An in-situ investigation of the anodic behavior of Ni and Ni-30 Cu using electrochemical and optical techniques

MISSING IMAGE

Material Information

Title:
An in-situ investigation of the anodic behavior of Ni and Ni-30 Cu using electrochemical and optical techniques
Physical Description:
vi, 126 leaves : ill. ; 28 cm.
Language:
English
Creator:
Smith, Randall J., 1959-
Publication Date:

Subjects

Subjects / Keywords:
Passivity (Chemistry)   ( lcsh )
Nickel films   ( lcsh )
Copper-nickel alloys   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1984.
Bibliography:
Includes bibliographical references (leaves 121-125).
Statement of Responsibility:
by Randall J. Smith.
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 - 000493683
notis - ACR2543
oclc - 11988429
System ID:
AA00003425:00001

Full Text










AN IN-SITU INVESTIGATION OF THE ANODIC BEHAVIOR OF
Ni and Ni-30 Cu USING ELECTROCHEMICAL AND OPTICAL TECHNIQUES







By


RANDALL J.


SMITH


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




UNIVERSITY OF FLORIDA


1984








ACKNOWLEDGEMENTS


The author wishes to express his sincere appreciation to his

advisor, Dr. Rolf E. Hummel,for his patience and confidence throughout

this work. Special thanks are extended to Dr. Ellis D. Verink, Jr.,

for his unwaivering confidence and guidance over the past four years.

The refreshing humor of Dr. Paul Holloway and his efforts in securing

financial support for the final year of research are also gratefully

acknowledged. Dr. John R. Ambrose and Mesuit Akkaya are also to be

thanked for their many stimulating conversations regarding the present

research. Sincere appreciation is expressed to each member of the

supervisory committee (R. E. Hummel, E. D. Verink, Jr., J. R. Ambrose,

P. H. Holloway, R. T. DeHoff and D. 0. Shah) for their support over

the past years.

The author would also like to thank Dr. Mike Kosinski for his

elegant and timely operation of the ESCA-AES system which were necessary

for the completion of this research.








TABLE OF CONTENTS



ACKNOWLEDGEMENTS ................................................

ABSTRACT......................................................

CHAPTER

1 INTRODUCTION.........................................

2 LITERATURE REVIEW ....................................
The Corrosion Behavior of Pure Nickel.................
Behavior of Nickel in Neutral Solutions................
The Behavior of Nickel in Alkali Solutions.............
The Transpassive Behavior of Nickel....................
The Effect of Chloride Ions on the Passive
Behavior of Nickel......................................
The Corrosion Behavior of Pure Copper.................
The Corrosion Behavior of Nickel-Copper Alloys..........
Summary ................................................

3 EXPERIMENTAL PROCEDURE ................................
Sample Preparation...................................
Solution Preparation .................................
The Corrosion Cell.....................................
Polarization Techniques ...............................

4 RESULTS AND DISCUSSION................................
Identification of Anodic Surface Films..................
Anodic Films on a Copper Electrode ...................
Anodic Films on a Nickel Electrode....................
Anodic Behavior of Nickel in 0.15 N Na2SO4............
Anodic Behavior of pH = 4.0..........................
Anodic Behavior of pH = 8.0..........................
Anodic Behavior of pH = 12.0.........................
Anodic Behavior of Nickel in 0.15 N Na2SO4 Containing
Chlorides ..............................................
Anodic Behavior at pH = 4.0..........................
Anodic Behavior at pH = 8.0..........................
Anodic Behavior at pH = 12.0.........................


Page
... i

iv


... 15

...18
... 20


... 24
...28
...31
...35

...38
...38
...41
...41
... 43

...44
...44
...44
.. .51
...66
...66
... 76
...81


...89
...89
...91
...97








CHAPTER Page

4 RESULTS AND DISCUSSION (continued)
Anodic Behavior of Nickel 30 Copper in 0.15 N Na2SO4.........97
Anodic Behavior at pH = 4.0................................... 97
Anodic Behavior at pH = 12.0.................................. 112
Summary.....................................................112

5 CONCLUSIONS AND SUGGESTIONS FOR FURTHER RESEARCH..............117

BIBLIOGRAPHY.........................................................121

BIOGRAPHICAL SKETCH..................................................126









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


AN IN-SITU INVESTIGATION OF THE ANODIC BEHAVIOR OF
Ni and Ni-30 Cu USING ELECTROCHEMICAL AND OPTICAL TECHNIQUES



By

Randall J. Smith

December, 1984


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


Although the passivation of nickel was observed by Faraday

over a hundred years ago, the composition of the passivating film

still has not been conclusively determined. This investigation

was undertaken, therefore, to positively identify the film respon-

sible for the passive behavior of nickel and to illustrate the

mechanism of film formation. Also, of particular interest were

the possible roles that alloying additions of copper and the

presence of chlorides in solution might have on the stability of

this film. To accomplish these goals standard electrochemical

polarization techniques were used in conjunction with ex situ

surface analytical techniques such as electron spectroscopy for

chemical analysis and Auger electron spectroscopy, as well as

Differential Reflectometry, which is capable of film identification

both in situ and ex situ.








On the basis of this research it is concluded that the film

which causes nickel to passivate in 0.15 N Na2SO4 (pH = 4.0 12.0)

is Ni(OH)2. It is suggested that the film forms via a precipitation

mechanism in the acid solutions while a solid state mechanism is

operative in the alkaline solutions. Both NiO and NiOOH are also

observed to form subsequent to Ni(OH)2 but only at the higher pHs.

The film responsible for the passivation of the Ni-30 Cu alloy

is again Ni(OH)2. In addition to dealloying, small amounts of Cu20

are detected to form in the acid solutions. However, the film that

passivates this alloy at the alkaline pHs is pure Ni(OH)2. Although

the presence of a 0.1 M chloride concentration does not alter the

composition of the film that forms on pure nickel, it does delay the

growth of this film at pHs of 4.0 and 8.0. In fact, in these solu-

tions film breakdown as well as severe pitting occurs subsequent to

passivation. However, at a pH of 12.0 the chlorides affect neither

the film formation process nor film stability.








CHAPTER 1
INTRODUCTION



Nickel and nickel base alloys are widely used to prevent or limit

corrosion in many environments. Nickel is quite resistant to alkalies

and is often used to solve corrosion problems involving caustic

solutions. Organic acids and compounds do not generally attack nickel;

this leads to its widespread use in the food industry. Nickel also

shows good resistance to neutral and slightly acid solutions but is

attacked by strongly oxidizing acids [1].

However, nickel is most commonly used in the form of an alloying

ingredient. While markedly improving the corrosion resistance, the

mechanical properties of the alloy are often improved as well. Nickel

is also metallurgically compatible with a wide range of important

elements (e.g., Cu, Cr,Fe, Mo, W) and can be used to bring together in

one alloy elements that are otherwise mutually immiscible (e.g., Cu and

Cr).

Perhaps the most commercially important nickel base alloys are the

Monels,* a class of alloys containing approximately 30% Cu along with

varying amounts of secondary additions (e.g., 1% Fe). These alloys,

while retaining the natural corrosion resistance of pure nickel to

alkalies, have improved resistance to chloride ions and reducing media.

They are commonly used to handle sulfuric acid, dilute hydrochloric and

hydrofluoric acid of all concentrations. They also find considerable

use in flowing seawater [2].




*Trademark of the International Nickel Company
1





2

Although the standard reduction potential of nickel (-.25 VSHE) in

the electromotive force series is less noble than that of hydrogen,

nickel often passivates in many deareated acids and alkalies. This

somewhat unexpected behavior is commonly believed to be the result of a

film formed at the metal-solution interface. The passive state of

nickel was observed by Faraday in 1844 and the concept of a passive film

was put forth at this same time; however, the exact composition and

nature of the film has still not been conclusively determined despite

literally hundreds of papers published on the topic [3,4]. Indeed,

without a basic understanding of the nature and composition of the

passivating film, it is difficult, if not impossible, to develop a means

of perfecting the performance of this film or to invent ways of

improving its stability in hostile environments (e.g., chloride

solutions).

The reason that an understanding of this passivating film has so

long eluded researchers is that to date no technique has been available

which can provide data suitable for making conclusive, unambiguous film

identification. Perhaps the most common method of film identification

uses thermodynamic parameters in an attempt to explain certain features

of experimentally obtained polarization curves. However, not only are

the thermodynamically calculated equilibrium potentials for the various

nickel oxides and hydroxides questionable [5,6], but also an objection

must be raised on the basis that, although a given film may at a certain

potential be thermodynamically possible, it may not form in sufficient

quantities for detection because of kinetic limitations. In addition,

it must be recognized that a considerable overpotential may be required

for the formation of those films that are not thermodynamically








prohibited. The combination of the debatable equilibrium potential

along with unknown required overpotentials makes film identification

using this technique rather tenuous.

A second method that has become increasingly popular with the

advent of surface sensitive analytical techniques is identification of

the corrosion film ex situ. After potentiostatically inducing

passivation in a suitable solution, the metal sample is removed and then

inserted into a vacuum chamber. One of several techniques (e.g.,

Electron Spectroscopy for Chemical Analysis (ESCA), Auger Electron

Spectroscopy (AES)) can then be used to unambiguously determine film

composition. However, upon removing the sample from solution and the

subsequent rinsing, drying and exposure to high vacuum, it cannot

reasonably be assumed that the film composition and/or thickness has not

changed. This uncertainty must be reflected in the conclusions drawn

from these data.

One in situ technique capable of providing corrosion data has been

used with increasing frequency over the past 25 years--ellipsometry.

The basis for ellipsometric studies is the influence that various

corrosion films have on the relative phase retardation and relative

amplitude dimunition of incident elliptically polarized light. Although

automated ellipsometers have simplified data collection, interpretation

of these data is by no means straightforward [7]. Often simplifying

assumptions are necessary in order to facilitate interpretation of these

data (e.g., assuming a smooth substrate), but the nature of these

assumptions renders many of the subsequent conclusions suspect.

A second in situ technique developed more recently is the A-C

impedence method. This method yields the value of the polarization








resistance which can be useful in determining corrosion mechanisms and

in corrosion rate monitoring [8]. However, this method is of no value

in determining film composition.

Recently a third in situ technique has been developed that allows

not only unambiguous film identification but at the same time is capable

of continuously monitoring corrosion film thickness. This technique is

based on the Differential Reflectometer (DR), which employs a

monochromatic light beam that is alternately reflected (at 60 Hz) from

two adjacent samples in solution, one corroding, one not corroding (the

standard). The intensity of the reflected light from each sample is

measured using a photomultiplier tube and this signal is subsequently

electronically processed to yield a plot of AR/R as a function of

incident light wavelength, where AR is the difference in reflectivity

between the two samples and R is the average of the reflectivity of two

samples. The monochromatic light is scanned from 200 nm to 800 nm in

about two minutes. Because of the differential nature of this technique,

the spectral response of both the cell window and the electrolyte

cancel. The operation of the DR has been discussed in detail elsewhere

[9,10].

Since each type of corrosion film has a unique band structure, the

spectral response of each film is different which is evidenced in the

reflectograms by a single peak or more often by a series of peaks. From

the magnitude of a given peak the film thickness can often be

determined.

Although the DR has an unequalled capacity for in situ film

characterization, the limitations of this technique should also be

considered. First, in order to be able to correlate a given set of






5

peaks in a reflectogram with a specific film composition, an independent

method of film identification must be used. Second, an independent

method must also be used to initially correlate a given peak height with

the film thickness. Third, it is necessary to maintain one sample in

solution at a standard reflectivity during the course of the experiment;

i.e. one sample must remain uncorroded while the second sample is

growing the corrosion film(s). These limitations are not nearly as

ominous as they might seem at first and the methods of dealing with them

are discussed in detail in the Experimental Procedure section.

The object of the present research is to obtain a general

understanding of the anodic behavior of nickel in aqueous solution

through corrosion film identification. Of particular interest are the

processes and mechanisms that induce and destroy passive behavior. It

is also of interest to establish an experimental basis for further

research concerning the role of copper in the passivation of nickel and

the role of chloride ions in the breakdown of this passive state.








CHAPTER 2
LITERATURE REVIEW


The Corrosion Behavior of Pure Nickel


Almost all corrosion studies directed at explaining the behavior of

nickel in aqueous solutions employ electrochemical techniques in

conjunction with independent analytical techniques such as ellipsometry,

impedence measurements, ESCA, etc. Some studies draw only on a

comparison of electrochemical behavior with thermodynamically predicted

reaction values in an attempt to determine the processes that occur at

the electrode-solution interface. However, some authors claim to have

identified the complete reaction scheme at the electrode interface,

including film compositions, using only electrochemical data.

In one such paper [11], MacDougall and Cohen anodize Ni in .15 N

H2SO4 (pH = 2.8) for short times--typically 90 seconds, and then monitor

the galvanostatic reduction of the surface films. Two arrests appear on

the reduction curve. The major arrest is attributed to NiO reduction

and the surface activity is recovered during this arrest. The rate of

recovery slows down after completion of this arrest leading the authors

to conclude that two different oxide species are present. Referencing

the behavior of Ni in vacuum, they conclude that oxygen chemisorbs onto

the surface as the first step in passivation. This chemisorbed

monolayer converts to NiOadsorbed and finally to NiO. The second

arrest, then, is due to the reduction of some unconverted, chemisorbed

oxygen. However, an arrest on a galvanostatic reduction curve merely

indicates that a charge consuming reaction is occurring at the surface.







The development of a detailed, mechanistic theory from such data alone

seems unjustified.

A second study [12] based solely on the electrochemical data was

performed by Vilche and Arvia in 1N NiSO4 + .05N H2S04, solution (pH =

1.75). Again, based on polarization curves alone, they conclude that

three films form on the surface. The first film, Ni(OH)2, is a pre-

passive film and is formed through the NiOH+ intermediate. The second

film, NiO, is formed by Ni(OH)2 dehydration while the final and truly

passivating film, Ni203, forms from Ni(OH)2 via the NiOOH intermediate.

Considering the limitations of polarization curves, these conclusions

should be viewed with suspicion.

Perhaps the uncertainty involved in interpreting polarization data

alone is best exemplified by examination of a paper by MacDougall and

Cohen [13] and a second paper by Vilche and Arvia [14]. In the paper by

MacDougall and Cohen, a series of polarization curves are obtained for Ni

in sodium sulfate solution (pH = 3.0) after cathodic pretreatments at

various N2 bubbling rates. As the bubbling becomes more vigorous or at

longer polarization times, the active-passive peak diminishes and a new

peak appears and grows at a more cathodic potential. The authors conclude

that this is the result of impurities in the solution adsorbing onto the

surface and blocking the normal film formation process. In a subsequent

paper Vilche and Arvia observed this same type of behavior but as a func-

tion of various sweep rates. They conclude that the "adsorption-of-

impurity" theory is an inadequate and hasty interpretation and point to the

"complexity of kinetic results. .and the difficulty of deriving from them

prompt mechanistic interpretations which are not fully substantiated."

[page 1063] It is interesting to note that this statement was made







by the same authors who used polarization curves alone to defend the

argument that Ni(OH)2, NiO, Ni203, NiOOH and NiOH+ are all formed on the

Ni surface.

A series of three papers by Okamoto and Sato [15-17] are perhaps

the most often referenced studies on the anodic behavior of Ni contained

in the corrosion literature. In the first paper of this series the

authors observe three distinct arrests on the potential decay curve of

Ni in H2SO4 (pH = 4.5). The first arrest has a corresponding potential

of E = .103 VSHE and compares favorably with the reversible potential

for .081 VSHE for the reaction


Ni + H20 NiO + 2H+ + 2 e".


The second and third arrests occur at E2 = .453 VSHE and E3 = 1.573 VSHE

and are stated to compare favorably with the thermodynamically expected

values for the reactions


E2 : 3NiO + H20 Ni304 + 2H+ + 2e

and

E3 : 2Ni304 + H20 3Ni203 + 2H+ + 2e-.


These values, however, are never actually stated for comparison.

The values found in Pourbaix's Atlas [18] for these reactions are E2

= .870 VSHE (cf .453) and E3 = 1.273 VSHE (cf 1.573) which differ

significantly from the experimental arrest values. The authors consider

the second arrest to correspond to the Flade potential and therefore

conclude.that the passivation of Ni is the result of a transformation of

NiO (prepassive film) to Ni304 (passive film).







Building on these somewhat questionable conclusions, Sato and Oka-

moto in the second paper [16] suggest the mechanisms by which the

various films are formed. They conclude that both NiO and Ni304 are

formed through NiOH intermediate. This argument would be more

persuasive if they had presented evidence for the existence of this

species.

In the third and final paper qualitative arguments are given to

show that passivation of nickel cannot occur either by a film

precipitation mechanism or by a direct oxidation film mechanism.

Instead the authors advance a new theory based on their two previous

papers, which they call "Higher Valence Oxide Film Theory." This theory

states that passivation can only occur via the conversion of a surface

intermediate, such as the NiOH+ complex. Unfortunately, again no

analytical evidence for the existence of NiOH+ was provided. Hence the

support for this theory rests solely on the correlation of a potential

arrest with an unstated reversible potential.

Cowan and Staehle [19] also attempted to identify surface corrosion

films by comparison of thermodynamically calculated reversible

potentials with the experimental arrest potentials obtained for Ni in

.050 N H2SO4 + 0.5 N K2SO4 (pH = 1.3) and .001 N H2SO4 + .099 N K2SO4

(pH = 3.4) solutions. Although the principal concern of their paper was

the establishment of high temperature Pourbaix diagrams, they conclude

from the single arrest observed in room temperature solutions that the

film is Ni304 since the experimental arrest was consistently within 20

mV of the thermodynamically predicted values.

Using a .15 N Na2SO4 solution (pH = 2.8), Mitrovic-Scepanovic and

Ives [20] obtained a potential plateau at -300 mVSCE which corresponds

well with the thermodynamically predicted reversible potential of -299.4








mVSCE which they quote for the Ni/NiO electrode. Other plateaus were

observed in addition to the one at -300 mVSCE but they did not lie close

to equilibrium values for any of the Ni/NiOn or NiO /NiOm electrodes.

Gromoboy and Shreir [21] conclude that up to four Ni oxides are

possible for a Ni electrode in a sulfuric acid solution (pH = 1.3)

containing various amounts of thiourea. Instead.of using potential

decay curves, the authors determined the experimental reversible

potentials from inflections in the potentiostatic and galvanostatic

polarization curves. The four inflections noted all occur within

millivolts of the reversible potentials for Ni/NiO, Ni/Ni304, Ni/Ni203

and Ni/NiO2. All four of these inflections were never observed on a

single polarization scan but rather on curves obtained in solutions of

various thiourea concentrations. Also it could be argued that many more

plateaus exist on the galvanostatic curve than were noted and explained

by the authors. Although the excellent agreement between inflections on

the polarization curve and the various reversible potentials is

remarkable, one must wonder what effects the thiorea have on the actual

film formation processes' and compositions. Unfortunately, no surface

analytical data were presented.

The adaptation of ellipsometry to in situ studies represents a

major advance in the methods available to the corrosion scientist.

Although only one paper was found which reported utilization of the

ellipsometer to study the anodic behavior of Ni in acid solutions [22],

it is regarded by many as a classic. The basis for the study is the

comparison of a polarization curve obtained in .01 N H2SO4 + .5 M K2SO4

(pH = 3.15) with ellipsometric data obtained concurrently. A

correlation is found between the onset of film formation (as evidenced





1I
by ellipsometry) and a slight inflection in the polarization curve. It

is suggested by the authors that since passivation does not occur

immediately that this film must be merely a "precursor film." Then, at

the current maximum, the absorption coefficient, K, begins to increase

indicating an increase in film conductivity which becomes roughly equal

to that of a semiconductor. From comparison of refractive indices, it

is concluded that this passivating film has a composition of NiO1.5-1.7.

Based on these data, Bockris and Reddy describe the passivation process

as follows: a thick (> 45 A) precursor film (Ni(OH)2) forms via a

precipitation mechanism, this film is subsequently converted to a

nonstoichiometric film at the passivation potential. Since the

electrical conductivity of this nonstoichiometric film is much higher

than that of the precursor, there exists no high field ionic

conductivity through this film--hence passivation.

The conversion of an insulating film to an electrically conducting

film as the cause of passivation is a unique explanation and capable of

explaining passivation behavior. However, several questions must be

raised as to the validity of the data upon which this theory is based.

First of all, in order to determine film thickness, certain assumptions

must be made as to the ranges of values for the refractive index, n, the

absorption coefficient, K, and even the composition of the film. These

values are then used in an iterative process that yields various A-i

plots, to which the experimental values must be compared for a "best

fit." This process led the authors to conclude that the film thickness
0
was approximately 60 A--a value which is much higher than generally

reported in the literature [13]. Perhaps one variable that was not

considered that could account for such an abnormally high value is the







roughening of the electrode surface during active dissolution. Such

roughening has been reported to influence ellipsometric data

significantly [7].

Secondly, it is not surprising that a film appears on the electrode

prior to passivation. It does not seem necessary that such a film be a

precursor that must be converted, but rather the passivation film could

form initially on grains more favorably oriented for such a process to

occur while less favorably oriented grains are still experiencing active

dissolution. In fact, there is evidence that film kinetics are indeed a

function of crystal orientation [23,24].

Finally, it is stated that ellipsometry is suitable for detecting

even monolayer changes. However, the authors then suggest


Ni + OH NiOHadsorbed + e


as the reaction that leads to the electrode dissolution, but there is no

change in ellipsometric response accompanying this proposed reaction.

It would therefore seem necessary that more substantial evidence be

obtained before giving credence to the ionic-electronic conduction

conversion theory.

Although impedence measurements do not result in data that can be

directly used for film identification, nevertheless it has become a

useful technique for determining important film characteristics.

Lovrecek and Lipanovic [25] concluded from a study in 1N H2SO4 that the

impedence response of the Ni electrode indicates that the surface film

cannot be electrically represented as a serial joining of resistive and

capacitive components. This indicates that relaxation phenomena must be

occurring within the film which suggests that the film is not merely an







insulator or simple semiconductor but, according to the authors, a

"quite complex film." However, a complex anodization scheme may be

partially responsible for their results. The electrode was initially

anodized at +1800 mVSCE (well above the oxygen evolution potential) for

15 minutes and then held at +450 mVSCE for ten days. This unusual

experimental procedure makes it difficult to compare their data with

that found elsewhere in the literature.

Turner, Thompson and Brook [26] also employed AC impedence

techniques to study the Ni electrode but in concentrations of H2SO4

varying from .5 18.8 M. The impedence data they obtained suggested

that the passivation film is thinner in the more highly concentrated

solutions. They also determined that Ni dissolves as a divalent ion in

the active region while passivation.is accomplished by a solid state

mechanism. This latter conclusion is in direct contrast with the

mechanism proposed by Bockris and Reddy [2.7].

The use of ex situ film identification procedures has also become

increasingly popular in the last ten to fifteen years. MacDougall and

Cohen [28] were able to elucidate several mechanisms responsible for the

electrochemical behavior of pure Ni. From cathodic decay curves in a

solution of pH = 2.8 obtained after various anodic treatments, the

authors deduced the following model: oxygen absorbs onto the surface as

a monolayer while NiO is concurrently nucleating at "active sites." The

NiO eventually spreads to complete coverage but with pores of "non-

stoichiometric or noncrystalline" NiO. This pore material subsequently

converts to stoichiometric NiO at more anodic potentials which accounts

for the small current passed in the passive region. The breakdown of

the film on the cathodic cycle occurs by dissolution of the pores and





14

then "undermining" of the remaining NiO. The authors also conclude,

from x-ray spectroscopy, that the anodic film thickness and composition

are independent of the anodic polarization potential.

It is somewhat surprising that the cathodic arrests are not

interpreted as being the interconversion of higher oxides, as previous

authors have suggested, but rather the dissolution of pore material.

Although film composition and thickness have been determined using ex

situ methods, the mechanistic model itself is based solely on the

electrochemical response of the electrode. There is no independent

analytical evidence offered to confirm the existence of the chemisorbed

oxygen, pores, non-stoichiometric pore material or undermining.

In the third paper concerning the anodic behavior of Ni, MacDougall

[29] concludes from galvanostatic charging profiles in conjunction with

solution analysis that film growth accounts for but a minor component of

the total anodic current. The majority of the current is consumed in

the production of the divalent nickel ion in solution. After the

passive film has formed, a residual film current persists and the

electrode potential increases. Citing the x-ray data which indicate no

film thickening, the author concludes that the residual current is being

used to repair defects in the passive film which also accounts for the

accompanying potential increase.

In yet another paper by MacDougall et al. [30], still more surface

analytical techniques are used to study the passive film. In .15 N

Na2SO4 solution (pH = 2.8), reflection high energy electron diffraction

(RHEED) indicates that the film is stoichiometric NiO having a thickness
0
of 9-12 A, which agrees with the values as determined from both

coulometry and 0 x-ray spectroscopy. A limiting thickness of 12 A is

noted even after 200 hour anodizations. In the transpassive region, the








film is no longer single crystal epitaxial but rather defective and

polycrystalline, as indicated by RHEED. It must be remembered, however,

that there was no way the authors could monitor changes that might occur

in the film upon removal from solution and placement in vacuum.

In a unique approach, Siejka et al. [31] used a 1 N H2S04 solution

enriched with 018 to passivate a Ni electrode and then examined the film

ex situ with nuclear microanalysis to determine the 018 content. They

concluded that the first layer of oxygen fixed on the surface causes the

anodic current to decrease by two orders of magnitude. After "unit cell

thickness" is reached the current is limited by field-assisted ionic

transfer. Current efficiency for film growth is low and growth seems to

occur at the metal-oxide interface by in-diffusion of oxygen. Upon

removal from solution the oxide layers are unstable and reoxidize.

Although the authors draw no conclusions as to the composition of the

passive film, the fact that the film is formed at the metal-oxide

interface excludes precipitation as the passivation mechanism. Also,

the evidence indicating that the ex situ film may not be in the same

oxidation state as the in situ films confirms that conclusions drawn

from ex situ film examinations may not be accurate.



Behavior of Nickel in Neutral Solutions


The behavior of Ni in neutral solutions has also been of

considerable interest to a great number of researchers. Sato and Okomoto

[15-17] are again responsible for the most often referenced work on this

topic. They employed a solution of of .5 M Na2SO4 + .1 M (KH2PO4 + K2

HPO4) (pH = 6.6) to obtain potential decay curves for a Ni electrode.







Comparison of the arrest potentials with calculated thermodynamic

reversible potentials led the authors to conclude, just as they had in

acid solutions, that three films are formed: NiO, Ni304 and Ni203. The

experimental arrest values are approximately -.2, +.1 and +1.1 VSHE

which do not compare well with calculated values of -.28, +.50 and +.91

VSHE. As in the acid solution, they attribute passivation to the

conversion reaction of NiO to Ni304.

Cowan and Staehle [19] concluded that the passive film formed on Ni

in .1 N K2S04 (pH = 6.3) is NiO. Their conclusion is also based on

comparison of an experimentally obtained arrest potential (-.204 VSHE)

with a calculated redox potential (-.254 VSHE). These values compare

only slightly more favorably than those of Okomoto and Sato. However,

in contrast to Okomoto and Sato, Cowan and Staehle do not observe any

arrests at more anodic potentials.

Okuyama and Haruyama [32], using a boric acid-borate buffer

solution (pH = 8.39), claimed to identify no less than three surface

oxide films formed on a Ni electrode. By comparing plateaus on both

anodic and cathodic charging curves with calculated redox potentials,

the authors concluded that at low potentials NiO forms while at

intermediate potentials Ni304 is formed. Both films are formed via a

direct oxidation mechanism. At higher potentials Ni304 transforms

completely to NiO2, which prohibits further divalent ion dissolution,

i.e., this conversion reaction is responsible for Ni passivation.

Ord et al. [33] studied the passive behavior of Ni in .15 N Na2SO4

solution (pH = 6.1 9.1) using ellipsometry. They reach the conclusion

that NiO is the initial film formed on an anodic potential sweep but

that this film converts to Ni(OH)2 and eventually to NiOOH as the








potential progresses to more anodic values. Also concluded is that the

film thickness is a direct function of electrode potential (contrary to

the finding of MacDougall and Cohen in ref. 34) and that film growth

is limited by the electric field within the film. The authors did not

complement these data with electrochemical measurements; they are

therefore unable to determine which film is responsible for passivation.

Also not discussed is the effect that surface roughening may have on

their results.

In a series of papers MacDougall et al. [27,30,34-37] arrive at

several conclusions regarding the anodic oxidation of Ni in both sulfate

and borate buffer solutions. The authors state that the film formed

in .15 N Na2SO4(pH-= 8.4) solution at potentials cathodic to the oxygen

evolution potential is NiO (determined by AES) having a maximum
0
thickness of 12 A. This film was said to be identical to the film they

found in acid solutions. In borate buffer solutions the film is again

determined by AES and RHEED to be NiO. At potentials anodic to the

oxygen evolution potential, this film becomes polycrystalline and highly

defective. However, at potentials anodic to oxygen evolution, a duplex
0
film is found to be somewhat thicker--about 16 A.

MacDougall et al. offer the following mechanism to explain the film

formation process in the borate buffer solution. Initially, the clean

Ni surface is oxidized to produce a passivating film of NiO. This film

is formed by a direct oxidation mechanism rather than by precipitation.

The NiO then begins to break down at defects with subsequent metal

dissolution. The divalent Ni ions that result from this dissolution are

oxidized at the oxygen evolution potential and precipitate out of

solution as a NiOOH film. The NiOOH film converts to Ni(OH)2 upon







cathodic polarization. No experimental evidence is provided to confirm

the existence of the Ni(OH)2. These authors propose that NiO cannot be

converted to higher oxides--an assumption that ultimately necessitates

the adoption of a comparatively complex reaction scheme, rather than the

simple conversion process suggested by others [15,22].

In a clever attempt to characterize the passive film on Ni in

neutral solution, Wilhelm and Hackerman [38] measured its

photoelectrochemical response. Because the absorption edge of the film

(3.7 eV) is the same as the known band gap of NiO (3.7 eV) the authors

conclude that the passive film is indeed NiO. From integration of the

anodic charging profiles it was also concluded that this passivating
0
film is 10 A thick. No higher oxides were detected by this technique.



The Behavior of Ni in Alkali Solutions


Sato and Okamoto [15] note three arrests on a potential decay curve

for a Ni electrode in .05 M Na2SO4 + .01 NaOH solution (pH = 11.7). As

in the acid and neutral solutions, the authors conclude that these

arrests are due to the formation of NiO, Ni304, and Ni203.

Unfortunately, the agreement between the experimental and calculated

potentials for these films is not convincing. In fact, a potential

arrest is not reported for the Ni/NiO reaction but nevertheless the

authors state that this reaction still occurs.

Davies and Barker [39] examined the behavior of Ni in .1 N NaOH

solution (pH = 13). By comparing the galvanostatic charging plateaus

with thermodynamic redox potentials, the authors conclude that Ni(OH)2

is formed. The Ni(OH)2 is subsequently converted to Ni203 and finally

partial conversion of the Ni203 to NiO2 results in a final composition








of NiO18. Based on current measurements, the Ni(OH) is determined to

be one monolayer in thickness while the Ni203 has a three monolayer

thickness. However, it is not considered that a component of the

current must be attributed to metal dissolution prior to passivation.

The final composition of NiO1.8 is determined by coulometry based on the

assumption that all current passed is used in the conversion of Ni203 to

Ni02--the possibility of new film growth is not discussed.

Using only a potentiodynamic polarization curve, Schrebler-Guzman

et al. [40] attempt to explain the behavior of Ni in .2 N KOH solution.

They note two peaks on this curve. The first peak (the common

passivation peak) is attributed to the formation of Ni(OH)2 through a

complex adsorption process. The second peak occurs just prior to oxygen

evolution and results from the conversion of Ni(OH)2 to B-NiOOH. The

compositions of these films are assumed without explanation.

T. S. Lee [41] used polarization curves as the basis for the

construction of an experimental Ni-H20 Pourbaix diagram. Agreement

between the experimentally determined and equilibrium (calculated)

diagrams is not generally obtained. The only good correspondence

between the kinetic studies of Lee and the original calculated diagram

of Pourbaix is between the zero current potential and the predicted

Ni/Ni(OH)2 equilibrium potential. There exists an obvious discrepancy

between the actual passivation potential and that predicted from

thermodynamic considerations. Also, the region of general corrosion is

found restricted to much lower values of pH and somewhat higher

potentials than predicted by Pourbaix. Lee was not successful in

identifying the films using x-ray techniques.





20

Hopper and Ord [42] used ellipsometry in an attempt to characterize

the film in 5 N KOH solution. They conclude that the film is B-NiOOH

and is formed by a direct oxidation mechanism. The authors state

candidly that unambiguous interpretation of ellipsometry data, including

their own, is generally impossible.

Ellipsometry was also the basis for a study by Lu and Srinivasan

[43] concerning the behavior of a Ni electrode in 1 N KOH electrolyte.
0
They concluded that a 190 A film of B-NiOOH is formed upon anodization

above the oxygen evolution potential. The authors note, however, that

it was necessary to lower the electrode potential below the oxygen

evolution potential in order to obtain their data. Therefore, if a

reversible conversion process is occurring at or near this potential it

may not be detectable because of this procedure.



The Transpassive Behavior of Nickel


At potentials anodic to the passive region,the protective film on

Ni often breaks down. This phenomenon is referred to as

transpassivation and usually occurs concurrently with oxygen evolution.

Although there is not general agreement as to the composition and

formation mechanisms of the passivation film, several studies have been

conducted in an attempt to understand the processes leading to the

demise of this film in the transpassive region.

Expanding upon an earlier paper, MacDougall and Cohen [44] discuss

the results of a cathodic charging profile for a Ni electrode in .15 N

Na2SO4 solution (pH = 2.8). After the potential has decayed through two

plateaus, a potential spike in the cathodic direction is observed.

Noting that the surface activity of the electrode increases from less







than 2% before the spike to 10% afterwards and that scanning electron

micrographs indicate the onset of pitting at this same potential, the

authors conclude that the spike results from localized breakdown of

highly defective sites within a NiO film. Once the breakdown has

occurred the dissolution of the Ni substrate proceeds, leading to the

undermining and eventual spelling of the remaining NiO layer. The

authors preference for this undermining and spelling scenario as opposed

to simple film thinning via dissolution and/or conversion is based on

the behavior of two electrodes upon different cathodic treatments. Both

samples were anodized under identical conditions. The first sample

underwent a 20-minute galvanostatic cathodic charging (20 iA cm2)

treatment immediately after anodization while the second sample received

this same cathodic treatment but only after it had decayed on open

circuit up to the point of the spike (ca. 30 min.). Since the final

recovered surface activities of the two electrodes differ by only 10%,

the authors conclude that the film thicknesses are identical, i.e., no

thinning of the film occurs on open circuit. An alternate explanation

may be that the processes that occur during the relatively severe

cathodic treatment predominate over those processes that may occur

during the relatively gentle open circuit treatment. No interpretation

is given by MacDougall and Cohen regarding the two potential plateaus

also present on the reduction curve.

From optical and electron microscopy studies on Ni electrodes in

sodium nitrate solutions,Datta and Landolt [45] also conclude that

transpassive behavior is attributable to localized breakdown of the

passive film. The breakdown is initiated at defects within the film

(e.g., grain boundaries and dislocations) resulting in a pit-like








appearance in the early stages of the process. Once local breakdown is

achieved the pit-like perforations expand allowing high rates of metal

dissolution. This overall process yields a severely etched metal

surface. The authors also conclude that dissolution occurs much more

rapidly at higher current densities and is a function of grain

orientations, and the process is virtually pH independent. Nickel 200

was used as the working electrode in these experiments. The authors do

not discuss the possible impurity effects on film breakdown.

In a subsequent study Landolt [46] employed AES to analyze the film

at various stages of transpassive breakdown. As in the case of iron and

chromium, the apparent anodic film thickness reaches a maximum at the

beginning of the transpassive region. The film thickness then decreases

as the potential progresses to more anodic values. Nitrogen is detected

within the transpassive film and its abundance is found to be a direct

function of increasing potential. The nitrogen is always found in its

highest concentrations at the metal-film interface. It is speculated by

the author that the nitrates diffuse through the film via micropores or

other defects and adsorb onto the metal surface. This process occurs to

a greater extent at higher potentials and eventually leads to the

destruction of the passive state. Neither thinning nor how the adsorbed

nitrates destroy the film is discussed.

Sato [47] presents an energetic approach to study transpassive

film behavior. This model is based on the formation of a cylindrical

pore of radius r. The work required to form such a breakthrough pore in

a passive film has two components as represented in the equation


Ab = {2Trha + r2(om -o)} .5r2CdAE2








where A = pore formation energy

h = film thickness

a = surface tension of the film-electrolyte interface

om = surface tension of the metal-electrolyte interface

Cd = difference in capacitance between the passivated metal

and metal with a porous film

AE = electrical potential difference between the metal and

electrolyte.


The first term on the right side of the equation represents the

variation in capillary energy which is dependent on surface tension

while the second term represents the variation of electrical energy due

to the potential field across the film. From equation (1) the author

derives


m -
AEb
b .5 Cd


where AEb is the lowest possible film breakdown potential. Based on an

analysis of this model Sato arrives at two major conclusions. First, it

is suggested that a potential range exists between the film formation

potential (i.e., passivation potential) and the lowest possible

breakdown potential (AEb) where the film is stable against

electrocapillarity breakdown. Secondly, from the viewpoint of film

dissolution kinetics, this model predicts that a critical potential

exists above which the passive film is electrochemically unstable (i.e.,

the transpassive potential). However, actual values are never

determined for AEb so that a comparison can be made directly with

experimental values.







The Effect of Chloride Ions on the Passive Behavior of Nickel


As just discussed, if the electrode potential is made too anodic

passive behavior can be lost resulting in extremely high rates of metal

dissolution. Passivity can also be destroyed by the addition of

aggressive anions to the electrolyte, most notably Cl- ions. This loss

of passivity can occur even though the electrode potential is well

within the normal passive range. Just as there is much confusion

regarding the nature of the passive film itself, there is again

considerable disagreement regarding the processes leading to chloride-

induced film instability.

T. S. Lee [41] reported that the presence of C1- ions in solution

produces an enlargement of the corrosion region to higher pH values and

more active potentials. The effect of the C1 is also manifested by

continued corrosion at quite noble potentials due to the absence of a

passivating film. Lee was not able, however, to identify the mechanism

by which the C1~ ions alter the corrosion process.

Kronenberg et al. [48] studied the behavior of a 99% pure Ni wire

in 1 N NaC104, 1 N NiSO4 and 1 KC1 solutions. From the Tafel slopes of

the experimental polarization curves the authors conclude that C1 ions

adsorb onto the surface and enhance the transfer of Ni from the metallic

state to an adsorbed state from which dissolution occurs more readily.

No further justification is given for this theory.

Based on stationary potentiostatic methods, potentiodynamic single

sweeps and triangular cyclic voltammetry, Vilche and Arvia [49] also

conclude that C1- ions adsorb onto the surface at low potentials within

the active dissolution range. Then, at more anodic potentials within

the active range, the surface becomes saturated with C1 ions resulting








in the formation of a layer of NiC12. No analytical confirmation of

such a NiC12 layer is provided, however.

Zamin and Ives [50] obtained polarization curves for a Ni electrode

in 1 N H2SO4 solution containing various amounts of 1 N NaC1. They

discovered that as the C1- ion concentration increased the passive

region on the polarization curve decreased and that in solutions with a

high enough Cl- ion concentration passivity was altogether unobtainable.

From accompanying metallographic observations, it was further shown that

in solutions of low C1- ion concentration the passive film breakdown is

initiated at the grain boundaries. However, at moderate concentrations

the attack is no longer localized at grain boundaries but rather occurs

in the interior of the grains as well. The authors conclude that

passivity is governed by two competing processes. The process of film

formation which occurs unhindered in C1 free environments is opposed by

film breakdown when C1- ions are present. In low concentrations film

formation is the predominant process and passivation proceeds, while at

higher C1- ion concentrations film breakdown occurs rapidly and

passivity is precarious. Once a critical Cl- ion concentration is

reached film breakdown occurs more easily than film formation at all

potentials and, as Lee [41] found, passivity is unobtainable. Although

this analysis is coherent and agrees well with their experimental data,

a mechanistic explanation is still lacking.

In a subsequent paper, Zamin and Ives [51] again utilize

polarization data along with microscopy studies to conclude that the

variation in grain size, amount of cold work and annealing atmosphere

have no effect on the critical pitting potential of a Ni electrode.

However, they did find that pitting occurs only along grain boundaries









in small grain samples while in larger grain size electrodes pitting

occurs both at the grain boundary and interior of the grain. It also

seems that pit nucleation occurs during active dissolution before

passivation and that these sites are merely reactivated when the

potential of the electrode becomes more anodic than the critical

potential.

Based on the Tafel behavior of a Ni electrode, Bengali and Nobe

[52] concluded that the kinetics and mechanism of anodic dissolution are

strongly dependent on the concentration of both the H+ and C1- ions. At

low concentrations of these ions the dissolution proceeds via an

adsorbed NiC1OH complex while at higher concentrations dissolution is

dependent on the formation of a NiCIH intermediate. Again, no

analytical evidence is provided to substantiate the existence of these

intermediates. Rather, the assumed existence of these species is

predicated solely upon analogy to the iron-chloride system.

Contrary to the papers just discussed, MacDougall [53] concludes

that C1 ions have no effect on the dissolution mechanisms of Ni. This

is based on the observation that polarization curves are identical in

the active region of Ni electrodes in .15 N Na2SO4 solutions (pH = 2.8)

with and without Cl- additions. However, the passivation current and

total dissolution current are several times greater in the Cl

containing solution which suggests that the role of the Cl- is to hinder

passive film formation. The author speculates that there is a

competition between adsorption of Cl ions and the oxygen containing

species necessary for passive film formation.

Also, in solutions containing C1 ions, transpassive behavior is

initiated at substantially lower potentials. MacDougall explains this








by postulating that a dynamic equilibrium exists in the passive state

between film breakdown at defects and subsequent film repair. The role

of the C1- ion then is not to accelerate breakdown but rather to

physically interfere with film repair. This idea is consistent with the

already mentioned fact that passivation occurs at much greater current

densities while at only slightly higher potentials.

In a second series of experiments described in this same paper, Ni

electrodes were passivated in solutions with and without Cl ions and

the subsequent open circuit behavior observed in solutions containing C1

ions. It was found that breakdown occurs much more rapidly for those

films grown in the Cl- containing solutions. This suggests that the

presence of C1 ions in the formation solution substantially decreases

the stability of the passivating oxide. This decreased stability is

attributed by MacDougall to a more highly defective film structure

rather than by direct incorporation into the film. X-ray emission

spectroscopy indicates that the C1 ions in solution affected neither

film thickness (still 9-12 A) nor film composition (still NiO). No C1-

ions were detected in the film but as MacDougall points out the

detection limit for C-1 is about 10%.

Sato [47] expands the energetic approach of transpassivity

described in detail above to also explain the effects of C1- ions on the

loss of passivity. He suggests that the adsorption of anions, in

particular C1 introduces new electron acceptance levels into the band

structure of the passive oxide film which tends to lower the

transpassive potential. Thus, the dissolution rate is faster at the

anion adsorption sites.







Although there is no general agreement regarding the mechanism, all

authors do at least agree that the presence of Cl ions in solution has

a deleterious effect on the passivation process.


The Corrosion Behavior of Pure Copper


The nature of the protective film that forms on copper is much

better understood than the film on nickel. It appears that only three

films form on a Cu electrode and each is formed in a fairly well defined

potential-pH regime.

In dilute acid solutions Cu dissolves to yield soluble salts and

does not develop a protective passive film [54]. However, in strongly

concentrated acid solutions, Leckie [55] has shown that a passive film

does indeed form. In 10 M H2SO4 solution the passivation potential was

observed to be .21 VSCE which agrees quite well with the

thermodynamically predicted reversible potential of.23 VSCE for the

Cu/Cu20 reaction.

T. S. Lee [41] employed a potentiokinetic technique to obtain

polarization curves which he used in the construction of experimental

Pourbaix diagrams. The experimental diagrams turned out to be quite

similar to the equilibrium (calculated) Pourbaix diagrams. In addition,

it was found that primary passivation occurred at potentials very close

to the Cu20/CuO transition while secondary passivation occurred near the

Cu20/Cu(OH)2 boundary.

Perhaps the most complete and coherent work published concerning Cu

passivation was offered in a series of four papers by Shoesmith et al.

[56-59]. By comparing open circuit potential-time transients with x-ray

and electron diffraction data along with SEM micrographs, the authors








were able to make several well substantiated conclusions. First, they

were able to determine the potential regions of stability for the

various corrosion films for a Cu electrode in 1 M LiOH solution. For

electrodes potentiostated between -.330 VSCE and -.280 VSCE the

resulting film is pure Cu20. A mixed film of Cu20 and Cu(OH)2 is formed

at potentials from -.280 VSCE to -.265 mVSCE range while mixed Cu(OH)2 +

CuO forms up to -.060 VSCE. At potentials anodic to -.060 VSCE a pure

CuO film results.

Although the authors offer no conclusions concerning the mechanism

of Cu20 formation, they are able to determine the formation process of

the intermediate Cu(OH)2 film. It appears that the initial Cu(OH)2 film

is formed via direct oxidation of the metal surface while subsequent

layers of Cu(OH)2 are the result of precipitation from solution. These

conclusions are based on SEM micrographs which show that the initial

film formed in the Cu(OH)2 region has a very smooth appearance. After

sequentially longer polarization times, the micrographs indicate that a

three dimensional network of crystals nucleates and grows on top of the

initial film. Without explanation it is also concluded that the CuO

layer formed at somewhat more anodic potentials occurs by the

dissolution of the lower oxides(s) and subsequent precipitation of CuO.

Burstein and Newman [60] employ a scratch test in an attempt to

understand the nature of the Cu film. After an anodic film developed on

a potentiostated Cu electrode, the surface was scratched and the

resulting current fluctuation monitored. By comparing current plateau

values with the standard reversible potentials the authors were able to

"determine" the composition and thickness of the film. The electrode

behavior was the same in both neutral and alkaline solutions. It seems








that a monolayer of CuOH is first formed over the scratch while a second

monolayer of Cu20 eventually develops on top of that. It is stated that

both monolayers are formed at potentials well cathodic to the Cu/Cu20

film ultimately responsible for passivation. No external analysis is

made either of film composition or thickness and no explanation is

offered as to why Cu20 is formed below its equilibrium potential.

Shanley et al. [61,62] attempted to identify the corrosion product

on copper using the DR. Ex situ reflectograms were obtained for a Cu

electrode potentiostated in a borate buffer solution (pH = 9.2). When

held in the Cu20 region (as predicated by the Pourbaix diagram), the

resultant reflectogram displayed a series of peaks which corresponded

quite well with the peak energies predicted by the band structure of Cu20.

From this correlation the authors made the reasonable assumption

that this reflectogram was indeed that of Cu20. When the electrode was

potentiostated at somewhat more anodic values a second, distinct

reflectogram was obtained. Although the band diagram for CuO was not

available, the optical constants of CuO were well known. From these

constants a theoretical reflectogram was calculated which is

practically identical to the experimental curve. Hence the conclusion

that this second reflectogram is indicative of CuO. Attempts to verify

film composition using x-ray techniques proved inconclusive, presumably

due to the lack of sufficient film thickness.

It was also shown in this work that the DR is capable of not only

detecting pure oxides but also of indicating when a mixed oxide is

present. The reflectogram of a mixed oxide, e.g., Cu20 and CuO, is

merely the superposition of the reflectograms for the pure components.








This is an extremely fortunate circumstance in that it permits the

detection and monitoring of the conversion of one oxide into a second

oxide. This conversion may conceivably result from either a change in

the electrode potential or a change in the electrode environment, e.g.,

the addition of C1- ions or even removal from solution.

Lastly, it was shown that peak height in a reflectogram is a direct

function of film thickness. That is, as the corrosion film thickens the

peaks on the reflectogram show a corresponding increase in magnitude.

It should be pointed out that all of the reflectograms presented in

this work were obtained by covering one-half of the electrode face with

collodion. This was necessary in order to maintain half of the sample

at a constant reflectivity. Although experiments showed that the

collodion itself has no spectral response within the wavelength range

tested by the DR, its use nevertheless provides an undesirable variable

in the interpretation of the reflectograms. This is particularly true

for those reflectograms obtained after long polarization times when the

collodion is no longer as protective. Also it should be recognized that

although the in situ reflectograms for a specific oxide display the same

general characteristics (e.g., peak positions) as do the ex situ

reflectograms, the in situ curves invariably exhibit a noticable "noise"

component. This, presumably, is also a direct result of the collodion

which is usually stripped before an ex situ curve is executed.


The Corrosion Behavior of Nickel-Copper Alloys


The most widely used family of Ni-Cu alloys are the Monels. Since

their introduction over half a century ago, the Monels have become

widely used in heat exchanger tubing and general marine environments.








However, in spite of this prevalence there has been a conspicuous lack

of basic research regarding Monel and particularly the parent Ni-Cu

binary system. As a result little is actually known about the

passivation process involved that makes these alloys so valuable. In

addition, there is no general agreement regarding the ideal composition

for optimum corrosion performance in many environments. In contrast to

pure Ni, the basic research that has been conducted on the corrosion

properties of the Ni-Cu system, although lacking in scope, is coherent

and on many aspects there is even general agreement.

From his work on experimental Pourbaix diagrams, T. S. Lee [41]

concluded that the binary alloy shifted from copper-like to nickel-like

behavior in the 20-45% nickel range. Lee also pointed out that the Ni-

80 Cu alloy had a zero current line near the Ni/Ni(OH)2 equilibrium

potential and that primary passivation occurred at nearly the same

potential as for pure Ni while secondary passivation coincided with the

passivation potential of pure copper. This of course suggests that

primary passivation is attributable of the nickel component while copper

oxide is responsible for any secondary passivation processes. However,

the author's attempts at independent verification of film composition

using x-ray diffraction techniques were unsuccessful because of the

thinness of the films.

Mansfield and Uhlig [63] studied polarization curves for various

Ni-Cu alloys in 1 N H2SO4. They found that primary passive behavior is

exhibited only by those combinations containing greater than 38% Ni

which is in agreement with the findings of Lee. It was noted that as

the nickel content is increased above this 38% threshold that primary

passive behavior becomes more pronounced. The authors reason that since









the d-band is unfilled in the 38% Ni alloy, these compositions can

more readily chemisorb oxygen which is in turn responsible for

passivation. Although it is conceded that metal ions may also be

present in the chemisorbed oxygen layer, it is concluded that they do

not exist in stoichiometric proportions.

Mansfield and Uhlig also observed secondary passivation. However,

it was shown that this behavior is independent of the Ni content and,

like primary passivation, is the result of a chemisorbed film. Although

both films are considered to be nonstoichiometric, the authors state

that upon cathodic polarization the secondary film converts into the

primary film which in turn decomposes at the Flade potential.

It was again found that passive behavior is lost in alloys

containing more than 70% Cu by Osterwald and Uhlig [64]. By comparing

the Flade potential with thermodynamic values,the authors reached the

conclusion that passivation is attributable to a stoichiometric NiO

surface film. The Flade potential is noted to increase to more noble

potentials as the copper content of the alloy increases. This is

explained in terms of the affinity for oxygen adsorption onto the

surface decreasing as the result of the higher Cu content. Finally, at

about 70% Cu, the stay time of the adsorbed oxygen is so short that it

is unable to convert into a protective oxide and metal dissolution

proceeds.

Due to the considerable cost of Cu and especially Ni, it has been

desirable to maintain or even increase the corrosion resistance of this

binary alloy through small third element additions. As noted by Lee

[41], the corrosion characteristics of a Ni-30 Cu alloy can actually be







enhanced by small (.9 w/o) additions of iron. Advantage of this

fortuitous improvement was taken in the development of Monels.

Kiyoshige and Yamane [65], however, reported that minor additions

(ca. 2%) of Mn, Cr and Fe had very little effect on the anodic or

cathodic polarization behavior of a Ni-30 Cu base alloy. On the other

hand, it was found that a homogenation heat treatment resulted in a

remarkable reduction in the corrosion current for all compositions.

This effect is presumably the result of the elimination of coring which

would essentially set up microscopic "galvanic" couples within'the non-

homogenized metal sample.

Several attempts have been made to identify the corrosion film

present on Monel samples. However, without exception, these experiments

are performed in a specialized solution or at high temperatures which

makes the data difficult to relate to the basic corrosion processes.

Hettiarachichi and Hoar [66] reported the formation of a uniform

"brownish gray" film on a Monel 400 electrode potentiostated in the

passive region. The electrolyte used was a chloride/bicarbonate

solution (pH = 8.3). Microscopic examination revealed that the film was

actually quite porous. No attempt was made to identify the composition

of the film.

McIntyre et al. [67] examined the film formed at open circuit on a

Monel 400 alloy in LiOH solution at 2850C. Using XPS it was determined

that at a pH of ~ 10 the film was predominantly Ni(OH)2. At a somewhat

higher pH (14), NiO was observed as the predominant film. Upon

sputtering the outer oxide layers, it was determined that a copper oxide

(valence not given) exists under the nickel oxide (hydroxide).







Ugiansky and Ellinger [68] studied a welded Monel screen that

failed while in service in a well pipe. Electron microprobe analysis

indicated that a copper-rich phase was present at the grain boundary

which led the author to suggest corrosion via a dealloying mechanism.

But as already pointed out by Kiyoshigea and Yamane [65], the heat

treatment can produce profound effects on the corrosion behavior of

Monel. Thus, an equally acceptable explanation is that the copper-rich

phase near the grain boundaries is the result of coring subsequent to

welding rather than dealloying or a solid state diffusion process.



Summary


As the result of extensive research, it has been "conclusively"

determined that the passivating film that forms on a nickel electrode in

acid solutions is Ni(OH)2, NiO, NiOOH, Ni304, Ni203, NiO2, Ni01.5-1.7 or

chemisorbed oxygen, which forms via the NiOads, NiOH+ or

NiOOH intermediates. Also, decisively shown is that the film is

formed by the precipitation and/or direct oxidation mechanisms and has a
0
thickness ranging from one monolayer to upwards of 90 A.

About 60% of the literature dealing with the passivation of nickel

in neutral solutions states that NiO is the passivating film. However,

other studies have suggested that NiOOH, Ni304, NiO2 or the

interconversion of these compounds may be responsible.

Regarding Ni in alkaline electrolytes, most researchers have

concluded that Ni(OH)2 causes passivity, but still others have

attributed the phenomenon to the formation of Ni304, Ni203 or NiOOH. So

it seems clear that in spite of the numerous research efforts, the

mechanism of nickel passivity is still a topic of much controversy.









In light of this controversy, it is somewhat surprising that the

few papers dealing with transpassive behavior seem to agree on the

general mechanism of film breakdown. Although no specific mechanisms

are generally mentioned, it is believed that breakdown occurs via

localized defects within the film. There is further agreement that Cl

ions in solution are able to accelerate this process. Although the

specific role of the C1~ ions is still unclear, it is believed that they

in some way cause a higher defect density within a passive film.

The behavior of copper in solution has been well documented with

some degree of accord. Only three films have been observed to form

anodically, CuOH, Cu20 and CuO.

When copper is alloyed with nickel, the alloy behaves as pure copper

if the copper content is greater than about 70%. If the copper content

is lower than this threshold value, the alloy behavior is better

approximated by the behavior of pure nickel. Small additions of iron

are found to improve the corrosion performance of pure Ni-Cu alloys but

detailed mechanisms have not been postulated.

In light of the controversy concerning the passivation of pure Ni

and the absence of basic research regarding the behavior of Ni-Cu

alloys, a concerted effort has been made in the present study to clarify

many of these points. Specifically, traditional polarization techniques

have been combined not only with dependable ex situ surface analytical

techniques (ESCA, AES) but also with Differential Reflectometry

utilizing a unique sample design which allows direct comparison of in

situ and ex situ results. This combination of techniques results not

only in positive in situ film identification but also provides much





37


valuable information which can be used to deduce much about the actual

mechanisms of film formation.







CHAPTER 3
EXPERIMENTAL PROCEDURE


Sample Preparation


As discussed in the Introduction, the DR measures the normalized

difference in reflectivity between two adjacent samples. For instance,

in order to obtain a reflectogram indicative of a specific corrosion

film it is necessary to compare a film free substrate to a second

identical substrate that has this surface film. Thus it is clear that a

special sample design is required to facilitate in situ reflectivity

measurements.

The electrode design that was adopted to achieve this situation in

situ is presented schematically in Figure 1. This electrode

configuration is hereafter referred to as a split insulated disk or SID

electrode. The "halves" of each SID electrode are identical in

composition, e.g., pure Ni. However, when in solution,one electrode is

cathodically protected, insuring no film growth while the adjacent

sample is held at an anodic potential which may induce a corrosion film.

In this manner an in situ reflectogram can be obtained for the given

corrosion film.

Specifically, each SID is prepared by obtaining two small

(typically 1 cm x .5 cm), flat pieces of a metal or alloy which is

generally achieved by halving a single piece of the metal or alloy to be

used with a diamond saw. Afterwards, an insulated wire (typically 18

gauge, copper multi-strand) is soldered to the back of each of the

halves. The two halves are then electrically insulated from one














To PotentiostatZZ'


Sample


Mylar film


mount


Figure 1. The Split Insulated Disk (SID) electrode used for in
situ optical measurements.







another by a thin (1 mil) sheet of Mylar* film after which the whole

assembly is set in Quickmount,** a self-setting acrylic epoxy.

Both the Ni and Cu used in the present work were obtained from

McKay Corp., New York. The Ni is 99.95% pure while the Cu has a purity

of 99.999%. The Ni-30 wt.% Cu alloy was produced by arc melting, in a

gettered He atmosphere, the appropriate amounts of the Ni and Cu just

described. The arc melting process was performed three times in all to

ensure thorough mixing of the two components. The resulting button was

then homogenized at 10000C for three days in a He atmosphere.

After mounting, each SID was ground with a series of 240, 320, 400

and 600 grit SiC papers. Microcut*** paper was then used to further

smooth the metal surface and, more importantly, remove any embedded SiC

particles. After a satisfactory surface finish was obtained,the SID was

polished with 6 pm and finally 1 pm diamond paste on Microcloth***

lubricated with Metadi*** fluid. The sample was subsequently washed in

a soap solution, thoroughly rinsed in distilled water and finally dried

in a gentle stream of dry air. All SID electrodes were inserted into

the electrolyte within five minutes after polishing.



* Trademark of DuPont

** Trademark of Fulton Metallurgical Products, Inc.

*** Trademarks of Buehler, Ltd.








Solution Preparation


All solutions used were prepared from triply distilled, deionized

water. Resistivity was typically greater than 106 ohms/cm. The

electrolyte used in all experiments was 0.15 N Na2SO4. The C1

solutions were produced by the addition of 5.84 g of NaC1 to one liter

of the sulfate solution producing 0.1 N concentration of the C1- ion.

Reagent grade Na2SO4 and NaC1 were used in all cases. The solution pH

was. adjusted by adding either 0.1 N NaOH or 0.3 N H2SO4 while deareation

of the solution was accomplished by purging for 18 hours with purified

N2. The solutions not deareated were simply left open to the atmosphere

during the course of the experiments and are referred to in the

remainder of this work as "open solutions."


The Corrosion Cell


The cell employed in these experiments was the same in situ cell

used by Shanely et al. [61] and is shown in Figure 2. As will be

discussed in the Results and Discussion, the quartz window is not trans-

parent at wavelengths less than 220 nm. In order to expand the range

over which data could be collected, the quartz was replaced by a

Spectrosil window which is transparent to wavelengths down to about

180 nm. It was found, however, that the effects of the electrolyte

must also be considered since, like quartz, it absorbs light of wave-

lengths shorter than 220 nm.

The counterelectrode was a one-inch square of platinum screen while

a saturated calomel electrode was used as a reference. All potentials

quoted are referred to this scale.





















































Figure 2. In situ corrosion cell-[61].








Polarization Techniques


After final polishing and drying the SID electrode was placed,

under potential control, into the electrolyte. The potential upon

immersion in each case was approximately 50 mV anodic to the reversible

hydrogen potential (i.e., -450 mV at pH = 4.0; -600 mV at pH = 8.0; -900

mV at pH = 12.0). Such a potential was found sufficiently cathodic to

prohibit film formation prior to the start of the experiment while not

being so cathodic as to induce H2 evolution which severely alters the

reflectivity of the electrodes.

The top electrode in the SID configuration was maintained at this

initial cathodic potential throughout the experiments. The potential of

the lower electrode on the other hand was increased by 50 mV increments

every two minutes until the oxygen evolution potential was reached.

Current measurements were taken at the end of each two minute interval

and reflectograms were taken every 100 mV throughout the experiments.

There was typically a ten minute interval between the insertion of the

SID electrode into solution and the initial potential step.

All experiments were repeated at least once with little variation

being observed (less than 10% in all cases) in either the polarization

or reflectivity data.

Ex situ reflectograms were obtained by removing the electrode from

solution, rinsing in distilled water and drying in a gentle stream of

dry air. The electrode was subsequently remounted on the stage of the

DR and a reflectogram obtained. The time elasped between the removal

from solution and the mounting on the stage was typically less than five

minutes.








CHAPTER 4
RESULTS AND DISCUSSION


Identification of Anodic Surface Films


Anodic Films on a Copper Electrode

Two types of surface films were obtained by anodic polarization of

a pure Cu electrode in 0.15 N Na2SO4. The first film type was obtained

at a potential of -250 mV in pH = 9.0 solution. The reflectogram

corresponding to 18 hours of polarization under these conditions is

shown in Figure 3a. This reflectogram displays a prominent maximum at

380 nm accompanied by two lesser maxima at 317 nm and 245 nm. In

addition, two shoulders are observed; one at 550 nm and the other at 460

nm. Subsequent to obtaining this reflectogram the Cu electrode was

removed from solution and a second, ex situ reflectogram obtained

(Figure 3b). Since the peak positions and relative magnitudes are

virtually identical, it can be concluded that the film undergoes no

compositional alteration upon removal from solution. The sample was

then promptly inserted into the vacuum chamber of a combination ESCA-AES

system. Magnesium Ka x-rays (1253.6 eV) were used to irradiate the

sample and the resulting data were standardized by adjusting the energy

scale so that the carbon peak would correspond to 285.0 eV. The

resulting ESCA data (Figure 4) show the Cu 2p3/2 peak occurring at 936.9

eV. After the carbon correction factor of 4.1 eV is taken into account

this peak value is adjusted to 932.8 eV which could belong to any of the

following copper compounds:










0












oco


-0 "




0 0
E


o C

0

II








4-J

o,





C
c--
CU



., o -r







OR
0 L-


0 LL
Ir: ii* + -






































































































c) 8
CO CO
CO (M


CD










Li


O
4-

CO
r--

L
s-
O




0

c-
-P
0






(0
O




0o



=3











OJ O
Oa







II
0




C
0r-










E
o


-- I
C0r
0)












I-I








Cu (metallic), BE = 932.4 eV;

Cu20, BE = 932.5 eV;

CuO, BE = 933.7 eV;

Cu(OH)2, BE = 934.7 [69].

However, it is observed that the shake-up peaks which normally

accompany the principle Cu 2p3/2 peak (15-20 eV higher) for both CuO and

Cu(OH)2 are absent. This indicates that this peak results from either

Cu20 or metallic Cu. When the Cu L3M4,5M4,5 AES peak is considered,

however, a positive identification of this film can be made. The

corrected AES peak for this film occurs at 916.1 eV (Figure 5) which

agrees quite well with the accepted value of 916.3 eV for Cu20 [70].

(The L3M4,5M4,5 AES peak occurs at 919.0 eV for pure copper metal.) The

combined ESCA-AES data indicate, therefore, that the film represented

in the reflectogram of Figure 3a is Cu20. A third reflectogram obtained

from this sample after removal from the vacuum chamber indicates that

neither the vacuum nor x-ray radiation altered the film composition

(Figure 3c).

The second type of film observed to form on the Cu electrode can be

obtained at a potential of +450 mV in a solution of pH = 4.3. The

series of reflectograms obtained for this film appear in Figure 6. Only

two features are distinguishable on the in situ reflectogram--a shoulder

at 575 nm and a "peak" at 225 nm. Comparison of the ex situ curves for

this sample mounted within and without (Figures 7a and b) the dry

corrosion cell indicate that this 225 nm "peak" is an artifact which

results from the absorption by the quartz window at wavelengths shorter

than 225 nm. This result is an indication that any true peak must be

off scale, i.e., at a wavelength shorter than 200 nm.



























































N -
0^,


co


O



I-

4.-
0

-c
0






0







V)
CD
O




3
V)
c -
O



EQ
Ou


LUL


-r-














<- (.
0I



-IE



( *I-





49




O
-O
0
-05









cn


0 0

o C


it;t +



II



\v0
ot






00
O o-
LO



oo





O-
E
\ \- 4-
o -u






O _0
na
E







io
S0 -







00 0-
co \D







\ \c
\~ \\ ll
\ \\ ^ rti =
\ \l s- ^
1 11 ("t a:n'^
\l| 0 0co

11 ^ =





gloc s











0
-0
O

.0
oo





o
-0 (3)
\\ rc






00

It





L)
C








LO



0 .
\E





S.-













r-






4-,
"o




















\\0








0 c
o











ao
*r--


\o





00
\ E

0 Li0

r-
\ a. .














|n: So







The data obtained from ESCA analysis of this film are presented in

Figure 8. The binding energy of the major peak (corrected value of

933.7 eV) agrees well with the accepted value for CuO [69]. The

presence of the shake-up peaks is also a clear indication that the film

is indeed CuO [69]. Again, comparison of the in situ, ex situ and post-

ESCA film shows that this film is also unaffected by removal from

solution and vacuum treatment. The correspondence of the reflectograms

of Figures 3 and 6 with Cu20 and CuO agree with the results of Shanley

[61] who used band structures and optical constants to arrive at the

same correlation.


Anodic Films on a Nickel Electrode

Three corrosion films are observed to form on a Ni electrode in

0.15 N Na2SO4 solution. A typical reflectogram corresponding to the

film most commonly formed appears in Figure 9. The only striking

feature is the "peak" (again artificial) at 225 nm. The curve drops off

smoothly from this "peak" to the upper limit of 800 nm. The 0 Is and Ni

2p3/2 peaks from ESCA analysis of this film are displayed in Figures 10

and 11. The Ni 2p3/2 peak occurs at 856.1 eV (corrected) which falls in

the middle of the range of values for Ni(OH)2 reported in the literature

(e.g., 855.6 eV [71], 855.9 eV [67], 856.1 eV [72], 856.45 eV [69]).

Also, see Table 1. The 0 Is value of 532.2 eV from Figure 10 also

agrees quite well with values reported for Ni(OH)2 by Barr [69] (531.95),

McIntyre et al. [67] (531.7 eV) and Ali [72] (532.1 eV). Several

researchers [73-75] have obtained ESCA structures very similar to the

ones in Figures 10 and 11 by bombarding a clean nickel surface with

oxygen in a vacuum system. They speculate that the resulting compound






















































0

*-




x ro
X Ln


(a



4-,


I
X L


LLJ
L)












0
Ln












r-f



o
4-,
4-)

CD






S-






0
OT


















4r-










u 0
0V





- 0
C-




Li-
Co








a)
5-



















U-
Ll_


0

I H


















0

v


-0
-l




0

II


o 2

-o cc

o< LO


o

-0
CD-

-0
0 "c
0 ,

o tD
\o .Y-









CO




0L.

co

\ o c >





\


C V
\ ~ I L




















d


x
8


CI
.-



(0



-0


II




z
OD Q <
1L,< ~






























8








C)




















-a






















4



N
zH --


2


))


4-'
E





















t_.






I)
0

lL

















r-





S-


4-


CL"


o
0P


















.r


Lt.)
L/)


I-


a)
--
U-
,r-







+-_
I-K





nr









TABLE 1

BINDING ENERGY DATA FOR NICKEL AND NICKEL -COMPOUNDS*




Binding Energy, (eV)

Ni NiO Ni304 Ni203 Ni02 Ni(OH)2


852.60 -- -- -- -- 856.10
852.50 854.00 -- -- 855.90
852.50 854.00 -- -- -- 855.60
852.75 854.60 -- 855.70 -- 856.45
852.90 854.50 -- 855.80 --
852.00 855.00 --
-- 854.00


* [reference 72]







is Ni203 rather than Ni(OH)2. There is, however, no reason to conclude

that the film formed in the present work in aqueous solution would be

similar to that formed under these extreme conditions, rather than the

more commonly formed Ni(OH)2. It is reasonably concluded, therefore,

that the film discussed above is indeed Ni(OH)2.

The second film type observed to form on a Ni electrode always


occurs as a minor constituent in a m

reflectogram indicating such a mix i

that a prominent shoulder occurs at

otherwise be indicative of Ni(OH)2.

of the peak at 225 and this shoulder

solution pH and electrode potential

mixed film. Although the relative ai

is interesting to note, when grown i


ix with Ni(OH)2. A typical

s displayed in Figure 12. Notice

about 320 nm on a curve that would

The fact that the relative height

at 320 nm varies as a function of

is a clear indication that this is a

mplitude can be varied somewhat, it

n solution, this shoulder never


develops into a true peak but rather is always a minor component of the

Ni(OH)2 peak. An ESCA analysis of two films that displayed a prominent

shoulder on a reflectogram indicated in both cases that the film was

still pure Ni(OH)2. Two possible explanations can be offered to explain

this result. First, it could be that the amount of the compound

responsible for the shoulder is present in an insufficient amount to be

detected by ESCA. Alternatively, it could be that this second compound

is present at the metal/Ni(OH)2 interface in which case the ejected

electrons needed for the analysis could well be completely attenuated by

the Ni(OH)2 overlayer, which is, in fact, sufficiently thick to prevent

detection of the underlying Ni substrate.

In order to identify this second Ni compound, a somewhat different

method of film preparation was attempted. Instead of growing the film











0
0







0.
II

0 CD
-0
co
CD
o E

o

O 0
0r





















C-)
CY



4'-
O o
























0)
0
0 "-




\ 0 c
I-





0


r-" *
O r






*r--*r--
S^ E
0 -o






.--4
o


r .CO


<1 C







under anodic polarization in an aqueous solution, half of the nickel SID

was removed from the epoxy mount after polishing and heated for three

minutes in a pure oxygen atmosphere. The final temperature was about

460C after which this half sample was left to cool to room temperature

in the passing oxygen (about 10 minutes). This sample was then

remounted with the as-polished specimen, which had been stored in room

temperature air, and the reflectogram in Figure 13 obtained. The Ni(OH)2

peak at 225 nm is again present just as in the solution grown film.

However, instead of just a shoulder, a well defined peak now exists at

330 nm and is actually greater in magnitude than the Ni(OH)2 peak. The

Ni 2P3/2 and 0 Is peaks obtained from the subsequent ESCA analysis are

presented in Figures 14 and 15. Close examination of both peaks reveals

that each is actually a doublet. A computer resolution of the doublets

appears in Figure 16 and 17. The Ni doublet deconvolutes into peaks

positioned at corrected energies of 854.1 eV and 856.0 eV while the

oxygen doublet yields peaks at 529.7 eV and 531.9 eV. The 856.0 eV

nickel peak and 531.9 eV oxygen peak agree well with the values of 856.1

eV.and 532.2 eV already obtained in this study for pure Ni(OH)2. The

remaining nickel peak at 854.1 eV agrees quite well with the published

value of 854.0 eV [67,71,76] for NiO while the oxygen peak at 529.7 eV

agrees exactly with published values [73] for NiO. This confirms that a

mixed film, namely Ni(OH)2 and NiO, is responsible for the reflectograms

in Figures 12 and 13. Moreover, since the "peak" at 225 nm has already

been identified as Ni(OH)2 it can then be concluded that the second peak

(or shoulder, depending on relative film composition) at 330 nm results

from NiO.





60




0
-0
c~

C(II





CD
4-,











\o
E


O .)

-0 00













LO 5r
oo--
o












4-,
^O 0
Sr




CD O




o^ +-















E
cu





S.-
4r--












0 C
o o







fO L










U)
\0 0


JIMO
\P O
\ v~

\ Ij- z L




































Q















-I
-J
-1
U

a
















0)


O
C\j



r-
C)












N
0













d
0c





S.-














o
4-














a
0








E
-a










S--
r-
0




CE
0

s-

4-





n3










CO-S
4-





0r
-c +-
o- (



























0

-j

i
Q
-a
U
0







>00

to
0
0(I

0

a-


t.4
U,





I a
z

N
1

N
8




I
8 -


2 0
SH


C'\J
o







r-




N
r-

CX
0
O o
ma)
cS-



-o







(0
Sa,




O E





x
0 E
-isS-











-0









I iL A
WE













U-
4-0
CLL
U..
13
J3
0 *










xL




























































C 'hi.







to*o
C C9
H100


L
S-


4-

U
4-

S-
SI
c-
C



O a
t--

i-i (






4- W
0 -












OO
S0U


=- S










Cr- 4--
4- )

0
jm
or
o L


4-
C=O



o-
L)
04-)







0




)
> Cn)


C 0


XOb


-1 S
U- 3t




U) 0
to o
0 ta'OD
tO i



ce d
C,


sw
lae

*-
>> >
00



5Q

fC








64










o
S-
C
4-)
0
(a
4-
C
o



S.-






,
I co









0

^ C
0f












"c-
'




C I D ox


Ob
CL Q --





ON 0


.4 0) (A



z t k *

_ 0 H OC 04-
-O q -

*
,r-






S0 0
N *CM > 0 *



"N l-







Based on calculations [77] involving the relative areas of the Ni

2p3/2 peaks it is determined that the oxidation film is approximately

52% NiO--the balance being Ni(OH)2. Comparison of the reflectogram peak

heights in Figures 12 and 13 indicates that the anodically formed film

is roughly 10% NiO, 90% Ni(OH)2. Since the detection limit of NiO in

Ni(OH)2 for ESCA is on the order of 10%, it is not surprising that the

anodic film was analyzed as being pure Ni(OH)2.

The third film type observed to form on a Ni electrode in sulfate

solution occurs only at a pH = 8 or greater and only on those electrodes

which have not experienced an active-passive transition. A rather

intriguing characteristic of this film is that it seems to form only by

conversion of Ni(OH)2. In fact, formation will not occur at all unless

a Ni(OH)2 layer is first grown on the electrode. A reflectogram

indicating the presence of this third film type is presented in Figure

18. Notice that as the peak at 625 nm increases, the Ni(OH)2 "peak" at

225 nm diminishes, which indicates clearly that a conversion reaction is

occurring. The magnitude of the NiO peak at 330 nm seems unaffected,

however. If the electrode potential is subsequently reduced to any

value cathodic to the oxygen evolution potential, the 625 nm peak

disappears within seconds while the Ni(OH)2 "peak" sharply increases.

This reverse conversion reaction also occurs if the electrode is removed

from solution. Since this type of film cannot be isolated ex situ,

positive identification is not possible.

It has been suggested by Jones and Wynne-Jones [78] and by

Lukovtsev [79] that Ni(OH)2 can be oxidized to NiOOH at higher

potentials via a deprotonation reaction:


Ni(OH)2 + OH NiOOH + H20 + e









0





^-- ~~^ := .--r-
-0
C'4







o ,
CO )
Q
0
4-

0 0
C\.J

0



oEo
oE4


r0





0.

o /\ o -P


O E




So-

oo





LL-
0 \ cn.









0.-


^ C ^0
0 0 L

O C-.
1^ E









Confirmation of this reaction in a nickel-cadmium cell was made by

Uno [80] who showed by x-ray measurements that Ni(OH)2 does indeed

convert to the oxyhydroxide upon charging. He further noted the reverse

reaction to occur upon discharge of the cell. Based on the agreement of

the present observations with those in the literature, it is assumed

that the peak at 625 nm on the reflectograms of Figure 18 is

characteristic of NiOOH.


Anodic Behavior of Nickel in 0.15 N Na2SO


Anodic Behavior at pH = 4.0

When a freshly polished Ni SID electrode is potentiostated between

-300 mV and -100 mV in .15 N Na2SO4 solution of pH = 4.0 in an open

corrosion cell, a negative peak appears on the resulting reflectogram

(Figure 19a and b). This negative peak is the mirror image of the curve

obtained for Ni(OH)2 growth which indicates that a Ni(OH)2 film is

dissolving from the surface of the working electrode, perhaps according

to the simple dissolution reaction:


Ni(OH)2 Ni+ + 20H-.


However, if oxygen reduction is occurring at the electrode surface,

2H20 + 02 + 4e 40H ,


the surface pH will no longer be that of the bulk solution but

substantially higher. This higher pH results in the stabilization of

the Ni(OH)2 film which, as shown in this study, is virtually insoluble

at pH values as low as 7.1 (Figure 20). Although oxygen reduction is





68



0
O

04





0
-0 0
CO



P-


S-


o
-C



\0
-0E


o E cr







o 2


LO
\0 0
CL C4
\ 0s

>) >E
\ 0

OO > -0
EcoE
o oo oas
O O



0 0\0

l w
<0
**Ji **





69



0
-0
N


-0









0
0'>)


r-0
o o



S4-3

S- 0.-
o CC















E
O 0
cI 0 u

0 U -'
0,



E E :-r
oo







*0




r'
[ o 0r

a co
U-



<








thermodynamically possible at any potential cathodic to +750 mV, it may

actually only occur at potentials substantially more cathodic to this

value due to the large overpotential required and the small exchange

current density for this process on nickel. If oxygen reduction ceases

at about -300 mV on the anodic scan, the surface pH of the working

electrode will begin to approach that of the bulk (4.0) and Ni(OH)2

reduction can proceed. However, the optical reference electrode is

still being maintained at a potential of -450 mV where oxygen reduction

easily occurs. This results in the continued stabilization of the "as

polished" Ni(OH)2 film on the optical reference electrode; hence, a

negative peak occurs on the reflectogram.

This explanation of film stability is supported further by the

results of the following experiment. After a thick Ni(OH)2 film

(relative to the "as polished" thickness) was grown on the working

electrode by potentiostating in the passive region, the electrode

potential was switched to -100 mV. As expected, the peak height

diminishes as a function of time indicating that the Ni(OH)2 film is

dissolving (Figure 21). If this same experiment is performed in a pH =

7.1 solution no film dissolution occurs even after 24 hours at this

reduction potential. These dissolution results are consistent with

those of MacDougall [29].

In the deareated solution, pH = 4.0, no negative peaks are observed

on the reflectograms (Figure 22). This behavior can be attributed to

the "as polished" Ni(OH)2 film dissolving simultaneously on both halves

of the electrode. This would be expected since the surface pH would be

that of the bulk solution (4.0) in the absence of oxygen reduction.









O

00






So
4-)
0

(3)








r

o -
0 -'

*r
4-



o E "






-o


.Cr




-E 0
0 0 \


' \I I 4-
0 0 \ 0 -








rnc
oo n0I






gjio
















0
.00

04

S o -a
-0 "






o -

C)
-0
*0-


-o




O







mo<

c -



0 0 4-
I II


S0 \0
to 0 (-I



S0\ I C









co
\\O \







Since the working electrode is free from surface films prior to

passivation, metal dissolution proceeds uninhibited. At about -150 mV

the current reaches a maximum and the Ni(OH)2 film has grown to a

substantial thickness (Figure 23). By + 50 mV the film has attained

maximum thickness and passivation is complete, the passivation process

occurring over a rather broad 400 mV span. The passivation reaction is

presumably


Ni++ + 20H- + Ni(OH)2.


In contrast, the passivation of the Ni electrode in the open

solution occurs over a narrower potential range of only 150 mV (Figure

24). Although the passivation current density is similar in the two

cases the passivation potential is approximately 100 mV more anodic in

the case of the open solution. These observations can be explained by

recalling that the surface film present on the working electrode in the

open solution is not completely dissolved until a potential of -100 mV

is reached on the anodic scan. The presence of this film retards

metallic dissolution as evidenced by the very low current values in the

potential range -450 mV to -100 mV. (Note that the absence of such a

film in the deaerated solution results in a substantially greater

current in this potential range.)

Finally, when the film is completely reduced, the working electrode

is at a potential well into the metal. dissolution region. This results

in rapid Ni++ ion dissolution as evidenced by the accompanying sharp

increase in anodic current at -80 mV (Figure 24). Recall that the

optical reference electrode is still at a potential where oxygen

reduction occurs resulting in the production of hydroxide ions. The









AR/R (225nm)


0 O0
L) 0
CI CO


0 0 0 0 0
1 0) 10 0 )O
CM4 Cq T- -,


0
0
O
(D


0



O

o
0


o

E
0 -
0
1-.





-





0






0
O
\
0 7


(Zwo/vri) /Ul!suaa


luajjiino








AR/R (225nm)


(ZWO/VrI) SoU!suaSc


0n 0 0 0 0 0
O 0 LO o O O
CN M I- V- I


luaJiin






++
high concentration of Ni ions coupled with the relatively high

concentration of OH- ions results in the rapid (as evidenced by the

sharp increase in AR/R) and uniform (as evidenced by the sharp decrease

in current) formation of the passivating Ni(OH)2 film. It is

interesting to note that the passive film in deareated solution is five

times thicker than the same passivating film formed in open solution

(c.f. reflectivity data of Figures 23 and 24). A possible explanation

for this lies in the fact that in the deareated solution Ni anodically

dissolves into solution at potentials between -400 mV and the passiva-

tion potential. It could well be that in this potential region Ni(OH)2

nucleates on those grains more favorably oriented for film growth

(c.f. references 23 and 24) and attains an appreciable thickness on

these grains prior to the passivation of the remaining surface. Reflec-

tivity measurements, in fact, show that Ni(OH)2 is indeed forming even

at potentials 150 mV cathodic to the passive potential. After the

passive potential is reached, the current drops off relatively slowly--

the passive current density not being attained for yet another 200 mV.

During this interim the reflectivity data indicate that the film

continues to thicken. This rather sluggish passivation behavior is

consistent with the suggestion that the film thickness varies over the

surface and explains why the average film thickness, as indicated by the

reflectivity data, is so great. Based on calculations involving optical

constants, this film appears to be approximately 130 A thick.

Once the passive plateau is reached the film is very stable and

there is no indication of any further thickening throughout the balance

of the experiment. The film formed in the open solution, however,

appears to thin, i.e., the peak height diminishes somewhat. An

alternative interpretation is that the film does not actually thin but







instead small amounts of Ni(OH)2 precipitate on the optical reference

electrode. This is quite possible because of the high hydroxide

concentration maintained at surface combined with Ni++ ions that could

easily diffuse to this surface from the adjacent working electrode.

Although the mechanism of film formation cannot be absolutely

determined from the present data several revealing observations can,

nevertheless, be made. Firstly, in the open solution the film is

observed to be rapidly formed at a potential of -50 mV while, as already

pointed out, the same type of surface film is unstable and dissolves at

a potential just 50 mV cathodic to this value. Secondly, film formation

is accompanied by very large anodic currents which would result in

extremely high Ni ion concentrations near the surface of the

electrode. Thirdly, although the magnitudes of the passive currents are

the same for the electrode in the deaerated and open solutions, the film

is five times as thick in the deaerated case. Moreover, the film in

bbth cases is observed to maintain a constant thickness (disregarding

the anomalous effect in the open solution) after formation, i.e., the

film thickness is not a function of electrode potential. All four

observations indicate clearly that the passivation film forms via a

precipitation mechanism.



Anodic Behavior at pH = 8.0

The behavior of the Ni electrode in pH = 8.0, deaerated solution

(Figure 25) is quite similar to the behavior already discussed for this

electrode in the pH = 4.0 deaerated solution. The similarities include

the magnitude of passivating current density (400 pA/cm2 at pH = 8.0)

versus 350 pA/cm2 at pH = 4.0) and the passivation potential (-200 mV










AR/R (225nm)


0
0





O
0
0
',4



E



C
0

0Q
0.

C*4
1

0



0
o





I
C8


La
CJ
0

o


CO 0

(ZWO/Vri/)


Cq C co
s0u
A!SUacJ IuGJJOno







versus -170 mV). A third striking similarity is the broad potential

range over which passivation occurs (400 mV in both solutions).

However, although the passivation film is Ni(OH)2 in both instances, it

is approximately twice as thick in the pH = 8.0 solution. Furthermore,

upon removal from the pH = 8.0 deaerated solution the working electrode

is observed to be partially covered with a milky-white film; the balance

of the surface film appearing clear as usual. (Both reflectivity

(Figure 26) and ESCA data indicate that this thicker film is still

Ni(OH)2.)

Once into the passive plateau the current density is nearly the

same for the nickel electrode in the pH = 8.0 deaerated solution as it

was in both the deaerated and open solutions of pH = 4.0. This

similarity is notable especially since the film thickness varies by a

factor of ten among these surfaces. This, of course, suggests that only

a relatively thin film is required to achieve passivity. Also note that

the current in all three cases (pH = 4.0 deaerated and open, pH = 8.0

deaerated) is constant throughout the passive plateau (a range of about

700 mV) and that there is no increase in film thickness. This indicates

that ionic transport through the film is not field dependent but perhaps

more a function of the gross structural characteristics, e.g., porosity.

In pH = 8.0, open solution, oxygen reduction is again evidenced by

a cathodic current below a potential of about -200 mV (Figure 27, c.f.

pH = 4.0, open cell). Above -200 mV the current density is quite low

and constant; apparently the electrode was passivated without

experiencing active dissolution. Reflectivity data shows that the film

grows steadily from the start of the experiment and, as expected, no





80




0
-0





o
Sco
C I



E\ I
-






*r-
O













o



CJ


000

a 0



r -
\000 0 o

0
cmJ
o o





\ 0 0 o a
0 \0 -0 on

\ ICI I a.


o .o\
[IC I <-
I I I II 1 c: -










AR/R (225nm)
0
t I


SI I I


1t 0


O o0


o
CM
0)
0 -
0
D LL-
I


(zulwo/Vtd) 4A!suea uaoJJno







reduction of the "as-polished" film occurs as it did in the pH = 4.0

open solution.

Near.-100 mV a small NiO component in the film develops (Figure

28). As the potential is increased this component becomes somewhat more

prominent although it never appears as more than a shoulder on the

Ni(OH)2 reflectivity structure.

The overall film thickness throughout the experiment appears to be

a linear function of potential. This linear dependency indicates that

field assisted ionic conduction through the film is likely responsible

for the passive current which is in contrast to the films already

discussed. Reflectivity data indicate that the film under present

consideration is some 60 times thinner than the film formed in the

deaerated solution. Both observations support the suggestion that the

conduction mechanism in the two films is quite different.

Also in contrast to the three films previously discussed is the

fact that this thin film converts to NiOOH at potentials near the oxygen

evolution potential. Since the composition of all the films in the

passive regions is virtually identical, the only explanation is that the

field across the thinner, more coherent film is large enough for the

transforming deprotonation reaction to occur. The field across a

bulkier film at a similar potential would, of course, be -considerably

less.



Anodic Behavior at pH = 12.0

The general behavior of the nickel electrode in pH = 12.0, open and

deaerated, is similar to that just discussed for pH = 8.0 solution. In

the pH = 12.0 open solution oxygen reduction.is again indicated at









0








go i
-0












o a

o
4--
CD
*r--






CO











00 o
r0








-P-
o
z













O EE

0 0 0 0 0


EEEE o
o 0
"T-





I I 0 4 N


Co
CO









CO -
lEEEE o o







potentials near the start of the experiment (Figure 29). Once oxygen

reduction has ceased (-600 mV), the current becomes slightly anodic and

maintains a rather modest current density of 4 pA/cm2 until the oxygen

evolution potential is reached. As was the case for the pH = 8.0 open

solution, passivation occurs without active dissolution. The film

thickness is again a linear function of electrode potential with the NiO

shoulder appearing at a potential of approximately 0 mV (Figure 30).

In the pH = 12.0 deaerated'electrolyte, the film growth is again a

linear function of potential (Figure 31) but thickness is approximately

half that obtained in the open solution. The NiO shoulder again appears

near 0 mV (Figure 32). Transformation of Ni(OH)2 into NiOOH is observed

for both the open and deaerated solutions at a potential of about

+450 mV.

In summary, the behavior of Ni in 0.15 N Na2SO4 can be divided into

two distinct categories. In solutions of low pH (pH = 4.0 open and

deaerated, pH = 8.0 deaerated) a very thick passivating Ni(OH)2 film

forms apparently by a precipitation mechanism. This film, once formed,

does not thicken throughout the balance of the experiment. It has also

been noted that neither NiO nor NiOOH is observed in these solutions.

Also, the current density throughout the passive plateau is relatively

high (8-10 pA/cm2).

These conclusions concerning film composition are quite different

than those reached by Okamoto and Sato [15-17] who stated that NiO, Ni203

and Ni304 are all formed in low pH solution (4.5). Although Bockris and

Reddy did identify Ni(OH)2 on a nickel electrode in a solution of pH =

3.15, they stated that it was merely a precursor film rather than the

film actually responsible for passivation. This of course is in direct









AR/R (225nm)


(uIo/V)I IAIsua u n
( MJ3/V71) AI!suecG 1UaJ~zno:


o
0





0





0
o





0 C

I O.
CO











O
SE












I
O
O






c o
I n



0
0




0


I





86



0
-0
cV



0
0

-0 V-
C
O

\~0

00



\ 0 \ 1

\o E




o o
O"









C









00 0
O OO





r0
II


(0 0 0
o





|co
<1Ir





87






AR/R (225nm) O






0 0
L O

S!r

0 0o

C\J



rI
S.-




c\J







S "E





O 0











4r-
o .








o co
N
*S-

0C
0











(~wU3:/V7l) Al!suea uuaaJn





88



0
0

c\


.0
00
O0



So




0.-
-E"
-





E > E z
o
(a) "-
o a





0000
e4O .o
)) )
oos, g o
co U-
SCO

EEE>>
EE 01


.. .. .. o **

W .0 O -0,
E5





0 10
Li
<,
Q Ir" in..








contrast with the evidence just presented which indicates that Ni(OH)2

formation is concurrent with the passivation process. There is also

notable contrast between the present work and the work published by

MacDougall et al. [13,27-30,34-37,44,53]. Employing virtually identical

experimental procedure (50 mV potential steps in 0.15 N Na2SO4 solution

of pH = 2.8 and 8.4) they concluded that NiO was the passivating film,

and, furthermore, that NiO was the only film to form on a Ni electrode.

A possible explanation for this disagreement lies in the fact that

MacDougall et al. electropolished and cathodically pretreated their

electrode before running the experiment while the "as polished"

electrode is used in the present case. Since it has been shown that the

"as polished" film-dissolves before the passivation process begins (c.f.

Figure 19 and discussion on pages 46-47) it is suggested that the NiO

film identified in those experiments by MacDougall et al. may actually

be formed in the pretreatment process. This would also explain why the
0
film is always reported to be 9-12 A thick independent of both potential

and solution pH.

Although Ord et al. [33] also report the formation of Ni(OH)2 on a

Ni electrode, they suggest that it forms by hydration of an initial

passivating NiO film on multisweep potentiodynamic experiments. In

fact, no study can be found in the literature that essentially agrees

with the present work as to the composition of the passivating film on

Ni in acid solutions.

In contrast to the behavior just discussed for acid solutions,

solutions of a high pH (pH = 8.0 open, pH = 12.0 open and deaerated)

produce a very thin passivating Ni(OH)2 film, apparently by a solid

state growth mechanism. This film is somewhat more protective as







2
evidenced by a low current density of 3-5 pA/cm in the passive region.

Although not observed in the low pH regime, NiO occurs at intermediate

potentials while the transformation of Ni(OH)2 into NiOOH is apparent

near the oxygen evolution potential.

Davies and Barker [39] also reported that a very thin (1 monolayer)

film of Ni(OH)2 was responsible for the passivation of Ni in high pH

solution (13.0). They also reported the development of two additional

film components at higher potentials but identified these as Ni203 and

NiO2, in contrast to the present results. Remarkable agreement is also

noted with the work of Schebler-Guzman et al [40] who report the

conversion of a passivating Ni(OH)2 film into NiOOH at a potential near

the oxygen evolution potential. The formation of Ni(OH)2 is also

suggested by Lee [41] although no higher oxides are observed in his

experiments.


Anodic Behavior of Nickel in .15 N Na2SO Containing Chlorides


Anodic Behavior at pH = 4.0

When Cl- ions are added to pH = 4.0 open solution, the prepassive

response of the Ni electrode is qualitatively unchanged. As before, a

negative peak occurs on the reflectogram (Figure 33) in the initial

potential region. The magnitude of the peak in this case is larger,

however, indicating that the dissolution of the "as-polished" film is

more complete in the C1 environment. Active dissolution along with

Ni(OH)2 film growth are observed at potentials slightly cathodic to the

initial film dissolution reaction with the film being 25% thicker in

this case. The passive potential in the C~- environment is slightly








0
AR/R (225nm)
__ o- -,,
O g
CO
cc
0r-


0
_g-


I0 -

: 1 I
I I o


SI 0
a
S0 -0


4 c o


O
SI4
I 01







)
/ C /
0 0 0 0








S0/ 0 ~c
/ o




c-






c o-O !
/ o 71 A I s





0 0 0 0 0 0-
0 0 7 0 0)D

/O /f 0- -
( W3/Vr/) A!SuaG luaJJfn3








higher (75 mV) while the current density at this potential is greater by

a factor of 16.

Immediately after passivation the current again rises rapidly and

is accompanied by a concurrent decrease in reflectivity with the spectra

being identical to that obtained in the nil Cl1 solutions. Upon removal

from solution, the surface of the electrode appears severely pitted.

The decrease in reflectivity during depassivation can be explained

either by the initiation and growth of pits, by film thinning, or both.

In the deaerated electrolyte, pH = 4.0, containing C1- ions, the

passivating current is again higher than in the chloride-free

environment (Figure 34). However, passivation in this case is never

completed. Rather, breakdown occurs prior to the electrode being truly

passive. As was the case in the open solution, Ni(OH)2 formation is

delayed to slightly more anodic voltages but the rate of formation, once

initiated, is considerably larger. The Ni(OH)2 film formed in this

deaerated solution is so thick that the DR saturates. Because of this,

it cannot be determined whether reflectivity drops off with the loss of

passivation as it did before.



Anodic Behavior at pH = 8.0

The response of the nickel electrode in the pH = 8.0 deaerated

solution containing chlorides is similar to the behavior just discussed

for the pH = 4.0 solutions. Subsequent to active dissolution,

passivation occurs but is lost almost immediately (Figure 35). The

passivating film is so thick that saturation occurs. A surprising

feature of this data, however, is that the passivating current density

is approximately 20% less for this solution than for the Cl- free





93

e
AR/R (225nm) 0
CM
S0
co



I 0
0 4-

I 0


o i
I --


0


1 S-
-o




S 0 >
/ I
Il O




-- 4








ao
= -'
/o


S0 0

.. O o



O c*r-





0 \ I c


0



CO CM


(Zwo/v'di) f~!suaal











AR/R (225nm)


Li



,I


Cs
AIIsuaa


CueJain


0
0






0
0
C0
0
U








o


0



c11
E






I
C*
1
0
0.

o








0
0d


0

r-
o


tz




co
13



."--
o
ro











4--







*I-- O
C0







*4-
O --
r-
5->
+-I






N *-
-4-
c I
fO (_


.--

0
o
o
CD


- ~s.~~cII
S.


U~- -


cO
qaJ

(ZUW3


0
0
iwl)