Citation
Differential capacity of stainless steel in potassium chloride solutions during potentiostatic and galvanostatic polarization

Material Information

Title:
Differential capacity of stainless steel in potassium chloride solutions during potentiostatic and galvanostatic polarization
Creator:
Fiorino, M. Elaine Curley-, 1944-
Publication Date:
Copyright Date:
1975
Language:
English
Physical Description:
xiii, 187 leaves : ill. ; 28cm.

Subjects

Subjects / Keywords:
Chlorides ( jstor )
Corrosion ( jstor )
Current density ( jstor )
Electrodes ( jstor )
Ions ( jstor )
Passivity ( jstor )
pH ( jstor )
Pitting ( jstor )
Potassium ( jstor )
Sulfates ( jstor )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Polarization (Electricity) ( lcsh )
Potassium chloride ( lcsh )
Steel, Stainless -- Testing ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Bibliography: leaves 179-186.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by M. Elaine Curley-Fiorino.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
025119554 ( AlephBibNum )
02573890 ( OCLC )
AAS3898 ( NOTIS )

Downloads

This item has the following downloads:


Full Text










DIFFERENTIAL CAPACITY OF STAINLESS STEEL
IN POTASSIUM CHLORIDE SOLUTIONS DURING
POTENTICSTATIC AND GALVANOSTATIC POLARIZATION











By

M. ELAINE CURLEY-FIORINO


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











UNIVERSITY OF FLORIDA


1975

































The author dedicates this dissertation to her

husband, John A. Fiorino.














ACKNOWLEDGEMENTS


The author would like to express her gratitude

to the members of her committee; Dr. E. D. Verink, Jr.,

Dr. J. D. Winefordner, Dr. W. S. Brey, Dr. R. Bates, and

especially to her research director, Dr. G. M. Schmid for

the interest and assistance given to her in the course

of this investigation and in the preparation of this

manuscript.

She would also like to thank Mr. R. Dugan and

Mr. A. Grant and their associates for their help in the

technical aspects of this work.

Steel samples were provided by the United States

Steel Corporation.

The author is also grateful for the financial assist-

ance received from the University of Florida in the form

of Graduate School Fellowships.

Finally, the author would like to thank

Dr. and I:rs. James L. Fortuna and Mr. and Yrs. Willis Bodine

for their friendship and encouragement during her

graduate career.
















TABLE OF CONTENTS


ACKNOWLEDGEENNTS. . . . .

LIST OF TABLES . . . .

LIST OF FIGURES . . . .

ABSTRACT. . . . . . .

Chapter

I. INTRODUCTION. . . .

II. EXPERIMENTAL. . . .

Experimental Design .


Experimental Technique. . . . .

Fotentiostatic polarization . .

Galvanostatic polarization . .

III. RESULTS . . . . . . . .

Current-Potential Behavior During
Fotentiostatic Polarization . . .

Primary solutions . . . . .

Secondary solutions . . . .

Capacity-Potential Behavior During
Potentiostatic Polarization .. ..

Primary solutions . . . . .


Secondary solutions . . . . .


Fage

. . . . . . iii

. . . . . . vi

. . . . . . viii

. . . . . . x



. . . . . . 1

. . . . . . 13

. . . . . . 13


. . 18

S . 21

. . 25

. . 34


. 92

. 92

. 102








Chapter

Galvanostatic Polarization . . . . .

Potential-time behavior . . . . .

Capacity-potential behavior. . . . .

IV DISCUSSION . . . . . . . . . .

Potentiostatic Polarization . . . . .

Active dissolution and passivation. . .

The hydrogen evolution reaction . . .

Rest Potential. . . . . . . .

Passive region. . . . . . . .

Pitting . . . . . . . .

Transpassive dissolution. . . . . .

Galvanostatic Polarization. . . . . .

V SUMMARY . . . . . . . . . .

LITERATURE CITED . . . . . . . . .

BIOGRAPHY . . . . . . . . . . .


Page

103

103

122

131

131

131

135

142

144

150

156

160

173

179

187








LIST OF TABLES


Table Page

1. Fotentiostatic current-potential behavior in
primary solutions (average curves) ..... 38

2. Fotentiostatic current-potential behavior in
primary solutions (individual experiments) . 39

3. Cathodic loop in primary solutions . . .. .40
4. Transpassive behavior in primary solutions . 41

5. Solution composition parameters investigated
in primary solutions . . . . . ... .42

6. Interrelationships in potentiostatic
experiments in primary solutions . . . 43

7. Potentiostatic current-potential behavior in
secondary solutions (average curves) ..... 69

8. Potentiostatic current-potential behavior in
secondary solution (individual experiments). 70

9. Interrelationships in potentiostatic
experiments in secondary solutions . . . 83

10. Cathodic loop in secondary solutions .... .95

11. Potentiostatic capacity-potential behavior in
primary solutions (average curves). ... 96

12. Potentiostatic capacity-potential behavior in
primary solution (individual experiments). . 97

13. Potentiostatic capacity-potential behavior in
secondary solutions (average curves) .... .98

14. Potentiostatic capacity-potential behavior in
secondary solutions (individual experiments) 99

15. Capacity-potential behavior in the potential
range of linearity of 1/C versus E . . .. 101

16. Galvanostatic potential-time behavior (0.0 M
potassium chloride). . . . . . .. 106

17. Galvanostatic potential-time behavior (0.100
[M potassium chloride). . . . . . ... 107










Table


Fage


1J. Galvanostatic potential-time behavior (0.303
M potassium chloride). . . . . . ... 108

19. Galvanostatic potential-time behavior (0.518
M potassium chloride). . . . . . ... 109

20. Tafel behavior during galvanostatic
polarization . . . . . . . . . 111

21. Average final breakdown potentials during
galvanostatic polarization . . . . .. 116

22. Open circuit decay behavior following
galvanostatic polarization . . . . ... .118










LIST OF FIGURES


Figure Page

1. Stainless steel electrode . . . . . 14

2. Electrochemical cell. ... . . . . 17

3. Kel-F electrode holder. ..... . . . . 20
4. Block diagram of the potentiostatic
polarization circuit . . . . . .. 23

5. Block diagram of the galvanostatic
polarization circuit . . . . . .. 27

6. Block diagram of the circuit employed in
capacity-potential data correlation . .. 31

7. Control panel ...... .. . . . 32
8. Schematic potentiostatic current-potential
behavior ... .............. * 36

9. Variation of the logarithm of the critical current
density with the concentration function,
pH + log ( [S042] + [C1-] ) . . . 48

10. Potentiostatic polarization curve showing
cathodic loop obtained in 0.334 M sodium
sulfate at pH 2.40 . . . . . . 51

11. Variation of the potential of total passivity
with chloride ion concentration . . . 54

12. Variation of the potential of total passivity
with the concentration function,
[C1-]/(2 [SC4 -] + [01- ) . . . . . 56

13. Variation of the potential of total passivity
with the concentration function,
[C1-]/(2 [S042-j + [C1-] ) ... . . . . 58

14. Variation of the potential of total passivity
with the concentration function, pH +
log ( [S42-] + [C1-] ) . . . . 60


viii









Figure Page

15. Variation of the potential of total passivity
with the cQncentration function,
log ( [S04 -] / [Cl-] ) . . . . . 62
16. Potentiostatic polarization curve showing the
influence of pitting obtained in 0.301 M
potassium chloride at pH 2.35 . . . . 65

17. Variation of the critical current density with
chloride ion concentration . . . . 72

18. Variation of the logarithm of the critical
current density with rest potential . .. 74

19. Variation of the logarithm of the critical
current density with the concentration function,
pH + log ( [S04-] + [C1-] ) ... . .. 76

20. Variation of the potential of total passivity
with rest potential . . . . . . 78

21. Variation of the logarithm of the critical
current density with the potential of total
passivity . .... . . . . . . 80

22. Variation of the primary passivation
potential with pH . . . . . * 85

23. Variation of the rest potential with pH . 87

24. Variation of the logarithm of the critical
current density with pH . . . * * 89

25. Variation of the logarithm of the critical
current density with the concentration function,
pH + log ( [S04j-] + [Cl-] ) . . . ... 91

26. Potentiostatic capacity-potential behavior
in the absence of hydrogen interference . 94

27. Schematic representation of the galvanostatic
potential-time curve in the presence (1)
and absence (2) of pitting breakdown. ... . 105

28. Schematic representation of the capacity-time
behavior of a system not subject to pitting
breakdown during galvanostatic polarization 124













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



DIFFERENTIAL CAPACITY OF STAINLESS STEEL
IN POTASSIUM CHLORIDE SOLUTIONS DURING
POTENTICSTATIC AND GALVANCSTATIC POLARIZATION

By

M. ELAINE CURLEY-FIORINO

June, 1975

Chairman Gerhard M. Schmid
Major Department: Chemistry

Potentiostatic and galvanostatic polarization of

AISI 304 stainless steel was performed in deaerated

solutions of 0.0 through 1.0 M potassium chloride at

pH 2.4. Sodium sulfate was added to an ionic strength

of one. Cylindrical electrodes were mechanically

polished and prepolarized at -0.700 V for twenty minutes.

The differential capacity of the metal-solution interface

was determined as a function of potential using the

single pulse technique.

All potentials are given versus the saturated calomel

electrode. Current densities and differential capacities

are referred to the electrode geometric area.

During potentiostatic polarization, an active-passive

transition is observed for all solutions. The rest

potential (-0.48 V) and primary passivation potential








(-0.42 V) are independent of solution composition. The

critical current density increases with the total anion

concentration. The capacity peak occurring slightly

negative to the primary passivation potential is attributed

to specific adsorption of the anions involved in the

dissolution-passivation mechanism. The potential at which

the electrode becomes totally passive shifts positive

with increasing chloride ion concentration and is dependent

on kinetic rather than thermodynamic factors.

Transpassive dissolution initiates between 0.57 and

0.64 V and exhibits Tafel behavior. Both the current

density maximum (10-4 A cm-2) and the corresponding

potential (0.95 V) associated with the onset of secondary

passivity are independent of solution composition. The

capacity peak observed in this region is attributed to

adsorption of the passivating species.

In solutions of ) 0.3 M potassium chloride, an

increase in current density caused by pitting starts

at potentials normally in the passive region. The pitting

potential depends on the relative amounts of chloride and

sulfate ions present. No capacity peak is observed

prior to pitting breakdown.

The capacity, between the potential of total passivity

and the capacity minimum occurring at 0.60 V, is approxi-

mately independent of the stability of the passive state.

This independence results from the domination of the

interfacial capacitance by that of the passive film, obscuring
xi









any changes in the film-solution capacity which may occur.

The sudden decrease in capacity observed at 0.30 V is

attributed to a change in the dielectric constant of the

film. Between 0.3 and 0.6 V, film growth follows an

inverse logarithmic law.

The potential-time behavior of samples subjected

to galvanostatic polarization depends on both current

density and solution composition. Systems not susceptible

to pitting reach and maintain a positive steady state

potential. In systems subject to pitting, the maximum

potential attained is unstable and a shift in potential

to more active values occurs.

In 0.3 M potassium chloride intermediate arrests

produced by a given current density as well as the

maximum potential achieved prior to breakdown correspond

to potential arrests in the non-pitting systems. In

0.5 M potassium chloride pitting breakdown occurs from

a potential maximum which is considerably below that

observed in 0.3 M but the behavior at more negative

potentials is similar.

It is assumed, therefore, that the initial effect

of the anodic current on the metal surface at a given

potential is the same in all cases and that pitting

succeeds through the perturbation of these initial

surface conditions by chloride ion.









Tafel behavior is associated with the arrests at

-0.4, 0.8, 0.85, and 1.12 V. Active dissolution at

-0.4 V exhibits a slope of 0.060 V decade-1 and is

thought to proceed by the Heusler mechanism. Arrests

at 0.8, 0.85 and 1.12 V correspond to transpassive

dissolution, secondary passivity, and oxygen evolution,

respectively. Capacity peaks are associated with the

latter two effects and with the arrest at -0.4 V but

not with pitting breakdown.


xiii













CHAPTER I

INTRODUCTION



A metal is in the passive state when it is inert

in an environment in which, on the basis of thermodynamics,

it should corrode readily. An ennoblement of the

potential of the metal-environment interface accompanies

the onset of passivation. Passivity has been recognized

since 1836, when Faradayl observed the stability of

iron metal immersed in concentrated nitric acid.

However, the nature of the surface species leading

to the onset and maintenance of the passive state is

still not well understood. The two major theories

advanced to explain the phenomenon are the bulk oxide

theory and the adsorption theory.

The oxide film theory proposes that a bulk oxide

is formed directly on the metal surface from the
2,3,4
products of the metal dissolution reaction.3 This

oxide then acts as a physical barrier between the metal

surface and the aggressive environment.

The adsorption theory of passivity attributes

passivation of the metal surface to the adsorption of

an "oxygen" species in less than or equal to monolayer

quantities.5, Two variations of the adsorption theory







2

have been described. In the chemical variation proposed

by Uhlig,8 adsorbed oxygen atoms are believed to

satisfy the surface valences of all atoms on the metal

surface. The correlation of the theoretical and

experimental concentrations of components of transition

metal alloys needed to bring about a sharp increase in

the ease with which the alloys are made passive lends

support to this hypothesis. The electrochemical

variation ascribes the retardation of the metal dis-

solution reaction to a change in the double layer

structure caused by the dipolar character of chemisorbed

oxygen.9 Orientation of the dipole with its positive

end towards the solution increases the activation

energy necessary for the metal dissolution reaction.

Hackerman10 has proposed a theory intermediate to

the above two. Here, the adsorption of oxygen atoms on

the metal results in a metastable state lending

temporary protection to the surface. Following electron

transfer from the metal to adsorbed oxygen atoms, an

amorphous bulk oxide is formed by cation migration through

the adsorbed array. It is this bulk oxide which provides

long-term protection. This theory, originally postulated

for metals in oxygen-containing solutions, has also been

proposed to explain the passivation mechanism of iron-

chromium alloys in deaerated acidic sodium sulfate

solutions.1 A similar mechanism has been proposed by

Frankenthall2 for the passivation of an iron-24 chromium








alloy. Electron diffraction studies of the surface of

stainless steels exposed to oxidizing acids or the atmosphere

for short times at intermediate temperatures (25 to 600C)

have shown that the passive film formed is non-crystalline.13

Transmission microscopy studies of films formed on an iron-24

chromium alloy during potentiostatic polarization in 0.5 M

sulfuric acid show that they are also amorphous.14

The corrosion resistance of austenitic stainless steels

in diverse environments is a result of the high degree of

passive state stability conferred by the presence of chromium

in amounts ) 12 percent.1 16 However, in specific media,

especially in halide solutions, this corrosion resistance is

lost, as evidenced by the onset of intense local attack,

i. e., pitting. The mechanism by which pits nucleate and

propagate in the presence of chloride ion has been investigated

in great detail. Excellent reviews are given by Kolotyrkin17

and Szklarska-Smialowska.18

The pitting phenomenon has been characterized by

three parameters the pitting potential, negative to

which no pits can nucleate; the critical chloride ion

concentration, the minimum concentration needed to

initiate pitting in a given system; and the induction

time, the time, at a given potential, which passes

prior to breakdown. Thus, studies have been carried out








19, 20, 21, 22, 23
to examine the effect of metal composition, 20, 21, 22, 23

defect density of a metal surface,24, 25 grain size,26

solution composition,27, 28, 29, 22 sulfide inclusions,30' 31, 32

and temperature28 on the pitting potential and/or the location

and number of pit sites. The critical chloride ion

concentration has been shown to depend strongly on the

concentration of inhibiting anions (e.g., sulfate, hydroxyl

and nitrate ions) in the solution33 35, 28 as well as

on alloy composition.36 The induction time has been found

to decrease with increasing chloride ion concentration,37
38, 39, 40 potential of passivation,39' 40 and

temperature.38

The influence of an induction time is seen both in

potentiostatic and galvanostatic polarization measurements.

Previous works on austenitic stainless steels in chloride-

containing solutions have shown that the application of a

constant anodic current to an initially active electrode

shifts the potential positive to a maximum value.41

Despite the continued application of the current, the

potential then decreases rapidly to a value normally in

the passive region. The length of time to potential

breakdown is a function of the current density and the

chloride ion concentration. Examination of the metal

surface shows that pitting has occurred.








5
The steady state potential achieved after breakdown

has been called the protection potential of the system
4?2
under study.42 At potentials between the protection and

pitting potentials previously nucleated pits can continue

to grow but no new pits may form. At potentials negative

to the protection potential, all existing pits become

inactive.35

Anodic potentiostatic polarization to values more

positive than the pitting potential also causes pit

formation. A change in potential to values just positive

to the pitting potential results in a decrease in current

density with time from the initially high value associated

with double layer charging.40 This current decrease

represents the readjustment of the electrode-solution

interface to maintain the passive condition. After a

time interval which depends on solution composition

and potential, the current decrease is replaced by current

oscillations signifying the pitting-induced breakdown of

passivity. As the potential is made still more positive,

the time for which passivity is maintained decreases

until an induction period is no longer apparent. 3

Mechanisms proposed to explain the dependence of

pitting parameters on experimental variables are a

function of the assumed nature of the passivating film.

On the basis of a bulk oxide, Hoar3 has described a

mechanical breakdown process in which the mutual repulsion









of anions adsorbed on the oxide surface leads to the

formation of cracks. A relationship between critical

breakdown stress and surface tension as influenced by

anion specific adsorption has been derived by Sato.44

Hoar and Jacob45 have also suggested that a metal cation

is dissolved from the oxide through the formation of a

metal-chloride complex containing 2.5 to 4.5 chloride

ions. Cation migration through the film then allows

continuation of the process. If the passive film is

assumed to be an adsorbed "oxygen" species, then its

replacement by chloride ions will lead to activation

of the metal when a critical surface concentration of
28, 17
chloride ion is attained.

The presence of an induction time associated with

pitting breakdown suggests that a time consuming change in

surface structure is occurring. If the specific adsorption

of chloride ion is involved in this change, then its

incorporation into the electrical double layer should

be reflected in the differential capacity-potential

behavior of the interface.

The electrical double layer, in its ability to store

charge, acts as a capacitor. The magnitude of the

capacity associated with it is a complex function of

potential and is therefore defined as a differential

capacity, q/6E. In the absence of specific adsorption,

only water molecules populate the compact double layer,










the region between the metal and the Outer Helmholtz

Plane. The capacity of the interface can then be

represented by two capacitors in series; that of the

region between the metal and the Outer Helmholtz

Plane (OHP); and that of the region between the OHP

and the bulk of the solution (the diffuse double layer).

In dilute solutions ( (0.001 M), the capacity of the

latter is small and dominates the interfacial capacity.

In concentrated solutions, all of the diffuse double

layer charge is located close to the OHP. The interface

then behaves like a single parallel plate condenser, i.e.,

its capacity is approximately independent of potential.

When specifically adsorbed ions populate the IHP

in a concentrated solution (> 0.001 M), the interface

again functions as two capacitors in series. Bockris and
46
Reddy, in their treatment of the effect of contact

adsorption on the total capacity of the metal-solution

interface, derive the equation

1/C = 1/KM-OHP (I- (qCA ) (1)
KM_-0OHP KM-IHP qao

where C is the total measured differential capacity and

KMOHp and KM-IHp are the integral capacities associated
with the metal to Outer Helmholtz Plane and metal to Inner

Helmholtz Plane regions, respectively. The function,

ZOqCA/'qOM, represents the change in the amount of

specifically adsorbed ion, qCA, with the charge on the

metal, qp. Thus, as the charge on the metal becomes more










positive with increasing potential, an increase in the

degree of specific adsorption with metal charge

(-2qCA/ qc2 >0) should occur, producing an increase

in the measured differential capacity of the interface.

As growth continues, however, the buildup of lateral

repulsion forces tends to decrease the degree to which

specific adsorption occurs at a given metal charge.

This inflection in the qCA_-q relationship results in a

peak in the differential capacity-potential curve

(-2qCA/ qRM2 = 0). Differential capacity measurements

have been successfully applied to the determination of

the specific adsorption of sulfate, perchlorate and

chloride ions on iron.47' 48, 49. 50

The presence of a positive charge on the metal

surface implies polarization at potentials positive to

its zero point of charge. Studies of binary alloys

suggest that the zero point of charge of an alloy

should approach that of its component with the most

negative zero point of charge if it is present in

sufficient concentration.51 The zero points of charge

for nickel, chromium and iron have been determined
52
in acidic sulfate solutions.52 The corresponding values

are -0.57, -0.69 and -0.62 V, respectively. It is

expected then, that the zero point of charge of active

stainless steel should be close to that of chromium,

i.e. -0.69 V.








9
In order to explain the potential dependence of the

activating effect of chloride ion in terms of specific

adsorption to a critical surface concentration on the

passive electrode, the zero point of charge of the

passive surface must lie in the passive region. For

passive iron, the zero point of charge occurs at -0.125 V

in 0.01 M sodium hydroxide.52 This positive shift in

value from that observed on active iron can be attributed

to an increase in work function resulting from the

presence of an adsorbed "oxygen" species or an oxide

film. The extension of this phenomenon to stainless

steel seems logical.

The primary purpose of the experiments conducted in

this study was the determination of the effect of

chloride ion on the capacity-potential behavior of

stainless steel observed during potentiostatic and

galvanostatic polarization in solutions initiating pitting.

In order to explain the potential arrests observed during

constant current polarization, the potentiostatic studies

were extended to cover the potential range from active

dissolution to oxygen evolution. Corresponding capacity

values were determined and the current-potential and

capacity-potential data correlated.

The relationship between the rate of an electrochemical

reaction and the potential difference across the interface

at which it occurs is discussed in detail by Bockris and










Reddy. When electron transfer is rate-determining,

i.e., the system is under activation control, the current-

potential relation is given by the general form of the

Butler-Volmer equation,

i = io exp 2 LF i exp ( F )i (2)
RT RT
where io is the exchange current density, the oC's are the

transfer coefficients, and n is the overvoltage. All other

terms have their usual meaning. The first term on the

right hand side of equation 2 is the current density resulting

from the oxidation anodicc) reaction. The second term

pertains to the reduction cathodicc) reaction.

At the equilibrium potential of the rate-determining

reaction, the overvoltage, which represents the potential

difference across the interface in excess of the equilibrium

potential difference ( E Eo), is zero. The net current

density observed (i) will therefore also be zero since at

equilibrium the rates of the anodic and cathodic reactions

will be equal. The transfer coefficients determine what

fraction of the potential difference across the interface

is operative in changing the energy barrier for the oxidation

and reduction reactions.

As the potential difference across the interface is

made more positive (v> 0), the contribution of the anodic

current density to the total current density will

increase. At a sufficiently positive overvoltage

(~0.120 V for a one-electron transfer reaction), the






11

influence of the cathodic current density becomes negligible.

The current-potential relationship is then given by

i =i exp aF_ (3)
RT (3)

which can be put in logarithmic form and rearranged to give

= 2.3 RT log i 2.3 RT log i (4)
SF ocF
A plot of overvoltage versus log i is therefore linear.

Such plots are known as Tafel lines. The slope of the

line contains the transfer coefficient which is a complex

function of the total number of electrons transferred

during the reaction as well as the mechanism by which the

reaction proceeds. From the slope of the line and its

intercept, the exchange current density for the reaction

can be calculated.

In systems in which a faradaic current can flow, i.e.,

charge can cross the metal-solution interface, the

electrical behavior of the interface can be represented by

a resistor in series with a capacitor and resistor in

parallel. The series resistor represents the resistance

of the solution to current flow. The capacity is the

differential double layer capacity. The parallel resistor

represents the polarization resistance of the faradaic

reaction, decreasing as the rate constant of the reaction

increases. In cases of low polarization resistance, the

determination of the differential capacity is difficult

since the polarization resistance can act as a leakage









path for the signal measuring the capacity. Faradaic

current also interferes indirectly by causing a change

in the true electrode surface area as well as in the solution

composition.

The use of the classical alternating current technique

in which the interface forms one arm of an impedance

bridge to determine the double layer capacity on a solid

electrode is precluded because of the dependence of the

capacity value measured on signal frequency.5 Most of

the direct interference can be eliminated, however, by

the use of the single pulse method of differential

capacity measurement developed by Riney, Schmid and
54
Hackerman. Analysis of the linear segment of the

potential transient resulting from a single current

pulse allows calculation of the capacity at the point

from which the pulse initiates since

C = i(dt/dE)t= (5)

where i is the pulse magnitude and dt/dE is the slope of

the Dotential-time transient evaluated at t = 0.













CHAPTER II

EXPERIMENTAL

ExDerimental Design

The material investigated was stainless steel, AISI

304, provided by the United States Steel Corporation. Its

composition was given as 0.03 C, 0.027 P, 1.10 Mn, 0.022 S,

0.43 Si, 9.26 Ni, 18.6 Cr, 0.39 Mo, and 0.04 N (weight

percent). Bar stock was machined to cylinders with a

diameter of 6 mm and a height of 9 mm (Figure 1). The

cylinders were tapped, threaded,and mechanically polished

at 2400 ram with 400 followed by 600 grit emery paper.

They were then degreased with spectral grade benzene in

an ultrasonic cleaner, rinsed with triply distilled water,

and stored in a closed polyethylene container until needed.

Primary studies were carried out in solutions

containing 0.0, 1.17 x 10-2, 9.97 x 10-2, 0.301, 0.508,

and 1.0 M potassium chloride. Solution pH was measured

with a Beckman pH meter and adjusted to 2.4 with

concentrated sulfuric acid. Sodium sulfate was added as

required to maintain an ionic strength of one (0.318, 0.313,

0.284, 0.232, 0.147, and 0.00 M, respectively). Secondary

experiments involved solutions of pH 2.4 containing 0.102,

0.123 and 1.0 M potassium chloride with no sodium sulfate'

additions as well as 0.3 M potassium chloride at pH 1.52

















































Figure 1. Stainless steel electrode.








15

and 6.22. All chemicals used in solution preparation were

reagent grade. Recrystallization of potassium chloride

from triply distilled water had no effect on experimental

results. The water employed was distilled from alkaline

potassium permanganate and then from a two-stage Heraeus

quartz still and collected in a two-liter Pyrex volumetric

flask. Its maximum conductivity, determined with a General

Radio Impedance Bridge, was 2 x 10-6 1 -lcm-.

Platinum electrodes, generally used in pre-electrolysis

to remove electroactive impurities from the solution, have

been shown to dissolve when polarized anodically in both

sulfate and chloride containing solutions.55 Because of

the possibility of contaminating both the solution and the

stainless steel surface with platinum, a pre-electrolysis

step was therefore omitted.

The electrochemical cell was made of Pyrex and was of

conventional design (Figure 2). A Luggin capillary

connected the saturated calomel reference electtode (SCE)

to the cell via two solution-lubricated mercury-seal
2
stopcocks and a potassium chloride salt bridge. A 1 cm

platinum flag auxiliary electrode was mounted on the cell

with a standard taper joint. For use in constant current

polarization and capacitance measurements, a platinum

gauze basket, approximately 100 cm2 in area (Engelhard

Industries), was mounted concentric to the test electrode.

The cell cap incorporated a 24/40 standard taper joint for

mounting the test electrode.



























Figure 2. Electrochemical cell.















SOLUTION INLET--
















GAS INLET


Pt FLAG -
ELECTRODE
COMPARTMENT


TEST ELECTRODE RECEIVER

-GAS OUTLET













REFERENCE ELECTRODE-
COMPARTMENT


Pt GAUZE ELECTRODE

Hg SEAL STOPCOCKS


LUGGIN CAPILLARY







18

The electrode holder (Figure 3) was made from a Kel-F

rod machined to approximately 1 cm in diameter in which a

3 mm center hole was drilled. The rod was then heated and

a threaded stainless steel rod inserted so that thread

protruded on both ends. The Kel-F rod was shrunk in an

ice bath to facilitate its insertion into a 1 cm ID glass

stirrer bearing sleeve which had a 24/40 standard taper

joint attached. The sample was then affixed to the

protruding steel rod. The electrode area exposed to
2
solution was approximately 2 cm2. The sample-sample

holder seal could be tightened by turning a finger nut

at the top of the assembly. Teflon washers were placed

between both the sample and holder and the nut and holder

to reinforce the seal. Spot checks of the shielded top

surface of the test electrode showed no evidence of

corrosion indicating the absence of leakage.

Experimental Technique

Solutions were deaerated with helium (99.99 percent)

for a minimum of eight hours prior to use. Prepurification

and water saturation of the gas was accomplished by passing
0
it through a 12 cm column of Linde 5 A molecular sieve

pellets into a gas wash bottle containing triply distilled

water. The gas then flowed into a two-liter reservoir

containing the solution and continued through a dispersion


































Figure 3. Kel-F electrode holder.










































24/40 STANDARD
TAPER JOINT
















TEFLON WASHER-----


FINGER NUT




TEFLON WASHER









KEL-F ROD




STAINLESS STEEL ROD







21

tube into the electrochemical cell. The cell gas outlet

terminated in a gas wash bottle to prevent contamination

of the cell contents with ambient air.

Immediately before each experiment, the cell was

washed with hot chromo-sulfuric acid cleaning solution

and rinsed with triply distilled water. Gas pressure was

then diverted to fill the cell with deaerated solution

from the reservoir. The stainless steel sample was

secured on the Kel-F holder, rinsed with triply distilled

water and the solution to be studied, and immersed in

solution. All samples were pretreated at -0.700 V for

twenty minutes to reduce air-formed surface films.

Solution stirring was accomplished with a magnetic

stirring bar and the helium flow continued throughout

the experiment. All solutions were at room temperature

and all potentials are reported relative to the saturated

calomel electrode. Current density and differential

capacity were calculated using electrode geometric

areas. The test electrode was grounded during all

experiments.

Potentiostatic polarization

The block diagram of the system employed in potentio-

static experiments is shown in Figure 4. Polarization

was accomplished with a slightly modified Harrar56

potentiostat. (Two each of obsolete transistors 2N333A

and HA7534 were replaced with the equivalent circuit elements

2N3568, 2N5869, and 2N3644,2N5867.)





















Figure 4. Block diagram of the potentiostatic polarization circuit.

A. Test electrode
B. Reference electrode
C. Platinum flag electrode
D. Platinum gauze electrode

















PULSE GENERATOR









After prepolarization at -0.700 V the potential of the

test electrode was shifted anodic in step-wise increments.

The magnitude of the imposed step depended on the potential

range under investigation and the electrochemical reactions)

associated with it. In regions where changes in potential

caused significant changes in current density, steps of 20

to 30 mV were generally employed. In regions of approx-

imately constant current, 50 mV steps were used. A

Keithley 610B electrometer was used to monitor the applied

potential. After an arbitrary time interval of 10 minutes,

the current flowing in the auxiliary-test electrode circuit

was determined from the potential drop across a 1.18 kGl

or a 10 k1 wire-wound precision resistor (1 percent)

shunting the input of a Keithley 660 electrometer (1014 _

input impedance).

The differential capacity of the stainless steel-solution

interface was determined as a function of potential using

the single pulse method.5 A voltage pulse from the gate

of a Tektronix 549 storage oscilloscope was used to trigger

a Tektronix 114 Pulse Generator. The 100 ,sec square wave

current pulse produced was applied to the platinum

basket-test electrode circuit. The resulting potential-time

transient was recorded on the oscilloscope operated in the
-l
storage mode at a sensitivity of 2 or 5 mV cm- and 2 or

5 ysec cm-1. Differential capacity values were calculated

from the linear segment of the transient slope during the

initial 10 usec of the pulse. Since the transient is







25

linear between 4 and 10 l sec, the slope determined between

these two points is equivalent to that of the tangent drawn

to the curve at t = 4 sec. Measuring times of 4,ysec

correspond to an alternating current frequency of

approximately 1.2 x 105 Hz. Since a frequency of 105 Hz
57
is required to eliminate faradaic interference completely,

very little interference is expected.

Prior to each experiment, the pulse magnitude was

calibrated using standard capacitors and resistors. The

current density of the pulse could be calculated from the

known value of capacity in the calibration circuit, the

measured slope of the transient, and the electrode geometric
-2
area. Pulse magnitude was approximately 5 mA cm and was

independent of capacity between 1 and 0ldfarads.

Galvanostatic polarization

After potentiostatic prepolarization the system was

switched to galvanostatic control using a two-position

toggle switch. A block diagram of the circuit used

during galvanostatic polarization experiments is given

in Figure 5. A constant current was supplied to the

platinum basket-test electrode circuit by a Hewlett Packard

881A power supply operated in the constant voltage mode

in series with a bank of resistors (4.7 to 47.8 kl ). The

positive terminal of the power supply was grounded to

establish the test electrode as the anode. The current

magnitude was determined from the voltage drop developed























Figure 5. Block diagram of the galvanostatic polarization circuit.

A. Test electrode
B. Reference electrode
C. Platinum flag electrode
D. Platinum gauze electrode



















ELECTROMETER


OSCILLOSCOPE 2


POTENTIOSTAT


ELECTROSCAN 30


POWER SUPPLY








across a lO lprecision resistor (1 percent) using a

Keithley 6-0 electrometer and was adjusted by varying

the series resistance and the voltage output of the

rower supply. Applied current densities ranged from

0.99 x 10 to 2.0 x In-3 A cm-2. The resulting

potential-time curve was displayed on the recorder of a

Beckman Electroscan 30. Potential values were monitored

with a Keithley 61(B electrometer.

A study of differential capacity as a function of

potential was also carried out during constant current

polarization. As described above, a square wave current

pulse was delivered from the pulse generator to the

platinum basket-test electrode circuit. The resulting

potential-time trace was displayed on the storage

oscilloscope and the capacity calculated from the initial

slope of the transient and the predetermined pulse

magnitude.

The development of the theory of the single pulse

method assumes that a perturbation current is applied

to a system under constant current polarization at a steady-

state potential so that the current before and after the

pulse remains the same. Since, during potentiostatic

polarization of austenitic stainless steel, either a steady

state current or a very slow decrease in current with time

is observed, the method should be directly applicable. For

galvanostatic polarization the constant current condition








29

is satisfied but a continuous change in potential with time

occurs.' However, for small measuring times ( 10 yOsec),

the change in potential will be negligible.

Because of the rapid change in potential with time

during constant current polarization, difficulties were

encountered in the correlation of capacity values and the

potentials at which they were determined. A system was

therefore developed which facilitated the sequential

storage of pulse-produced transients and supplied a marker

signal to the Electroscan recorder whenever the pulse

generator was triggered (Figure 6).

Two oscilloscopes were used sequentially. Depression

of PG 1 or PG 2 on the control panel (Figure 7) first

changed the DC level of the input of the oscilloscope in

use by 10 mV. This provided automatic downward displacement

of the successively produced potential-time transients.

A two-step generator was associated with PG 1, an eight-

step generator with PG 2. In most cases, the oscilloscope

sensitivity needed for precise determination of the

transient slope (2 or 5 mV cm-) precluded the use of

automatic stepping and vertical displacement was accomplished

manually instead.


























Figure 6. Block diagram of the circuit employed in capacity-potential data
correlation.
























3 CHART
RECORDER













O VERT.
INPUT
(OSCILLISCOPE)













00000
SET CLIP TP PG2 PG, RESET
REC CELL VERT TI T2 PG
000000


Figure 7. Control panel.









Then, after a 50osec delay, the horizontal sweep

of the oscilloscope was triggered. Simultaneously, 15 V,

100 msec pulse was supplied to the external trigger

of the pulse generator to produce the current pulse

needed to determine the capacity. At the same time a 50 mV,

100 msec pulse was superimposed on the galvanostatic

potential-time response signal from the cell to the recorder.

Since the potential values recorded during constant

current polarization reached 1.4 V, while the marker pulse

superimposed on this signal was only 50 mV, the cell signal

was passed through a buffered amplifier and could be

clipped at a preset value. The recorder sensitivity could

therefore be adjusted to display the pulse marker while the

entire potential-time curve remained on chart.













CHAPTER III

RESULTS

Current-Potential Behavior During
Potentiostatic Polarization

The response of the stainless steel samples to poten-

tiostatic polarization is given schematically in Figure 8.

The region ABC represents the transition from the active

to the passive state characteristic of this material in

the solutions employed. An adjoining region of approximately

constant current is then usually observed (CC'D). The

increase in current on further polarization is caused by

loss of passive state stability as the result of pitting (DE)

or of transrassivity (FG). The potential and current

density notations on the schematic are defined below. They

are presented in the text in approximately the same order

in which they arise during polarization.

Possible interrelationships between experimentally

determined current densities, potentials, and/or solution

composition variables were examined. The various data

were plotted. If a plot appeared to be linear, then a

linear regression analysis was performed. Only curves

with correlation coefficients (r2) of ) 0.90 were

reported as linear.

























Figure 8. Schematic potentiostatic current-potential behavior,

















[crtl,2
B
cri t, /I



ifp


i A




V-i


Etr Epp,2
I I


POTENTIAL


r pp,
Er E,,
I


I 1 I '









Primary solutions

Characteristic potentials and current densities

determined from the potentiostatic polarization curves

obtained in primary solutions are presented in Tables 1 4.

The solution composition parameters investigated are listed

in Table 5. The logarithms of these functions have also

been considered. Relationships with correlation coefficients

) 0.90 are reported in Table 6.

After the twenty-minute prepolarization period at

-0.700 V, net currents are cathodic with values between

2.1 x 10 3 and 5.3 x 10-3 A cm2. No correlation between

the hydrogen evolution current and solution composition at

constant pH and constant potential is apparent (Table 1).

As the potential is shifted anodic, the measured current

decreases and becomes zero at the rest potential, Er.

The rest potential is defined as that potential at which

the absolute values of the internal anodic and cathodic

currents become equal, resulting in a net external current

flow of zero. Values of the rest potential determined here

are -0.450 to -0.510 V. The maximum average deviation for

data obtained for a given solution composition is 16 mV for

0.0 M potassium chloride (Table 2). For all other

compositions it is substantially less ( < 7 mV).

Polarization at potentials positive to Er results in

a net anodic current which increases with potential to a

maximum, signifying the onset of passivation. This maximum

current density and its corresponding potential are the










Potentiostatic current-potential
i-70Q
[C-"] [S042-] (A cm ) Er
x 101 x 101 pH x 103 (V,SCE)


Table 1

behavior in primary solutions (average curves).
icrit i
Epp,l (A cm (A pm-2) ETP Epit(V,SCEI"
(V,SCE) x 105 x 107 (V,SCE) inc gr


12.90 -0.285

5.25 -0.255

6.40 -0.242

7.85 -0.230
-0.205

-0.160


0

0.117

0.997

3.01

5.08

10.0


3.34

3.29

3.00

2.48

1.63

0.16


2.40

2.40

2.40

2.35

2.42

2.40


-2.20

-3.10

-3.50

-2.23

-2.06

-5.31


-0.453

-0.475

-0.480

-0.470

-0.510

-0.460


-0.420

-0.425

-0.432

-0.420

-0.400

-0.400


2.26

2.73

3.13

3.43

8.49

6.73


+0.395

0.00

-0.090
-o. 090o


+0.360

-0.012

-0.155











Potentiostatic
experiments).


Table 2

current-potential behavior in primary solutions (individual


i-700
(A cm-2)
pH x 103


0 3.34 2.40 -2.3


-2.1

0.117 3.29 2.40 -2.7

-3.5

0.997 3.00 2.40 -3.8
-3.2

3.01 2.48 2.35 -1.8
-2.7

5.08 1.63 2.42 -2.4
-1.9

-1.9

10.0 0.160 2.40 -4.9

-5.8


ocritsl
Er Epr,l (A cm-lt
(V,SCE) (V,SCE) x 105


-0.420

-0.460

-0.460

-0.473

-0.473

-0.480

-0.475

-0.475

-0.480

-0.508

-0.520

-0.520

-0.460


-0.400

-0.430

-0.410

-0.410

-0.420

-0.430

-0.430

-0.420

-0.415

-0.400

-0.400

-0.400

-0.400


1.31

2.65

1.90

3.4

2.3

3.3

2.9

3.55

3.39

8.3

7.6

9.5

7.3


i
(A cm-2)
x 107


10.7

21.5

65.9

5.32

5.18

7.19

5.01

7.23

8.48


x[ 1 J [S041l
x d1 x i1


Epit(V,SCE)
inc gr


ET1p
(V,SCE)

-0.278

-0.287

-0.287

-0.255

-0.265

-0.237

-0.250

-0.230

-0.226

-0.200

-0.200

-0.210

-0.197


0.365

0.360

-0.020

-0.047

-0.020

-0.145 o


- -0.197 -0.05 -0.175


-0.460 -0.400 6.2


0.420

0.370

0.00

0.00

0.00

-0.10










Cathodic loop

70Q2
(A cm ) C-7002
pH x 103 (if cm2)


Table 3

in primary solutions.

Negative
Cl loop
(pf cm-2) (V,SCE)


0 3.34 2.40 -2.31



-2.10

0.117 3.29 2.40 -2.67

-3.54

0.997 3.00 2.40 -3.83

-3.18

3.01 2.48 2.35 -1.8

-2.67

5.08 1.63 2.42 -2.45

-1.89

-1.P5

10.0 0.160 2.40 -4.86

-5.77


-0.30 to 0.00 -13.6



-0.250 -2.27



-0.27 to -0.20 -3.89










-0.20 to -0.15 -4.25


] 01 SO 21
x x 0


ic
(A cm-2)
x 107


1-200-2
(A cm
x 10


13.4

27.3

-9.13

4.54

3.10

10.4

-0.09

9.66

11.5

5.97

8.15

10.6

-4.25

3.42








Table 4

Transpassive behavior in primary solutions.

ba E Ep icrit,
pH V decade (V, E) (V,bCE) (A cm-)
x 104

2.40 0.114 0.671 0.950 1.12

2.40 0.114 0.641 0.950 1.02

2.40 0.150 0.575 0.950 1.03


rso
x


0.0

0.117

0.997


3.34

3.29

3.00


A i,
A cm-2)
x 106

21

14

17








Table 5

Solution composition parameters investigated in primary solutions.

Cla S04/C1 C1/(C1 + 04) C1/(C1 + 2 S04) C1/(C1 + S04) pH + log(Cl + S04)


10-2

10-2


28.1

3.00

0.82

0.32

0.016


0

0.034

0.249

0.548

0.755

0.984


0.017

0.142

0.377

0.609

0.968


0.066

0.400

0.707

0.861

0.982


1.92

1.93

2.00

2.09

2.25

2.40


a
Cl and S04 represent the analytical concentration of chloride and sulfate, respectively.
The brackets and charge designations have been omitted for clarity.


0.0

1.17 x

9.97 x

0.301

0.508

1.00








Table 6

Interrelationships in potentiostatic experiments in primary solutions.


Independent
vr Qhl1 0


---;aba


pH + log(S04 + Cl)a

Cl

C1/(C1 + SO4)
Cl/(Cl + 2 SO4)

pH + log (SO4 + Cl)

log(S04/Cl)





Cl

C1/(C1 + SO4)
C1/(C1 + 2 SO4)

C1/(C1 + S04)
log (SO/Cl)

pH + log (S04 + Cl)


Dependent
variable


log icrit

ETP

ETp

ETp
ETp

ETp






ip

ip
ip
i
P

log ip

log ip


Correlation
Slope coefficient


1.10

0.108 V mole-1

0.103 V

0.110 V

0.218 V decade-1

a. 2 segments -
-0.016 V decade
-0.039 V decade-1

b. 1 segment -1
-0.030 V decade

8.68 x 10-7 A cm2 mole-1

5.03 x l0-7 A cm"2

7.09 x 10" A cm2

4.05 x l0-7 A cm-2

-0.11

1.05


(0.83)

0.93
0.92

0.94

0.94


0.98
0.99

0.94

0.97
1.00

0.99

0.99

0.99

0.99


Data
points


variabl










Table 6 continued


Independent
variable


Dependent
variable


Correlation
Slope coefficient


log Cl

SO0/Cl

C1/(C1 + S04)
log{Cl/(Cl + SO4))

log[Cl/(Cl + 2 SO4)

C1/(C1 + j SO4)

log Cl/(Cl + i S04)1

pH + logS04 + Cl))

SO /Cl

log1Cl/(C1 + S04)}

logfCl/(Cl + 1 SC 4)


Data
points


Epit(gr)

Epit(gr)

Epit(gr)
Epit(gr)

Epit(gr)

Epit(gr)

E .
pit(gc)
Epit(inc)

Epit(inc)
Epitine)


-0.96
0.65

-1.17
-2.09

-1.26

-1.90

-3.73
-1.67

0.62

-1.98

-3.54


decade1


V decade-1

V decade-1

V
decade-1
V decade-1

V

V decade-1

V decade-1


0.91

0.99
0.92

0.97
0.94

0.97

0.97

0.95
0.96

0.92

0.93










Table 6 continued


Independent
variable


Dependent
variable


Correlation
Slope coefficient


C1/(C1 + 2 SO4)

pH + log(SO4 + Cl)


Epit(inc)

Epit(inc)


-1.80 V


-1.57 V decade-1


a
Cl and S04 represent the analytical concentration of chloride and sulfate, respectively.
The brackets and charge designations have been omitted for clarity.


Data
points


0.92

0.95









critical current density (icrit) and the primary passivation

potential (E p ), respectively. All values of ic

measured here are between 1.3 x 10-5 and 9.50 x 10-5 A cm-2.

Although, in some cases, the lowest value of critical

current density observed at a given chloride ion concentration

approximates the highest seen in the next lower chloride ion

concentration, there is a definite trend in the average

values toward a reproducible maximum at 0.508 M potassium

chloride (Table 2).

The only relationship for which linearity is suggested

is that between the logarithm of critical current density

and the solution composition function, pH + log ( [SO42- +

[Cl1] ), where [S 42-] and [C1-] are the analytical

concentrations of sulfate and chloride, respectively. A

slope of 1.10 is measured (Figure 9). No linearity is

found between the logarithm of icrit and E contrary to
crit r
the work of Wilde.5 However, the most negative rest

potential and highest critical current density values

occur in the same system.

From its value at the maximum the current density

then decreases to a low value ( 10 A cm2 ) characteristic

of the passive region, i and is approximately independent

of potential. In some cases, prior to the attainment of

the passive current value, a negative current loop is

observed (CicC'). Experimentally determined characteristics

of the loop are presented in Table 3 and an experimental





























Figure 9. Variation of the logarithm of the critical
current density with the concentration
function, pH + log ( [S042-] + [ClI- ).


























-4.0-

0


0







0
-4.5-



0 0



0




-5.0-







2.0 2.5



pH + Log ([SOP + [ci])







49

curve containing a cathodic loop is presented in Figure 10.

The potential range in which it occurs, -0.30 to 0.0 V, is

independent of solution composition. The maximum

cathodic current density associated with the loop, ic

(1.36 x 106 A cm-2), is largest in solutions with no
chloride ion (0.334 M sulfate). Solutions containing

chloride ion show significantly lower current values

(2.3 x 10-7 to 4.2 x 10-7 A cm-2) which increase slightly

with increasing chloride ion concentration.

To minimize the effect of the negative current loop,

the magnitude of i is evaluated at 0.3 V except in 0.508

and 1.0 M potassium chloride where pitting occurs below

this potential. The dependence of the passive current

density on solution composition is qualitatively similar

to that of the cathodic loop current. The addition of

1.17 x 10-2 M chloride ion causes a substantial lowering

of i from the value observed in 0.334 M sulfate (5.25 x 10-7
P
versus 1.29 x 10-6 A cm-2, respectively) (Table 1). A

further increase in chloride content is accompanied by

a slight increase in passive current density. In this

region of increasing current, several linear relationships

between i and solution composition are observed (Table 6).

Because of the narrow range covered by the three values of

the passive current density available (5.25 x 10-7 through

7.85 x 10-7 A cm-2), the validity of these relationships

is questionable.

























Figure 10. Potentiostatic polarization curve showing cathodic loop obtained
in 0.334 M sodium sulfate at pH 2.40.










































-..
- -7.0





-5.0-





-3.0- /


-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2


POTENTIAL (V, SCE)








The classical evaluation of the potential of total
passivity, ETp, involves the determination of the point
at which the current increases from its value in the
passive region into the anodic loop of the active-passive
59, 60
transition during a cathodic potential scan. In
the present work the potential of total passivity is
defined as the intersection of the line defined by the
decrease in anodic current from the maximum anodicc
potential scan) with the value of the passive current at
0.30 V. This method gives more reproducible results because
of the random cathodic loop observed here.
Values of ETp observed undergo a positive shift (from
-0.285 to -0.160 V) as the chloride ion concentration of
the solution is increased (Table 1). Examination of the
interrelation of E and solution parameters suggests
TP
several possibilities (Figures 11-15):


Relationship Slope r2

1. -ETp// [ClI] 0.108 V mole 1 0.93
2. -TEp/ [Cl-] /( [S042- + [C1-] ) 0.103 V 0.92
3. -Elp/ [Cl /(2[S0 2-+ [C1-] ) 0.110 V 0.94
4. -Ep/ pH + log ([042-] +[C1] ) 0.218 V 1 0.94
S+decade1

5. ETp/ I log ( [042-] / [Cl )o

a. two segments with slopes -0.016 and -0.039 V

decade-1 and correlation coefficients of 0.98

and 0.99, respectively or





























Figure 11. Variation of the potential of total passivity with chloride ion
concentration.






















0.15










c.O
LL 0.25-





I I I I I I -
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0























Figure 12. Variation of the potential of total passivity with the concentration
function, [Cl] /( Is42] + [Cl )
( S0 + [Cl7 ]











015











0.20- -


FC0




0. 0
w

0

0.25--










01 02 03 0.4 05 0.6 0.7 0.8 0.9 1.0
[c ] /([so^] + [CL-])
[C L71 1 4J






















Figure 13. Variation of the potential of total passivity with the concentration
function, [C1-] /(2 [s0421 + [Cl] )*









0.15- -











0.20--
0

0

00

I-



0.25-
0








0.1 0.2 0.3 04 0.5 0.6 0o7 0.8 0o9 10 00
[c E-] / (2[so 0l + Ec ])





















Figure 14. Variation of the potential of total passivity with the
concentration function, pH + log ( [S04 -] + [Cl-] ).

































0.20+


pH + Log ([S0o:]+ [CL- )


I

























Figure 15. Variation of the potential of total passivity with the
concentration function, log ( [S042-] / [C1- )'
























0 15-






O 0.20--




.I-
I 0.25-







-1.0 0.0 1.0

Log 0LS4 r]/[Cl-









b. one segment with slope -0.030 V decade-I

and correlation coefficient 0.94.

The latter plot lends itself to two interpretations. It

may be considered to have two linear segments whose point

of intersection occurs at ETp = -0.230 V and 0.301 M

potassium chloride which is the smallest chloride ion

concentration investigated in which pitting occurs.

However, evaluation of data in terms of one single line is

also feasible. Differentiation between the two possibilities

is difficult because only five data points are available.

In solutions whose sulfate to chloride ion concentration

ratio is greater than one ( 0.3 M chloride), a continuing

shift to more anodic potentials results in the breakdown

of the passive state with the onset of pitting. An

experimentally determined polarization curve showing the

influence of pitting is given in Figure 16. Two values

are recorded for the potential at which pitting initiates,

Epit Epit(inc) is the most negative potential at which
any increase in current density with time is observed within

the ten minute waiting period. At potentials slightly

positive to Epit(inc) ( 0 to 150 mV), an initial decrease

in current followed by current spikes is usually observed.

At more positive potentials an immediate increase in

current density occurs. Epit(gr) is the potential obtained

by extrapolation of the increasing pitting current density-

potential curve back to the point at which it intersects

the value of the passive current at 0.30 V.

























Figure 16. Potentiostatic polarization curve showing the influence of
pitting obtained in 0.301 M potassium chloride at pH 2.35.























-3.0-






-5.0-






-7.0--
c'J
E
o




-7.0


0

-1

-5.0-







-3.0-


-0.6 -0.4 -0.2 0.0 0.2 04 0.6


POTENTIAL (V, SCE)










Both values of E pit are arbitrary because of the
pit
imposed time limitation for current measurement and the

heavy fluctuations in current of up to an order of

magnitude observed in this potential range, respectively.

Thus, the induction period for pitting may be greater than

ten minutes at potentials more negative than Epit(inc) and

the current magnitude, after ten minutes at a given potential,

is a random function of time as a result of the fluctuations

present. The graphically determined value of the pitting

potential, Epit(gr), is always the more negative. A

three-fold increase in chloride ion concentration shifts

Epit(inc) from 0.395 to -0.090 V and Epit(gr) from 0.360
to -0.155 V.

Several linear relationships are observed between

Epi and the solution composition parameters examined.
pit
Their corresponding slopes, intercepts, and correlation

coefficients are reported in Table 6. Because only three

data points are available, it is difficult to judge which

of these relationships, if any, are valid. It does appear,

however, that the potential at which pitting initiates

shifts toward more active values as the chloride ion

concentration of the solution is increased.

In systems not subject to pitting attack (< 0.3 M

potassium chloride), loss of passive state stability occurs

with the onset of transpassive dissolution (Table 4). The

disolution reaction in this region involves oxidation of










chromium(III) in the passive film to chromium(VI) and

exhibits Tafel behavior over 1 to 1.5 decades of current.

An anodic Tafel slope of 0.114 V decade-1 is observed in

0.0 M and 1.17 x 10-2 M potassium chloride, while a value

of 0.150 V decade-1 is found for 9.97 x 10-2 M potassium

chloride. The potential at which transpassivity initiates,

Etr, is taken as the intersection of the computed Tafel
line with the passive current density at 0.30 V. Etr shifts

negative from 0.671 to 0.575 V with increasing chloride

ion concentration.

Continued anodic polarization produces a slight current

maximum as a result of secondary passivity. Neither the

current magnitude at the maximum, irit,2 (1 x 10 A cm),
crit, 2
nor its associated potential, E pp,2 (0.95 V), is affected

by solution composition at constant pH. The current then

decreases to a minimum prior to oxygen evolution. The

degree of stability of secondary passivity is represented

by the difference in the magnitude of the maximum and

minimum current values, a i. Present results indicate that

the stability decreases slightly on addition of chloride

ion to solutions originally 0.334 M in sulfate ion at

pH 2.4. Current differences,A i, of 14 to 17 pA cm-2 and

21 A cm-2 are observed for 1.17 x 10-2 to 9.97 x 10-2 WM
potassium chloride and 0.0 M potassium chloride (0.334 M

sulfate), respectively.










Secondary solutions

A brief survey of the effect of chloride ion concentra-

tion in the absence of sulfate ion on the polarization

curve was carried out at pH 2.4 (hydrochloric acid) in

solutions containing 0.102, 0.123 and 1.0 M potassium chloride.

Characteristic potentials and current densities obtained

from the polarization curves are presented in Tables 7 and

8. Linear relationships with correlation coefficients

) 0.90 are reported in Table 9 with the corresponding

slopes.

Cathodic current magnitudes at -0.700 V are signifi-

cantly lower than those in sulfate containing solutions,

varying from 1.25 x 10-3 to 7.06 x 10- A cm-2 in the

chloride range 0.102 to 1.0 M. An increase in concentration

from 0.102 to 0.123 M causes only a slight shift in rest

potential (-0.470 to -0.480 V) while the rest potential

for 1.0 M potassium chloride occurs at -0.505 V, a negative

shift of 45 mV. The primary passivation potential is not

a function of chloride concentration. In the absence of

sulfate ion icrit changes linearly with chloride ion

concentration (Figure 17) and with the rest potential

(Figure 18). The corresponding slopes are 3.11 x 10-5 A

cm-2 mole-1 1 (r2 = 0.98) and -15.7 decades V-1 (r2= 1.0),

respectively. A linear relationship is also observed

between the logarithm of the critical current density and

the concentration function, pH + log ( [S04 2- + [Cl )

(Figure 19). The corresponding line has a slope of 0.48














[Cl-] [s4 2-]
x 101 x 101


Table 7

Potentiostatic current-potential behavior in secondary solutions
(average curves).

i-700 icrit 1 1
(A cm-2) Er Eppl (A cm- ) (A cm-2) Ep Ept(V,SCE)
pH x 104 (V,SCE) (V,SCE) x lo5 x 107 (V,SCE) inc gr


2.40

2.35

2.42


Polarization

-12.5 -0.470

-12.5 -0.480

-7.06 -0.505


in the

-0.420

-0.400

-0.420


Polarization

1.52 -203.0 -0.420 -0.370

2.35 -22.30 -0.470 -0.430

6.22 -0.0013 -0.760 -0.500


absence

1.18

1.69

4.17


of sulfate ion

-0.287

-0.273

-0.250


at varied pH


4.65

3.43

0.147


-0.681

7.85


-0.205

-0.230


1.02

1.23

10.0


3.01

3.01

3.01


0.233

0.248

0


0.140

-0.025

-0.100


0.070

-0.120

-0.132




0.397

0.360


0.425

0.395








Table 8

Fotentiostatic current-potential behavior in secondary solutions
(individual experiments).


[C 1 [S04x
x 101 x 101


i-700
(A cm 2)
pH x 104


icrit,l p
Er Epp,l (A cm ) (A cm-2)
(V,SCE) (V,SCE) x 105 x 107


ETp Epit(V,SCE)
(V,SCE) inc gr


Polarization in the absence of sulfate ion


1.02 0 2.40 -12.3

-12.6

1.23 0 2.35 -15.3

-9.61

10.0 0 2.42 -5.71

-8.42



3.01 2.33 1.52 -189.0

-196.0

-224.0

3.01 2.48 2.35 -18.0

-27.0

3.06 0 6.22 -0.0013


-0.460 -0.420


-0.480

-0.480

-0.480

-0.500

-0.500


-0.420

-0.400

-0.400

-0.420

-0.420


Polarization at

-0.400 -0.350

-0.440 -0.370

-0.400 -0.370

-0.475 -0.420

-0.480 -0.415

-0.760 -0.500


0.881

1.49

1.7

1.68

4.69

3.70


varied pH

4.1

5.8

5.25

3.55

3.39

0.147


-0.287

-0.282

-0.285

-0.270

-0.250

-0.235



-0.240

-0.230

-0.155

-0.230

-0.226


7.61

7.49

2.80

7.23

8.48


- 0.075

- 0.120

-0.05 -0.107

0.00 -0.120

-0.10 -0.155

-0.10 -0.120


0.450

0.400

0.400

0.420

0.370


0.335

0.400

0.460

0.365

0.360


































Figure 17. Variation of the critical current
density with chloride ion concentration.




































4.0--


2.0-4-


I II I I


[crt-


6.0--









74



















-4.8--







-4.6- -







-4.4--



0.50 0.49 0.48 0.47



-4.2








0.50 0.49 0.48 0.47


-Er (V,SCE)






























Variation of the logarithm of the critical
current density with the Boncentration
function, pH + log ( [S04 9 + [Cl-] ).


Figure 19.


















































2:0 2:

pH + Log ( [S04o]+ [I-] )

































Figure 20. Variation of the potential of total passivity
with rest potential.































0.24-t-


0.26--


0.28-
c 0.28-


0.30--


I I I I


-Er (V,SCE)



































Figure 21. Variation of the logarithm of the
critical current density with the
potential of total passivity.























































0.25


0.28 0.27 0.26


-ETP (V,SCE)








81
and a correlation coefficient of 0.93. No cathodic current

loop is observed in these solutions.

Little change in the potential of total passivity is

observed in going from 0.102 to 0.123 M potassium chloride

when data from individual runs are examined. Values are

-0.287 V and -0.282 V for the former and -0.285 V and

-0.270 V for the latter. However, average values show a

definite positive trend with increasing chloride ion

concentration in agreement with the value of -0.250 V

found in 1.0 M potassium chloride.

A one-to-one correlation between ET and Ep is

suggested (Figure 20). The corresponding slope is -1.03.

The log of icrit increases as Ep becomes more positive

at the rate of 15.1 decades V-1 (Figure 21).

In the absence of sulfate ion, pitting occurs at all

chloride ion concentrations investigated. A negative shift

in the pitting potential occurs with increasing chloride

ion concentration but no statistically valid variation in

this relationship is apparent.

The effect of pH on polarization curve parameters was

also examined in 0.3 M potassium chloride at pH 1.52, 2.35

and 6.22 (Table 7,8). In the former two cases, sodium

sulfate additions of 0.233 M and 0.248 M, respectively,

maintained the ionic strength at one. No sulfate was

added to the latter and polarization in this system was










terminated after passivation occurred. Possible linear

relationships are presented in Table 9.

Net currents at -0.700 V are cathodic and decrease

with decreasing hydronium ion activity. Values of

2.03 x 10-2, 2.23 x 10-3 and 1.34 x 10-7 A cm-2 are recorded

for pH 1.52, 2.35, and 6.22, respectively. A linear

relationship is observed between both the rest potential

and pH (-0.073 V pH-1) (Figure 22) and the primary passiva-

tion potential and pH (-0.025 V pH-1) (Figure 23). The

slope of the straight line resulting from a plot of log

icrit s pH is -0.328 (Figure 24). A linear relationship
is also observed between the logarithm of the critical

current density and the concentration function, pH +

log ( [S0 2-] + [Cl-] ) (Figure 25). A slope of -0.35

and a correlation coefficient of 0.99 are associated with

the line.

No cathodic loop is found in the pH 2.35 solutions

but one does occur between -0.15 and 0.0 V in two of three

experiments at pH 1.52 (Table 10). The maximum associated
-7 -2
cathodic current density is 4.5 x 10 A cm-

The increase in pH from 1.52 to 2.35 shifts the

potential of total passivity to more active values (-0.205

to -0.230 V) and changes the current density at 0.300 V from

cathodic to anodic values. An active shift in pitting

potential also occurs.








Table 9

Interrelationships in potentiostatic experiments in
secondary solutions.

Independent Dependent
Variable Variable Slope

[C1- i crit 3.11 x 10-5 A cm-2 moles
Er log i crit 15.7 decades V-1

Er ETP -1.03

ETp log i crit 15.1 decades V-1
pH + log ([O042-] + [Cl] ) log i crit 0.48

pH Er -0.073 V pH-1

pH Epp -0.025 V pH-1

pH log i crit -0.328

pH + loo ([SO42-j + [Cl] ) log i crit -0.35


Correlation
Coefficient

-11 0.98

1.0

0.99

0.99

0.93
1.0

0.91

0.99

0.99


Data
Points

3

3

3

3

3

3

3

3

3


~


























-Hd 1q;TmM T'ET9uaGod uoTOATlsspd ReuTjid eq. jo uoTeBaltjA "'3 ajnrTJ
















0.30--


0.40--


0.50--


I 1























*H-d M1qTM -[-reTueaod qseJ GqO Jo UOTLBTJeA *CZ aJnSTj



























0.5 0--


0,704-




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EOCPAC6RA_01GJXP INGEST_TIME 2017-07-17T20:19:15Z PACKAGE UF00098874_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

DIFFERENTIAL CAPACITY OF STAINLESS STEEL IN POTASSIUM CHLORIDE SOLUTIONS DURING POTENTIOSTATIC AND GALVANOSTATIC POLARIZATION By M. ELAINE CURLEY-FICRINO A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1975

PAGE 2

The author dedicates this dissertation to her husband, John A. Fiorino.

PAGE 3

ACKNOWLEDGEMENTS The author would like to express her gratitude to the members of her committee; Dr. E. D. Verink, Jr., Dr. J. D. Winefordner, Dr. W. S. Brey, Dr. R. Bates, and especially to her research director, Dr. G. M. Schmid for the interest and assistance given to her in the course of this investigation and in the preparation of this manuscript. She would also like to thank Mr. R. Dugan and Mr. A. Grant and their associates for their help in the technical aspects of this work. Steel samples were provided by the United States Steel Corporation. The author is also grateful for the financial assistance received from the University of Florida in the form of Graduate School Fellowships. Finally, the author would like to thank Dr. and Mrs. James L. Fortuna and Mr. and Mrs. Willis Bodine for their friendship and encouragement during her graduate career. iii

PAGE 4

TABLE OF CONTENTS Fage ACKNOWLEDGEMENTS iii LIST OF TABLES vi LIST OF FIGURES viii ABSTRACT x Chapter I. INTRODUCTION 1 II. EXPERIMENTAL 13 Experimental Design 13 Experimental Technique 18 Fotentiostatic polarization 21 Galvanostatic polarization 25 III. RESULTS 3^ Current-Potential Behavior During Fotentiostatic polarization 3^ Frimary solutions 3? Secondary solutions 68 Capacity-Potential Behavior During Fotentiostatic Folarization 92 Frimary solutions 92 Secondary solutions 1°2 iv

PAGE 5

Chapter Page Galvanostatic Polarization 103 Potential-time behavior 103 Capacity-potential behavior 122 IV DISCUSSION 131 Potentiostatic Polarization 131 Active dissolution and passivation 131 The hydrogen evolution reaction 135 Rest Potential 142 Passive region 144 Pitting 150 Transpassive dissolution 156 Galvanostatic Polarization. .... 160 V SUMMARY 173 LITERATURE CITED 179 BIOGRAPHY 187

PAGE 6

LIST OF TABLES Table pa S e 1. Fotentiostatic current-potential behavior in primary solutions (average curves) 38 2. Fotentiostatic current-potential behavior in primary solutions (individual experiments) . . 39 3. Cathodic loop in primary solutions ^0 I±. Transpassive behavior in primary solutions . . 41 5. Solution composition parameters investigated . in primary solutions ^2 6. Interrelationships in potentiostatic experiments in primary solutions ^3 ?. Potentiostatic current-potential behavior in secondary solutions (average curves) 69 B. Potentiostatic current-potential behavior in secondary solution (individual experiments). . 70 9. Interrelationships in potentiostatic experiments in secondary solutions 83 10. Cathodic loop in secondary solutions 95 11. Potentiostatic capacity-potential behavior in primary solutions (average curves) 96 12. Potentiostatic capacity-potential behavior in primary solution (individual experiments) ... 97 13. Fotentiostatic capacity-potential behavior in secondary solutions (average curves) 98 Ik. Potentiostatic capacity-potential behavior in secondary solutions (individual experiments) . 99 15. Capacity-potential behavior in the potential range of linearity of 1/C versus E 101 16. Galvanostatic potential-time behavior (0.0 K potassium chloride) 106 17. Galvanostatic potential-time behavior (0.100 M potassium chloride) 107

PAGE 7

Table Fa ? e IS. Galvanostatic potential-time behavior (0.303 M potassium chloride) 108 19. Galvanostatic potential-time behavior (0.518 M potassium chloride) 109 20. Tafel behavior during galvanostatic polarization HI 21. Average final breakdown potentials during galvanostatic polarization Ho 22. Open circuit decay behavior following galvanostatic polarization 118

PAGE 8

LIST OF FIGURES Figure Page 1. Stainless steel electrode 1^ 2. Electrochemical cell 17 3. Kel-F electrode holder 20 k. Block diagram of the potentiostatic polarization circuit 23 5. Block diagram of the galvanostatic polarization circuit 2? 6. Block diagram of the circuit employed in capacity-potential data correlation 31 7. Control panel 3 2 8. Schematic potentiostatic current-potential "behavior 3& 9. Variation of the logarithm of the critical current density with the concentration function, pH + log ( [S0 4 2-] + [ci-] ) ^8 10. Potentiostatic polarization curve showing cathodic loop obtained in 0.33^ M sodium sulfate at pH 2.^0 51 11. Variation of the potential of total passivity with chloride ion concentration 5^ 12. Variation of the potential of total passivity with the concentration function, [Cl~]/(2 [SC^i +[C1-J ) 56 13. Variation of the potential of total passivity with the concentration function, [Cl-]/(2[S0^2-] + [C1-] ) 58 Ik. Variation of the potential of total passivity with the concentration function, pH + log ( [S0^2-J + [C1-] ) 60 vm

PAGE 9

Figure Page 15. Variation of the potential of total passivity with the concentration function, log ([SO^ 2 -] / [CI"] ) 62 16. Potentiostatic polarization curve showing the influence of pitting obtained in 0.301 M potassium chloride at pK 2.35 65 l?. Variation of the critical current density with chloride ion concentration 72 18. Variation of the logarithm of the critical current density with rest potential 7^ 19. Variation of the logarithm of the critical current density with the concentration function, pH + log ( [SO^-] + [C1"J ) 76 20. Variation of the potential of total passivity with rest potential 78 21. Variation of the logarithm of the critical current density with the potential of total passivity 80 22. Variation of the primary passivation potential with pH 85 23. Variation of the rest potential with pH . . . 87 2h. Variation of the logarithm of the critical current density with pH 89 25. Variation of the logarithm of the critical current density with the concentration function, pH + log ( [SO4 2 -] + [CI"] ) 91 26. Potentiostatic capacity-potential behavior in the absence of hydrogen interference ... 9^ 27. Schematic representation of the galvanostatic potential-time curve in the presence (1) and absence (2) of pitting breakdown 105 28. Schematic representation of the capacity-time behavior of a system not subject to pitting breakdown during galvanostatic polarization . 12^

PAGE 10

Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DIFFERENTIAL CAPACITY OF STAINLESS STEEL IN POTASSIUM CHLORIDE SOLUTIONS DURING POTENTICSTATIC AND GALVANCSTATIC POLARIZATION By M. ELAINE CURLEY-FIORINO June, 1975 Chairman: Gerhard M. Schmid Major Department: Chemistry Potentiostatic and galvanostatic polarization of AISI 304 stainless steel was performed in deaerated solutions of 0.0 through 1.0 M potassium chloride at pH 2.4. Sodium sulfate was added to an ionic strength of one. Cylindrical electrodes were mechanically polished and prepolarized at -0.?00 V for twenty minutes. The differential capacity of the metal-solution interface was determined as a function of potential using the single pulse technique. All potentials are given versus the saturated calomel electrode. Current densities and differential capacities are referred to the electrode geometric area. During potentiostatic polarization, an active-passive transition is observed for all solutions. The rest potential (-0.48 V) and primary passivation potential

PAGE 11

(-0.42 V) are independent of solution composition. The critical current density increases with the total anion concentration. The capacity peak occurring slightly negative to the primary passivation potential is attributed to specific adsorption of the anions involved in the dissolution-passivation mechanism. The potential at which the electrode becomes totally passive shifts positive with increasing chloride ion concentration and is dependent on kinetic rather than thermodynamic factors. Transpassive dissolution initiates between 0.57 and 0.64 V and exhibits Tafel behavior. Both the current density maximum (10 -4 A cm" 2 ) and the corresponding potential (0.95 V) associated with the onset of secondary passivity are independent of solution composition. The capacity peak observed in this region is attributed to adsorption of the passivating species. In solutions of >0.3 M potassium chloride, an increase in current density caused by pitting starts at potentials normally in the passive region. The pitting potential depends on the relative amounts of chloride and sulfate ions present. No capacity peak is observed prior to pitting breakdown. The capacity, between the potential of total passivity and the capacity minimum occurring at 0.60 V, is approximately independent of the stability of the passive state. This independence results from the domination of the interfacial capacitance by that of the passive film, obscuring xi

PAGE 12

any changes in the film-solution capacity which may occur. The sudden decrease in capacity observed at 0.3° V is attributed to a change in the dielectric constant of the film. Between 0.3 and 0.6 V, film growth follows an inverse logarithmic law. The potential-time behavior of samples subjected to galvanostatic polarization depends on both current density and solution composition. Systems not susceptible to pitting reach and maintain a positive steady state potential. In systems subject to pitting, the maximum potential attained is unstable and a shift in potential to more active values occurs. In 0.3 M potassium chloride intermediate arrests produced by a given current density as well as the maximum potential achieved prior to breakdown correspond to potential arrests in the non-pitting systems. In 0.5 K potassium chloride pitting breakdown occurs from a potential maximum which is considerably below that observed in 0.3 M but the behavior at more negative potentials is similar. It is assumed, therefore, that the initial effect of the anodic current on the metal surface at a given potential is the same in all cases and that pitting succeeds through the perturbation of these initial surface conditions by chloride ion.

PAGE 13

Tafel behavior is associated with the arrests at -0.il-, 0.8, 0.85, and 1.12 V. Active dissolution at -0.4 V exhibits a slope of 0.060 V decade" and is thought to proceed by the Heusler mechanism. Arrests at 0.8, 0.85 and 1.12 V correspond to transpassive dissolution, secondary passivity, and oxygen evolution, respectively. Capacity peaks are associated with the latter two effects and with the arrest at -0.4 V but not with pitting breakdown.

PAGE 14

CHAPTER I INTRODUCTION A metal is in the passive state when it is inert in an environment in which, on the basis of thermodynamics, it should corrode readily. An ennoblement of the potential of the metal-environment interface accompanies the onset of passivation. Passivity has been recognized since I836, when Faraday observed the stability of iron metal immersed in concentrated nitric acid. However, the nature of the surface species leading to the onset and maintenance of the passive state is still not well understood. The two major theories advanced to exolain the phenomenon are the bulk oxide theory and the adsorption theory. The oxide film theory proposes that a bulk oxide is formed directly on the metal surface from the 2,3,^ products of the metal dissolution reaction. This oxide then acts as a physical barrier between the metal surface and the aggressive environment. The adsorption theory of passivity attributes passivation of the metal surface to the adsorption of an "oxygen" species in less than or equal to monolayer quantities.-" 5 ' Two variations of the adsorption theory 1

PAGE 15

2 have been described. In the chemical variation proposed 9 R by Uhlig, adsorbed oxygen atoms are believed to satisfy the surface valences of all atoms on the metal surface. The correlation of the theoretical and experimental concentrations of components of transition metal alloys needed to bring about a sharp increase in the ease with which the alloys are made passive lends support to this hypothesis. The electrochemical variation ascribes the retardation of the metal dissolution reaction to a change in the double layer structure caused by the dipolar character of chemisorbed oxygen. ^ Orientation of the dipole with its positive end towards the solution increases the activation energy necessary for the metal dissolution reaction. Hackerman 10 has proposed a theory intermediate to the above two. Here, the adsorption of oxygen atoms on the metal results in a metastable state lending temporary protection to the surface. Following electron transfer from the metal to adsorbed oxygen atoms, an amorphous bulk oxide is formed by cation migration through the adsorbed array. It is this bulk oxide which provides long-term protection. This theory, originally postulated for metals in oxygen-containing solutions, has also been proposed to explain the passivation mechanism of ironchromium alloys in deaerated acidic sodium sulfate solutions. A similar mechanism has been proposed by Frankenthal for the passivation of an iron-24 chromium

PAGE 16

3 alloy. Electron diffraction studies of the surface of stainless steels exposed to oxidizing acids or the atmosphere for short times at intermediate temperatures (25 to 60 C) 13 have shown that the passive film formed is non-crystalline. Transmission microscopy studies of films formed on an iron-24 chromium alloy during potentiostatic polarization in 0.5 M 14 sulfuric acid show that they are also amorphous. The corrosion resistance of austenitic stainless steels in diverse environments is a result of the high degree of passive state stability conferred by the presence of chromium in amounts > 12 percent. 1 ^' However, in specific media, especially in halide solutions, this corrosion resistance is lost, as evidenced by the onset of intense local attack, i. e., pitting. The mechanism by which pits nucleate and propagate in the presence of chloride ion has been investigated • 17 in great detail. Excellent reviews are given by Kolotyrkm 1 P, and Szklarska-Smialowska. The pitting phenomenon has been characterized by three parameters! the pitting potential, negative to which no pits can nucleate; the critical chloride ion concentration, the minimum concentration needed to initiate pitting in a given system; and the induction time, the time, at a given potential, which passes prior to breakdown. Thus, studies have been carried out

PAGE 17

19. 20, 21, 22, 23 to examine the effect of metal composition, defect density of a metal surface, ' D grain size, solution composition, 2 ?' 28 ' 29 ' 22 sulfide inclusions, 30, 3 ' and temperature 28 on the pitting potential and/or the location and number of pit sites. The critical chloride ion concentration has been shown to depend strongly on the concentration of inhibiting anions (e.g., sulfate, hydroxyl and nitrate ions) in the solution 33, 3 ' 35 ' * ' as well as on alloy composition. 3 The induction time has been found 37 to decrease with increasing chloride ion concentration, 38, 39. ^0 potential of passivation, 39, and temperature. 3 The influence of an induction time is seen both in potentiostatic and galvanostatic polarization measurements. Previous works on austenitic stainless steels in chloridecontaining solutions have shown that the application of a constant anodic current to an initially active electrode shifts the potential positive to a maximum value. Despite the continued application of the current, the potential then decreases rapidly to a value normally in the passive region. The length of time to potential breakdown is a function of the current density and the chloride ion concentration. Examination of the metal surface shows that pitting has occurred.

PAGE 18

5 The steady state potential achieved after breakdown has been called the protection potential of the system under study. ' At potentials between the protection and pitting potentials previously nucleated pits can continue to grovi but no new pits may form. At potentials negative to the protection potential, all existing pits become inactive. J Anodic potentiostatic polarization to values more positive than the pitting potential also causes pit formation. A change in potential to values just positive to the pitting potential results in a decrease in current density with time from the initially high value associated with double layer charging. ^ This current decrease represents the readjustment of the electrode-solution interface to maintain the passive condition. After a time interval which depends on solution composition and potential, the current decrease is replaced by current oscillations signifying the pi tting-induced breakdown of passivity. As the potential is made still more positive, the time for which passivity is maintained decreases until an induction period is no longer apparent. J Mechanisms proposed to explain the dependence of pitting parameters on experimental variables are a function of the assumed nature of the passivating film. On the basis o^ a bulk oxide, Hoar^ has described a mechanical breakdown process in which the mutual repulsion

PAGE 19

of anions adsorbed on the oxide surface leads to the formation of cracks. A relationship between critical breakdown stress and surface tension as influenced by anion specific adsorption has been derived by Sato. Hoar and Jacob ^ have also suggested that a metal cation is dissolved from the oxide through the formation of a metal-chloride complex containing 2.5 to k.5 chloride ions. Cation migration through the film then allows continuation of the process. If the passive film is assumed to be an adsorbed "oxygen" species, then its replacement by chloride ions will lead to activation of the metal when a critical surface concentration of 28, 17 chloride ion is attained. The presence of an induction time associated v/ith pitting breakdown suggests that a time consuming change in surface structure is occurring. If the specific adsorption of chloride ion is involved in this change, then its incorporation into the electrical double layer should be reflected in the differential capacity-potential behavior of the interface. The electrical double layer, in its ability to store charge, acts as a capacitor. The magnitude of the capacity associated with it is a complex function of potential and is therefore defined as a differential capacity, "Sq/bE. In the absence of specific adsorption, only water molecules populate the compact double layer,

PAGE 20

7 the region between the metal and the Outer Helmholtz Plane. The capacity of the interface can then be represented by two capacitors in series; that of the region between the metal and the Outer Helmholtz Plane (OHP); and that of the region between the OHP and the bulk of the solution (the diffuse double layer). In dilute solutions ( < 0.001 M), the capacity of the latter is small and dominates the interfacial capacity. In concentrated solutions, all of the diffuse double layer charge is located close to the OHP. The interface then behaves like a single parallel plate condenser, i.e., its capacity is approximately independent of potential. When specifically adsorbed ions populate the IHP in a concentrated solution (> 0.001 M), the interface again functions as two capacitors in series. Eockris and Reddy,^ 6 in their treatment of the effect of contact adsorption on the total capacity of the metal-solution interface, derive the equation i/c = iAm-ohp
PAGE 21

8 positive with increasing potential, an increase in the degree of specific adsorption with metal charge ( -Z) 2 qQ A /"c>q M ~ > 0) should occur, producing an increase in the measured differential capacity of the interface. As growth continues, however, the buildup of lateral repulsion forces tends to decrease the degree to which specific adsorption occurs at a given metal charge. This inflection in the QQ^-q^ relationship results in a peak in the differential capacity-potential curve ( -2> 2 q CA /"2>q M = 0). Differential capacity measurements have been successfully applied to the determination of the specific adsorption of sulfate, perchlorate and chloride ions on iron. 4 ?' 48 ' **9» 50 The presence of a positive charge on the metal surface implies polarization at potentials positive to its zero point of charge. Studies of binary alloys suggest that the zero point of charge of an alloy should approach that of its component with the most negative zero point of charge if it is present in sufficient concentration. * The zero points of charge for nickel, chromium and iron have been determined in acidic sulfate solutions. J The corresponding values are -0.57, -0.69 and -0.62 V, respectively. It is expected then, that the zero point of charge of active stainless steel should be close to that of chromium, i.e. -0.69 V.

PAGE 22

9 In order to explain the potential dependence of the activating effect of chloride ion in terms of specific adsorption to a critical surface concentration on the passive electrode, the zero point of charge of the passive surface must lie in the passive region. For passive iron, the zero point of charge occurs at 0.125 V in 0.01 M sodium hydroxide. ^ 2 This positive shift in value from that observed on active iron can be attributed to an increase in work function resulting from the presence of an adsorbed "oxygen" species or an oxide film. The extension of this phenomenon to stainless steel seems logical. The primary purpose of the experiments conducted in this study was the determination of the effect of chloride ion on the capacity-potential behavior of stainless steel observed during potentiostatic and galvanostatic polarization in solutions initiating pitting. In order to explain the potential arrests observed during constant current polarization, the potentiostatic studies were extended to cover the potential range from active dissolution to oxygen evolution. Corresponding capacity values were determined and the current-potential and capacity-potential data correlated. The relationship between the rate of an electrochemical reaction and the potential difference across the interface at which it occurs is discussed in detail by Bockris and

PAGE 23

10 Reddy. When electron transfer is rate-determining, i.e., the system is under activation control, the currentpotential relation is given by the general form of the Butler-Volmer equation, i = i exp «JL« i exp ( -S-£ )«1 (2) RT l RT L where i is the exchange current density, the ot's are the transfer coefficients, and n is the overvoltage. All other terms have their usual meaning. The first term on the right hand side of equation 2 is the current density resulting from the oxidation (anodic) reaction. The second term pertains to the reduction (cathodic) reaction. At the equilibrium potential of the rate-determining reaction, the overvoltage, which represents the potential difference across the interface in excess of the equilibrium potential difference ( E E Q ), is zero. The net current density observed (i ) will therefore also be zero since at equilibrium the rates of the anodic and cathodic reactions will be equal. The transfer coefficients determine what fraction of the potential difference across the interface is operative in changing the energy barrier for the oxidation and reduction reactions. As the potential difference across the interface is made more positive (*|>0), the contribution of the anodic current density to the total current density will increase. At a sufficiently positive overvoltage (~ 0.120 V for a one-electron transfer reaction), the

PAGE 24

11 influence of the cathodic current density becomes negligible. The current-potential relationship is then given by i = i exp -S*J_ri (3) o RT v. which can be put in logarithmic form and rearranged to give = 20 RT lQg . . 2.? RT log io (4) L 5f *f A plot of overvoltage versus log i is therefore linear. Such plots are known as Tafel lines. The slope of the line contains the transfer coefficient which is a complex function of the total number of electrons transferred during the reaction as well as the mechanism by which the reaction proceeds. From the slope of the line and its intercept, the exchange current density for the reaction can be calculated. In systems in which a faradaic current can flow, i.e., charge can cross the metal-solution interface, the electrical behavior of the interface can be represented by a resistor in series with a capacitor and resistor in parallel. The series resistor represents the resistance of the solution to current flow. The capacity is the differential double layer capacity. The parallel resistor represents the polarization resistance of the faradaic reaction, decreasing as the rate constant of the reaction increases. In cases of low polarization resistance, the determination of the differential capacity is difficult since the polarization resistance can act as a leakage

PAGE 25

12 path for the signal measuring the capacity. Faradaic current also interferes indirectly by causing a change in the true electrode surface area as well as in the solution composition . The use of the classical alternating current technique in which the interface forms one arm of an impedance bridge to determine the double layer capacity on a solid electrode is precluded because of the dependence of the 53 capacity value measured on signal frequency. Most of the direct interference can be eliminated, however, by the use of the single pulse method of differential capacity measurement developed by Riney, Schmid and Hackerman. J Analysis of the linear segment of the potential transient resulting from a single current pulse allows calculation of the capacity at the point from which the pulse initiates since C = i(dt/dE) (5) ' t=0 where i is the pulse magnitude and dt/dE is the slope of the potential-time transient evaluated at t = 0.

PAGE 26

CHAPTER II EXPERIMENTAL Experimental Design The material investigated was stainless steel, AI5I 304, provided by the United States Steel Corporation. Its composition was given as 0.03 C, 0.02? P, 1.10 Kn, 0.022 S, 0.43 Si, 9.26 Ni, 18.6 Cr, O.39 Mo, and 0.04 N (weight percent). Ear stock was machined to cylinders with a diameter of 6 mm and a height of 9 mm (Figure 1). The cylinders were tapped, threaded, and mechanically polished at 2400 rpm with 400 followed by 600 grit emery paper. They were then degreased with spectral grade benzene in an ultrasonic cleaner, rinsed with triply distilled water, and stored in a closed polyethylene container until needed. Primary studies were carried out in solutions containing 0.0, 1.17 x 10" 2 , 9-9? x 10" 2 , 0.301, O.508, and 1.0 M potassium chloride. Solution pH was measured with a Beckman pH meter and adjusted to 2.4 with concentrated sulfuric acid. Sodium sulfate was added as required to maintain an ionic strength of one (O.3I8, 0.313* 0.284, 0.232, 0.14?, and 0.00 M, respectively). Secondary experiments involved solutions of pH 2.4 containing 0.102, 0.123 and 1.0 M potassium chloride with no sodium sulfate* additions as well as 0.3 M potassium chloride at pH I.52 13

PAGE 27

m Figure 1. Stainless steel electrode

PAGE 28

15 and 6.22. All chemicals used in solution preparation were reagent grade. Recrystalli zation of potassium chloride from triply distilled water had no effect on experimental results. The water employed was distilled from alkaline potassium permanganate and then from a twostage Heraeus quartz still and collected in a two-liter Pyrex volumetric flask. Its maximum conductivity, determined with a General Radio Impedance Bridge, was 2 x 10" J fL " cm" . Platinum electrodes, generally used in pre-electrolysis to remove electroactive impurities from the solution, have been shown to dissolve when polarized anodically in both sulfate and chloride containing solutions. 55 Because of the possibility of contaminating both the solution and the stainless steel surface with platinum, a pre-electrolysis step was therefore omitted. The electrochemical cell was made of Pyrex and was of conventional design (Figure 2). A Luggin capillary connected the saturated calomel reference electrode (SCE) to the cell via two solution-lubricated mercury-seal 2 stopcocks and a potassium chloride salt bridge. A 1 cm platinum flag auxiliary electrode was mounted on the cell with a standard taper joint. For use in constant current polarization and capacitance measurements, a platinum gauze basket, approximately 100 cm in area (Engelhard Industries), was mounted concentric to the test electrode. The cell cap incorporated a 24/^0 standard taper joint for mounting the test electrode.

PAGE 29

o rH o •H 6
PAGE 30

17

PAGE 31

18 The electrode holder (Figure 3) was made from a Kel-F rod machined to approximately 1 cm in diameter in which a 3 mm center hole was drilled. The rod was then heated and a threaded stainless steel rod inserted so that thread protruded on both ends. The Kel-F rod was shrunk in an ice bath to facilitate its insertion into a 1 cm ID glass stirrer bearing sleeve which had a 24/^0 standard taper joint attached. The sample was then affixed to the protruding steel rod. The electrode area exposed to solution was approximately 2 cm . The sample-sample holder seal could be tightened by turning a finger nut at the top of the assembly. Teflon washers were placed between both the sample and holder and the nut and holder to reinforce the seal. Spot checks of the shielded top surface of the test electrode showed no evidence of corrosion indicating the absence of leakage. Experimental Technique Solutions were deaerated with helium (99-99 percent) for a minimum of eight hours prior to use. Prepurif ication and water saturation of the gas was accomplished by passing o it through a 12 cm column of Linde 5 A molecular sieve pellets into a gas wash bottle containing triply distilled water. The gas then flowed into a two-liter reservoir containing the solution and continued through a dispersion

PAGE 32

Figure 3« Kel-F electrode holder.

PAGE 33

20 24/40 STANDARD TAPER JOINT — TEFLON WASHERFINGER NUT ^ s *s £ TEFLON WASHER § S 1 s £ I KEL-F ROD STAINLESS STEEL ROD 1 ty

PAGE 34

21 tube into the electrochemical cell. The cell gas outlet terminated in a gas wash bottle to prevent contamination of the cell contents with ambient air. Immediately before each experiment, the cell was washed with hot chromo-sulfuric acid cleaning solution and rinsed with triply distilled water. Gas pressure was then diverted to fill the cell v/ith deaerated solution from the reservoir. The stainless steel sample was secured on the Kel-F holder, rinsed with triply distilled water and the solution to be studied, and immersed in solution. All samples were pretreated at -0.700 V for twenty minutes to reduce air-formed surface films. Solution stirring was accomplished v/ith a magnetic stirring bar and the helium flow continued throughout the experiment. All solutions were at room temperature and all potentials are reported relative to the saturated calomel electrode. Current density and differential capacity were calculated using electrode geometric areas. The test electrode was grounded during all experiments . Potent! os tatic polarization The block diagram of the system employed in potentiostatic experiments is shown in Figure 4. Polarization was accomplished with a slightly modified Harrar 5 potentiostat. (Two each of obsolete transistors 2N333A and HA753'twere replaced with the equivalent circuit elements 2N3568, 2N5869, and 2N36*J4,2N5867 . )

PAGE 35

o

PAGE 36

23

PAGE 37

24 After prepolarization at -0.?00 V the potential of the test electrode was shifted anodic in step-wise increments. The magnitude of the imposed step depended on the potential range under investigation and the electrochemical reaction(s) associated with it. In regions where changes in potential caused significant changes in current density, steps of 20 to 30 mV were generally employed. In regions of approximately constant current, 50 m V steps were used. A Keithley 610B electrometer was used to monitor the applied potential. After an arbitrary time interval of 10 minutes, the current flowing in the auxiliary-test electrode circuit was determined from the potential drop across a 1.18 kCL or a 10 kfi. wire-v/ound precision resistor (1 percent) shunting the input of a Keithley 660 electrometer (10 A input impedance). The differential capacity of the stainless steel-solution interface was determined as a function of potential using the single pulse method. ^ A voltage pulse from the gate of a Tektronix 5^9 storage oscilloscope was used to trigger a Tektronix 114 Pulse Generator. The 100 usee square wave current pulse produced was applied to the platinum basket-test electrode circuit. The resulting potential-time transient was recorded on the oscilloscope operated in the storage mode at a sensitivity of 2 or 5 mV cm and 2 or 5 usee cm . Differential capacity values were calculated from the linear segment of the transient slope during the initial 10 usee of the pulse. Since the transient is

PAGE 38

25 linear between 4 and 10 usee, the slope determined between these two points is equivalent to that of the tangent drawn to the curve at t = k jisec. Measuring times of 4^.sec correspond to an alternating current frequency of approximately 1.2 x 10^ Hz. Since a frequency of 10' Hz is required to eliminate faradaic interference completely, very little interference is expected. Prior to each experiment, the pulse magnitude was calibrated using standard capacitors and resistors. The current density of the pulse could be calculated from the known value of capacity in the calibration circuit, the measured slope of the transient, and the electrode geometric _2 area. Pulse magnitude was approximately 5 mA cm and was independent of capacity between 1 and lO^farads. Galvanostatic polarization After potentiostatic prepolarization the system was switched to galvanostatic control using a two-position toggle switch. A block diagram of the circuit used during galvanostatic polarization experiments is given in Figure 5. A constant current was supplied to the platinum basket-test electrode circuit by a Hewlett Packard 881A power supply operated in the constant voltage mode in series with a bank of resistors (if.? to ^?.8 kH ). The positive terminal of the power supply was grounded to establish the test electrode as the anode. The current magnitude was determined from the voltage drop developed

PAGE 39

p

PAGE 40

2?

PAGE 41

28 across a lOOitprecision resistor (1 percent) using a Keithley £60 electrometer and was adjusted by varying the series resistance and the voltage output of the power supply. Applied current densities ranged from 0.99 x 10~" r to 2.0 x 10'-v A cm"'". The resulting potential-time curve was displayed on the recorder of a Beckman Electroscan 30. Fotential values were monitored with a Keithley 6lOB electrometer. A study of differential capacity as a function of potential was also carried out during constant current polarization. As described above, a square wave current pulse was delivered from the pulse generator to the platinum basket-test electrode circuit. The resulting potential-time trace was displayed on the storage oscilloscope and the capacity calculated from the initial slope of the transient and the predetermined pulse magnitude. The development of the theory of the single pulse method assumes that a perturbation current is applied to a system under constant current polarization at a steadystate potential so that the current before and after the pulse remains the same. Since, during potentiostatic polarization of austenitic stainless steel, either a steady state current or a very slow decrease in current with time is observed, the method should be directly applicable. For galvanostatic polarization the constant current condition

PAGE 42

29 is satisfied but a continuous change in potential with time occurs/ However, for small measuring times ( ^ 10 ^isec), the change in potential will be negligible. Because of the rapid change in potential with time during constant current polarization, difficulties were encountered in the correlation of capacity values and the potentials at which they were determined. A system was therefore developed which facilitated the sequential storage of pulse-produced transients and supplied a marker signal to the Electroscan recorder whenever the pulse generator was triggered (Figure 6). Two oscilloscopes were used sequentially. Depression of PG 1 or PG 2 on the control panel (Figure 7) first changed the DC level of the input of the oscilloscope in use by 10 mV. This provided automatic downward displacement of the successively produced potential-time transients. A two-step generator was associated with PG 1 , an eightstep generator with PG 2. In most cases, the oscilloscope sensitivity needed for precise determination of the transient slope ( 2 or 5 mV cm -1 ) precluded the use of automatic stepping, and vertical displacement was accomplished manually instead.

PAGE 43

p H cd •H P QJ P O P I >3 -P >>

PAGE 44

I
PAGE 45

32 o SET CLIP TP PG 2 PG, RESET REC CELL VERT T, T 2 PG Fiffure 7. Control panel,

PAGE 46

33 Then, after a 50/jisec delay, the horizontal sweep of the Oscilloscope was triggered. Simultaneously, 15 V, 100 msec pulse was supplied to the external trigger of the pulse generator to produce the current pulse needed to determine the capacity. At the same time a 50 mV, 100 msec pulse was superimposed on the galvanostatic potential-time response signal from the cell to the recorder. Since the potential values recorded during constant current polarization reached 1.4 V, while the marker pulse superimposed on this signal was only 50 mV, the cell signal was passed through a buffered amplifier and could be clipped at a preset value. The recorder sensitivity could therefore be adjusted to display the pulse marker while the entire potential-time curve remained on chart.

PAGE 47

CHAPTER III RESULTS Current-Fotential Behavior During Potentiostatic Polarization The response of the stainless steel samples to potentiostatic polarization is given schematically in Figure 8. The region ABC represents the transition from the active to the passive state characteristic of this material in the solutions employed. An adjoining region of approximately constant current is then usually observed (CC'D). The increase in current on further polarization is caused by loss of passive state stability as the result of pitting (DE) or of transnassivity (FG). The potential and current density notations on the schematic are defined below. They are presented in the text in approximately the same order in which they arise during polarization. Possible interrelationships between experimentally determined current densities, potentials, and/or solution composition variables were examined. The various data were plotted. If a plot appeared to be linear, then a linear regression analysis was performed. Only curves p with correlation coefficients (r ) of ^ 0-90 were reported as linear. 3^

PAGE 48

o > •H P f.: +-' o a, i •p C a> U Sm o cd •P tfl o P C Q) -P O Ph

PAGE 49

36 oiqonv ; 6o-| oiaoHivo

PAGE 50

37 Primary solutions Characteristic potentials and current densities determined from the potentiostatic polarization curves obtained in primary solutions are presented in Tables 1-4. The solution composition parameters investigated are listed in Table 5« The logarithms of these functions have also been considered. Relationships with correlation coefficients ^ 0.90 are reported in Table 6. After the twenty-minute prepolarization period at -0.700 V, net currents are cathodic with values between 2.1 x 10 and 5-3 x 10 A cm . No correlation between the hydrogen evolution current and solution composition at constant pH and constant potential is apparent (Table 1). As the potential is shifted anodic, the measured current decreases and becomes zero at the rest potential, E„. The rest potential is defined as that potential at which the absolute values of the internal anodic and cathodic currents become equal, resulting in a net external current flow of zero. Values of the rest potential determined here are -0.^50 to -0.510 V. The maximum average deviation for data obtained for a given solution composition is 16 mV for 0.0 M potassium chloride (Table 2). For all other compositions it is substantially less ( ^ 7 mV). Polarization at potentials positive to E results in a net anodic current which increases with potential to a maximum, signifying the onset of passivation. This maximum current density and its corresponding potential are the

PAGE 51

38 ,

PAGE 52

39 cd 3 'C •H > H c •H rH

PAGE 53

<*0 CI Oo£ o (VOH I •H < X CM I CNS O O rH o H < X

PAGE 54

CM I ^o ' f* 2 •H O i— I

PAGE 55

kz

PAGE 56

^3 6 •H ft d • H CO d 0) S •H S-i CD ft X O •H P cd P CO O .H P c
PAGE 57

44 w p> cd a P.H a3 o Q ft c-p

PAGE 58

45 CO p cd C P -H Ctf o Q ft CP o fl •H CD P .H Cd O H -H 0)
PAGE 59

46 critical current density (i cr i t ) and the primary passivation potential (E , ), respectively. All values of i . y PP»1 r crit _ 2 measured here are between 1.3 x 10" 5 and 9*50 x 10-5 A cm" . Although, in some cases, the lowest value of critical current density observed at a given chloride ion concentration approximates the highest seen in the next lower chloride ion concentration, there is a definite trend in the average values toward a reproducible maximum at 0.508 M potassium chloride (Table 2). The only relationship for which linearity is suggested is that between the logarithm of critical current density and the solution composition function, pH + log ( [S0^ J + [CI"] ), where [SO^ 2 "] and [Cl~] are the analytical concentrations of sulfate and chloride, respectively. A slope of 1.10 is measured (Figure 9)No linearity is found between the logarithm of 1 and E r , contrary to the work of Wilde. However, the most negative rest potential and highest critical current density values occur in the same system. From its value at the maximum the current density then decreases to a low value ( /v» 10" A cm" ) characteristic of the passive region, i , and is approximately independent of potential. In some cases, prior to the attainment of the passive current value, a negative current loop is observed (Ci c C')« Experimentally determined characteristics of the loop are presented in Table 3 and an experimental

PAGE 60

Figure 9. Variation of the logarithm of the critical current density with the concentration function, pH + log ( [SO^ 2] + [Cl~] ).

PAGE 61

48 -4.0 — -5.0-2.0 2.5 pH + Log ([so!] + [ Cr ])

PAGE 62

49 curve containing a cathodic loop is presented in Figure 10. The potential range in which it occurs, -0.30 to 0.0 V, is independent of solution composition. The maximum cathodic current density associated with the loop, i c (1.36 x 10 A cm ) , is largest in solutions with no chloride ion (0.334 M sulfate). Solutions containing chloride ion show significantly lower current values (2.3 x 10"7 to 4.2 x 10" 7 A cm -2 ) which increase slightly with increasing chloride ion concentration. To minimize the effect of the negative current loop, the magnitude of i is evaluated at 0.3 V except in O.508 and 1.0 M potassium chloride where pitting occurs below this potential. The dependence of the passive current density on solution composition is qualitatively similar to that of the cathodic loop current. The addition of 1.17 x 10" 2 M chloride ion causes a substantial lowering of i from the value observed in 0.334 M sulfate (5-25 x 10" 7 P versus 1.29 x 10" 6 A cm" 2 , respectively) (Table 1). A further increase in chloride content is accompanied by a slight increase in passive current density. In this region of increasing current, several linear relationships between i and solution composition are observed (Table 6). P Because of the narrow range covered by the three values of _n the passive current density available (5*25 x 10 through 7.85 x 10 -7 A cm' 2 ), the validity of these relationships is questionable.

PAGE 63

TJ

PAGE 64

(eUjo/v) ; &°i

PAGE 65

52 The classical evaluation of the potential of total passivity, E Tp , involves the determination of the point at which the current increases from its value in the passive region into the anodic loop of the active-passive 59, 60 transition during a cathodic potential scan. In the present work the potential of total passivity is defined as the intersection of the line defined by the decrease in anodic current from the maximum (anodic potential scan) with the value of the passive current at 0.30 V. This method gives more reproducible results because of the random cathodic loop observed here. Values of E Tp observed undergo a positive shift (from -0.285 to -0.160 V) as the chloride ion concentration of the solution is increased (Table 1). Examination of the interrelation of E and solution parameters suggests TP several possibilities (Figures 11-15): 2 Relationship Slope r 1. ^E Tp /c> [C1"J 0-108 V mole _1 l 0.93 2. ^E Tp / T, j[Cl1/(rS0^ 2 1 + [CI"] )j 0.103 V 0.92 3. ^E Tp / c>|[Cl-]/( 2 [S0 4 2 1+ [CI ]) 0.110 V 0.9^ 4. ^E Tp /^)[pH + log ([S0 4 2 -] +[C1-] ) 0.218 V 0.9^ ih L 1 decade -1 5. ^E Tp /a[iog ( [so 4 2-] / [cl -j J a. two segments with slopes -0.016 and -0.039 V decade" 1 and correlation coefficients of O.98 and 0.99, respectively or

PAGE 66

c o •H Q) -O •H ^ O rH x: o p •rH >> P •H > to CO nJ Ph P o P O rH cd •H P C 0) p o cu s: p • c o bO • H

PAGE 67

54 ODS'A) dl 3-

PAGE 68

a o •H P rt U •> c o c o o CD CO H n) p o p •h r CdCM P c — i o C » o C •H O P -H cd P •H O cd 3 > «H o GO a) •H

PAGE 69

56 OOS'A)

PAGE 70

CO

PAGE 71

(3DS 4 A) di 3-

PAGE 72


PAGE 73

60 0.15" 0.20-Q. H UJ 0.252.0 2.2 2.4 pH 4Log ( [SO^ + [cr] )

PAGE 74

0) x: p x: p •H • •P I* •H rH > o cd rHCM P o O 00 -P' — i o rH O CO rH •H -P » c c Q) O P -H O P ft O o rH u bo

PAGE 75

62 OOS'A) dl-a-

PAGE 76

63 b. one segment with slope -0.030 V decade and correlation coefficient 0.9^-« The latter plot lends itself to two interpretations. It may be considered to have two linear segments whose point of intersection occurs at E Tp -0.230 V and 0.301 M potassium chloride which is the smallest chloride ion concentration investigated in which pitting occurs. However, evaluation of data in terms of one single line is also feasible. Differentiation between the two possibilities is difficult because only five data points are available. In solutions whose sulfate to chloride ion concentration ratio is greater than one (^0.3 M chloride), a continuing shift to more anodic potentials results in the breakdown of the passive state with the onset of pitting. An experimentally determined polarization curve showing the influence of pitting is given in Figure 16. Two values are recorded for the potential at which pitting initiates, E ... E -.u/inM is the most negative potential at which pit pix v int. ; any increase in current density with time is observed within the ten minute waiting period. At potentials slightly positive to E pit ( inc ) ( to ^ 150 mV), an initial decrease in current followed by current spikes is usually observed. At more positive potentials an immediate increase in current density occurs. L^/ j is the potential obtained by extrapolation of the increasing pitting current densitypotential curve back to the point at which it intersects the value of the passive current at 0.30 V.

PAGE 77


PAGE 78

65 -NH — ^^H( 2 luo/V) ? Bon

PAGE 79

66 Both values of E ., are arbitrary because of the pit imposed time limitation for current measurement and the heavy fluctuations in current of up to an order of magnitude observed in this potential range, respectively. Thus, the induction period for pitting may be greater than ten minutes at potentials more negative than E -w^v and the current magnitude, after ten minutes at a given potential, is a random function of time as a result of the fluctuations present. The graphically determined value of the pitting potential, E it / \, is always the more negative. A three-fold increase in chloride ion concentration shifts E pit(inc) from °'395 to -0.090 V and E pit(gr) from O.36O to -0.155 V. Several linear relationships are observed between E .^. and the solution composition parameters examined, pit Their corresponding slopes, intercepts, and correlation coefficients are reported in Table 6. Because only three data points are available, it is difficult to judge which of these relationships, if any, are valid. It does appear, however, that the potential at which pitting initiates shifts toward more active values as the chloride ion concentration of the solution is increased. In systems not subject to pitting attack (< 0.3 M potassium chloride), loss of passive state stability occurs with the onset of transpassive dissolution (Table 4). The disolution reaction in this region involves oxidation of

PAGE 80

6? chromium(IIl) in the passive film to chromium(VI) and exhibits Tafel behavior over 1 to 1.5 decades of current. An anodic Tafel slope of 0.114 V decade" 1 is observed in 0.0 M and 1.17 x 10" 2 M potassium chloride, while a value of 0.150 V decade" 1 is found for 9. 97 x 10~ 2 M potassium chloride. The potential at which transpassivity initiates, E-t r , is taken as the intersection of the computed Tafel line with the passive current density at O.30 V. E^ shifts negative from 0.6?1 to 0.575 V with increasing chloride ion concentration. Continued anodic polarization produces a slight current maximum as a result of secondary passivity. Neither the -4 -2 current magnitude at the maximum, i cr ^ t ? ^ x 10 ~ A cm ~ '* nor its associated potential, E^ (0.95 V), is affected PP» <~ by solution composition at constant pH. The current then decreases to a minimum prior to oxygen evolution. The degree of stability of secondary passivity is represented by the difference in the magnitude of the maximum and minimum current values, A i. Present results indicate that the stability decreases slightly on addition of chloride ion to solutions originally 0.334 M in sulfate ion at pH 2.4. Current dif f erences , A i, of 14 to \1 yk cm and -2 M 21 uA cm -2 are observed for 1.17 x 10" 2 to 9-97 x 10 potassium chloride and 0.0 M potassium chloride (0.334 M sulfate), respectively.

PAGE 81

68 Secondary solutions A brief survey of the effect of chloride ion concentration in the absence of sulfate ion on the polarization curve was carried out at pH 2.4 (hydrochloric acid) in solutions containing 0.102, 0.123 and 1.0 M potassium chloride. Characteristic potentials and current densities obtained from the polarization curves are presented in Tables 7 and 8. Linear relationships with correlation coefficients ^ 0.90 are reported in Table 9 with the corresponding slopes. Cathodic current magnitudes at -0.700 V are significantly lower than those in sulfate containing solutions, varying from 1.25 x 10 J to 7-06 x 10" A cm in the chloride range 0.102 to 1.0 M. An increase in concentration from 0.102 to 0.123 M causes only a slight shift in rest potential (-0.470 to -0.480 V) while the rest potential for 1.0 M potassium chloride occurs at -O.505 V, a negative shift of 45 mV. The primary passivation potential is not a function of chloride concentration. In the absence of sulfate ion i cr i+ changes linearly with chloride ion concentration (Figure 17) and with the rest potential (Figure 18). The corresponding slopes are 3*11 x 10"5 A cm-2 mole-1 1 (r 2 = O.98) and -15-7 decades V-l (r2=1.0), respectively. A linear relationship is also observed between the logarithm of the critical current density and the concentration function, pH + log ( [S0^ J + [Cl~] ) (Figure 19). The corresponding line has a slope of 0.48

PAGE 82

69 to C o •H P H O to p

PAGE 83

?0 p c Q> -P o • i to p p C C (1) 0) ^ E t-i -H 3 u O CD Ph O X •H C «H a 1 -a p C O -H en M PhO o CO

PAGE 84

Figure 17. Variation of the critical current density with chloride ion concentration .

PAGE 85

72 6.04.02 0O 02 0.4 0.6 0.8 [cr]

PAGE 87

74 -4.6-4.20.50 0.49 0.48 -E r (V,SCE) 0.47

PAGE 88

Figure 19. Variation of the logarithm of the critical current density with the concentration function, pH + log ( [SO^ 2 ""] + £ci~] ).

PAGE 89

-4.0-4.5o o -5.076 1.5 2.0 pH + Log ( [SO4"] + [CI"] ) it

PAGE 90

Figure 20. Variation of the potential of total passivity with rest potential.

PAGE 91

0.240.26-O V) 0.28CL ILU i 0.300.50 49 0.48 -E h (V V SCE)

PAGE 92

Figure 21. Variation of the logarithm of the critical current density with the potential of total passivity.

PAGE 93

so -4.8--4.6-4.2vO 0.28 0.27 0.26 0.25 -E TP (V.SCE)

PAGE 94

81 and a correlation coefficient of O.93. No cathodic current loop is observed in these solutions. Little change in the potential of total passivity is observed in going from 0.102 to 0.123 M potassium chloride when data from individual runs are examined. Values are -0.287 V and -0.282 V for the former and -0.285 V and -0.270 V for the latter. However, average values show a definite positive trend with increasing chloride ion concentration in agreement with the value of -0.250 V found in 1.0 M potassium chloride. A one-to-one correlation between E and E r is suggested (Figure 20). The corresponding slope is -1.03. The log of i cr jt increases as E becomes more positive at the rate of 15. 1 decades V (Figure 21). In the absence of sulfate ion, pitting occurs at all chloride ion concentrations investigated. A negative shift in the pitting potential occurs with increasing chloride ion concentration but no statistically valid variation in this relationship is apparent. The effect of pH on polarization curve parameters was also examined in 0.3 M potassium chloride at pH 1.52, 2.35 and 6.22 (Table 7,8). In the former two cases, sodium sulfate additions of 0.233 M and 0.248 M, respectively, maintained the ionic strength at one. No sulfate was added to the latter and polarization in this system was

PAGE 95

82 terminated after passivation occurred. Possible linear relationships are presented in Table 9* Net currents at -0.700 V are cathodic and decrease with decreasing hydronium ion activity. Values of 2.03 x 10' 2 , 2.23 x 10~ 3 and 1.3^ x 10" 7 A cm -2 are recorded for pH 1.52, 2.35, and 6.22, respectively. A linear relationship is observed between both the rest potential and pH (-0.073 V pH" 1 ) (Figure 22) and the primary passivation potential and pH (-0.025 V pH" 1 ) (Figure 23). The slope of the straight line resulting from a plot of log i .. v8 pH is -0.328 (Figure 24). A linear relationship crit r is also observed between the logarithm of the critical current density and the concentration function, pH + log ( [S0 4 2_ ] + [CI "J ) (Figure 25). A slope of -0.35 and a correlation coefficient of 0.99 are associated with the line. No cathodic loop is found in the pH 2.35 solutions but one does occur between -0.15 and 0.0 V in two of three experiments at pH 1.52 (Table 10). The maximum associated -7 -2 cathodic current density is 4.5 x 10 A cm . The increase in pH from 1.52 to 2.35 shifts the potential of total passivity to more active values (-0.205 to -0.230 V) and changes the current density at 0.300 V from cathodic to anodic values. An active shift in pitting potential also occurs.

PAGE 96

R3 p c ai P o ft to en c ft o •H -H £ -P m 3 C H o o •H CO P cc! >> rH ^ u x: i* C cd o -p o H CO

PAGE 97

(to c *i pP> <-h HO o 3CD H> 3 V 03 0) H<: O •a o c+ HH« P)

PAGE 98

*5 o x Q. (3 OS 'A) 66 3-

PAGE 99

c CD o c+

PAGE 100

87 OOS'A) J 3-

PAGE 101

Figure 24. Variation of the logarithm of the critical current density with pH.

PAGE 102

89 -5.4•t: -5.0-4.62.0 4.0 6.0 PH

PAGE 103

Figure 25. Variation of the logarithm of the critical current density with the concentration function, pH + log ( [S0 4 2 "] + [CI"] ) .

PAGE 104

5.0" 91 2.0 6.0 pH + Log ( [S0 4 2_ ] + [Cl~] )

PAGE 105

Capacity-Potential Behavior During Potentiostatic Polarization The capacity-potential "behavior determined from the average curve for a given solution composition is used as the basis for data analysis. A schematic representation of the dependence of capacity on potential in the absence of hydrogen interference (see below) is given in Figure 26. In only a few cases are the characteristic potentials determined from the average curve not representative of values found on individual runs (Tables 11-14). Primary solutions After prepolarization at -0.?00 V for twenty minutes, capacity values of 40 to 60jutf cm are typically observed. _2 The one exception is 0.5 M potassium chloride (22 / )jtf cm ). A definite increase in capacity with increasing potential occurs in the region of active dissolution of the metal surface. In systems relatively free from interference (see below), a single capacity peak (C^) is then described at a potential (E-.) which is usually to 40 mV more negative than the corresponding primary passivation potential. The capacity maximum ranges from 62 to 73jJif cm and is independent of solution composition. In some cases, however, capacity peak values of greater than 100 uf cm -2 are observed in this region. They appear to be associated with the presence of the cathodic loop described above or with anodic current densities (~5 x 10 A cm" 2 ) between -0.3 and 0.0 V which are lower than the 92 -2

PAGE 106

-3 cd ro o hj Htj o c+ & c+ hj H . o o (TO M CD c+ 3 P o c+ (D O M P HjTj CD P 3 O CD h" 3 c+ O «< CD I . >d o c+ CD c+ HP CD 3" < O 3 H3 c+ 3" CD P cy w CD 3 o CD

PAGE 107

94 E_^ AllDVdVO

PAGE 108

95 .ccj r cm O I oo so (\J OH I •H < X CVJ EO 0) > -* •H W a) o to bO o « CD rH > ccv Ol rE i o o K ..-( ex
PAGE 109

w C o • H rH o E H ft

PAGE 110

CM 3,1 CO o TO to CM On CM 9?

PAGE 111

98

PAGE 112

99

PAGE 113

100 -6 o usual values of i ( ***> 10 A cm ) found in this potential P region. An assessment of the fine structure of the general capacity-potential "behavior here is hindered by scatter. Although scatter is present in all systems studied, the effect is much more pronounced in regions where the capacity is large. Resolution of the data into two peaks, one in the vicinity of the primary passivation potential, the other slightly positive to it, seems possible, but is not unequivocal. In the vicinity of the potential of total passivity, either a capacity peak or a sudden decrease in capacity is observed for all systems (E^, G 2 )» Further anodic polarization results in a smooth decrease in capacity to 20 to 25jli f cm -2 for 0.0 M through 0.3 M _2 potassium chloride solutions and 13 to 15 Uf cm for 0.5 and 1.0 M potassium chloride. The capacity in the passive region then continues to decrease slowly until a potential of 0.28 to 0.31 V (Ev) is attained (in 0.0 through 0.3 M potassium chloride) or until measurements are terminated at potentials positive to the pitting potential (in 0.5 and 1.0 M potassium chloride). Polarization beyond E^ causes a small drop off in capacity followed by a slow decrease to a minimum value of 11 to 15/Jtf cm" 2 (C min#1 ) at 0.6 to 0.65 V (E min#1 ). Plots of 1/C vs E in the potential region from E^ to E min ^ are linear (Table 15). Slopes of 0.06, 0.09, 0.06, and 0.04 V" 1 cm 2 Uf~ 1 are observed for 0.0, 1.17 x 10" 2 , 9.97 x 10* 2 ,

PAGE 114

101 p o o> tc C rH a) • H P c rQ

PAGE 115

102 and 0.3 M potassium chloride, respectively. The corresponding intercepts and correlation coefficients obtained by linear regression analysis are also presented in Table 1$. The increase in current observed in the transpassive region is accompanied by an increase in capacity to a maximum (C_) of 27 to 53 uf cm at a potential of 0.85 to 0.90 V (E ). The capacity then decreases to a minimum of 19 to 22uf cm" 2 (C . 2 ) at 1.20 V (E min 2 ) and increases sharply at still more positive potentials. Secondary solutions The change in capacity with potential for solutions of varying chloride ion concentration in the absence of sulfate ion and those at pH 1.52 (0.3 M potassium chloride, 0.233 M sulfate) corresponds closely to that described above over the entire investigated potential range (Tables 13 and 14). However, a capacity peak is associated with the potential of total passivity only for the 0.102 M potassium, chloride, pH 2.4 system. Pitting occurs in all instances here. The capacity values observed prior to and in the pitting region are similar to those seen in the same potential regions in _ p nonaggressive solutions (11 to 20 u f cm ) where the surface remains passive. Polarization, continued for one experiment at pH 1.52, shows a capacity minimum at 0.6 V followed by an increase in capacity in the transpassive region. No capacity increase is associated with the increase of current caused by the onset of pitting.

PAGE 116

Galvanostatic Polarization Potential-time behavior After potentiostatic prepolarization at -0.?00 V, control was switched to the galvanostatic circuit. The potential-time response resulting from the application of -4 -3 -2 anodic current densities from 1 x 10 to 2 x 10 J A cm was observed. Only primary solutions (constant ionic strength, pH 2.4) were investigated. Chloride ion concentrations were varied from 0.0 to 0.518 M. A schematic representation of the current-induced potential transient is given in Figure 27. The arrests observed and the stability of the maximum potential achieved for a given system are a function of the current density and solution composition employed. Chronopotentiometric techniques of curve analysis are used to determine plateau length (t) and associated potential values (E , ), although the theory is not directly applicable to systems with forced convection. ' *" Results have been tabulated (Tables 16-19) • In some cases, current polarization was repeated using the same electrode arid solution. Even though prepolarization at -0.700 V was carried out before each use, the potential value of the first arrest, E. , was found to shift anodic with continued use while the arrest length became shorter. Little effect on the potential of the more noble arrests was observed but plateau lengths became irreproducible. Data analysis is therefore based, 103

PAGE 117

Figure 2?. Schematic representation of the galvanostatic potential-time curve in the presence (1) and absence (2) of pitting breakdown.

PAGE 118

1 .40 — 105 .10-0.80-ai O 0.50 CO 0.20-•0.10 -0.40100 200 300 400 TIME (sec) 500 600

PAGE 119

106 CO •P o ft CD £ H P I rH Cti •H P C Q> P O Ph o •H P rt • -p -— to (u o t3 C-H > o iH rH as x: o p o ° PQ m P W men cntO o o o >3 >> >3 >> >> >> >2 T3 (DTD (DTD CD T3 CD -O CD TD (D 'O CD cti-PaJ-PaS-PaSPal-Pai-PaJ-P CDctSCDajCDaJcDaScDajcDcticDas cocococococococococococococo NO

PAGE 120

10? co w cO -P o p, U o •H > o a> to cm o o co P W mo > w rQ

PAGE 121

108 o •H > CO x:
PAGE 122

109 .3 VTA O CD T3 •H U o r-i x: c £ 3 •H 10 CO cri -P o ft E 00 ^23•3" CO On CO rH o CD 6 •H V I H cd • H -P c o P o P, o cMr.o > CM o o 3Pd iH On CM CM O rH -^ ^ -3" E4o o On

PAGE 123

110 in most cases, on the potential transient resulting from the first current application. Open circuit behavior and galvanostatic polarization from open circuit are also investigated . All systems studied exhibit an initial potential arrest near -0.4 V. The reaction occurring is assumed to be dissolution of iron from the alloy as iron(II). For a given solution the current-arrest potential relationship shows Tafel behavior (Table 20). A slope of 0.06 V decade is found for chloride ion concentrations of 0.0 through 0.303 M. Exchange current densities, calculated from the slopes and intercepts of the Tafel lines (see Introduction), range from 2.27 x 10" 6 to 1.11 x 10" 3 A cm" 2 (Table 20). In O.5I8 M potassium chloride a change in Tafel slope to 0.014 V decade"-*occurs. The variation in E, with solution composition for a given current density is a function of the current density applied. Solutions containing chloride ion (0.100 through _4 _? 0.518 M) at low current densities ( < 3 . 77 x 10 A cm ) exhibit a potential arrest which is negative to that observed in 0.337 M sulfate, but is approximately independent of chloride ion concentration. At higher current densities an increase in chloride concentration causes a negative shift in Et. Arrest potentials resulting from the application of 1 x 10"3 a cm -2 vary from -O.365 V in 0.0 M potassium chloride to -0.409 V in O.518 M potassium chloride.

PAGE 124

Ill p cd P W o c cvJ > rH cd c u •a o • H > cd P ft -a cd CD o ft 0) O 'C rH 00 > (^ 3ce CN ON CO o o CO Ox UN oOn CO ON OON •4i o U>1 ON I c C\J cn CN I O I o X ON 00 I o H X -3X CM 1 o

PAGE 125

112 Plateau lengths (t) associated with this arrest are irreproducible. In general it appears that t decreases with increasing current density. The charge needed to achieve the passive state , i.e., the product of current density and arrest time (indicated by a rapid change in potential to more noble values) also decreases with increasing current density. The potential arrest found at 0.0 V, E , is also observed in all the systems investigated. At the chart speeds (10-20 sec in" ) used to display the major portions of the potential response curve, its characteristic length and associated potential can only be determined approximately. The data suggest a positive shift in E with increasing current density accompanied by a decrease in plateau length. Potential values associated with arrests occurring at still more noble potentials as well as the stability of the maximum potential achieved are a function of current density and solution composition. In solutions which cause pitting (0.303 and O.5I8 M potassium chloride), the maximum potential attained is unstable and a rapid decrease in potential to more active values occurs, despite the continued application of anodic current. Systems not susceptible to pitting attack (0.0 M and 0.100 M potassium chloride) reach and maintain a constant maximum potential. Potential arrestcurrent relationships are tabulated (Tables 16-19).

PAGE 126

li 3 The arrest described at 0.8 V, E„., occurs in solutions 3A containing 0.0, 0.100, and O.303 M potassium chloride and is visible at the recorder speeds employed (10 to 20 sec in" ) for current densities of ^ 3.7 x 10 A cm . In O.303 M potassium chloride this arrest is also seen in two of five -3 -2 experiments conducted at 1 = 1.0 x 10 ^ A cm . The current density-arrest potential relationship associated with this plateau follows Tafel behavior in all three systems. Tafel parameters are presented in Table 20. The reaction associated with this arrest is assumed to involve the oxidation of chromium(III) in the passive film to chromium(IV) . From the slopes and intercepts of the Tafel lines, exchange current densities of l.?4 x 10~ 18 , 2.57 x 10" , and 2.69 x 10~9 a cm are calculated for 0.0, 0.100, and O.303 M potassium chloride, respectively. In solutions containing O.303 M potassium chloride polarized by current densities < 1.0 x 10"3 a cm, the potential of this plateau is the maximum achieved and breakdown to more active values ensues. The next arrest observed, E~, occurs in the vicinity of 0.85 V in 0.0 M (i > O.99 x 10" A cm" 2 ), 0.100 M (i > 1.93 x 10' 4 A cm" 2 ), and in O.303M (i>1.0 x 10" 3 A _2 cm ) potassium chloride. The presence of a maximum in the current-potential arrest behavior is suggested in 0.100 and O.303 M potassium chloride. The data in 0.100 M potassium chloride solutions prior to the potential maximum are insufficient for Tafel analysis. However, Tafel behavior

PAGE 127

114 is exhibited in 0.0 and O.303 M potassium chloride. Exchange current densitites of 4.57 x 10~ 23 and 4.1? x 10" 10 A cm' 2 are calculated for the two systems, respectively. In only _3 two of five experiments at current densities of 1.0 x 10 _2 A cm does breakdown occur from this plateau in O.303 iv; potassium chloride. In all other cases, as with non-pitting systems, further polarization to more noble potentials corresponding to oxygen evolution occurs. The potential arrest associated with oxygen evolution, E Q , is seen in 0.0 M (i > 1.94 x 10 _Z| A cm" 2 ), 0.100 Ivi (i > 3.72 x 10" 4 A cm" 2 ), and O.303 Ni (i > 1.0 x 10~ 3 A -2 X cm ) potassium chloride. Attainment of this plateau in O.303 M potassium chloride is followed by pitting breakdown. Exchange current densities of I.38 x 10" , 6. 61 x 10"" 10 , -9 ? and 1.12 x 10 7 A cm are calculated for the oxygen evolution reaction for the three solutions, respectively. Constant current polarization of specimens in O.5I8 M potassium chloride initiates breakdown from potentials substantially below those observed in O.303 M potassium chloride. Breakdown occurs from 0.44?, 0.471, O.508, O.683, -4 and 0.?98 v during the initial imposition of O.997 x 10 , 1.93 x 10' 4 , 3.89 x 10~\ 0.995 x 10* 3 , and 2.02 x 10~ 3 A _2 cm , respectively. Repolarization results in successively more positive potential maxima. Values of O.683, 0.722, and O.763 V are observed as a result of the first, second and third application of O.995 x 10" 3 A cm -2 , respectively.

PAGE 128

115 The steady-state potential obtained after pittinginduced potential breakdown during constant current polarization has been called the protection potential of the system under study. rfc To determine this characteristic value long-term (10 to 90 min) constant current polarization behavior was examined for samples in O.303 and O.518 M potassium chloride. The history of the samples used varied widely. However, the results of the recycling experiments indicate that after passivation is achieved, especially in higher chloride ion concentration solutions, the absolute value of potential arrests is only slightly affected by history while the general potential-time response remains approximately the same. The precise definition of the course of the potentialtime curves during breakdown and of the steadystate potential reached is hindered by potential oscillations varying in magnitude from +10 to + 60 mV and in frequency from 4 to 6 per 100 seconds. Therefore, only the average values of the final potentials achieved for a given solution composition and current density are presented in Table 21. Also included are the maximum potentials reached prior to breakdown of the passive state. In all cases in O.303 M potassium chloride rapid potential breakdown occurs to approximately O.50 V. A slow negative shift in potential follows. The average final potential attained is a function of current density, ranging from 0.085 V at 1.93 x 10"^ A cm" 2

PAGE 129

116 c o • H cd .H o ft -P cd P W o p > rH a W) tiD P •H T3 P P cr> p o ft c o TO w w w o o T^ 00 P 3" 00 bf

PAGE 130

117 to O.36 to 0.39 V at 1.0 x 10~3 A cm -2 . When oxygen evolution potentials are reached prior to breakdown, slightly more negative values result (0.31 and 0.33 V for 1.01 x 10"3 and 1.97 x 10"7 A cm , respectively). Considerably more negative average final potentials are observed in 0.518 M potassium chloride over the entire current range studied. The potential magnitude shifts in the positive direction (from -0.043 to 0.045 V) with increasing current density for the three lower currents employed. A marked positive shift to 0.2^5 V occurs when the applied current density is increased to O.995 x 10"^ -2 A cm . A potential value of 0.230 V is observed at a current density of 2.02 x 10"^ a cm -2 . Following constant current polarization, the galvanostat was removed from the circuit and the resulting open circuit potential transient was observed. Open circuit behavior is a strong function of solution composition and electrode surface history. The representative data presented in Table 22 are obtained from samples of diverse history and therefore are discussed only qualitatively. Galvanostatic parameters from the immediately preceding polarization are also included. In solutions which do not initiate pitting breakdown during constant current polarization (0.00 and 0.100 Ivi potassium chloride) three open circuit arrests are described.

PAGE 131

118 (in M O Pm do > J x CM

PAGE 132

119 w

PAGE 133

120 Arrest potentials of 0.4 to 0.5 V, E,, 0.12 to 0.19 V, E 2 , and -0.4 to -0.5 V, E~, are observed in 0.00 M potassium chloride. In 0.100 M potassium chloride the first two arrests occur at slightly more negative values, 0.38 to 0.40 V and 0.04 to 0.15 V, respectively. The third is found between -0.4 and -0.5 V. Although the most negative potential achieved remains constant for up to five minutes, a subsequent slow increase in potential with time occurs. Thus, after six hours at open circuit in 0.0 M potassium chloride, following constant current polarization at i = -4 ? 3.8 x 10 A cm, the electrode potential has shifted from -0.480 to -0.393 V. After twelve hours, a value of 0.228 V is recorded for the open circuit potential of a sample subjected previously to a constant current of I.65 x 10"-' A cm . Systems subject to pitting attack exhibit complex open circuit behavior. In both O.303 and O.5I8 M potassium chloride, an apparent steady state potential occurs between -0.44 and -O.50 V. Following polarization at i = 3.8 x 10" A cm , the potential remains constant in this range for four hours in O.303 M solutions. In O.5I8 M potassium chloride, however, the open circuit potential shifts positive with time from the most negative potential observed. Values of -0.128 to -0.013 V are recorded after four to ten hours at open circuit in this system. An additional arrest

PAGE 134

121 is observed at -0.27 to -0.32 V in 0.518 M potassium chloride, regardless of the preceding polarization current density. This arrest is also visible in some cases in O.303 ft! potassium chloride. Other open circuit arrests are described between the average potential achieved after breakdown during constant current polarization and the arrest at -0.27 V. The variation in the potential values associated with these arrests with solution composition and the maximum potential achieved during constant current polarization as well as their relation to the arrests observed in non-pitting systems cannot be evaluated on the basis of the data available. Following attainment of the apparent steady-state open circuit potential, the constant current was reapplied to determine the difference, if any, in the potential-time response resulting from open circuit pretreatment. Regardless of the length of time at open circuit, the most negative potential reached, or the potential arrests observed during decay, reapplication of a constant anodic current produces no observable arrests below 0.8 V. In non-pitting systems the behavior above 0.8 V is quite similar to that achieved during the initial polarization from -0.700 V. The potential arrests observed in O.303 M potassium chloride are slightly more negative but the general pattern follows that of the initial polarization. More noble breakdown potentials are observed in O.5I8 M potassium chloride.

PAGE 135

122 Capacity-potential behavior The differential capacity of the stainless steelsolution interface was determined as a function of potential during constant current polarization. Because of the manual technique employed in data correlation and the rapidity of the potential-time change, capacity behavior is not precisely defined. Although, in some cases, re-use of an electrode causes a decrease in the magnitude of the capacity associated with a given potential, the qualitative capacitypotential behavior remains unaffected. Therefore, the data accumulated for any one arrest, regardless of electrode history and polarization current density, are used to define the course of the capacity-potential (time) curve for a given solution composition. Capacity values, when given, represent initial polarization data or values not influenced strongly by electrode re-use. A capacity-time curve for a system not subject to pitting (0.0 and 0.100 M potassium chloride) is superimposed on its associated potential-time curve and is presented in Figure 28. Steady-state capacity values determined at the end of the twenty-minute prepolarization period range from _2 21 to 18uf cm (C-700). These values are typical of capacities associated with the double layer in solutions of high ionic strength.

PAGE 136

o CtJ -H -P Ch Cti O -P W ^ o o C •H cti £ K) •H ^ -P 3 I 'C >> -P c •H 5 o o CO T3 cd 03 o a) U -P W) o P> -P C-H O PL, P O nJ -P -P C -P CD o W Q) CD •<-* u & . (U id C ^ o •P -H O O -P •HC(D -P N e a) jh 3 o Ul CD ft a: CM 3 fc,H H

PAGE 137

12^ ( 2 .ujo iri) AllOVdVO OOS'A) IVIlNBlOd

PAGE 138

125 In all systems subjected to galvanostatic polarization (0.0 to O.5I8 M potassium chloride), a capacity peak, C-^, is associated with the first potential arrest observed, E-, . Solutions containing chloride ion exhibit a slightly 2 greater peak magnitude (38 to 58 u f cm"'-) than those with none (35 to 48 yx f cm' 2 ). Capacity values observed during the rapid potential transition from the active to the transpassive state range from 12 to 22jaf cm" 2 in 0.0 and 0.1 M potassium chloride. This represents a substantial decrease from the values in the active region. Similar capacity decreases occur prior to and during potential breakdown from the transpassive region in O.303 M potassium chloride. The second capacity peak observed in systems containing 0.0 and 0.1 M potassium chloride is associated with the potential transition from the arrest at 0.8 V to that at O.85 V. Peak values of 20 to kO uf cm" 2 are measured. The continued change in potential to oxygen evolution values is accompanied by a capacity peak only when potentials ^1.3 V are attained. The maximum potentials achieved in these solutions are stable with respect to time. No pittinginduced potential breakdown occurs. Corresponding steadystate capacity values range from 15 to 26 uf cm" for final potentials^ 1.26 V and from 21 to 39 jxf cm -2 for final potentials of > 1.3^ V.

PAGE 139

126 In O.303 M potassium chloride the potential arrests observed and the potential from which breakdown occurs coincide with arrests occurring in non-pitting systems. Capacity data, although not conclusive, suggest that the corresponding capacity-potential behavior is also similar. In O.303 M potassium chloride solutions a definite capacity increase is associated with potential arrests and/or maxima between 0.72 and 1.04 V. However, the relative position of the capacity peak(s) and characteristic potential(s) are not well defined. A capacity maximum occurs prior to breakdown from potentials in the oxygen evolution region (1.3 to 1.4 V). The magnitudes of the largest measured capacities are 15 to 24 and 28 to 63 Uf cm" , respectively, in good agreement with those found in non-pitting systems in similar potential regions. The rapid potential decrease to approximately 0.5 V caused by pitting in O.303 M potassium chloride is accompanied by a capacity decrease to values of 11 to 14 uf -2 -2 cm and 8 to 11 uf cm following breakdown from the transpassive and oxygen evolution regions, respectively. An examination of long-term constant current polarization data shows that, in most cases, the capacity remains constant as the slow negative shift in potential with time is achieved. Two instances of diverse behavior are observed

PAGE 140

127 An increase in capacity to 20 uf cm"' occurs at 0.12 V as the average potential (O.O85 V) resulting from 1-93 x 10" ^ ? -3 A cm"' is approached. For polarization with 1.02 x 10 o A cm"', the capacity increases slowly from 14 to 18 u f cm as the potential decreases from 0.461 V to the average final value, 0.391 V. The largest potentials attained in O.5I8 M potassium chloride for the three lower current densities employed (0.447 to O.508 V) are substantially more negative than any of the potential arrests characteristic of non-pitting systems. The largest potentials reached at the two highest current densities more closely approximate non-pitting values (0.683 and 0.798 V). The capacity-potential behavior resulting from the application of 0.997 x 10"^ and I.93 x 10*^ A cm" 2 exhibits a well-defined, smooth decrease in capacity from the prepassive region across the potential-time peak. Typical capacity values associated with these potential maxima range from 13 to 1.5 }x f cm" 2 . For 3.89 x 10"^' A cm -2 , the number of data points available is much smaller. The shape of the capacity-potential curve is similar to that described above but with slightly higher capacity values (16 and 18 uf cm" 2 at potential maxima of O.568 and O.588 V, respectively) .

PAGE 141

128 Since only two data points are available at the two higher current densities, no conclusions can be drawn with respect to the presence or absence of a capacity peak associated with the potential-time maximum achieved prior to breakdown. However, capacity values determined for these higher potential maxima approach those observed in nonpitting systems. As discussed above, subsequent polarizations of an electrode result in more noble potential maxima in 0.518 M potassium chloride. Thus, capacity values of 21 to 52 uf cm -2 are recorded for potential maxima of O.758 to 0.823 V. The capacity-potential behavior following potential breakdown during long-term constant current polarization in O.5I8 M potassium chloride was also observed. For the two lower current densities, an increase in capacity of 2 to 3 u f cm" 2 to an approximately constant value of -2 • 14 to 18 y. f cm occurs. A similar constant capacity region is observed during polarization at 3*89 x 10" A cm"' which for all three currents persists to approximately 0.3 V. Subsequently, the capacity increases to between 20 and 29]J-f cm within the potential range 0.27 to 0.09 V. The presence of a well-defined peak at 0.268 V in one experiment (0.518 M potassium chloride, 3*89 x 10 A cm"^) supports the implication of this capacity increase.

PAGE 142

129 During polarization by 0.995 x 10"1 A cm , the constant capacity region (16 to 21 uf cm" 2 ) is observed to a potential of 0.25 V and is followed by an increase in value to 28 uf cm -2 at 0.202 and 0.129 V. No peak is defined, however. Since the average final potential attained with 2.02 x 10~ 3 A cm" is near 0.3 V, only the constant capacity region (14 to 22 u f cm~^) is seen. Capacity-potential behavior was also observed during open circuit decay from final state potentials induced by constant current polarization. A decrease in capacity to a constant value after removal of constant current control is well defined in 0.0 and 0.100 M potassium chloride and is strongly suggested in O.303 M potassium chloride. The negative boundary of this constant capacity region lies between 0.20 and 0.30 V. Capacity values range from 11 to 16 u f cm" 2 and 7 to 14 jj. f cm" 2 for systems whose preceeding current induced potential maxima occur in the transpassive and oxygen evolution regions, respectively. In O.303 and 0.518 M potassium chloride, in most cases, final potentials achieved after pitting breakdown under the continued influence of anodic current are in or negative to this constant capacity region. At more negative potentials, capacity values were determined only at widely spaced time intervals. However, the existence of two capacity peaks of about 30 to 40 jj. f cm is suggested, occurring between 0.10 to -0.10 and

PAGE 143

130 -0.20 and -0.40 V, respectively. In all cases, the potential arrest observed in the range of the former peak occurs before the capacity maximum. The second capacity peak is observed both in the presence and the absence of the potential arrest sometimes visible in the same potential range. The capacity peak precedes the potential arrest when both occur simultaneously.

PAGE 144

CHAPTER IV DISCUSSION Potentiostatic Folarization The results of the potentiostatic experiments are discussed in terms of the phenomena represented by characteristic regions of the polarization curve. The topics to he considered in detail include: the anodic dissolution-passivation mechanism(s) ; the hydrogen evolution reaction; the open circuit or rest potential as it relates to the rates of the active dissolution and hydrogen evolution reactions; the passive region; the pitting reaction; and transpassive dissolution. Active dissolution and passivation The transition from the active to the passive state observed for all systems studied here is characterized by the critical current density, the primary passivation potential, and the potential of total passivity. The presence of a current maximum in the polarization curve prior to the onset of passivation suggests a competition between the reaction leading to dissolution and the reaction leading to passivation of the metal surface. Both reactions are assumed to occur on bare metal surface sites. At the primary passivation potential the passivation reaction has succeeded to the extent that the fractional 131

PAGE 145

132 coverage ( ) of active surface sites (kinks, ledges, etc.) has reached 0.5* Further polarization results in a continuous increase in until the surface becomes totally passive at E TF The magnitude of the critical current density in any given solution should depend on the effect of the solution components on the kinetics of the two competing reactions. Present results on stainless steel indicate an increase in the dissolution rate by both chloride and sulfate anions. In all solutions studied in the present work, it has been observed that the logarithm of the critical current density is proportional to the logarithm of the concentration function, pH + log ( TSO^ 2 "] + [Cl~] ), where [SO// ] and [CI ] are the analytical concentrations of the sulfate and chloride ions present, respectively (Figures 9, 19, and 25). This implies that chloride, bisulfate and sulfate ions are equivalent in their effect on the dissolution-passivation reaction and that it is the total anion concentration which determines the critical current density. 64 Similarly, Florianovich et al. have shown a linear relationship between the logarithm of the anodic current density for iron dissolution and the logarithm of the total anion activity in acidic sulfate solutions. Since linearity is obtained only if the total sulfate activity is considered,

PAGE 146

133 these authors assume that sulfate and bisulfate ions function equally. Reaction orders of +1 for both the hydroxyl and the sulfate ion are observed. Chloride ion, between 0.01 and 1.9 M, has also been demonstrated to have a positive 65 reaction order for iron dissolution. Little work has been done to evaluate the effect of solution composition on the rate of the reaction leading to passivation, although sulfate ion has been shown to depolarize the passivation reaction for iron. Fresent results indicate that, at constant pH, the primary passivation potential is independent of solution composition in the range studied (Tables 1 and 7). This behavior implies that both the dissolution and passivation reactions are depolarized equally in a given solution. The positive shift in primary passivation potential observed when the pH is decreased suggests that hydroxyl ion, while participating in both reaction, accelerates the dissolution reaction more so than the passivation reaction. A capacity peak is associated with the active to passive transition in these studies and occurs, in most cases, to 40 mV negative to the primary passivation potential. In view of the arguments given above it is likely that this peak indicates the specific adsorption of the anions participating in the dissolution and passivation reactions (hydroxyl, sulfate, bisulfate and

PAGE 147

13** chloride ions). The adsorption of these anions on iron in acid solutions has been demonstrated by Hackerman et al/ 7, 47 ' 4R ' 5 ° The evaluation of the effect of solution composition on capacity peak magnitude is hindered by the presence of absorbed hydrogen and hydrogen bubbles. In systems exhibiting no evidence of interference (see below), values of 62 to 73 U f cm -2 are observed in all solutions. The majority of the relationships observed between the potential of total passivity and the solution in which it is determined are linear with respect to the solution composition rather than the logarithm of solution composition. This suggests that the potential of total passivity is dependent on kinetic, rather than thermodynamic, factors. It is unlikely that the positive shift in the potential of total passivity with increasing total anion concentration can be directly attributed to an increase in film solubility in the concentration range studied. If this v/ere the case, the primary passivation potential ( 6 = 0.5) would also undergo a positive shift and a monotonic increase in the passive current density would occur. Neither effect is observed (Tables 1 and 7). It is possible, however, that expulsion of specifically adsorbed anions from the inner double layer, caused by a change in the zero point of charge of the metal surface upon passivation, may increase the local aggressiveness of the solution sufficiently to decrease passive film stability.

PAGE 148

135 The number of specifically adsorbed anions on a metal surface increases with the concentration of adsorbable anions in the solution. A positive shift in the potential of total passivity with adsorbable anion concentration would therefore be expected. A decrease in the rate at which the applied potential is changed would then be expected to cause a negative shift in E T p since additional time would be provided at each potential for adjustment of the double layer structure to that representative of the bulk solution, but this effect was not studied. The hydrogen evolution reaction The net cathodic current density measured at -0.700 V is the algebraic sum of the internal anodic and cathodic currents flowing at this potential. The maximum anodic current density observed during polarization (i r ;+) is approximately two orders of magnitude less than the cathodic current observed at -0.700 V (-~'10"; A cm ) and occurs in the vicinity of -0.40 V. Since the anodic current density is expected to decrease exponentially with potential as the potential is made more negative, it can be assumed that the magnitude of the residual internal anodic current at -0.700 V is negligible with respect to the cathodic current. Any change in the net cathodic current density with solution composition is therefore attributed to a variation in the kinetics of the hydrogen evolution reaction, only.

PAGE 149

136 Chloride ion has been shown to depolarize the hydrogen 6 ° f. Q evolution reaction on stainless steel.' ' "' A similar effect has been observed on chromium in 0.5 M sulfuric acid where the addition of hydrochloric acid decreases the reduction reaction overvoltage. ' ° The negative shift in the potential of the Outer Helmholtz Plane produced by the presence of specifically adsorbed chloride ion is believed to accelerate reduction of the hydronium ion by decreasing the activation energy for the reaction. The effect of sulfate on the kinetics of the hydronium ion reduction reaction on ferrous metals has not been studied, although sulfate ion has been shown to be specifically adsorbed on an active iron surface. ' Since the zero point of charge of an active stainless steel surface is assumed to be approximately -0.69 V, specifically adsorbed chloride ions are expected to be present in the inner double layer at -0.700 V. However, the increase in cathodic current density at -0.700 V with increasing chloride ion concentration expected on the basis of literature results is not observed in the present studies. In primary solutions at pH 2.^ and at constant ionic strength with chloride ion concentrations between 0.0 and 0.508 M (sulfate ion concentrations from 0.33^to 0.16 3 M), the cathodic current density is approximately constant. The variations observed within this concentration range (Table 1) are most likely a result of differences in the true electrode

PAGE 150

137 area of the electrodes employed. A variation in the true electrode area by up to a factor of 1.5 would account for the changes observed. Parallel changes occurring in the corresponding differential capacity values (Table 11) support this conclusion. It therefore appears that sulfate ion counteracts the depolarizing effect of chloride ion if it is present in sufficient quantities. The large increase in cathodic current density occurring when the sulfate ion concentration is decreased from 0.163 to 0.016 M supports this contention (Table 1). Data accumulated in systems of varying chloride ion concentration and ionic strength at constant pH (2.4) differ considerably from those at constant ionic strength discussed above. The cathodic current density at -0.700 V for a given chloride ion concentration is two to three times lower than that observed in solutions containing sulfate ion and a decrease in current density with increasing chloride ion concentration is observed. That these phenomena are not artifacts is supported by the good agreement between results obtained two months apart for 0.102 and 0.123 M potassium chloride (Table ?). The lower current densities observed in 0.102 and 0.123 M potassium chloride may be caused, in part, by differences in roughness factor, i.e., the ratio of the actual surface area to the apparent geometric area from

PAGE 151

138 one electrode to another. If the differential capacities measured at -0.700 V in 0.1 M potassium chloride solutions -2 -2 in the presence (63 U f cm ) and absence (21 to 31 U f cm ) of sulfate ion are compared, ratios of two to three are also observed. Since both current density and differential capacity are calculated on the basis of geometric area, the parallel variations in these two parameters support the idea of an area effect. The behavior of the 1.0 M potassium chloride system is reproducible but cannot be explained in terms of the data accumulated here. The presence of 0.016 M sulfate would be expected to exert little influence but the values obtained in the presence and absence of this species are 7 Li -2. widely divergent (5*31 x \0~ D versus 7«06 x 10 A cm , respectively) . Insufficient data are available from present experiments to determine the mechanism by which hydrogen is evolved at the stainless steel surface in acid solutions. Tafel slopes of 0.087 to O.I33 V decade" 1 , 71, 7? " 5 '~ exchange 6 ? 58, 72 current densities of 2 to 33 x 10 A cm" , ' ~ and a 71 reaction order of O.85 for hydronium ion have been reported in the literature. The electrochemical discharge step, H + e — * H , (Volmer reaction) (6 ) where H a( ^ s is an adsorbed hydrogen atom, has been shown to be

PAGE 152

139 73 rate-determining in basic solutions. This step is also rate-determining for the hydrogen evolution reaction on iron in acid solutions. r ' Correlation of the capacity-potential and current density-potential data observed here suggests that the hydrogen evolution reaction is also associated with the random cathodic loop seen during anodic polarization. *9 Wilde has studied the effect of cathodic pretreatment on the capacity-potential behavior of AISI 30^ in 0.5 M sulfuric acid. Since austenitic stainless steels are known to absorb hydrogen, he attributes the increase in the magnitude of the capacity peak in the active-passive transition range with increasing pretreatment time to an adsorption pseudocapacitance. In the majority of cases in the present results in which the loop occurs or low anodic net currents are observed in the potential range of the loop, significantly higher capacity values and broader capacity peaks are measured (Tables 3 and 10). It appears likely, therefore, that these high capacity values are caused by the presence of absorbed hydrogen atoms in the stainless steel. The reversible potential of the hydrogen evolution reaction is a function of the pH at the electrode-solution interface, E = -0.06 pH 0.2^-5 ( 7 ) where E is the reversible potential in V versus SCE and

PAGE 153

]>0 0.24-5 V is the potential of the saturated calomel electrode on the hydrogen scale. As the potential is shifted positive past the reversible potential, oxidation of the absorbed hydrogen atoms from the metal will occur, causing an increase in the capacity value measured. The accumulation of the resulting hydronium ion at the interface will cause a decrease in pH and a produce a positive shift in the reversible potential of the reaction. If the interfacial pH becomes low enough, hydrogen evolution will again become energetically feasible. The interfacial pH necessary to allow hydronium ion reduction in the potential range of the cathodic loop (-0.30 to 0.0 V) is calculated from equation 7 as 0.91 to 4-. 10. If it is assumed that all of the hydrogen produced at -0.700 V ( ,^.2.4 coul) is absorbed by the stainless steel and then emitted into an approximately . 100 A double layer, an interfacial pH of -4.1 is calculated. Since only a fraction of the hydrogen produced is expected to be absorbed, this is obviously an approximation. With the present data it is not possible to make the distinction between one broad capacity peak or two separate, narrower peaks in the active-passive transition region. Spurious results are caused, in part, by the presence of hydrogen bubbles on the horizontal surface of the electrode. 76

PAGE 154

]>1 The influence of chloride ion on the degree of hydrogen absorption by stainless steel is open to question. Present data suggest that in 0.3 and 0.5 M potassium chloride (pH 2.4, ionic strength = 1) as well as in 0.102, 0.123, and 1.0 K potassium chloride (pH 2.4, variable ionic strength) absorption is considerably diminished when compared to lower chloride ion concentrations in the presence of sulfate. Wilde's data, 9 on the other hand, show an enhancement in the degree of absorption in the presence of chloride ion with respect to that observed in dilute sulfuric acid. The appearance of a cathodic loop in the polarization curves of chromium and chromium alloys following passivation in deaerated acid solutions has been noted by many investigators . 7?' ? ' '^Attempts ^o study the kinetic parameters associated with this loop for nickel-chromium '9 fin and cobalt-chromium' alloys have been unsuccessful. p-i However, Wilde and Hodge have examined the hydrogen evolution reaction on active and passive chromium in dilute sulfuric acid in detail. The substantial decrease in exchange current density on the passive surface o p (5.3 x 10 A cm" ) from that on the active surface n p (3-5 x 10"' A cm"'') is attributed to the presence of a chromium-deficient semiconducting oxide on the metal surface. The Tafel slope corresponding to hydrogen evolution on the passive metal is given as 0.06? V decade .

PAGE 155

l42 The exchange current density reported in the literature for hydrogen evolution on active stainless steels is approximately 10~5 A cm" (see above). If, similar to chromium, the exchange current density on the passive steel surface is assumed to be two orders of magnitude -7 -2 less than that on the active surface (i.e.,^10 A cm ) and a Tafel slope of 0.06? V decade" is valid, the current density at a given overvoltage in the cathodic loop can be calculated. Applying the high-field modification of the Butler-Volmer equation and assuming an -6 -2 overvoltage of -0.2 V, a value of 2.0 x 10 A cm is obtained. This agrees very well with the value, 1.36 x 10 _2 A cm , measured in 0.33^ M sodium sulfate at pH 2.4. Rest potential It appears from the present results that the rates of both the hydrogen evolution reaction (i^-.^) and the dissolution reaction (i_) depend on chloride as well as on sulfate ion concentration. At the rest potential these two reaction rates must be equal. Therefore, it is difficult to evaluate changes in the rest potential with solution composition when both sulfate and chloride ion concentrations are varied simultaneously (primary solutions). Rest potential values observed under these conditions range from -0.453 to -0.510 V. The negative shift in rest potential with increasing chloride ion concentration seen in the absence of sulfate

PAGE 156

143 ion interference (secondary solutions) is similar to that reported for AISI 304'°' 2 and AISI 304L 8 3 in solutions of constant sulfate concentration. Chin and Nobe,°5 using constant ionic strength perchlorate solutions to eliminate possible sulfate interference, report a negative shift in rest potential of 0.020 V per decade increase in chloride ion concentration. On the other hand, numerous investigators have found the rest potential to be independent of chloride ion concentration. Thus, for AISI 304,3^' iron, 3 and an iron-25 nickel alloy in sulfuric acid solutions as well as for iron in sodium chloride at pH 2.58 5 and iron-chromiui alloys in saline solutions varying in pH from 4.5 to 11,°' no change in rest potential as a function of chloride ion concentration has been observed. The data obtained in the present work suggest 1.) that the rest potentials observed in the secondary solution experiments fall within the range of values found for the primary solutions; 2.) that the rest potential values in 0.1 M potassium chloride with and without sulfate additions (-0.4-70 and -0.480 V, respectively) are approximately the same; and 30 that the addition of only 0.016 M sulfate ion to 1.0 M potassium chloride is unlikely to produce any change in the rest potential, although a 40 mV positive shift is observed. From these observations it is concluded that for AISI 304 in the solutions employed,

PAGE 157

144 chloride ion exerts no influence on the measured rest potential. This implies that both the cathodic and anodic reactions are depolarized by chloride ion to a similar extent. The above data are not entirely conclusive, however, and two facts must be noted. The typical variation in rest potential values for any given solution is only 5 to 10 mV (Tables 2 and 8). Also, in both cases in which rest potentials of ^ -0.50 V are observed, correspondingly low cathodic current densities are recorded at -0.?00 V, suggesting polarization of the cathodic reaction in these systems. Passive region The current flowing across a passive metal-solution interface may have three components! an electron current, flowing through the passive film, e.g., if a redox couple is present; an ionic current, caused by the passage of ions through the film resulting in film growth; and an equivalent ionic current, arising from dissolution of the passive film. The low current densities observed in these studies in the passive region in systems not subject to pitting breakdown are independent of potential. In the absence of an electron current, this implies that the rate of film growth is independent of potential since, for short times (< 1 hour), the ionic current of film dissolution has been shown to be

PAGE 158

145 negligible with respect to the film growth current. 8 ? An increase in potential in the passive region must therefore induce a proportional increase in film thickness, resulting in the maintenance of a constant field across the film. For an 18-8 stainless steel in 0.5 M sulfuric acid, Schwenk and Rahmel have shown that the logarithm of the rate of passive film growth is inversely proportional to the thickness of the film, i.e., growth occurs in accordance with the inverse logarithmic law, i = A exp (BV/d) (8) where i is the current density, V is the potential drop across the film, d is the film thickness, and A and B are constants. This behavior satisfies the constant current densityconstant field relation and can be explained in terms of the Mott-Cabrera ^ theory of high field film growth. This theory, originally derived for valve metals such as tantalum, has been extended to include the films QO formed on passive metals. Growth occurs by the high field (^ 10 6 V cm -1 ) conduction of metal cations through the film. The current-potential behavior is therefore exponential rather than ohmic. If the presence of a constant current in the passive region implies a proportionality between potential and passive layer thickness and if the interfacial capacitor can be represented as a parallel plate condenser with the passive

PAGE 159

146 layer as its dielectric, then a plot of reciprocal capacity (1/C) with respect to potential (E) should be linear. The contribution of the film-solution capacity to the total interfacial capacity is assumed to be small (see below). This theory does not apply to the entire potential range in which the passive current density remains constant or in the potential region preceding the break in capacity at E^. However, in the potential range bounded by E, and the potential at which the capacity minimum occurs, excellent correlation between the two functions is exhibited (Table 15). The capacity of a parallel plate condenser is given by c = ep /0.113d ( 9 ) _2 where C is the capacitance in jj. f cm , e is the dielectric constant, o , the surface roughness factor and d, the o distance between the capacitor plates in A. The constant, 0.113, has the units cm u f % . The film thickness may therefore be calculated from the measured value of capacity at a given potential in the potential range in which the condition of linearity between 1/C and E is satisfied. Since the measured capacities are not steady state values, only an approximation of the film thickness corresponding to a given potential is obtained. Evaluation of the passive layer thickness is complicated by the lack of knowledge concerning the exact nature of the passive film and its dielectric constant.

PAGE 160

14? Engell and Ilschner, 91 using a value of 10 and roughness factor of 4, have calculated the thickness of the passive layer formed on iron in 0.5 M sulfuric acid at potentials of 0.614 and 1.13 V. They obtain thicknesses of 10 and 20 A, respectively, after subtracting 2 A from the calculated value to account for the effect of the film-solution capacity on the total measured capacity. The use of 2 A is valid since, in concentrated solutions, the Outer Helmholtz plane is expected to be located within a few angstroms of the passive metal surface. The field o strength corresponding to film thicknesses of 10 and 20 A at potentials of 0.614 and 1.13 V is calculated as 5.14 x 10 ' V cm -1 . A similar analysis can be applied to the capacitypotential behavior observed in the present experiments. A sample calculation is presented for data obtained on AISI 304 in 0.334M sodium sulfate at pH 2.4. Using the roughness factor of 1.2 reported for an iron surface mechanically polished through 600 grit emery92 and the dielectric constant, 15»6, determined for films formed on AISI 304in 0.5 M sulfuric acid,-' film thicknesses of 5«9 and 10.7 A at 0.31 and 0.60 V, respectively, are obtained. The passive current density evaluated at 0. 30 V is 1.26 x 10 A cm The field required to maintain this rate of film formation is therefore, (0.60 0.31)/(10.? 5-9). 6.8 x 10 8 V A*" 1 , 6 _i or 6.8 x 10 V cm .

PAGE 161

14 8 The capacity values associated with 18-8 stainless _2 steel in the passive region prior to E^ (^20 uf cm ) agree well with those observed for platinum in the passive region. ^4 Similar agreement, coupled with the lack of hysteresis in the capacity-potential behavior for an 18-8 95 stainless steel, have lead Popat and Hackerman to assume that, like platinum, stainless steel has no bulk oxide on its surface in the passive region. The break in the capacity-potential relationship occurring at 0.3 V implies a change in the nature of the passive film resulting in a decrease in the conductivity or dielectric constant of the film. The ionic current resulting in film growth, i , remains constant during the capacity change P and passive films on stainless steel have been shown to be good electron conductors. ° Consequently, it is more likely that a decrease in dielectric constant occurs. Studies of the passive state of AISI 304in 0.5 M sulfuric acid show an increase in the protective nature of the passive film at 0.4 V.° f This has been attributed to a change in film structure through the loss of bound water. Values of the dielectric constant of bulk iron (III) oxide have been shown to increase significantly with an increase in iron (II) content. ° Either the elimination of a lower valence ion or of bound water from the passive film could, therefore, account for the decrease in capacity at 0.3V observed here.

PAGE 162

149 It appears, therefore, that the nature of the passive film on stainless steel in the solutions studied here is similar to that observed on platinum in the passive state up to a potential of 0.3 V. The passivity of platinum is attributed to an adsorbed "oxygen" species. At 0.3 V a change occurs in the structure of the film on stainless steel. Although this change may correspond to an increase in the degree of crystallinity of the passive film, it is unlikely . that a film of only 5 to 10 A m thickness would possess the properties of a bulk oxide. The passive current density observed in 0.33^ M sodium sulfate (1.26 x 10" A cm" ) is significantly higher than that observed in solutions containing chloride ion. A similar observation has been made by Moshtev, 99 who attributes the phenomenon on iron to the incorporation of sulfate anion into the passive film. The resulting increase in film defect density brings about an increase in the ionic conductivity of the film. The slight increase in passive current density with increasing chloride ion concentration observed here (Table 1) implies an increase in the solubility of the film. An increase in passive current density from -3 -3 -2 0.3 x 10 J to 1.3 x 10 A cm is also observed by Lebet and Piotrowski 10 for an 18-9 steel when K sodium chloride is added to 0.5 M sulfuric acid. (The relatively high current densities reported by these investigators are most likely the result of the high (0.330 V sec -'-) scan rate employed . )

PAGE 163

150 Fitting The increase in anodic current density caused by the onset of pitting dissolution occurs at a well-defined potential, E . , which is a function of the metal composition, pit its history, and the composition of the solution in which polarization is carried out (see Introduction). The active shift in pitting potential with increasing chloride ion concentration observed in the present experiments is well supported by literature data. ' ' Values of 7> E . V"2>log a-,,-, where a ,_ is the activity of pit' 01 ci the chloride ion, of -0.0? and -0.09 V decade" have been q Q on found by Leckie and Uhlig and by Leckie for AISI 304 in acid solutions. Various anions, e.g., the hydroxyl, perchlorate, sulfate and nitrate anions, have been shown to have an inhibiting effect on the activating action of 102 33 103 chloride ion. ' A critical activity of the inhibiting anion above which no pitting can occur for a given chloride ion activity has been proposed by Leckie 28 and Uhlig. For sulfate ion the relationship is given by log a cl _ = 0.85 log a go " O.05. (10) Application of this equation to the present results _2 accurately predicts the observed behavior. For 1.17 x 10 and 0.3 M potassium chloride, the calculated critical _2 sulfate concentrations are 7.6 x 10 and 0.295 M, respectively. Since the sulfate content of the 0.3 M potassium chloride solution is only 0.2^8 M, pitting breakdown is expected

PAGE 164

151 and is seen to occur. The inhibiting effect of sulfate ion in the present systems is demonstrated by the susceptibility of the steel to pitting breakdown in 0.1 M potassium chloride when no sulfate ion is present. No pitting is observed in 0.100 M potassium chloride with 0.300 M sulfate. Present results indicate a slight positive shift (~30 mV) in the pitting potential when the pH is decreased from 2.35 to 1.52 in 0.3 H potassium chloride. Since hydroxyl ion has been shown to act as a pitting inhibitor 28 for pH> ?, the effect of a decrease in pH in the acid range, if any, should be a negative shift in pitting potential. Literature data suggests, "' ' however, that the pitting potential is independent of solution pH in the acid range (pH<7). The positive shift in pitting potential observed is, therefore, attributed to the inherent irreproducibility of the pitting potential as it is determined here. The presence of an induction time is evidenced in the present studies by the current-time behavior observed at potentials positive to the pitting potential. A positive shift in potential results in an initially high current density attributed to charging of the double layer. When the potential is stepped to potentials positive to the pitting potential, an initially steady decrease in current density with time is observed signifying the continuation of the passivation process. The increase in film thickness caused by the ionic current flow results in a decrease in

PAGE 165

152 the magnitude of the field across the film. The current density therefore decreases until the steady state film thickness corresponding to the applied potential is reached. The fluctuations in current which follow this current decrease at a given potential have been attributed to the activation and subsequent repassivation of pits on the metal surface. Recent studies have associated the observed current oscillations with the increase and decrease in the resistance of the solution within pits caused by the successive formation and release of hydrogen bubbles. ' -' The formation of a bubble in a pit causes a negative shift in the potential of the pit base by increasing the ohmic (iR) potential drop across the pit solution. Since the potential at the pit base has been shown to be in the region of active dissolution, a decrease in anodic current density results. Release of the bubble reverses the effect. The most widely accepted theory suggests that passivity of the metal surface is destroyed by the local replacement of an adsorbed "oxygen" species by chloride ions at areas of high metal activity. Rapid metal dissolution ensues. An increase in chloride ion concentration and electrode potential favors adsorption of chloride ion and stabilizes the pitting reaction.

PAGE 166

153 The addition of sulfate ion inhibits the pitting reaction. The occurrence of sulfate-chloride ion competitive adsorption on a passive metal surface has been demonstrated by Maksimchuk and Rozenfel'd. ' The amount of chloride ion adsorbed on a passive porous chromium surface decreases with sulfate concentration and is completely eliminated above a critical sulfate concentration ( Cso^ ~] / [ci~J>5)» The relative adsorption characteristics of sulfate ion and chloride ion have been predicted from considerations based on Freundlich isotherms. 28, 103 The observed induction time for pitting breakdown has been attributed to the time-consuming reaction of competitive adsorption. Golovina et al. have shown that the induction time is related to the relative rates of the passivation and pitting reaction at a given potential. On the other hand, cyclic voltammetric behavior of an 18 10 stainless steel in solutions of sulfate and chloride ions shows no evidence of adsorption 109 of chloride on the passive surface. 7 For iron, the zero point of charge of the passive surface has been shown to occur at 0.10 V in 0.01 M sodium hydroxide. ^ 2 This value is considerably more positive than that of the assumed bare active metal in 5 x 10"^ M sulfuric acid (~-0.6l V). The zero point of charge of active iron is controversial because of the participation of adsorbed hydroxyl ions in the active dissolution reaction. If passivation results in a similar positive shift in the zero

PAGE 167

15^ point of charge for the stainless steel surface, then no specific adsorption of chloride or sulfate ions would be expected until potentials are slightly negative to the zero point of charge. In solutions containing only chloride ion, the negative shift in pitting potential would then be attributed to the known variation in zero charge potential in the presence of a specifically adsorbed species. How positive the pitting potential is shifted with respect to the zero point of charge by the addition of sulfate ion would be determined by the relative preferential adsorption of the two species since a critical surface concentration of chloride ions is necessary to induce pitting. If pitting breakdown does indeed occur through the specific adsorption of chloride ion to a critical surface concentration, the effect should be reflected in the behavior of the differential capacity-potential curve. If the interface can be represented in terms of equipotential planes parallel to the metal surface then it is unlikely that only local changes in double layer composition occur. However, no capacity-potential peak is observed between the potential of total passivity and the pitting potential or in the capacity-time relationship at potentials exhibiting an induction period prior to breakdown. This does not necessarily imply the absence of specific adsorption. The capacity of the electrical double layer must be considered to be in series with that of the passive film. From the

PAGE 168

155 expression relating total capacity to the individual values of capacities in series, 1/C = i/c 1 + i/c 2 , (ll) it can be seen that the presence of a relatively small passive film capacity would obscure changes in the capacity of the film-solution interface caused by specific adsorption. If the passive film capacity (C ) is assumed to be 15 U f _2 cm , then a change in the film-solution capacity from 30 to 60uf cm" as a result of specific adsorption would only increase the total measured differential capacitance of the interface from 10 to 12 U f cm" 4 . A similar domination of the interfacial capacity by that of the passive film has been reported by Fosey and Sympson. The rate of copper(II) reduction on a passive stainless steel electrode has been shown to increase in the presence of chloride ion as a result of the specific adsorption of the anion. However, no change occurs in the capacity-potential relationship in the corresponding potential region. Analysis of the capacity-potential curves presented here shov/s that the same behavior persists both in the presence and absence of pitting breakdown. The break in capacity at E^ occurs at 0.3 V whether the pitting potential is positive or negative to this value. This behavior supports the concept that the capacity of the passive film dominates the interfacial capacity. Since changes in the passive film caused by pitting breakdown are only

PAGE 169

156 local, they will not be reflected in the measured capacityvalues which are an average property of the interface. In conclusion, it can therefore be stated that differential capacity measurements cannot differentiate between the presence or absence of the specific adsorption of chloride ion prior to the onset of pitting corrosion. The domination of the interfacial capacitance by that of the passive film acting in series with the film-solution capacity obscures any changes which may occur in the latter. Since pitting induces only local changes in the passive film, these changes will not be reflected in the measured capacity which is a property of the average condition of the film. Transpassive dissolution Transpassivity is common to passive alloys one or more of whose components exhibit a second stable oxidation state. For alloys containing chromium, the current increase in the transpassive region is attributed to the oxidation of chromium (III) in the passivating film to chromium (VI) 112 which dissolves as the chromate or hydrogen chromate 79 111 anion. The standard potentials for the oxidation of various forms of chromic oxide to either of these species 114 range from 0.872 to 1.141 V. Dissolution in the transpassive region exhibits Tafel behavior over 1 to 1.5 decades of current. A Tafel slope of 0.114 V decade -1 is observed for 0.0 M and 1.17 x 10~ 2 M potassium chloride. This value is considerably higher than

PAGE 170

157 those reported in the literature for stainless steels in acid solutions (0.055 to 0.090 V decade" 1 ). 115, 116 ' 117 Values determined for pure chromium range from O.O38 to 0.048 V a -1 5, P0 decade x . The increase in slope to 0.150 V decade" in 9«97 x _2 10 M potassium chloride implies a change in dissolution mechanism. This is unexpected in light of the results of 11? Heumann and Fanesar ~ for chromium and iron-chromium alloys in which neither chloride ion nor sulfate ion participates in the dissolution reaction. Although the Tafel slope for transpassive dissolution of nickel in sulfuric acid solutions has been shown to be 0.150 to 0.172 V decade" 1 , 79 ' llS the transpassive current-potential relationship for chromium 79 dissolution from both nickel-chromium binary alloys y and iron-chromium alloys containing nickel is independent of nickel content. Therefore, the Tafel slope of 0.150 V decade" observed in the present work cannot be attributed to an effect of nickel on the transpassive dissolution of stainless steel. The onset of secondary passivity following the region of Tafel behavior is evidenced by a current maximum. Neither the current maximum (1 x 10 ^ A cm" ) nor its associated potential (O.95 V) is affected by solution composition (Ln non-pitting systems) at constant pH. The current then decreases to a minimum followed by an increase which is the result of oxygen evolution. The degree of

PAGE 171

stability of the secondary passive state is represented by the difference in magnitude of the current density value of the maximum and subsequent minimum, A i. Present results indicate that stability decreases slightly on the addition of chloride ion (1.1? x 10 M) to a solution originally containing sulfate only at pH 2.4. Current differences, Ai, of 14 x 10" 6 to 1? x 10 -6 and 21 x 10 -6 A cm" 2 , respectively, are observed. These results agree well with those of Szklar ska -Smialowska and Janik-Czachor, who, for an iron-13 chromium alloy, find a decrease in A i on adding 0.05 M sodium chloride to 0.035 M sodium sulfate but see no change in the parameters of secondary passivation. The presence of a differential capacity peak at potentials in the vicinity of the secondary passivation potential observed in the present work suggests the involvement of an adsorption step in the passivation mechanism. A capacity maximum in the region of secondary passivity has also been 119 observed by Frazak and Cihal for an 18 9 stainless steel in 0.5 M sulfuric acid. The cause of secondary passivity has been attributed to the adsorption of an "oxygen" species which is facilitated by the presence of carbon in the alloy. Carbon is thought to increase the affinity of the alloy for 120 ' 121 adsorbed oxygen.

PAGE 172

159 Transpassive dissolution initiates between O.57 and 0.64 V in the systems studied here. These values are substantially negative to the standard potentials for chromium oxide oxidation defined above. This is explained, in part, by the negative shift of the reversible potential for the transpassive dissolution reaction with increasing pH. Variations of-O.058 V pH" 1 and -O.O65 V pH" 1 have been 117 observed for a stainless steel in nitric acid and for 112 iron chromium alloys m sulfuric acid, respectively. If the cold work generated by mechanical polishing during electrode preparation causes an increase in the diffusion 13 of chromium to the alloy surface as suggested by Rhodin, then the negative shift in E, with increasing chromium concentration reported for stainless steels in 6 M sulfuric acid could also contribute. ' The negative shift observed in E, with increasing tr chloride ion concentration (Table 1) has also been observed 122 by Shock et al. for AISI 304 in 67 percent sulfuric acid. No effect of chloride ion concentration on E tr for AISI 304 in 5 M sulfuric acid is observed by Acello Q p and Greene," however. In summary, in systems not subject to pitting corrosion , transpassive dissolution occurs, starting at 0.57 to 0.64 V. Dissolution in this region is attributed to the oxidation of chromium(III) in the passive film to chromium(VI) . Tafel behavior is observed. A slope of 0.114 V decade occurs in 0.0 and 1.17 x 10~ 2 M potassium chloride. The

PAGE 173

160 change in slope to 0.150 V decade" when the chloride -2 -1 concentration is increased to 9-97 x 10 V decade implies a change in the mechanism by which transpassive dissolution proceeds. The subsequent maximum in the current density-potential relationship indicates the onset of secondary passivity. Both the current density at the maximum and its corresponding potential are independent of solution composition. The capacity peak observed in this region is attributed to the adsorption of the passivating species . Galvanostatic Polarization The potential-time behavior of stainless steel samples subjected to galvanostatic polarization in the solutions studied has been shown to depend both on current density and solution composition. Systems which are not susceptible to pitting corrosion reach and maintain a positive steady state potential under influence of an anodic current. In systems subject to pitting breakdown, the maximum potential attained is unstable and a shift in potential to more active values is observed. Similar results have 123, 42, 41, 38, 40 been reported by several investigators. From the values of the potential arrests and maxima as well as from the presence of an induction time prior to pitting breakdown, it appears that the initial action of the anodic current is the same in all cases. Schmid has suggested that this initial action involves the

PAGE 174

161 conversion of adsorbed water dipoles to an adsorbed "oxygen" species . This hypothesis is supported by the presence of the initial potential arrest, E, , leading to passivation of the alloy surface in all the solutions studied here. In 0.0 through O.303 M potassium chloride, a Tafel slope of 0.06 V decade" is associated with the reaction occurring at E, . Very little work has been done to elucidate the dissolution mechanism of stainless steels in acid solution. Accordmg to Mueller, most of the dissolution current observed on stainless steels containing nickel should represent the oxidation of iron. Chromium and nickel should passivate as soon as they are exposed to the solution by removal of iron from the lattice. If steady-state polarization is approximated, only one passivation peak will be observed as soon as the fraction of the surface consisting of passivated chromium and nickel exceeds 0.5It is possible, therefore, that the effect of solution composition on the kinetics and mechanism of stainless steel dissolution may be similar to that on pure iron. For iron in acidic solutions of weakly adsorbable anions (e.g., sulfate and perchlorate ions), dissolution is believed to proceed through the adsorption of water dipoles which dissociate to form specifically adsorbed hydroxyl ions. Two mechanisms have been proposed to account for the kinetic parameters observed: the

PAGE 175

162 125 126 1?? Bockris-Kelly ' ' mechanism and the Heusler mechanism. The reaction sequence representing the Bockris-Kelly mechanism is Fe + H 9 = Fe(H 9 0) (12) c ^ ads Fe(H 2 0) adg Fe(OH-) ads + H + (13) Fe(OH") ads = Fe(OH) + e (14) Fe(OH) n -* FeOH + + e (15) ads FeOH + + H + = Fe 2+ + HO (16) where reaction 15is the rate-determining step. A Tafel slope of 0.040 V decade and a reaction order with respect to hydronium ion of -1 are characteristic of this mechanism. Dissolution by the Heusler mechanism proceeds through the steps, Fe + HO = Fe(H o 0) A (1?) 2 2 'ads v ' ' Fe(H 2 0) ads = Fe(OH-) ads + H + (18) Fe(OH-) ads = Pe(OH) ads + e (19) Fe(OH) ads + Fe = Fe(FeOH) ads (20) Fe(FeOH) + OH" — FeOH + + Fe(OH) . + 2e (21) FeOH + + H + = Fe 2+ + H 0. (22) Reaction 21 is the rate-determining step and involves the surface catalyst, (FeOH) , . The value of the kinetic parameters characterizing this reaction sequence depends on the manner in which they are measured. Steady-state polarization results in a Tafel slope of 0.030 V decade and

PAGE 176

163 a hydronium ion reaction order of -2. Non-steadystate measurements yield values of 0.060 V decade and -1, respectively. In both cases, an increase in pH accelerates the 12P dissolution reaction. Recent work has suggested that the Eockris-Kelly mechanism predominates at low energy surface sites while the Heusler mechanism is favored by a high imperfection density and grain boundary to grain area ratio. In solutions containing a specifically adsorbable ion, e.g., chloride ion, participation of the ion in the 129 130 dissolution mechanism has been indicated. The influence of chloride and hydroxyl ions on dissolution depends on the relative concentrations of the two species. 131 McCafferty and Hackerman have shown that, in solutions of high chloride content (6r.'i)» hydronium ion, in excess of 2.^ M, accelerates rather than inhibits the dissolution reaction. Similar results have been reported by Darwish, .. i 132 et al. In solutions of low pH (^ 1.52 to 2) and moderate -2 133 chloride ion concentrations (^10 fv]), Lorenz has shown that chloride ion has an inhibiting effect on iron dissolution. A change in reaction mechanism characterized by a change in Tafel slope to 0.060 V decade" and a reaction order of -1 for both hydronium and chloride ions is

PAGE 177

164 o observed. However, below 10 M potassium chloride at low pH (< 1.52 to 2) ** as well as in 1 M potassium 130 chloride at pH 2, iron dissolution again proceeds according to the Bockris-Kelly or Heusler mechanisms. The mechanism by which dissolution of stainless steel occurs in the chloride-sulfate containing solutions employed here cannot be determined precisely from available data. However, because of the increase in critical current density with increasing total anion concentration observed during potentiostatic polarization, it is assumed that, in the pH range studied (1.52 to 6.22), chloride ion does not inhibit the dissolution process. The Tafel slope of 0.060 V decade" observed in 0.0 through O.303 H potassium chloride is determined galvanostatically, i.e., using a non-steady-state technique. It therefore seems likely that the dissolution reaction associated with the potential arrest at E, proceeds according to the Heusler mechanism. The high number of 12 -2 surface imperfections (~10 cm ) expected on the mechanically polished (cold worked) surfaces used here supports this conclusion. Based on the results of impedance measurements, a similar conclusion has been reached by Schwenk and Buhler. 1 35 Verification of this hypothesis awaits determination of the reaction orders of the species involved in the dissolution reaction.

PAGE 178

16 5 The change in Tafel slope from 0.060 to 0.014 V decade -1 observed on increasing the chloride concentration from 0.303 to 0.516 f.j indicates that a change in dissolution mechanism has occurred. The significant depolarization of the reaction implied by this Tafel slope (0.014 V decade" 1 ) results in the high value of the critical current density seen in 0.5 M potassium chloride during potentiostatic polarization (Table 1). The charge necessary to bring about passivation in any given solution determined from the product of the applied current density and the length of the arrest at E, appears to decrease with increasing current density. This is probably related to the corresponding positive shift in E, which occurs. The more positive the potential in the active dissolution region, the higher the internal anodic current will be. Since it is the total anodic charge passed which determines plateau length, a decrease in t with increasing applied current density will therefore occur as the potential of the arrest becomes more positive. Potential arrests observed at more noble potentials ( ^ 0.8 V) agree well with characteristic potential regions observed during potentiostatic polarization. The plateau occurring at 0.3 V is attributed to the transpassive dissolution of chromium from the passive film. The Tafel slope observed in O.337 M sulfate (O.058 V decade" 1 ) is in

PAGE 179

16 6 much better accord with literature values (see above) than that determined potentiostatically (0.114 V decade" 1 ). This maybe attributed, in part, to the different measuring techniques employed. In potentiostatic polarization, the time spent at each potential permits the surface to approach the steadystate conditions corresponding to that potential. This results in the film's being thicker at 0.8 V during potentiostatic polarization than at the same potential during galvanostatic polarization. The presence of a thicker film will inhibit the transpassive dissolution reaction. A higher overvoltage is therefore necessary to produce a given current density, as shown by the change in the Tafel slope to 0.114 V decade . A change in dissolution mechanism in 0.1 H potassium chloride is reflected both in the galvanostatic and the potentiostatic results as shown by an increase in Tafel slope from 0.058 and 0.114 V decade" 1 to 0.121 and 0.150 V decade -1 , respectively. A value of 0.161 V decade is determined galvanostatically in O.303 M potassium chloride. -1 Q The increase in exchange current density from l.?4 x 10" to 2.69 x 10"° A cm over the chloride concentration range 0.0 to O.303 M suggests that the reaction is depolarized by chloride ion. During the change in potential from the plateau at 0.8 V to that at O.85 V, a peak in the capacity-potential relationship occurs. The presence of a potential maximum

PAGE 180

16? at /\/0.90 V in the current-potential arrest relationship associated with the plateau at O.85 V also indicates its relation to the secondary passivation process. The value of the current density maximum associated with secondary -4 -2 passivity in potentiostatic experiments is 1 x 10 A cm The current density at which the potential maximum is -3 -2 produced during galvanostatic polarization, 1 x 10 J A cm , is higher. This emphasizes the importance of the age of 37 the film on its protective properties. The Tafel slopes, calculated from the current values prior to the potential maximum, range from 0.047 V decade in 0.337 Ivi sulfate to 0.145 V decade" in 0.3 M potassium chloride. This again suggests a change in mechanism in the Z4--1 presence of chloride ion. Schmid has also reported a Tafel slope of 0.147 V decade" in this potential region. -23 An increase in exchange current density from 4.57 x 10 10 — 2 to 4.17 x 10 A cm with increasing chloride concentration is also observed for this arrest. The oxygen evolution reaction occurring in 0.0 through O.303 f>'i potassium chloride appears to be unaffected by solution composition at constant pH in the range studied. In 0.100 I.; potassium chloride solutions, no pitting breakdown occurs, while samples subjected to constant current polarization in O.303 M potassium chloride solutions undergo pitting corrosion. However, comparison

PAGE 181

168 of the more noble potential arrests observed in these two systems supports the conclusion that the initial surface states produced by the applied anodic current are similar and that it is a perturbation of these initial surface states which results in the onset of pitting. In both systems transpassive dissolution (Tafel slopes of 0.121 and 0.161 V decade" , respectively), secondary passivity, and oxygen evolution (Tafel slope of 0.022 V decade" in both cases) are observed (Tables 17 and 18). In 0.3 I", potassium chloride, however, the chloride concentration is high enough to produce, by competitive adsorption, the critical surface concentration needed to induce pitting breakdown. Noble potentials in the transpassive dissolution or oxygen evolution regions are then no longer necessary to maintain the imposed current density, and a decrease in potential to values associated with the pitting reaction occurs. In 0.5 K potassium chloride solutions pitting breakdown occurs from potentials considerably below those observed in 0.3 M potassium chloride (0.44-7 to 0.798 V versus O.765 to 1.39° v » respectively). It appears that the increase in chloride ion concentration, coupled with the decrease in sulfate concentration, allows the critical surface concentration of chloride needed to cause pitting to be reached at a smaller positive charge on the metal and with a shorter induction time.

PAGE 182

169 Analysis of the capacity-potential behavior observed in both 0.303 and 0.518 M potassium chloride solutions shows that the only capacity peaks occurring are those associated with the primary and secondary passivation reactions (and, in some cases, the oxygen evolution reaction). No capacity peak appears to be associated with the timeconsuming alteration of the surface occurring prior to pitting breakdown. This behavior, coupled with the capacity values observed during breakdown from potentials below 0.8 V (R to 15 yf cm ) , support the idea that the capacity of the interface is dominated by that of the passive film, as discussed above. The steady-state potential achieved after pittinginduced potential breakdown has been described as the 42 protection potential of the system. Above this potential, previously nucleated pits can grow. In the presence of inhibiting ions in concentrations insufficient to prevent pitting, the determination of the value of the steady-state potential is hampered by oscillations in the potential with time. The phenomena leading to current oscillations are 136 described by Rozenfel'd and IVIaksimchuk. ' To maintain current flow through the solution, chloride ion migrates into the pits. The resulting depolarization of the dissolution reaction in the pit causes a shift in the potential within the pit to active values. Because of local cell

PAGE 183

1?0 action, this induces a negative shift in the potential of the entire surface. This causes partial desorption of chloride ions. Under the continued influence of the applied current, a positive shift in potential then occurs, causing re-adsorption of the chloride ion. The resulting cyclic behavior is reflected in potential oscillations. The potential at the base of a pit has been shown to be in the region of active dissolution of the metal under study. -' The positive shift in the average potential attained after pitting breakdown with increasing current density at a given chloride ion concentration is representative of the behavior expected if the potential within the pit is in the active dissolution region of the pit metalsolution interface. The composition of the anolyte in the pit has been shown to be significantly higher in chloride ion and hyclronium ion because of the influence of chloride ion migration and metal ion hydrolysis, respectively . "^7 Because of the depolarizing effect of the chloride ion on active metal dissolution, the average potential observed in 0.513 K potassium chloride is considerably more active than that in O.303 M (Table 21). The pit density on a metal surface has been shown to be a function of potential. f0 The slight negative shift in the average final potential when oxygen evolution potentials are reached prior to breakdown may therefore be attributed to an increase in the number of pits formed

PAGE 184

171 on the surface. This results in a decrease in the current density associated with each pit. Very few data are reported in the literature on the open circuit decay behavior of stainless steels in acid solutions. During forced decay (constant current reduction) of the passive film on iron-chromium alloys in 0.100 \<] sodium sulfate at pH 2.2, Aronowitz and Hackerman find three potential arrests occurring at approximately 0.48, 0.05 and -0.1 V, respectively. Because no interrelation in plateau lengths is found, it is assumed that the arrests do not involve stepwise reduction of a single passivating species hut are a result of the reduction of three independent surface entities. The first arrest (0.48 V) is shown to result from reduction of a species produced during transpassive dissolution. For alloys with < 12 percent chromium, reduction of the surface material associated with the second arrest results in activation of the metal surface. When the chromium content of the alloy is increased to > 12 percent, the surface species reduced during the third -2 arrest, corresponding to a charge of 0.7 mcoul cm , is sufficient to maintain the alloy in the passive state. The arrests observed in the present experiments during open circuit decay in non-pitting systems show similar potential values. The one between 0.4 and 0.5 V occurs

PAGE 185

172 only when transpassive dissolution occurs before constant current control is removed. This supports its association with a substance produced in the transpassive dissolution region. The two lower arrests observed in the present results occur at more negative potentials than those seen during constant current reduction. This is, in part, a result of the increased stability observed in the passive films formed on nickel-containing stainless steels. Capacity peaks of 20 to 4-0 y. f cm seen in conjunction with these two lower arrests suggest the participation of a desorption reaction related to the activation of the alloy surface. The most negative potential observed during open circuit decay is not a steady state value. A positive shift in potential occurs over extended periods of time (Table 22). This positive shift is probably a result of the equilibration of the double layer and bulk solution which is expected to contain oxidizing species (Fe^ + , Cr 2 ~, ) produced during the polarization process.

PAGE 186

CHAPTER V SUMMARY The mechanism by which chloride ion acts to initiate pitting breakdown of a passive metal surface is unknown. If specific adsorption of chloride ion to a critical surface concentration is a necessary precursor to pitting, then its occurrence should cause a change in the differential capacity of the passive film-solution interface. To examine this possibility, the differential capacity of an AISI 30^ stainless steel-solution interface was determined during potentiostatic and galvanostatic polarization in solutions of potassium chloride at pH 2.4. Sodium sulfate was added to maintain an ionic strength of one. The current density-potential and potential-time behavior resulting from constant potential and constant current polarization, respectively, were also observed. During potentiostatic polarization, the transition from the active to the passive state is obtained for all solutions studied (0.0 through 1.0 M potassium chloride). The rest potential ( /v -0.48 V versus SCE) and the primary passivation potential ( /%/ -0.42 V versus SCE) are independent of solution composition. The critical current density increases linearly with the concentration function, pH + log ( [S0^ J + [C1"J ), 173

PAGE 187

174 where [sc^ "] and [ci~] are the total analytical concentration of the sulfate and chloride species present. This behavior implies that sulfate and chloride ions are equivalent in their influence on the dissolutionpassivation mechanism(s) . The invariance of the primarypassivation potential at constant pH suggests that the reaction leading to dissolution and the reaction leading to passivation are equally depolarized at any given solution composition. The capacity peak occurring slightly negative to the primary passivation potential is attributed to the specific adsorption of the anionic species involved in the reaction (hydroxyl, sulfate, bisulfate and chloride ions). The potential at which the electrode becomes totally passive is a linear function of solution composition, rather than of the logarithm of solution composition. It appears, therefore, that its value is determined by kinetic rather than thermodynamic factors. In systems not subject to pitting ( <0«3 M potassium chloride), transpassive dissolution occurs, initiating between O.57 and 0.64 V. Dissolution in this region is attributed to the oxidation of chromium (III) in the passive film to chromium (VI). Tafel behavior is observed. A slope of 0.114 V decade -1 occurs in 0.0 and 1.17 x 10~ 2 M potassium chloride. The change in slope to 0.150 V decade -1 when the chloride concentration is increased to 9.9 x 10 -2 M implies a change in the mechanism by which transpassive dissolution proceeds.

PAGE 188

175 The subsequent maximum in the current density-potential relationship indicates the onset of secondary passivity. Both the current density at the maximum ( 1 x 10"^ A cm" 2 ) and its corresponding potential (0.95 V) are independent of solution composition. The capacity peak observed in this region is attributed to the adsorption of the passivating species. In solutions of > 0.3 M potassium chloride, the passive state is not stable and pitting corrosion occurs at potentials normally in the passive region. Sulfate ion is found to inhibit the activating effect of chloride ion. The pitting potential, therefore, depends on the relative amounts of chloride and sulfate ions in the solution, shifting toward more active values as the chloride ion concentration is increased. No capacity peak is observed prior to pitting breakdown. However, this does not imply the absence of a specific adsorption event. Rather, it is found that the capacity, between the potential of total passivity and the capacity minimum at 0.60 V, is approximately independent of the stability of the passive state. This independence results from the domination of the interfacial capacitance by that of the passive film, obscuring any changes in the filmsolution capacity which may arise. In both pitting and non-pitting systems, the capacity changes suddenly at 0.3 V from a value of ~ 18 uf cm" 2 to ~15^f cm" . This change is attributed to a decrease in

PAGE 189

176 the dielectric constant of the film caused either by the elimination of bound water or of low valence ions (e.g., Fe(II)) from the film. Between 0.3 and 0.6 V, a linear relationship is observed between the reciprocal of the differential capacity and the potential. Film growth, therefore seems to follow an inverse logarithmic lav/. The passive film formed in 0.33^ K sodium sulfate at 0.3I and 0.6 V is found to have thicknesses of 5*9 and 10.7 A, respectively. The potential-time behavior of samples subjected to galvanostatic polarization depends on both current density and solution composition. Arrests are observed at -0.4, 0.0, 0.8, 0.85, and 1.12 V. Systems not susceptible to pitting (<0.3 r potassium chloride) reach and maintain a steady-state potential between 0.8 and 1.4 V for -4 -3 -2 the current densities employed (1 x 10 to 2 x 10 A cm ). In systems subject to pitting, the maximum potential attained is unstable and a shift in potential to more active values occurs. The final average potential observed following breakdown is shown to be in the potential range of active dissolution for the metal in contact with the solutions in the pit.

PAGE 190

177 In 0.303 l : . potassium chloride, intermediate arrests produced by a given current density as well as the maximum potential achieved prior to breakdown correspond to potential arrests in non-pitting systems. In O.5I8 l.i potassium chloride, pitting breakdown occurs from a potential maximum which is considerably below that observed in O.303 ]V] but the behavior at more negative potentials in the two solutions is similar. It is assumed therefore, that the initial effect of the anodic current on the metal surface at a given potential is the same in all cases and that pitting succeeds through a perturbation of these initial surface states by chloride ions . Tafel behavior is associated with the arrests at -0.4, 0.8, 0.8.5 and 1.12 V. Active dissolution at -0.4-0 V exhibits a Tafel slope of 0.06 V decade" in 0.0 through O.303 M potassium chloride. A mechanism change occurs in 0.518 U potassium chloride where a slope of 0.014 V decade" is observed. In all cases, a capacity peak is associated with this arrest. Since the critical current density is shown to increase with chloride concentration and galvanostatic polarization is a non-steadystate technique, dissolution is assumed to proceed through the Heusler mechanism. The arrest at 0.8 V results from transpassive dissolution. The discrepancy between the Tafel slope observed in O.337 M sodium sulfate during potentiostatic polarization (0.114 V

PAGE 191

178 decade ) and that observed during galvanostatic polarization (O.O58 V decade -1 ) is attributed to the thicker film expected at this potential under potentiostatic conditions. Tafel slopes of 0.121 and 0.161 V decade* 1 are observed in 0.100 and O.303 M potassium chloride, respectively, confirming the presence of a different reaction sequence at higher chloride ion concentrations. Arrests at O.85 and 1.12 V correspond to secondary passivity and oxygen evolution, respectively. Capacity peaks are associated with the onset of secondary passivity and in some cases with oxygen evolution. However, no capacity peak is associated with pitting breakdown per se.

PAGE 192

LITERATURE CITED 1. M. Faraday, Phil. Mag. ,_Q_, 57, 122, 153 (I836). 2. K. Vetter, Z. Elektrochem., 62, 642 (1958). 3. T. F. Hoar, Corrosion Sci . , £, 3^1 (1967). 4. U. R. Evans, J. Chem. Soc, 1024 (1927). 5. Ya. M. Kolotyrkin, Z. Elektrochem., 62, 664 (1968). 6. B. Kabanov, R. Burstein, and A. Frumkin, Disc. Faraday Soc, 1, 259 (19W. 7. H. G. Feller and H. H. Uhlig, JElectrochem. Soc, 107, 864 (I960). 8. H. H. Uhlig, J. Electrochem. Soc, 21, 215c (1950). 9. B. N. Kabanov and D. I. Leikis, Zh. Fiz. Khim., 20, 995 (1946). 10. N. Hackerman, Z. Electrochem., 62, 632 (1958). 11. G. Aronowitz and N. Hackerman, J. Electrochem. Soc, 110 , 633 (1963). 12. R. P. Frankenthal, J. Electrochem. Soc, 114, 5^2 (I967) 13. T. N. Rhodin, Jr., Corrosion, 12, 123 (1956). 14. C. L. McBee and J. Kruger, Electrochim. Acta, 17 , 1337 (1972). 15. H. H. Uhlig, Corrosion, 1£, 231t (1963). 16. K. Osozawa and H. J. Engell, Corrosion Sci., 6, 389 (1966). 17. Ya. M. Kolotyrkin, Corrosion, 1£, 26lt (1963). 18. Z. Szklarska-Smialowska, Corrosion, 2£, 223 (1971). 19. H. Bbhni and H. H. Uhlig, Corrosion Sci., 9_, 353 (1969). 20. S. A. Glazkova, L. I. Freiman, G. L. Shvarts, and G. S. Raskin, Zashch. Metal., 8, 660 (1972). 179

PAGE 193

180 21. A. F. Bond, J. Electrochem. . Soc, 120 , 603 (1973). 22. A. F. Bond and E. A. Lizlovs, J. Electrochem.. Soc, 11.5 , 1130 (1968). 23. V. Cihal and P. Frazak, J. Iron Steel Inst., 193, 3^0 (1959). 24. P. Fujii and W. Kurnada, Nippon Kinzoku Gakkaishi, 3.4, 1001 (1970). 25. A. Randak and F. W. Trautes, Werkst. Korros., 21, 97 (1970). ~ 26. N. T. Tomashov and 0. N. Parkova, Zashch. Petal,, 6, 21 (1970). 27. H. F. Leckie, J. Electrochem... Soc, 117, 1152 (1970). 28. H. F. Leckie and H. H. Uhlig, J. Electrochem. Soc, 113 , 1262 (1966). 29. V. Hospadaruk and J. V. Petrocelli, J. Electrochem. . Soc, 113, 878 (1966) . 30. B. Pondot, I". Da Cunha Belo and J. Montuollo, C. R. Acad. Sci., Ser. C, 274, 1028 (1972). 31. G. S. Eklund, J. Electrochem.. Soc, 121, 467 ( 197*0 . 32. P. Smialowski, Z. Szklarska-Smialowska, M Rychcik, and A. Szummer, Corrosion Sci., £, 123 (1969). 33. L. L. Rozenfel'd and V. F. Paksimchuk, Zh. Fiz. Khim., 3_5, 2561 (1961). 3^. H. Uhlie; and J. Oilman, Corrosion, _20, 289t (196*1). 35. P. Pourbaix, L. Klimzack-Pathieiu, Ch. Pertons, J. Peunier, CI. Vanleugenhaghe, L. DePunck, J. Laureys, L. Neelmans, and M. V/arzce, Corrosion Sci.,_3_, 329 (I963). 36. R. F. Jackson and D. Van Rooyen, Corrosion, 27, 203 (1971). 37. N. Stolica, Corrosion Sci.,_9, 205 (1969). 38. Z. Szklarska-Smialowska and M. Janik-Czachor, Corrosion Sci., 7, 65 (1967). 39. N. Stolica, Corrosion Sci., 9_, 455 (1969). 40. W. Schwenk, Corrosion, 20, 129 (1964).

PAGE 194

181 41. G. M. Schmid and N. Hackerman, J. Electrochem. Soc, 108 , 741 (1961). 42. Z. Szklarska-Smialowska and M. Janik-Czachor, Corrosion Sci., 11, 901 (1971). 43. K. Nobe and R. F. Tobias, Corrosion, 20, 263t (1964). 44. N. Sato, Electrochim. Acta, 16, 1683 (1971). 45. T. P. Hoar and W. R. Jacob, Nature, 216, 1299 (1967). 46. J. O'M. Bockris and A. K. N. Reddy, "Modern Electrochemistry," Vol. 1, pp. 749-769, Plenum Press, New York (1970). 47. T. Murakawa, T. Kato, S. Nagaura, and N. Hackerman, Corrosion Sci., £, 657 (1967). 48. H. Vai dyananthan and N. Hackerman, Electrochim. Acta, 16, 2193 (1971). 49. T. Murakawa, T. Kato, S. Nagaura, and N. Hackerman, Corrosion Sci., 8, 34l (1968). 50. T. Murakawa and N. Hackerman, Corrosion Sci., 4, 387 (1964). 51. B. S. Krasikov and V. V. Sysoeva, Proc. Acad. Sci. USSR, Phys. Chem. Sect., 114, 363 (1957). 52. R. S. Perkins and T. N. Andersen, "Modern Aspects of Electrochemistry," Vol. 5, J. O'M. Bockris and B. E. Conway, Editors, pp. 203-290, Plenum Press, New York (1969). 53« R« de Levie, Electrochim. Acta, 1_0, 113 (1965). 54. J. S. Riney, G. M. Schmid, and N. Hackerman, Rev. Sci. Instr., ^2, 588 (1961). 55R« P. Frankenthal and H. W. Pickering, J. Electrochem. Soc, 112, 514 (1965). 56. S. Harrar, Anal. Chem., ^4, 1036 (1963). 57> G. M. Schmid and N. Hackerman, J. Electrochem. Soc, 115 , 1033 (1968). 58. B. E. Wilde, Corrosion Sci., £, 315 (I967). 59' P. F. King and H. H. Uhlig, J. Phys. Chem., 63, 2026 (1959).

PAGE 195

182 60. H. J. Rocha and G. Lennartz, Arch. Eisenhuttenw. , 26, 117 (1955). 61. R. N. Adams, "Electrochemistry at Solid Electrodes," pp. 183-184, Marcel Dekker, Inc., New York (1969). 62. P. Delahay, "New Instrumental Methods in Electrochemistry, pp. 20--209, Interscience, New York (1954). 63. F. Delahay, "Double Layer and Electrode Kinetics," pp. 33-°0, Interscience, New York (I965). 64. G. V.. Florianovich, L. A. Sokolova, and Ya. M. Kolotyrkin, Electrochiin. Acta, 12, 879 (1967). 65. R. J. Chin and K. Nobe, J. Electrochern . Soc . , 119, 1457 (1972). 66. g. Kreysa, U. Ebersbach, and K. Schwabe, Electrochim. Acta, 16, 1489 (1971). 67. N. Hackerman and S. J. Stephens, J. Fhys. Chem., 58, 904 (1954). 68. r.'i. Fourbaix and F. Vander ' velden, Corrosion Sci., 5, 81 (1965). * 69. CD. Kim and B. E. Wilde, Corrosion, 28, 26 (1972). 70. Ho Ngok Ea and Nguyen Dyk Vi, Soviet Electrochern., 4, 894 (1968). 71. L. Feraldo Bicelli and C. Romagnani, Chim. Ind . (Mian). 53, 1128 (1971). 72. L. Troselius, Corrosion Sci., 11, 473 (1971). 73R. N. O'Brien and F. Seto, J. Electrochern Soc, 117 , 32 (1970). 74. I". A. V. Devanathan and Z. Stachurski, J. Electrochern. Soc, 111, 619 (1964). 75. C. D. Kim and B. E. Wilde, J. Electrochern. Soc, 118, 202 (1971). " 76. E. NcCafferty and N. Hackerman, J. Electrochern. Soc. 119, 146 (1972). 77. F. G. Hodge and B. E. Wilde, Corrosion, 26, 146 (I970).

PAGE 196

183 78. H. '.'/. Pickering and R. P. Frankenthal, J. Electrochem. Soc, 112, 761 (1965). 79. J. R. Myers, F. H. Beck and M. G. Fontana, Corrosion 21, 277 (I96.5). 80. G. T. Seaman, J. R. Myers, and R. K. Saxer, Electrochim. Acta, 12, 855 (1967). 81. B. E. V/ilde and F. G. Hodge, Electrochim. Acta, 14, 619 (1969). ~ 82. S. J. Acello and N. D. Greene, Corrosion 18, 286 (I962). 83. N. D. Greene and G. Judd, Corrosion, 21, 15 (I965). 84. D. 0. Condit, Corrosion, 28, 95 (1972). 85. S. E. Trauenberg and R. T. Foley, J. Electrochem. Soc, 1 18 , 1066 (1971). p 6. E. D. Verink and ]'.;. Pourbaix, Corrosion, 2_7, 495 (1971). q 7. G. K. Bulman and A. C. C. Tseung, Corrosion Sci., 12, 415 (1972). — PP. W. Schwenk and A. Rahmel, Electrochim. Acta, 5, 180 (1961). 2 89. N. Cabrera and N. F. Mott, Rents. Pros. Fhys . , 12. 163 (1948-9). — D 90. J. L. Ord. and J. H. Bartlett, J. Electrochem. Soc, 112 , 160 (1965). 91. H. J. Enpcell and B. Ilschner, Z. Elektrochem. , 59, 716 (1955). -** 92. S. Thibault, J. M. Godot, J. Pagetti, and J. Talbot, Froc. Int. Cong. Metal. Corrosion, 4th, I969, 594 (1972), 93. Go. Okamoto and T. Shibata, Corrosion Sci., 10, 371 (1970). — 9 Z *. H. A. Laitenen and C. G. Enke, J. Electrochem. Soc, 107 , 773 (I960). 95. P. Popat and N. Hackerman, J. Phys. Chem., 65, 1201 (1961). 96. M. Stern, J. Electrochem. Soc, 104, 559 (1957).

PAGE 197

184 97. Go. Okamoto and T. Shibata, Proc. Int. Congr. Metal. Corrosion, 4th, 1969, 620 (1972). 98. 0. Jantzen, Z. Angew. Fhysik, 18, 560 (1965). 99. R. V. r.'oshtev, Ber. Bunyen^es. Fhys . Chera., 72, 452 (1968). ~ 100. R. Lebet and A. Fiotrowski, Corrosion, 22, 257 (I966). 101. Z. Szklarska-Smialowska and M. Janik-Czachor, Froc. Int. Congr. Metal. Corrosion, 4th, 1969, 651 (1972). 102. I. L. Rozenfel'd and V. P. Maksimchuk, Z. Fhysik. Chem. (Leipzig), 21£, 25 (I960). 103. M. Frazak, V. Spanily and J. Tousek, Corros. Week Manifestation Eur. Fed. Corros., 4lst, 1968, 776 (1970). 104. H. W. Pickering and R. P. Frankenthal, J. Electrochem. Soc, 119, 1297 (1972). 105. A. A. Seys, M. J. Brabers, and A. A. VanHawte, Corrosion, 30, 47 (1974). 106. Ya. M. Kolotyrkin, J. Electrochem. Soc, 108, 209 (196I). 107. V. F. Maksimchuk and I. L. Rozenfel'd, Dokl. Akad. Nauk. SSSR,131» 354 (i960). 10^. G. V. Golovina, G. M. Florianovich, and Ya. M. Kolotyrkin, Soviet Electrochem., 1, 8 (1965). 109. P. Forchhammer and H. J. Engell, Corrosion Sci., 11, 49 (1971). ~ 110. W. J. Forenz and H. Fischer, Electrochim. Acta, 11, 1597 (1966). ~ 111. F. A. Fosey and R. F. Sympson, J. Electrochem.. Soc, 109, 716 (1962). 112. Th. Neumann and H. S. Panesar, J. Electrochem. Soc, 110, 628 (1963). 113. R. Olivier, Thesis, Leiden (1955). 114. M. Pourbaix, "Atlas of Electrochemical Equilibria in Aqueous Solutions," pp. 256-261, Pergamon Press, New York (I.966).

PAGE 198

185 115B. E. Wilde and J. S. Arm! jo, Corrosion, 24, 393 (1968). 116. N. A. Bozin and M. Fi. Kurtepov, Zh. Fiz. Khim., 85, 152 (1961). 117. I. K. Burtseva, G. F. Plotnikova, and A. I. Krasil ' shchikov , Zh. Fiz. Khim., 3_8, 1956 (1964). 118. F. Zainin and F. B. Ives, J. Electrochern. Soc, 121, 1141 (1974). 119. M. Frazak, V. Frazak, and V. L. Cihal, Z. Electrochern., £2, 739 (1958). 120. J. deBecdelievre, A. M. deBecdelievre, and C. Courty, C. R. Acad. Sci., Ser. C, 2_5_2, 1333 (1964). 121. I. Epelboin, Bull, Soc. Chim. France, 475 (1964). 122. D. A. Shock, J. D. Sudbury, and 0. F. Riggs, Jr., Froc. Int. Congr. Fetal. Corrosion, 1st, I96I, 336 (1961). 123. F. Janik-Czachor and Z. Szklarska-Smialowska, Corrosion Sci., 8, 215 (1968). 124. ':/. A. Mueller, Corrosion, 18, 73t (1962). 125. J. O'M. Bockris, D. Drazic, and A. R. Despic, Electrochim. Acta, 4, 325 (I96I). 126. E. J. Kelly, J. Electrochern. Soc, 112, 124 (1965). 127. K. E. Heusler, Z. Elektrochern . , 62, 582 (1958). 12S. f. Filbert, Y. Miyoshi, G. Eichkorn, and W. J. Forenz, J. Electrochern. Soc, 118, 1919 (I97I). 129. A. J. Arvia and J. J. Podesta, Corrosion Sci., 8, 203 (1968). 1-30. F. Felloni, Corrosion Sci., 8, 133 (1968). 131. E. FcCafferty and N. Hackerman, J. Electrochern. Soc, 119 . 999 (1972). 132. N. A. Darwish, F. Hilbert, W. J. Lorenz, and H. Rosswag, Electrochim. Acta, 18, 421 (1973). 133. W. J. Forenz, Corrosion Sci., j5, 121 (I965).

PAGE 199

186 13 ; +. 1". Schorr and J. Yahalom, Corrosion Sci., 12 , 867 (1972). 135. W. Schwenk and H. E. Buhler, Corrosion Sci., 3, 261 (1963). 136. I. L. Rozenfel'd and V. P. Maksimchuk, Zh. Fiz. Khim., 35, 1832 (1961). 137' T. Suzuki, M. Yamabe, and Y. Kitamura, Corrosion, 29 , 18 (1973).

PAGE 200

BIOGRAPHY M. Elain Curley was "born in Dorchester, Massachusetts, on June 3» 19^« She attended Saint Gregory's Grammar and High Schools, graduating in June of 1962. In June, 1966 she received her Bachelor of Arts in Chemistry Degree from Regis College, Weston, Massachusetts. She entered graduate study at the University of Florida in September, I966. While there, she was the holder of Teaching and Research Assistanceships and the recipient of two Graduate School Fellowships and the Dupont Teaching Award. She married John A. Fiorino in February, 1972. She re-enrolled at the University of Florida in March, 1975, to complete the requirements for her graduate degree. 187

PAGE 201

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ^J£J frl, j^U^f^ erhard M, Schmid , Chairman associate Professor of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Ellis D. Verink, Jr. Professor of Materials/ 3ci and Engineering I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. 'It** , n. ^w>James D. Winefordj/er professor of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Wallace S. Brey Professor of Chemistry tf

PAGE 202

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. K. P. U Assistant Professor of Chemistry This dissertation was submitted to the Graduate Faculty of the Department of Chemistry in the College of Arts and Sciences and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. June, 19?5 Dean, Graduate School

PAGE 203

UNIVERSITY OF FLORIDA 3 1262 08666 276 3