•

·'

EVALUATION OF THEORIES

FOR THE INITIAL STAGES

OF PITTING CORROSION

by .

Pierre G.R. Zaya, Ing. A&M, M.Sc.(Eng.), P;Eng.

j'

A Thesis

Submitted to the School of Graduate &tudies

in Partial Fulfilment of the Requirements

for the Degree of

Doctor of Philosophy

McMaster University

©. October 1984

.' .~

EVALUATION OF THEORIES lOR PIT7ING CORROSION

. ./,

:

/

DOCTOR OF PHILOSOPHY (1984)

(Metallurgy)

McMASTER U~IVERSITY

Hamilton, Ontario

TITLE: EVALUATION OF THEORIES FOR THE INITIAL STAGES OF

PITTING CORROSION

AUTHOR: Pierre Gabriel Robert Zaya

Ing~nieur Arts et H~tiers (Paris, France)

M.S~.(Eng.), (L~ndon, England)

SUPERVISOR: Professo~ M.B. Ives

NUMBER OF PAGES: xi, 217.

~

(Ii)

"~.

I

)

o

ABSTRACT

It is assumed in this study that the protection

imparted by passive laye~ is dependent upon the balance

between the breakdown of the protective film and the healing

of the produced flaws. Therefore, the occurrence of a pit

(

must coincide with a change in the local conditions in a

flaw which did ~ot heal.

Using this hypothesis, five proposed ~heories for-pitting corrosion are compared using data published in.the

literature. The processes assumed to occur during pit, <,

nucleation are simulated. The solution compositions and

potentials 'from reported experimental detertnations of

pitting potentials are used as ~tiai conditi~ns. The

po~ential and concentrations at the base of the flaw are

computed by using mass transport equations.

The values obtained are then used to compare.the

five models. For.each theory, an expression depending on\

potential and concent~ations at ~he metal surface was iden-

tified, so that where a theory is correct, thii expression

is smaller (or la'rger)' than a parameter., . Thill parameter,

should be constant, and independent of the particular set of

experimental conditions (e.g. pH, chloride concentration in

(11 i)

the bulk solution), as all the calculations are m~de f9r

conditions corresponding to the pitting potential, i.e.

/

conditions borderlining between pit~l~g and passivity.-

The particular Case of iron" in borate buffers is

considered here.' The computed results indicate that the

most probable mechanism controlling pitting is the adsorp- -. ~

tion of the chloride ions at the metal surface. The repas-

sivation by precipitation of ferrous hydroxide is always

thermodynamically fav9ured, but apparently slow.

(iv)

•

\

ACKNOWLEDGEMENTS"

I.wish to thank the following people for their help

with this work:

Dr. M.B. Ives, my-supervisor, for his support and

understanding during the various stages of the project and

for his generosity in facilitating encounters with

specialists in the field.

Dr. J. Kirkaldy and Dr. W.K. Lu" the members of my Ph.D.

Supervisory Committee, for their encouragement, critiques

and f~cusing suggestions.

The staff of the McMaster Computer Centre,

,.

for their

help in deciphering the "messages· of The Machine and for

their suggestions which facilitated and expedited its

operation.

Drs. R.C. Alkire, M. Janik-Czachor. B. MacDougall and

H.H. Strehblow who. at various times during the development,

of this work, hsve patiently listened to my ideas and came

•forward with valuable suggestions and counter-arguments.

This thesis would have been very different without their

help.

(v)

Angela Bull and Gail MacLean, for their skill in typing

and their patience with the countless revisions.

And Phyllis Gordan and Claire ~o~ their love and,

,.

\.

understanding during all those evenings .

.'.

., ,;

(vi)

,

r/

•

'.

•

.t.

•

. .

"

,

..

•

We dance around in a ring and suppose,

But the Secret sits in the midd~e and kpOW5 •

I\

I(vii )

-

,

, .

..).

TABLE OF CONTENTS

I INTRODUCTION

II CHARACTERISTICS AND THEORIES OF PITTING CORROSION

Stages of Pit Development

Characteristics of Pitting Corrosion

Defects in th~'Passive Film

•2-4 Dissoiution of th~Passive Layer

. I2-5 Film Breaking

2-6 Breakdown and Repair

2-7 Competitive Adsorption

f- 8

2-9

Ion Penetration

Local Changes in the Environment0

2-9-1 'Critical Concentration..2-9-2 Localized Acidification

2-9-3 Resistance Polarization

2-10 Salt Film ..,.....

2-11 Cmncluding Remarks

•

CORROSION

techniques3-·2 Su'r£ace

III" METHODS OF INVEST~IO/T 0";31 PIT.TING). ~'~

3-1 Electrochemic 1 Methods

(viii) '..

•

)

3-3 Probabilis~ic Approach

3-3-1 Time Distrib~tion (stochastic approach)

3-3-2 Spatial Distr~bution

3-4 Mass~;ansfer Cal~ula~ions

"

IV MECHANISMS OF IRON DISSOLUTION

.4-1 . Mechanisms Involving Only OH-

4-2 Mechanisms in the Presence ~f Cl-

4-3 Mechanisms on Scratched Electrodes.

4-4 Mechanisms Relevant to Pitting ~o~rosiqn• -.

t:

V PROPOSED MOD~L FOR PIT INITIATION

5-1 Definition.of the~pr~blem

5-2 Chemical Syste~.

5-3 Transport Equations

5-4 Boundary Conditions. .•

5-5 . comptitat~ro~ of ctj:'yent-. " .':..

...,'.

",

.-"">. ~_..., - .

.... ~...-:.- ...'~~ ..

•5~6 Charge of the Ele~tricai Double Layer

5-7; Numeri~al Values of the Parameters

5-~-1 Anodic Dissolution of Iron

5-7-2 Other Parameters

5-8 Numerical, Solution "f the Eqcatlons

~f# ••

•

: 5-9 Possible Criteria for Pitting•

5-9-1 Repassivation

5-9~2' Salt Film Precipitation

..(ix)

..

..

I .

-'1J

',. 5-9-3 Adsorption

5-9~4 Critical Chloride Concentration

;,"'

I"

. VI' RESULTS

6-1 General Results

6-2 Setisitivity of the Results to ~6i r~pbt Values.' :

6-3 Precision

6-4 Stability . ',j

6-5 Testing of the Program." ....'c,"

6-6" Pi.tt~ng. in N110H-NaCI Solutions

6-7 Pitting in NaOH-NaCI-B(OH)3 Solutions

6-7-2

6-7-1 Sources of Data

:C'\Analysis o~ the Res~lts...............' ~.,

VII DISCUSSION; 0

7-1 Ideal Solution

7-2 One Dimensional Diffusion

" " ,

7-3 Inert Film

7-4 No ~onvection~:.

' .. -,.,,~, .1

. -::- ,. .,.--

r<,

No ·Precipit.ation

Supporting Electrolyte

7-5

7-6

~

()..."'.

7-8 Steady State

.' 7-7 Chemical Reactions

. , .....,,- .~. "\ ;

.J

7-9 Breakdown and Repair'; .

, '

7-10 Tem~erature.

(x)

.,, ,

J

,(~

. ..,'7-12 Cathodic Reaction

7-11 Fixed.G~ometry

..0{ -~.•

7-13 Comparison of Criteria

VIII CONCLUSIONS

IX FUTURE ,WORK \ ,

X REFERENCES

-APPENDIX A: DESCRIPTION OF THE COMPUTER PROGRAM

A-1 Program MAIN

A-2 Sub'routine DRIVER

A-3 Subroutine CALC

A-4 SUbroutin~IN

A-S Subroutine DERIV'

A-6 Subrou'tine INTEG

A-7 Subroutine CONC

A-a Subrou'tine DCDT

, "

I • APPENDIX B: LIST OF SYMBOLS .' • \

"

,."

, . (xi )

•

. CHAPTER I

- ,IliTRODUCTION

It has been estimated that the cost of corrosion in

the United States is approximately 70 billion dollars a

year, ot4.2% of the Gross National Product (Payer, Boyd,

Dippold and Fisher, 1980). Comparable percentages of the

GNP have been obtained in other countries where~evaluations

have been undertaken (Behre~s, 1975i Hoar, 1971~ Rene and)

Uhlig , ..1974). Pitting corrosion is certaiqly responsibleI'

for a re~pectable fraction of this cost, as judged by its•

impact on certain sectors, such as the pulp and paper indus-

try'. One chemical manufacturer found that over a four-year

period, pitting corrosion was respons~ble for about 8% of,all failures of the metallic piping .and equipment (Collins

and Monack, 1973).

Pitting can be defined as a form of corrosion "in

which only small areas of the metal surface are attacked

whilst th'e remainder is largely unaffected" (Shreir, 1976a).

This can occur· on many common· metals such as iron_, nickel,,titanium, aluminum, zin~, an·d· t.heir alloys'. The danger of

this type of attack is that it is impossible to monitor Its

I

2

progress by visual inspection until the metallic part has

been completely perforated.

The s~udy of pitting cort~iiori has received great

iapetus from 'the development of commercial potentiostats and

their systematic use in experimental electrochemistry since,

the early .1960' s. (Macdonald, 1977). These instruments, by

controlling the po~ential between the sample under study and

a reference electrode in the solution of interest, can

rapidly lndicate some of the conditions necessary for the

occurrence of pitting corrosion. Numerous experimental

results have been obtained: Shreir (1976b) gives a list of

269 articles published on the subject between 1960 and

1974.

The present work is motivated by a desire to gain

theoretical insights by using the already accumulated exper-

imental data. Starting with the assumption that pitting

proceeds through processes of breakdown. and repair, the

elemental physico-chemical processes 'are'simulated, using

contemp9rary numerical'techniques. These permit' the solu-

tion of the relevant differen~ial equations with a minimum

•of restrictive assumptions.

1By simulating the behaviour of

a passive metal. during a poten'tiostatic experiment, it is

possible to use the experimental pitting potentials and the~ ,

•solutions in which they were determined to compute local

electrochemical conditions. These results in turn allow the

quantitative evaluation of five different theories

As iron is the

3

attempting to explain pitting corrosion.

After a short description of the characteristics of

corrosion pits, the different theoretical models which have

been proposed to elucidate this phenomenon are discussed.

The mai~ approaches of. current research are then reviewed in

order ·to situate jJ'e methodology used here.

metal which has been selected for the application of the

proposed methodology, the concluding section of the litera-

ture review discusses the most important theories of iron

dissolution •

..