or plug type. According to Fontana and

Greene, "Uniform, or layer type, dezincification seems

to favor the high brasses (high zinc content) and de-

finitely acid environments. The plug types seem to occur

more often in the low brasses (lower zinc content) and

neutral, alkaline, or slightly acidic conditions. These

are general statements, and many exceptions occur"

(Ref. 1, p. 329).



II-2 Mechanisms

Calvert and Johnson [18] were the first to identify

dealloying as a unique corrosion process in 1866. The

mechanisms) of dezincification of alpha-brasses, which

has received a great amount of theoretical and experi-

mental research, has been elucidated by the work of Verink

and Heidersbach [4,19]. Prior to this research two

theories were proposed to explain the selective removal

of the zinc component of alpha brasses [20-23]. One

theory states that zinc is selectively dissolved from the

alloy leaving vacant sites in the copper-zinc lattice

structure causing a weak, porous residue of copper [24-

30]. The alternative theory contends that brass first

dissolves stoichiometrically, then the zinc ions remain

in solution while the copper ions replate back on the













brass surface [31-43]. To further complicate matters

some researchers reported that apparently both mechanisms

occur simultaneously [44-50].

The absence of a standardized testing techniques)

which would accurately reproduce the dealloying phenomenon

rapidly under field conditions led to enormous variation

in published results. The slow kinetics of dezincification

have encouraged many researchers to use accelerated

tests to evaluate susceptibility of alloy systems to de-

alloying. These include alteration of electrolyte compo-

sitions by increasing the concentration of certain species,

addition of specific ions, and electrochemical stimulation

of alloys. The credibility of accelerated tests, however,

is limited by the requirement that their use not bias

the experimental results. Unfortunately, it has been

shown [4,19] that many of the accelerated dealloying

tests significantly alter the controlling reactions.

The results of Verink and Heidersbach's [19], and

Verink and Parrish's [49] research showed that by

organizing the published dealloying data in an orderly

manner, through the use of Pourbiax diagrams (potential

vs. pH), the seemingly ambiguous results obtained by

previous investigations can be more clearly understood.













II-2-1 Selective Leaching Mechanism

Pickering and others [51-57] have stated that pre-

ferential removal of the less noble component of a binary

alloy undergoing anodic dissolution requires (1) that

the difference between the single electrode potentials

of the two constituent metals in the electrolyte is suf-

ficiently large, i.e. several times greater than RT/F,

and (2) that the potential of the dissolving alloy is

higher (more positive than) than that of the less noble

metal and significantly lower (more negative than) than

that of the more noble metal.

Pickering and Wagner [56] outlined the various

transport mechanisms capable of accounting for the move-

ment of the less noble metal species to the surface and

the aggregation of the more noble species on the surface.

According to Pickering and Wagner, the two mechanisms

likely to be operating when one metal is preferentially

dissolved from a binary alloy are:

A) Surface Diffusion Mechanism only the

less noble species ionizes and enters the

solution while the atoms of the more noble

metal aggregate by surface diffusion.

B) Volume Diffusion Mechanism only the less

noble species ionizes and enters the













solution and atoms of both metals move in

the solid phase by volume diffusion.

These authors favor an enhanced volume-diffusion mechanism

controlled by divacancy transport. Skuratnik [57] studied

the role of nonsteady-state volume diffusion of negative

(less noble) components from the bulk to the surface of

the alloy/solution interface as the controlling reaction

in selective alloy component dissolution (dealloying)

in the In-Zn system. The results obtained agree well with

the experimental data on the ionization of zinc during

anodic dissolution.


II-2-2 Dissolution/Replating Mechanism

This mechanism can best be described by examining

the reactions involved in the simultaneous dissolution of

copper and zinc and the replating of copper ions [56]. The

anodic dissolution reactions are

Zn (alloy) Z Zn+2 (aq) + 2e- (1)

Cu (alloy) Cu+2 (aq) + 2e- (2)

These reactions are independent of each other and also

independent of the cathodic redeposition reaction

Cu+2 (aq) + 2e- I Cu (solid) (3)

The equilibrium potential of reaction (2) in a solution

of fixed copper ion concentration is more noble than that













of reaction (3) because the activity of copper in a Cu-Zn

alloy is less than unity, whereas the activity of Cu

(solid) in reaction (3) is 1. Therefore, dissolution but

no replating of copper ions from a copper-zinc alloy

will occur when the electrode potential is more noble

than that established by reaction (2) [561.

When reactions (1) and (2) are not independent

(coupling), the equilibrium electrode potential for the

copper dissolution reaction will be shifted in the active

direction and may attain values less noble than reaction

(3). Under these conditions any copper ions entering the

solution by anodic dissolution of the brass surface may

be redeposited at a copper site according to reaction (3).

The availability of copper sites is determined by the

extent of adsorption of copper atoms on the surface and

their subsequent migration to form aggregates and grains

of pure copper. An activity gradient may be the driving

force for the aggregation of adsorbed copper atoms; Cu

adsorbedd) a >1 and Cu (solid) a = 1.

In their investigation into the dezincification of

70-30 Cu-Zn alloy in 0.1 M chloride solutions, Verink

and Heidersbach [19] outlined the conditions (potential

and pH) which determine what mechanism, selective leaching

or dissolution/redeposition, is predominant (see Fig. 2.1).





13



















1,4 CuO hy )r






A8-
3Cu(OH)2-Cu CI "


S CuO




14
-2z










"-z Z n



1---(------------"-L._ I




3 4 I 9 t i

pH





Figure 2.1. Potential-pH diagram for Cu-30 Zn alloy in
0.1 M chloride solution, at 25C (from
Ref. 4).













In solutions with a pH <7 and in the potential range

-0.940 to 0.000 VSHE, the brass alloy exhibited preferen-

tial dissolution of zinc. The kinetics of this reaction

were controlled by the residual copper at the surface

which acted as a barrier. Replating of copper ions or

other reducible species will occur if added to the solu-

tion.

The dezincification behavior of 70-30 Cu-Zn in the

potential range 0.000 to +0.200 VSHE was shown to be

simultaneous dissolution of both constituents. The ratio

of copper/zinc in solution is influenced by the oxidation

state of the predominant copper ion. Copper ions in

solution in this potential range will deposit or redeposit

from unstirred solutions or in occluded cells.

In the most noble potential range (above +0.200

VSHE) the dealloying behavior shows both zinc and copper

dissolve stoichiometrically and no deposition of copper

ions occurs from 0.1 M chloride solutions. Additionally,

specimens which have been exposed at potentials above

+0.200 VSHE and subsequently potentiostated below +0.200

VSHE in 0.1 M chloride solutions exhibit copper ion re-

deposition.

The controversy concerning which mechanism (selective

leaching or dissolution/replating) in alpha-brasses was













the result of the absence of a reliable standardized

testing method. It has been shown that each of the pro-

posed mechanisms may operate individually or simultaneous-

ly, depending upon the solution and electrochemical condi-

tions.



II-3 Techniques

Numerous surface and bulk analysis techniques have

been used to elucidate the controlling mechanism of de-

alloying in the Cu-Zn system. X-ray and electron dif-

fraction [19,41,48,56,58-64] have been among the most

widely used to detect the changes in the lattice parameters

of alloys undergoing dealloying. The electron probe

microanalyzer [19,58,60,65] has been used increasingly

in corrosion research because of its unique capability

for chemical analysis of very small areas. Additionally,

scanning electron microscopy [19], Auger electron spectro-

scopy [66], and electrochemical polarization methods [19,

67-78] have been employed to characterize this corrosion

phenomenon.


II-3-1 Optical Methods

Dealloying was observed and identified as a corrosion

process during the mid-nineteenth century [18]. The












increasing reddish hue of yellow brass exposed to de-

alloying environments became the most characteristic

symptom of this corrosion process. The only instruments

available for a mechanistic investigation at this time

were the microscope and metallograph. However, these

tools were only sufficient to classify the dezincification

phenomenon into two distinct classes, plug type and

layer type. The appearance of a reddish surface during

dezincification encouraged many investigators, using

metallographic instruments, to speculate that an apparent

increase in copper at the surface was the result of a

selective removal process.


II-3-2 X-Ray and Electron Diffraction

Through the use of diffraction techniques [58-64] re-

searchers could determine the extent and origins) of

copper enrichment during dealloying. The diffraction

pattern for each alloy of the Cu-Zn system has a unique

structure. The extent of copper enrichment on and near

the surface may be obtained by comparing its characteristic

peaks with those of the recorded standards. The origins

of dezincification may also be inferred by the resultant

diffraction pattern. A diffraction pattern which shows

only the peaks of the original alloy and pure copper (see













Fig. 2.2) indicates a dissolution/redeposition mechanism,

while the detection of an alloy of intermediate composi-

tion, as shown in Fig. 2.3, would suggest that selective

leaching is the dealloying mechanism. The changes in

the lattice parameter that are measured by this technique

(see Fig. 2.4) are restricted in the alpha-brass system.

At room temperature the limit of solubility as shown in

Fig. 2.5 occurs at approximately 37 wt.% Zn. This con-

fines the changes in the x-ray and electron diffraction

angles to a small range. The major limitation of this

technique is that it is not a surface analysis method.

Therefore, a substantial amount of dezincification must

occur to be detected over the background signal.

Additionally, no estimation of dezincification zone

thickness can be made with this technique.


II-3-3 Electron Probe Microanalyzer

In the electron probe microanalyzer (EPMA), also

known as the electron microprobe, the primary signals of

interest are the characteristic x-rays which are emitted

by bombarding the specimen with electrons. The analysis

of these characteristic x-rays yields both qualitative

and quantitative compositional information. The spatial

resolutions of the electron microprobe are of the order















U)







4-
a)
04
04
0
rcl
r.(
(TJ









0
U)
0-.



a)
La





C)













N4-4
a


040



















44-I
0
x a



0 w




* H



o U)



C4I

C4

kQ)
7r4

54






19






I


0-

E
0O
O






,~--I





>-1












O -.
0





4-i











mO
Nrd-
< 0







-,-
Q4 i
Sr-l 44


T-l (









(Ol 1
rrn r

LU)



*H0

-H

N>i



C4)


-P




(- U

rH

C--p











-H(
I-4

























0cc
L= 3.66




-3.64

o 3.64
I-
I-
-J


ATOMIC PERCENTAGE ZINC


Figure 2.4. Lattice parameter of alpha-brass as a function
of atomic percent zinc (from Ref. 4).









































30 40 50 60 70
WEIGHT PERCENTAGE ZINC


Figure 2.5. Copper-zinc phase diagram (from Ref. 4).












1 1m. The analysis is nondestructive with an accuracy

of approximately 2% of the amount present for a given

element. The electron probe microanalyzer is also capable

of generating x-ray signal scanning pictures, which show

the elemental distribution in a given area. These

pictures can be correlated with photographs obtained by

optical metallography to provide additional information

about surface topography and local compositional varia-

tions. However, the use of electron microprobe in studies

on dealloying is limited by the necessity of establishing

a calibration curve for a given alloy system. The

calibration curve is obtained by measuring the intensities

of a set of standard alloys of known composition, having

identical physical characteristics as the unknown. The

dezincification process can significantly alter the

physical nature of the unknown alloy surface, density

changes, and surface roughening, which may lead to er-

roneous quantitative chemical analysis. Other factors

that restrict the use of this technique in dezincification

studies are the necessity for special data manipulation

to get meaningful results and the requirement of analyzing

specimens under vacuum conditions.


II-3-4 Electrochemical Methods

A great deal of electrochemical data has been ac-

cumulated since the identification of dezincification as












an electrolytic corrosion process [19,20,27,49,65,67,68,

79-891. Some researchers have tried to determine the

role of anodic and cathodic currents and the species

involved, others have stressed the role of single elec-

trode and alloy equilibrium potentials, and others the

potential-pH relationships of the proposed mechanisms.

Electrochemical techniques allow control of the

mechanism of dezincification attack and thus identify

the parameters affecting the alloy's behavior. Some

of the more significant results of electrochemical tests

are identification of the effects of anodic films formed

during dezincification [65], influence of zinc composition

and alloy structure on dezincification behavior [67,82],

and rates of dissolution of noble and active components

during dezincification [27,83]. Additionally, Verink

and others [4,19,74,75] have elucidated the dezincifica-

tion behavior of yellow brasses in terms of equilibrium

potential-pH diagrams.

The major shortcoming of electrochemical testing

is that it is not a surface analysis technique. There-

fore, it cannot, of itself, provide the necessary compo-

sitional information needed for evaluation of the nature

of the dealloying phenomenon.












II-3-5 Differential Reflectometry

It has been shown that the differential reflectometer

[90-94] is capable of providing useful information con-

cerning the electron properties of metals and alloys not

previously available using conventional methods. This

technique enables the researcher to evaluate and identify

changes in surface composition and formation of oxidation

products by enhancing the spectral reflectivity of metals

and alloys.

In the analysis of differential reflectograms ob-

tained on alloys of the copper-zinc system, it has been

observed [5,92] that certain peaks, which are designated

A, B, C, and D, shifted noticeably with a change in zinc

concentration. In particular, peak A shifts to lower

wavelengths with increasing zinc concentration, whereas

peaks B and C shift to higher wavelengths. It has been

determined that this technique is capable of detecting

changes in composition as small as 0.1 at.%. Dif-

ferential reflectometry does not require a vacuum. The

dealloying experiments can be conducted in situ. It

has been shown by Shanley, Hummel, and Verink [6] that

corrosion products only a few Angstroms thick can be

readily identified. These superior capabilities make dif-

ferential reflectometry an ideal technique for the












investigation of dealloying. Details on the differential

reflectometer are contained in Chapter III.



II-4 Oxides

The oxide films that are formed on metals upon exposure

to aggressive solutions play a fundamental role in their

corrosion behavior. Copper, aluminum, and silver are among

the metals which form thin, semipassive layers that isolate

the metal substrate from the solution. The films may be

pure oxides, hydroxides, chlorides, or of a mixed nature,

depending on exposure conditions. The reaction products

contribute to the films' insulating or semiconducting

properties. Thus, the corrosion product layers affect the

electrode potential of the surface and influence the trans-

port of metal ions at and near the reaction interface.

Kruger [95,96] and Kruger and Calvert [97] studied the

nature, growth, and effects of light on the oxides formed on

copper. Kruger [95] exposed copper single crystals to dis-

tilled water with free access to atmospheric air, with and

without added oxygen. He observed: (1) Cu20 was formed

in the absence of added oxygen; CuO was formed when

oxygen was bubbled through the solution; (2) at room

temperature, the reaction kinetics were considerably












faster in water than in air; (3) the rate of oxidation

was different on different crystallographic planes;

(4) the oxide films were not continuous; (5) the degree

of orientation and epitaxy of the oxide films depended

on the crystal faces upon which they were grown; (6) light

had a marked influence on the oxidation process in

water.

Kruger [96] observed interference colors beneath

a loosely-held layer of CuO. He explained this observa-

tion on the basis that CuO and Cu20 both were growing

simultaneously but at different rates.

Kruger and Calvert's [97] measurements of the effect of

light on copper in solution confirmed the theory of Mott

[98] that when a Cu-Cu20 couple is illuminated, an EMF

develops. The electrons produced by the light flow

through the blocking layer (stoichiometric Cu20), adjacent

to the copper substrate, from the cation-deficient Cu20

at the solution interface. The illumination will have

its most pronounced effect upon the electrons, the minority

carriers, since Cu20 is a p-type semiconductor. This

countercurrent electron flow reduces the anodic current.

This reaction suppresses the formation of Cu20 (i.e.

reduces its rate of growth).

Maja, Ginatta, and Spinelli [99] studied the effects

induced by light on the copper-copper oxide system in













alkaline solutions using photopotential measurements.

Their results indicated that the photovoltaic process

at the copper/oxide functions were dependent upon the

oxidation procedure (thermal, fused salt, or anodic

oxidation), while photoelectrochemical processes at

oxide/electrolyte interfaces were mainly influenced by

the solution.

Shoesmith, Rummery, Owen, and Lee [100] studied the

nucleation and growth of cupric hydroxide films on copper

in alkaline solutions. They observed that Cu(OH)2 forms

in two layers: a base layer grown by a solid-state

mechanism and an upper layer of individual crystals

nucleated and grown from solution. The size and number

of upper layer crystals are dependent on electrode

potential. More anodic potentials produce a large number

of randomly deposited crystals, whereas less anodic

potentials result in fewer, more highly developed crystals.

Increased stirring results in a greater loss of material

into solution, and in the extreme, nucleation and growth

are completely prevented. For sufficiently low crystal-

lization rates, obtained by potentiostating, the thermo-

dynamically stable phase, CuO, is formed. At higher

crystallization rates the formation of Cu(OH)2 dominates.

Akimov, Astaf'ev, and Rozenfel'd [101,102] using an

electroreflection method studied the sequential oxidation












of a copper electrode in alkaline solutions. Their inter-

pretation of the results of the electroreflection spectra

obtained as the potential was shifted to more noble

values was that the experimentally determined phase

boundaries agreed with those described by Pourbaix [86]

for the copper-water system at 250C.

Shanley, Hummel, and Verink [103,104] reported

optical spectra for cuprous oxide (Cu20) and cupric

oxide (CuO) using the differential reflectometer.

Shanley further reported [103] that the intensity of the

structure in these reflectograms is related to the

thickness of the film, and can be used as a quantitative

measure of oxide growth up to approximately 500 A.

Subsequent studies [104] on copper, potentiostated in

pH %9.2 solutions, revealed that stepping the applied

potential from the CuO region to the Cu20 region resulted

in the growth of Cu20 beneath the existing layer of CuO

without significantly altering the CuO.

Gabel, Beavers, Woodhouse, and Pugh [105] observed

the same type of mixed oxide formation, inner layer of

Cu20 and outer layer of CuO, by exposing several alloys

of Cu-Zn, Cu-Al, and Cu-Ni to an ammoniacal solution.

Additionally, they reported that the thick tarnish films

were essentially depleted with respect to the alloying













elements (i.e. they exhibited features expected of

dealloyed structures).

The electronic structure and properties of zinc and

ZnO [106-111] have been extensively studied. Unertl and

Blakely [106] reported that the mean thickness of the

initial oxide (ZnO) increases linearly, as a function of

time, with oxygen exposure until a few monolayers of

oxygen are taken up; this oxidation step is believed

by Unertl et al. to be unactivated and highly disordered.

In any event, the oxide appeared to grow heterogeneously

in this initial stage.

Chelikowsky [107] calculated the band structure,

electronic density of states, reflectivity spectrum, and

valance pseudocharge density for ZnO using an oxygen

nonlocal ionic pseudopotential model.

Dahlberg [111], using optical modulation of the

transmission of incident low energy electrons, studied

the behavior of etched single crystals of ZnO. At low

incident electron energies the observed structure, oc-

curring at approximately 3.3 eV, correlates well with the

conduction band density of states of ZnO and is believed

to originate from either energy matching and/or energy

loss processes.












II-5 Summary

There is considerable evidence in the literature

that the formation of a protective oxide on a copper or

brass surface proceeds in a complex manner. The initial

reaction at the metal/solution interface produces cuprous

oxide (Cu20), and if the equilibrium potential is suf-

ficiently noble an outer layer of cupric oxide (CuO) also

will form. Oxides on alloys in which copper is the major

constituent (>60 at.%) seem to be partially or completely

depleted in the minority species.

From the work of Shanley, Hummel, and Verink [103,

1041, and Akimov et al. [101,102], a more detailed picture

of the sequential oxidation process can be constructed.

The first reaction between solution and metallic copper

produces a diffuse zone of Cu(OH); this subsequently forms

the inner oxide layer of Cu20. Under certain conditions

of potential and pH an additional zone of metastable

Cu(OH)2 may form. This may produce an additional outer

layer of cupric oxide (CuO).

The dezincification process in alpha brass is in-

exorably linked to the oxidation reactions occurring on

the surface. The deficiency of Zn and ZnO in the corrosion

products of brass, and presence of Zn+2 ions in samples of

dealloying solution suggest that Zn+2 ions are mobile in












the corrosion products and contribute to the corrosion

current.

Through an understanding of Pourbaix diagrams [74,

75,85,86] and their use in illustrating a tendency for

dealloying, by the superposition of the constituent

elements of an alloy to reveal regions of potential and

pH where one element may preferentially corrode, the

plethora of dezincification data can be put into an

orderly and consistent framework.

Differential reflectometry and particularly the

technique of compositional modulation offer the corrosion

researcher an opportunity to observe and record, for the

first time, the initial surface reactions occurring as

dezincification proceeds. Additionally, the inherent

sensitivity of this instrument affords the investigator

the ability to detect possible dezincification of red

brasses previously thought to be immune.















CHAPTER III
EXPERIMENTAL PROCEDURE


III-1 Introduction

The use of the differential reflectometer to investi-

gate the dezincification behavior of alpha-brasses was

predicated on its demonstrated ability to detect and

characterize the formation of thin film corrosion products

on copper [103,104] as well as changes in the structure

in the differential reflectivity of copper upon alloying

zinc [90-93].

The present investigation also required the use of

electrochemical polarization techniques to stabilize and

control the mechanisms of the rate determining reactions.

Because of the importance of sample and solution pre-

paration upon experimental results, special procedures

were followed to reduce experimental scatter.



III-2 Differential Reflectometer

Differential reflectometry is a modulation spectro-

scopy technique where alloy composition is the modulated

parameter. Detailed information on the functioning of the













differential reflectometer can be found in papers by

Holbrook and Hummel [93] and Nastasi-Andrews and Hummel [5].

A brief summary of the functioning of the instrument and

its capabilities will be presented here.

The instrument is used to measure the difference in

the reflectivities of a dealloyed specimen with respect

to a standard, i.e. a freshly polished specimen of the

original composition. The specimens are arranged side by

side and illuminated alternately with monochromatic light

by means of an oscillating mirror.

A xenon light source provides the broadband input

for the scanning monochromator (Fig. 3.1). A scanning
0
rate of 2000 A/min was used in all measurements and the

wavelength range extended from 2000 A (6.2 eV) in the
0
ultraviolet to 8000 A (1.55 eV) in the near infrared.

The monochromatic light is reflected by a flat mirror

to a concave mirror oscillating at a frequency of 60 Hz.

The oscillating light beam is thus focused alternately

on the two halves of the specimen. The total area scanned

on the sample is approximately 2 mm x 4 mm. The indivi-

dual reflectivities of the two halves, R1 and R2, are

focused on a frosted quartz plate in contact with the

face of the photomultiplier to distribute the intensities

of R1 and R2 uniformly. The output signal of the

























3.?

0 V


'A

CL
InP


0LA
-j
Ut
o


0

E



C
0
c
0


r


rh.
=0

6L












photomultiplier is split into two paths; the first branch,

which contains the lock-in amplifier tuned to the fre-

quency of the oscillating mirror, produces the AR =

R1 R2 signal. The second branch contains a low pass

filter whose output provides the R = (R + R2)/2 signal.

The operating voltage of the photomultiplier is elec-

tronically varied in order to hold the output voltage

at 5 volts. This stabilizes the R signal. The two

signals, R and AR, then enter a divider circuit, which

produces the recorder Y-axis input of AR/R. The X-axis

input is obtained from a variable resistor connected to

the scanning monochromator and is calibrated in units of

wavelength.



III-3 Specimen Preparation

The copper 7, 10, and 20 at.% zinc specimens were

taken from alloys already available from previous experi-

ments. The copper-5% zinc specimens have been prepared

exclusively for this study. The preparation of the alpha-

brass alloys was made from starting materials of the fol-

lowing purity: copper 99.999% and zinc 99.9999%.

Prior to melting, the metals were initially degreased with

acetone and then cleaned in a 50% nitric acid solution to

remove surface contaminants and oxides.












The compositions of the alloys used in this study

are listed in Table 3.1 and were determined by electron

probe microanalysis. The ingots were melted in helium-

filled ceramic tubes using an induction furnace. They

remained in the molten state for approximately 5 minutes.

After cooling, the alloys were homogenized for 14 days at

approximately 10000C, subjected to a 50% reduction, and

then given an annealing heat treatment at 6000C for 1

hour. This produced a relatively uniform grain size of

approximately 0.8 mm2, as shown in Fig. 3.2. Specimens

were then cut from the rolled strips using a diamond

wheel.

The specimens were mounted using a two-part metal-

lographic compound and cast into 1-inch diameter by 0.5-

inch thick discs. A small hole was drilled in the back

of each sample prior to mounting and an electrode wire

was permanently attached and sealed to insure intimate

electrical contact necessary for potentiostating.

The metallographic polishing procedure for the

specimens was as follows:

1. Successive grinding with 180, 320, 400,

and 600 grit silicon carbide paper

lubricated with liquid soap and water.


















































Figure 3.2. Photomicrograph of specimen grain size.



















Type

Red Bras

Red Bras

Red Bras

Red Bras

Yellow B


Table 3.1
Compositions of Copper-Zinc Alloys Used

Designated Actual
Composition Composition
at.% Zn at.% Zn Spec

s 5 5.33

s 7 6.79

s 10 9.18

s 10 9.62

rass 22 21.42


imen Code

B

C

D

E

F


Note: Cu-20 at.% Zn alloy = (Cu-20.6) wt.% Zn alloy.













2. Hand polishing with Microcut* paper (which

is a napped paper with very fine silicon

carbide particles) to reduce possible

embedment of large abrasive particles.

This polishing step was repeated on several

pieces of Microcut paper, lubricated with

liquid soap and water, until virtually

all silicon carbide contamination was

eliminated, as confirmed by microscopic

examination.

3. Fine polishing was performed using 6

micron and 1 micron diamond paste re-

spectively on Microcloth* using Metadi*

fluid as a lubricant. After each polishing

step the samples were rinsed with methanol

and dried in a warm air stream. This

polishing procedure produced a uniform,

mirror-like finish.

4. The polished specimens were then potentio-

stated for various times to induce dezinci-

fication.

5. Upon termination of the electrochemical

test, one-half of the sample was polished


*Trademark of Buehler, Ltd., Evanston, Illinois.













with 1 micron diamond paste, as in step 3,

and rinsed with methanol prior to analysis.



III-4 Solution Preparation

All solutions were prepared using reagent grade

chemicals and distilled water, which had been additionally

passed through two ion exchange columns. The chemical

compositions of the chloride-free solutions used in this

study are shown in Table 3.2. The selected buffering

solutions, 0.5 Molar potassium acid phthalate (KHP) and

0.7 Molar boric acid (H3BO3), were used to stabilize

the pH of the 4.0 and 9.2 solutions respectively. Pre-

vious work has shown that these buffers do not affect the

reaction products which form [55,58].

Solutions prepared for potentiostatic studies were

vacuum deaerated prior to testing. The test cell and

solution were purged with hydrogen gas during the course

of the experiment. Two grades of hydrogen gas were used

to determine the sensitivity of the electrochemical

reactions to low-level oxygen and possibly carbon dioxide

contamination.

Standard welding-grade hydrogen gas whose oxygen

impurity concentration was given by the manufacturer to be

%330 ppm was used in most dezincification experiments.





41







Table 3.2
Constituents of Electrolyte
Used in Chloride-Free Experiments

pH Buffer NaOH (0.5 M) Distilled Water

4.0 200 ml, 0.5 M KHP -- 1800 ml

9.2 143 ml, 0.7 M H3B03 117 ml 1740 ml













Selected experiments were repeated with a special "grade

5" hydrogen gas with an oxygen impurity level of less

than 1 ppm. All glassware and electrodes were used

exclusively for chloride-free work and received periodic

cleaning with a 50% nitric acid solution to minimize

chloride contamination.



III-5 Electrochemical Equipment and Operating Conditions

All studies were conducted in a standard electro-

chemical cell shown schematically in Fig. 3.3. The

volume of the cell was approximately 1 liter and was

sealed with a hard rubber stopper with entrances for a

platinum auxiliary electrode, Luggin probe, sample and

gas diffuser, and outlet. The Luggin probe was placed

approximately 2-3 mm from the specimen surface and was

connected to an external, 70 ml electrode reservoir which

contained a saturated calomel reference electrode. The

solutions were continuously stirred with a Teflon*-en-

capsulated magnetic stirrer to reduce the possibility of

local concentration gradients.

The potentiostat used in these experiments was

constructed in-house and is shown schematically in



*Trademark E. I. du Pont de Nemours, Inc., Wilmington,
Delaware.









AUXILIARY
ELECTRODE


SAMPLE


MAGNETIC


STIRRER



Figure 3.3. Schematic of electrochemical cell (from
Ref. 6).


SAMPLE













Fig. 3.4. These units maintained a constant potential

difference (+1 my) between the exposed sample and the

saturated calomel reference electrode. A block diagram

of the potentiostat and an operating electrochemical

cell is shown in Fig. 3.5.

In the dezincification experiments the samples were

potentiostated at -350 my SCE in pH %4 solutions to ac-

celerate the selective removal of zinc from the alloys.

The oxide formation experiments were conducted in solu-

tions of pH %9.2 and the specimens potentiostated at -225

my SCE and +350 my SCE corresponding to the domains of

Cu20 and CuO respectively. In selected cases, specimens

were prepotentiostated at approximately -1300 my SCE to

retard any surface oxidation reactions prior to testing.

All electrochemical experiments were conducted under

conditions of constant fluorescent lighting and room

temperature (<70F).







IN 4006
-
-I- V


CE


Figure 3.4.


Schematic of potentiostat used
cation experiments.


in dezincifi-


100
K








POTENTIOSTAT


METER


Counter electrode


Working electrode


Reference electrode








Specimen

.- Platinum


CORROSION
CELL AND
SALT
BRIDGE


Block diagram of electrochemical system.


Figure 3. 5.















CHAPTER IV
EXPERIMENTAL RESULTS


IV-1 Preface

The purpose of this study was to detect and charac-

terize dezincification of red brasses. Yellow brasses

of 22 and 30 at.% have been chosen as "baseline" materials

for this investigation, since it is well established in

the literature that dezincification takes place in these

alloys (see Chapter II).

Theoretical Pourbaix diagrams [68] for the copper-

water and zinc-water systems are shown in Figs. 4.1 and

4.2 respectively. In the acid region (pH %4) at a

potential of approximately ESHE = -100 mv, the stable

solid species for copper is Cu(solid); however, the

stable species of zinc, under the same conditions of

potential and pH, is the zincic ion (Zn+ ). These are

the electrochemical conditions that should produce de-

zincification of brasses by a selective leaching mechanism.

Additionally, these electrochemical conditions were

selected to preclude the equilibrium cathodic hydrogen

evolution reaction and simplify the analysis of the de-

zincification reactions.



















U,

0





z
w
I-
0
CL


-2 0 2 4 6 8 10 12 14 16

pH


Figure 4.1.


Potential-pH equilibrium diagram for the
system copper-water at 25C (ionic species
a = 10-6) (from Ref. 86).




49






2-

I I
HZnO2

S b ZnO2













O .
Z -
w. -- I
I- "--- -.





0 I

-2 0 2 4 6 8 10 12 14 16

pH


Figure 4.2. Potential-pH equilibrium diagram for the
system zinc-water at 250C (ionic species
a = 10-6) (from Ref. 86).













IV-2 Yellow Brass Dezincification

In Fig. 4.3, the differential reflectogram (AR/R

versus wavelength A) of a copper-zinc alloy is shown

(average composition of 20.8 at.% Zn). This reflectogram

was obtained by scanning samples of the following compo-

sitions: Cu-19.3 at.% Zn and a Cu-22.3 at.% Zn alloy

(compositional modulation). Figure 4.3 has all the fea-

tures known from previous compositional modulation

studies [90-94]. The reflectogram is characterized by

several features: two maxima; designated peak A (509

nm); and peak C (358 nm); a minimum, designated as peak

B (420 nm); and a shoulder D around 250 nm.

Figure 4.4 depicts a differential reflectogram of

a copper-21.4 at.% zinc alloy. One-half of the specimen

was potentiostated for various times in a pH -4 (KHP)

solution at a potential of ESHE = -100 my (see Fig. 4.5).

The other half, the "reference," was in the unpotentio-

stated state. Comparison of Figs. 4.3 and 4.4 shows that

potentiostating, as described above, leads to a change

in zinc content, i.e. compositional modulation. (If

the composition of the two halves would have remained

equal, a horizontal "zero line" would have been the

result.)

A word should be said about the sign (convention

used in describing features) of a given peak in a


























to




0
Cl

>1
















E:
N










r(
U








0





0


44J
U













-i



Cd

.1i
WSI


4-)



Q)
L




14









a


0
0
C0d


0
-0
ro

^0
.0





LO
0





O




0
.0



-0
^---_____--
I







0

10

0 0

<

























04





-H
a)


U)
0
04






x
a)
:s






0








CN
u










4-4



0





0
4-4
u
a)
H















r-H



4J-4
a)






(da)


4~





a)

4-4
44
ri U




4



































crIio


0


0
0


0
0
f()

to


0

0


0
0



0
(D



r,-

0
0
OD
O)
a,



































8O
800


700 600 500 400 300 200


X (nm)



Figure 4.5. Montage of Cu-22 Zn, yellow brass, dezinci-
fication reflectograms.












differential reflectogram. The nomenclature used in

compositional modulation denotes that (Fig. 4.3) a

positive peak "A" (maximum) is always obtained if the

alloy with the higher solute concentration is placed into

the instrument as the upper part of the couple. In the

present dealloying experiments, the "reference" half of

the specimen was always placed into the instrument as the

upper part of the couple. Therefore, following previous

practice, a positive peak "A" in a differential reflecto-

gram of dezincification indicates loss in zinc content

in the lower part of the specimen.

The question immediately arises as to how much zinc

is lost due to dezincification occurring under these

experimental conditions. At attempt to estimate this

will now be made. Earlier investigations of the composi-

tional modulation behavior of copper-zinc alloys [5] have

shown that the position (wavelength) of peak "A" decreases

with increasing average zinc concentration, as shown in

Fig. 4.6. Peak "A" in Fig. 4.4 was found to be located

at approximately 506 nm which is equivalent, according

to Fig. 4.6, to an alloy of average zinc concentration of

19.2 at.% after dezincification. By knowing the original

composition of the specimen (21.4 at.% Zn) and the mea-

sured average alloy composition of the couple, one can
















x
(nm)

520 r -.____




500 I i
0 2 4 6 8 10 12 14 16 18 20

ATOMIC % ZINC
Figure 4.6. Position of peak "A" versus average zinc
content in copper-zinc alloys (from Ref.
5).













calculate the zinc content of the surface of the dealloyed

specimen. In the case of the specimen shown in Fig. 4.4,

the dealloyed half was calculated to have a zinc content

of approximately 17 at.% Zn, which corresponds to a loss

of about 4.4 at.% Zn. This result is consistent with

published data [66]. The loss of zinc can be readily

observed by visual means.

Inspection of the individual curves in Fig. 4.5

(which have been obtained by varying the time of

potentiostating) leads to several additional observations.

a) The peaks are much broader in the case of

dezincification (Fig. 4.5) compared to those where two

alloys of different composition (Fig. 4.3) were scanned.

The observed broadening phenomenon may derive from several

independent sources. First, a given peak in a dif-

ferential reflectogram is much sharper when the compo-

sitional difference between the two alloys is small

[92]. This is illustrated by the zinc concentration

indicated in the case of Fig. 4.4. The compositional dif-

ference in Fig. 4.4 is approximately 4.4 at.% Zn while

that of Fig. 4.3 is only 3.0 at.% Zn. However, this re-

latively small increase in compositional difference would

not solely account for the broadening of the peaks in

Figs. 4.4 and 4.5. It must be assumed, therefore, that

additional mechanisms are operating.













The results of compositional modulation of a couple

made up of two homogeneous alloys of different composi-

tions are shown in Figs. 4.7A and B (e.g. 22 and 20 at.%

Zn). The measured differences in reflectivities as a

function of wavelength (Fig. 4.3) provide information

about changes in the electron band structure as the

result of addition of X at.% solute. In the case of

dezincification (Figs. 4.7C and D), the reference half

of the couple is still homogeneous in composition but

the dealloyed half may possess a composition gradient of

unknown shape and magnitude. Therefore, the degree of

dezincification can be determined only with respect to

the apparent composition of the dealloyed sample since

the degree may vary with depth. Superposition of one or

more signals upon a compositional modulation spectrum

can broaden or skew an cl-type spectrum (a minimum fol-

lowed by a maximum as in peaks B and C). This effect is

particularly severe when a minimum and maximum from two

superimposed signals occur at similar wavelengths. The

result may be the loss of both peaks, or a broadening,

i.e. the shifting of both the observed maximum and minimum

in opposite directions.

b) Figure 4.5 also shows that the positions of

peaks A through D do not vary systematically with











22

AT. %
ZINC


20


SURFACE


IcL


22

AT. %
ZINC


22


BULK DZ


18
SURFACE


Figure 4.7.


Schematic of zinc concentration gradient in
dezincification zone.


[A IB













potentiostating time. Some variation in peak positions

is observed; however, the fluctuations seem to be random

about an average value. The average wavelengths of these

peaks (Table 4.1) were determined from Fig. 4.5 to be

A = 506 nm, B = 397 nm, C = 295 nm, and D = 240 nm. The

absence of a systematic variation of the peak positions

with potentiostating time leads to the conclusion that

after 1 hour the final composition of the surface of the

dezincified alloy has essentially been established, and

no significant additional change in composition is ob-

served in the reflectogram during subsequent potentio-

stating.

c) Figure 4.5 shows that with increasing potentio-

stating time the absolute difference in reflectivity,

between the potentiostated half and the reference half

of the sample, becomes successively larger. This is

particularly evident for the experiments conducted for

48 and 73 hours (note the change in scale for these

spectra).

In previous studies on the corrosion behavior of

copper [6], it was concluded that the peak intensity in-

creases with increasing corrosion film thickness. It

seems reasonable to assume, therefore, that the source

of the increasing dezincification signal during the first













Table 4.1
Peak Positions for a Copper 20.8 at.% Zinc
Alloy (Average Composition) and a Dezincified
Yellow Brass Alloy (Starting Composition, 21.4 at.% Zn)


A20.8 E20-8


A E


Peak (nm) CeV) (nm) (eV)

A 509 2.44 506 2.45

B 420 2.95 397 3.12

C 358 3.46 295 4.20

D 250 4.96 240 5.17












48 hours is related to an increase in the depth of

penetration of the corrosion attack. The dezincification

spectra do not increase significantly after 48 hours.

This is interpreted to be a result of the maximum depth

of penetration of light (-100 A), which limits the

ability to detect any additional dealloying, rather than

a limitation on the progress of dezincification.

d) Figure 4.5 reveals that with increasing potentio-

stating time an upward trend in the spectra at wave-

lengths below 300 nm can be observed. It is known that

cupric oxide (CuO) exhibits such behavior [6] (see

Fig. 4.8). This suggests that for potentiostating times

greater than 48 hours, CuO is formed as a thin corrosion

product layer.

e) Finally, Fig. 4.7 shows that in all the dif-

ferential reflectograms (with the exception of the one

obtained for 1 hour of potentiostating time), so-called

"satellite" peaks occur. The principal features of the

satellite structure are at 613, 545, and 475 nm. The

curves for 10 and 14 hours show an increase in the relative

intensity of the satellite structure, with respect to

the compositional modulation signal. After 48 hours, the

intensity of the satellite peaks has not increased pro-

portionately with the dezincification structure.







0
-0
\J




0
-0



O







oo
'n-I
U



X0 0

\-0 E



0
o -


C 0




0

O c4
\ a)



l4
-,-4





I I *
00 O
0


4-4
\ VA/ O
\ u
\r











The origin of the satellite peaks is not known at

this time. They could be caused by surface roughening,

corrosion products of unknown nature, or by certain

peculiarities of the differential reflectometer. It

should be noted that they are not apparent in the de-

zincification studies of red brasses. Further investi-

gations have to shed some light on this point (see

Chapter VI).



IV-3 Red Brass Dezincification

In Fig. 4.9, a differential reflectogram of a red

brass alloy containing copper 6.8 at.% Zn is shown.

This reflectogram was produced by scanning between a

copper 4.4 at.% Zn and a copper 9.2 at.% Zn alloy (compo-

sitional modulation). The wavelengths of the major peak

positions are listed in column 2 of Table 4.2. These data

serve as a reference for the subsequent dezincification

experiments with red brass.

Figure 4.10 depicts a differential reflectogram of

a copper-zinc alloy which has initially 9.4 at.% zinc.

One-half of it was potentiostated for 1 hour in p1H 4

(KIP) solution at a potential of ESHE = -100 mn. The

other half remained in the unexposed condition. The light

beam of the differential reflectometer was scanned between












Table 4.2
Peak Positions for a Copper 6.8 at.% Zinc Alloy
(Average Composition) and a Dezincified Red Brass
Alloy (Starting Composition, 9.2 at.% Zn)


6.8
6.8


E6.8


A E


Peak (nm) (eV) (nm) (eV)

A 540 2.30 543 2.28

B 348 3.56 327 3.79

C 303 4.09 252 4.92





























U)
















4-4
U)

0














a)
N





C4












0-i
fo
0














a)







Q)
A













4 J
C)






L44





rC)

H
4A







0
0




0
0o. -0
ro


0
0
;I-



0
-0



0
LO



-0
(CD


0
0
r---





r-, 0























a)





04
r-







0
4











0
'-4





0











0

4J
U)
a




-4
c4
a)
roC:


a)




















.r-i C~)
0

















a)
44
44
r- U





4






0
0(
0C\


0o
-o
r()
0

-0
O



0
-0




(D
\0

-0
CO


0


S0
SOO
1II
mO-_____ i 1 ___0













these two specimen halves. Figure 4.10 has all the major

features of the compositional modulation (i.e. peaks A,

B, and C), as comparison with Fig. 4.9 reveals. The

peak positions are listed in columns 4 and 5 of Table 4.2.

From this result it is concluded that (similarly to

yellow brasses) red brass undergoes dezincification if

potentiostated at the appropriate pH and potential. This

observation is clearly in contrast to numerous earlier

statements in the literature reporting that red brass

does not dezincify (see Section IV-2).

Similarly as in Section IV-2, an attempt was made to

estimate the approximate surface composition of the de-

zincified alloy. From the position of peak A (543 nm)

and referring to Fig. 4.6, an average composition of

7.3 at.% Zn of the dezincified and unexposed parts of the

sample can be deduced. This results in a zinc concentra-

tion of the dealloyed part of about 5.2 at.% zinc or a

difference of 4.0 at.% zinc. Zinc loss in both yellow

(4.4 at.%) and red (4.0 at.%) brass seems to be nearly

the same under the same conditions of pH and potential.

By comparing Fig. 4.9 with Fig. 4.10, one can

observe that the peaks are much less pronounced in the

dezincified case, i.e. they are flatter and less intense.

In the case of compositional modulation (Fig. 4.9), the













difference in reflectivity between the two sample halves

is about 6% for peak A, whereas for the case of dezinci-

fication this difference is less than one-half of 1

percent (Fig. 4.10). As shown in Section IV-2, peak

height is a function of the corrosion film thickness.

Therefore, for red brass, the dezincified layer after 1

hour potentiostating time is very thin and certainly less

than 50 A. Such a thin film cannot be detected by the

unaided eye. This may explain why dezincification in

red brass generally has escaped detection in the past.

Figure 4.11 shows the chronology of the dezincifica-

tion behavior of a Cu-10 at.% Zn alloy. The uppermost

curve repeats the 1 hour data in pH %4 (KHP) solution at

an electrode potential of ESHE = -100 my but plotted

on a scale comparable with the rest of the figures. Peaks

A, B, and C gradually grow in intensity through 37 hours.

This is interpreted as resulting from an increase in

depth of the dezincified layer. As in the case of yellow

brass, the wavelength corresponding to the peak positions

in Fig. 4.10 does not vary with potentiostating times,

thereby suggesting that the alloy composition of a

dezincified layer does not vary (only the thickness of

the layer seems to change).

The average wavelengths for the interband transitions

associated with .peaks A, B, and C are listed in Table 4.2.








































800 700 600 500 400 300 200

X (nm)



Figure 4.11. Montage of Cu-6.8 Zn red brass dezincifi-
cation reflectograms.












The broadening of the B and C peaks in Fig. 4.11 (parti-

cularly for longer potentiostating times) is not as pro-

nounced as that observed for yellow brass. This may be

a consequence of a shallower composition gradient and/or

the relative insensitivity of the B-C interband transi-

tion to compositional variation in this range of compo-

sition.

Figure 4.12 illustrates the effect of various oxygen

concentrations on the dezincification behavior of copper-

10 at.% Zn. Figure 4.12C was obtained by potentio-

stating for 48 hours in a pH %4 (KHP), deaerated solution

and purging continuously with high purity hydrogen (grade

5, 99.999% H2). This reflectogram again exhibits the

structure typical of compositional modulation for red

brass. The middle curve shows the effects of oxygen

contamination on the dezincification behavior. This

reflectogram was obtained by potentiostating the same

alloy 48 hours in pH %4 (KHP), deaerated solution purged

with welding grade hydrogen; which has an abnormally

high oxygen level (>300 ppm). Figure 4.12B does not show

a substantial amount of structure, particularly not the

compositional modulation peaks A, B, and C. In Fig. 4.13

another reflectogram of the same specimen is shown, but

in this case the reflectogram was taken using a higher











30


20


(%o)


800 700 600 500 400 300 200
X (nm)


Figure 4.12.


Differential reflectograms of oxygen con-
centration effects on dezincification be-
havior of red brass, 48 hours exposure (curve
A air saturation, B >300 ppm oxygen, and C
<10 ppm oxygen).

















o
C




4-)






0



0o


-p
ld






C1


o
4

,

0



0






t4



0








.--H
r 0
CO
CO



M o
CU
4- 1

















tn
-,-
U 0
(U -
a-l +
4-1 n3










n c
(l (
4- C
















0
re-.


0
rO


C\j
_ '


0
O
I1)


[LIir












sensitivity of the differential reflectometer in order to

reveal more details of possible structure. The reflecto-

gram shows a weak maximum at about 370 nm, a minimum at

about 320 nm, and a shoulder at 530 nm. Furthermore,

there is an upward trend towards shorter wavelengths.

This upward trend is characteristic of CuO (Fig. 4.8).

The maximum near 370 nm is characteristic of Cu20 (see

Fig. 4.14). Thus, Fig. 4.12B and Fig. 4.13 show a super-

position of a very weak compositional modulation signal

on a mixed oxide spectrum. A decrease in the intensity

of the compositional modulation signal (as observed in

Fig. 4.12B) indicates that the presence of oxides on red

brass may tend to stifle dezincification.

The uppermost curve in Fig. 4.12 illustrates the

effect on dezincification of red brass of air-saturation

of the test solution. Figure 4.12C was obtained by purging

a pH '4 (KHP) solution with filtered, compressed air

during potentiostating at ESHE = -100 my for 48 hours.

The usual compositional modulation peaks A, B, and C are

evident together with a general upward trend towards

shorter wavelengths. This is believed to result from

superposition of compositional modulation and cupric

oxide (CuO) spectra. The most stable oxide of copper in

the presence of excess oxygen is cupric oxide [86].








0
-0

or


J ( O
0

0 0



CN
O0 0
r. j. 0, E



OO-0



O 0 00
SO 0 0


t4
s \\ -


O. o
O
S4-4




0)
o0.





00





I 0 CO -
00)
4-i






Sf C












Therefore, the absence of any significant amount of Cu20

after 48 hours exposure, under these conditions, is to be

expected. However, CuO is not a protective oxide species

and the increased intensity of the dealloying signal

confirms this postulation.

Figure 4.12 shows that the presence of a small amount

(<300 ppm) of oxygen may be advantageous in inhibiting

dezincification of red brass through the formation of

a very thin Cu20 oxide film adjacent to the metal surface.

Exposure of red brass to a higher concentration of dis-

solved oxygen may convert the (protective) cuprous oxide

to cupric oxide and destroy the protective character of

the film.

Figure 4.15 shows the behavior of Cu-10 at.% Zn to

moderate concentrations of oxygen. The upper curve

depicts a specimen which was potentiostated 72 hours

at ESHE = -100 my in a pH %4 (KHP) solution and was

purged with high purity (grade 5) hydrogen. The familiar

differential reflectogram for compositional modulation is

observed. The lower curve in Fig. 4.15 depicts a dif-

ferential reflectogram of Cu-10 at.% Zn which was potentio-

stated as above, but in a nondeaerated pH n4 (KHP) solu-

tion. This spectrum exhibits a weak maximum in the

vicinity of peak A, a maximum near 375 nm, a minimum at



















R
2
0



(%)










2


<10 ppm 02


02 saturated


0 -


800 700 600 500 400 300 200

X (nm)



Figure 4.15. Differential reflectograms of oxygen con-
centration effect on dezincification be-
havior of red brass, at 72 hours exposure.












305 nm, and an upward trend towards shorter wavelengths.

As before, a superposition of Cu20, CuO, and weak compo-

sitional modulation is evident in this curve.

A critical range of oxygen concentrations seems to

exist which significantly affects the surface reactions

of red brass dezincification. In the presence of suf-

ficient oxygen, the following initial anodicc) dezincifi-

cation reactions are believed to occur in red brass:

2 Cu 2 Cu+ + 2e- (4)

Zn Zn++ +2e- (5)

Reaction (4) is probably quite slow due to a smaller

thermodynamic driving force, and reaction (5) provides

the major component of the anodic current. The cathodic

reaction of copper with dissolved oxygen will give

2 Cu+ + 1/2 02 + 2e- + Cu20 (6)

The cuprous oxide formed in reaction (6) will (in the

presence of excess dissolved oxygen) become cupric oxide

through reaction (7).

Cu20 + 1/2 02 2 CuO (7)


IV-3-1 Copper-5.3 at.% Zn

Figure 4.16 depicts a differential reflectogram of

a copper-zinc alloy which initially contained 5.3 at.%

zinc. One-half of it was potentiostated for 10 hours at



















?r -H(
-4-4
vl-H

2C,

N~
-4


4::

0

J

(I)
0

04
4-4
EnV




0-
O







4-4


ur) N
10)











U
o -4







QH
~ H
uro



-4X


0
'-U






CO
-H



1)0
S-i






14-4 (1)
) 0
O rJ2-r
.rA

r.

4-4 J
4- rA
r- 0r-
(2 n4







0
-0
O




0





.0
0
.0
oo




0






So









40
0O
-0



SJ 0 0 C OM 00













SHE = -100 my in pH A4 (KHP) solution, purged with

grade 5 hydrogen. The other half remained in the un-

exposed condition. The reflectogram in Fiq. 4.16

exhibits a maximum (peak A) at 559 nm, a broad minimum

(peak B) at 362 nm, and a shoulder near 275 nm, plus a pro-

nounced upward trend in the spectrum toward shorter

wavelengths (indicating CuO). Figure 4.16 exhibits

peaks A and B and also has the characteristic shape of

compositional modulation. This strongly suggests that

even red brass with only 5.3 at.% Zn undergoes some

dezincification if potentiostated at the proper pH and

potential. Some subtle distinctions to the 10% zinc

alloy are observed. First, peak C is "missing," i.e.

it is masked in the strong upward signal at shorter

wavelengths which is believed to indicate the presence

of CuO. Secondly, it requires substantially longer

times (approximately 10 hours) to obtain a characteristic

compositional modulation spectrum. Even after 10 hours,

it was found that a curve as depicted in Fig. 4.16A is

obtained only by scanning the light beam on selected

areas of the specimen, whereas scanning the light over

other areas may yield a structureless, almost horizontal

line (Fig. 4.16B). This suggests that dezincification of

this alloy (for the experimental conditions given above)













occurs only in localized areas which are not visible to

the unaided eye.


IV-3-2 Cupric Hydroxide Cu(OII)2

Figure 4.17 depicts a differential reflectogram of

a Cu-9.2 at.% Zn alloy, one-half of which was potentio-

stated for 20 hours at ESHE = +10 mv in a pH %9.2

(H3BO3) solution. The solution system was stirred and

continuously purged with welding-grade hydrogen gas. The

other half was masked with collodion during potentio-

stating. The coating was removed before measurement in

the reflectometer. The light beam was scanned between

the corroded and uncorroded halves.

The major features in Fig. 4.17 are the sharp

minimum located at 276 nm, the broad maximum near 465 nm,

and a general downward trend towards shorter wavelengths.

Several minor peaks are located around 555, 510, 355, and

328 nm.

Previous studies on copper using the differential

reflectometer [6,104] have identified CuO (Fig. 4.8) and

Cu20 (Fig. 4.14). In contrast to these corrosion pro-

ducts, the species whose reflectogram is shown in

Fig. 4.17 exhibited a light blue color which corresponds

to the strong structure in the blue region of the spectrum



































0


C7i

a)

x
0

>1
U


U






0-4
04
u









'4-4
0





a)
E







4J
a)
0








'4-)

4-4
a)





4-'











L4
ID)
D4














C\D


LO
- )
ro


IO
(D


IT)


in
In


,\
* ..


p I


0
0
CNJ



0
0




0
0
ir
OC



OC$


E
sz.
c


0
0




0
0
r,-


0
0
CD















(near 276 nm). The electrode potential and pH were

chosen to be in the vicinity of the Cu20/Cu(OH)2 boundary

of the copper-water Pourbaix diagram. Pourbaix [861 lists

the species, cupric hydroxide or hydrated cupric oxide,

Cu(OH)2, as a metastable species of CuO which is formed

under the conditions shown in Fig. 4.18, and notes that

it is light blue in color. Verink and Lee [112] also

found, while determining the experimental potential-pH

diagram for the copper-water system (Fig. 4.19), that

a secondary passivation event occurs close to the ex-

pected position of the Cu20/Cu(OH)2 coexistence potential.

This domain corresponds exactly to the potential-pH

conditions used to form the corrosion product whose

reflectogram is shown in Fig. 4.17. It is, therefore,

suggested that Fig. 4.18 represents the species Cu(OH)2.

Several attempts were made to characterize the corrosion

product through analysis by x-ray diffraction. However,

insufficient quantities of the dried powder and its

amorphous character produced inconclusive results. On

the other hand, no rigorous calculation of Cu(OH)2 nor

conventional optical data are available in the literature

as of this writing.