Skip to main content

Full text of "NASA Technical Reports Server (NTRS) 19920010065: Development of an Accelerated Test Method for the Determination of Susceptibility to Atmospheric Corrosion"

See other formats


1991 NASA/ASEE SUMMER FACULTY FELLOWSHIP PROGRAM 


JOHN F. KENNEDY SPACE CENTER 
UNIVERSITY OF CENTRAL FLORIDA 


DEVELOPMENT OF AN ACCELERATED TEST METHOD FOR THE DETERMINATION 
OF SUSCEPTIBILITY TO ATMOSPHERIC CORROSION 


PREPARED BY: 

ACADEMIC RANK: 

UNIVERSITY AND DEPARTMENT: 

NASA/KSC 

DIVISION: 

BRANCH: 

NASA COLLEAGUE: 

DATE: 

CONTRACT NUMBER: 


John R. Ambrose, Ph.D., P.E. 

Associate Professor 

University of Florida 
Department of Materials Science 
and Engineering 

Materials Science Laboratory 

Failure Analysis and 
Materials Evaluation 

Coleman J. Bryan 

July 26, 1991 

University of Central Florida 
NASA-NGT-60002 Supplement: 6 


1 



ACKNOWLEDGEMENTS 


I would like to express my appreciation for being selected to 
participate in the 1991 NASA/ASEE Summer Faculty Fellowship 
program, to Ray Hostler [University of Central Florida] and Mark 
Beymer [NASA] for their capable administration Q f the program, 
and to my NASA colleagues Cole Bryan and Rupert Lee for their 
assistance and encouragement. Special thanks must go to Jim Jones 
for his assistance in expediting the various chemical analyses 
which 1 requested and to Helein Hitchcock, Hae Soo Kim, Stan 
Young and Bill Carman for their capable technical assistance* And 
to Lee Underhill, without whose assistance in finding the various 
apparatus with which to set up my experiments i would never have 
even begun this project, I offer a nomination for the "Golden 
Scrounge" award* 

And most of all, my gratitude to the Materials Testing 
Laboratory staff - fo Steve Mcpanel , whose progress from a former 
student to a capable failure analyst makes this teacher proud, to 
Scott Murray for his patience and cooperation in allowing this 
curious child of 51 years to tag along, seeing things again for 
the first time , and most especially to Peter Marciniak, whose 
instruction and encouragement allowed me to find satisfaction in 
doing things undone for some 25 years . May Scott always find one 
last jelly doughnut each morning and Peter his film for "just one 
more" . 

...and to KSC and its surrounding Merritt Island Wildlife 
Refuge for being what and where they are* 1 can sincerely say 
that no summer has meant more to me or will mean more as time 
goes by . Even though endless summers must always end, thanks for 
the memories ... 


2 



ABSTRACT 


The theoretical rationale is presented for utilization of a 
repetitive cyclic current reversal vol tammetric technique for 
characterization of localized corrosion processes, including 
atmospheric corrosion. Applicability of this proposed 
experimental protocol is applied to characterization of 
susceptibility to crevice and pitting corrosion, atmospheric 
corrosion and stress corrosion cracking. Criteria upon which 
relative susceptibility is based have been determined and tested 
using two iron-based alloys commonly in use at NASA/KSC - A36, a 
low carbon steel and 4130, a low alloy steel. 

Practicality of the procedure has been demonstrated by 
measuring changes in anodic polarization behavior during high 
frequency current reversal cycles of 25 cycles per second with 
1 mA/cm* current density amplitude in solutions containing C'l’*. 
The results demonstrated that, due to excessive polarization 
which affects conductivity of barrier corrosion product layers, 
A36 was less resistant to atmospheric corrosion than its 4130 
counterpart - behavior which has also been demonstrated during 
exposure tests. 


3 


SUMMARY 


Based on an analysis of factors which are known to contribute 
to the overall corrosion process, gal vanostatic electrochemical 
procedures show greater promise for creating an environment 
during accelerated testing which more closely simulates natural 
environments than do conventional potentiostatic methods. 
Similarly, since both anodic and cathodic reactions occur 
simultaneously on the surfaces of freely corroding metals, it is 
necessary to periodically reverse the direction of current flow. 
On the basis of these modelling simulation requirements, the 
Repetitive Current Reversal Voltammetry technique has been 
developed . 

Using a galvanostat to supply current, and a storage 
oscilloscope to measure potential transients as a function of 
time, a methodology and associated predictive criteria have been 
generated using two representative alloys found in common use at 
NASA\KSC - A36 steel and 4130 low alloy steel. By trial and 
error, the optimum combination of impressed current amplitude and 
current reversal cycle frequency has been found which will 
differentiate environmental degradation behavior for these two 
metals - +/- 2 milliamperes current amplitude with a 25 cycles 
per second current reversal frequency. 

Results to date suggest that susceptibility to various 
localized corrosion forms of attack may be associated with "over- 
polarization", that is, too high an electrical resistivity for 
insoluble corrosion product barrier layers which form at the 
metal/environment interface. In the presence of aggressive ions 
such as the chloride ion, these high potential drops or fields 
across the barrier layers "draws M these aggressive anions into 
the film, causes its decrease in resistivity and eventually leads 
to the inability of passivity being maintained. 

This report summarizes many of the insights, opinions and 
perspectives of the author which may prove useful to others as 
they contemplate developing experimental procedures for the study 
of environmental degradation of materials. 


4 


TABLE of CONTENTS 


1.0 INTRODUCTION 

1.1 EXPERIMENTAL BASIS FOR EXPERIMENTAL PROCEDURE 

1.1.1 CRITERIA FOR CORROSION SUSCEPTIBILITY 

1.1.2 EXAMPLES OF CRITERIA APPLICATION 

1.1. 2.1 ATMOSPHERIC CORROSION 

1.1. 2. 2 CREVICE CORROSION 

1.2 PROPOSED EXPERIMENTAL PROTOCOL 

2.0 EXPERIMENTAL - MATERIALS AND EQUIPMENT 

3.0 EXPERIMENTAL RESULTS AND DISCUSSION 

3.1 LOW FREQUENCY CURRENT REVERSAL CYCLING 

3.2 HIGH FREQUENCY CURRENT REVERSAL CYCLING 

3.2.1 POLARIZATION IN STOCK SOLUTION 

3.2.2 EFFECT OF CHLORIDE ADDITIONS 

4.0 CONCLUSIONS 

5.0 RECOMMENDATIONS FOR FUTURE WORK 


5 


Figure 

1 

2 

3 

4 

5 

6 

7 

8 

9 

10 

11 

12 

13 

14 

15 

16 
17 


LIST OF ILLUSTRATIONS 
Title 

Schematic showing physical sense of current . 

Schematic representation of the Separation of Siicide 
and cathode in an electrochemical cell utilizing 
potentiostatic control . 

Schematic representation of limiting cases in 
polarization behavior - metal dissolution v . 
ihSolUble corrosion product . 

Schematic representation of how increased solute 
concentrations can be uSed to simulate evaporation of 
condensed moisture 1 ayeis . 

Schematic representation of how a crevice leads to 
1 ocal ized breakdown of passivity . 

Photograph of machined samples assembled into the bi- 
specimeh electrode configuration. 

Photograph of electrochemical cell used. 

Schematic of experimental procedure With reSpect to 
the RcRv control variables . 

Conventional Polarization Plot [ I v . V] showing the 
effects of oxygen depletion on corrosion kinetics . 

RCRV plot for A36 alloy in air-saturated stock. 

Polarization kinetics fof A36 alloy in the stock 
solution; experimental parameters shown oh graph . 

Replot of Figure 11 , plotting -Lh(voits) v . time . 

A comparison of polarization kinetics for the A36 and 
4130 alloy steels. 

A comparison of polarization kinetics for A36 and 
4i30 Steels With chloride ions present . 

A comparison of polarization behavior for A36 with, 
and without, chloride ions. 

Polarization behavior of 4130 with and without 
chloride ions. 

Effect of 5 minutes exposure to chloride ions on the 
polarization behavior of A36 and 4130 alloys. 


6 



1.0 INTRODUCTION 

The vast majority of electrochemical test methods used in 
corrosion science experimentation derive from potentiostatic 
polarization techniques - i.e. a measurement of current required 
to maintain the potential of a metal at some reference value. As 
the name implies, the electronic instrumentation fixes potential 
by regulation of current. As a consequence of this functionality, 
corrosion rates [current] vary with time as the externally 
maintained potential remains fixed [potentiostatic] or is varied 
[potentiodynamic sweep] during the course of the experiment. 

A number of "accelerated" corrosion test procedures are based 
upon potentiostatic polarization procedures - linear current 
versus applied potential relationships or "polarization curves", 
low amplitude linear polarization for determination of 
"polarization resistance", low amplitude cyclic polarization or 
"AC impedance" and linear cyclic voltammetry or "LCV" to name but 
a few. Although a great deal of information has been obtained 
through such experimentation, some of it even useful, attempts to 
correlate long term corrosion behavior with the results of such 
procedures have often been fraught with inconsistencies, 
irreproducibi 1 ities and difficulties with interpretation. The 
problem, as I see it, is that the basis for these techniques - 
potential or polarization control - is not a suitable or accurate 
model for the way metals behave in "real life" situations. The 
true basis is, in fact, just the opposite. 

When a metallic material is exposed to a corrosive 
environment, there develops, rather quickly, an electrochemical 
potential across the metal /envi ronment interface which serves to 
drive those chemical reactions which contribute to that 
particular electrochemical potential. Corrosion processes consist 
of one or more anodic or oxidation reactions which, in a physical 
sense, carry current out from the surface of the metal and into 
the environment. Anodic current is balanced by a cathodic current 
of equal amplitude which flows into the metal from the 
environment, carried by one or more reduction reactions. The 
respective current flows produce, due to a electrical resistance 
which opposes the flow of current [polarization resistance], a 
shift in potential or polarization equal to the product of the 
current flow and the resistance through which that current flows. 
The resulting so-called "mixed" or corrosion potential of the 
material is positioned between the reversible potentials for the 
various electrochemical processes, the displacement being 
dependent upon the polarization resistance for the various 
contributing electrochemical reactions. Mathematical models of 
electrochemical kinetics [e.g. Butler-Volmer equation] have been 
used to characterize the relationship between current and applied 
potential, resulting in what have come to be known as "Evans 
Diagrams". A schematic representation of current flow across the 
metal /environment interface is shown in Figure 1. 



uniform anodic current density 
-J(A): M 0 =M +Z +z«- 


uniform cathodic current density 
-J(C): 0 2 +2H^)=40H" 


insoluble corrosion product 

uniform corrosion current density 

- M^-zOH^MfOH)^ 

- electrical resistance 


Figure 1 - Schematic showing physical sense of current 
direction - anodic current leaving metal and cathodic current 
entering it. 

The point to be made is simply this: Since the corrosion 
potential of the material is determined by polarization, 
specifically polarizibility of the material as influenced by the 
contributing electrochemical reactions, then it is the flow of 
current which controls potential, not the other way around. 

Fixing or controlling the potential of the material will cause 
current flow to conform to the degree of polarization - if there 
is any change in resistance to current flow which occurs as a 
natural consequence of the chemical reactions taking place during 
service exposure, then current will change accordingly. One such 
occurrence, the evolution of insoluble reaction products which 
block the flow of current across the interface, is also shown in 
Figure 1. Under "natural", free corroding conditions, it would 
seem that development of insoluble corrosion products with 
whatever inherent intrinsic electrical and ionic transport 
resistivity they possess would cause the corrosion potential to 
shift in one direction or another with respect to a film-free 
metal/electrolyte interface. Since many metal oxides and 
hydroxides [iron based, e.g.] provide greater resistance to the 
anodic current flow [ion transport control], the corrosion 




8 




potential under natural conditions tends to shift toward more 
positive values. When experimental simulation of corrosion 
processes are under potentiostatic control, the potential, of 
course, is fixed. There are two consequences of this control. 

One, the corrosion current would not be expected to be the same 
as under natural control conditions, resulting in lack of 
correspondence with long term corrosion rates. Secondly, since 
the potential of the metal determines what electrochemical 
reactions can occur, differences in potential may result in 
altogether different reaction products being generated. Since it 
is the reactions products which determine the nature of insoluble 
reaction products, and it is the insoluble reaction products 
which affect both degree of polarization and amplitude of the 
corrosion current, the end result is most likely a continual 
divergence of experimental corrosion rates from the real ones. 

A second negative characteristic associated with 
potentiostatic control of electrochemical processes derives from 
the configuration of the system used to control the process being 
studied. Under anodic polarization - when the potential of the 
material is driven to higher [more positive] values than its 
corrosion or steady-state value - the current which produces this 
polarization leaves the material [working electrode], flows 
through the electrolyte and enters a remote auxiliary [counter] 
electrode. In a freely corroding metal, on the other hand, 
current leaves and 
enters the same exposed 
surface, albeit not 
necessarily at the same 
point on that surface. 

What this means is that 
cathodic and anodic 
reactions do not occur 
concurrently on the 
same metal surface - 
more significantly, the 
reaction products of 
these electrochemical 
processes are not 
produced at a common 
surface. Since most 
insoluble corrosion 
products which form in 
aqueous environments 
contain metal cations, 
hydroxide ions and 


ting 

anode from cathode in the electrochemical cell changes the 
localized chemistry associated with the natural or free corrosion 
process [Figure 2]. 


other constituen 
anions [SO, , Cl" 

r*r\ "2 _ i. _ T - 1 



9 




1 . 1 THEORETICAL BASIS for PROPOSED EXPERIMENTAL 
PROCEDURE 

Measurement of degree of polarization which result from 
impressed current is riot a new technique, having been used for 
some time as an analytical chemistry technique - 
chronopotentiometry , Stripping voltammetry, etc. It has not been 
used to any great extent in corrosion science applications. It 
would seem, however , to be most appropriate in measuring the 
behavior of a material in response to flow of current across the 
metal /environment interface . By impressing a constant 
[gal vanostatic] current between an inert electrode [platinum 
counter] and the material being characterized [working 
electrode] , the potential change or polarization can be measured 
as a function both of time and of the amplitude of the impressed 
current . In order to simul ate "natural " conditions , the current 
direction should be regularly reversed in order to develop 
concentrations of both kinds of reaction products at the metal 
interface - anodic and cathodic . Thus evolves the name of the 
technique - "CYCLIC CURRENT REVERSAL VOLTAMMETRY" . The 
experimental variable to be measured will be the rate of change 
in material potential , or "polarization rate". 

1.1.1 CRITERIA for CORROSION SUSCEPTIBILITY 
EVALUATION 

Polarization in response to current flow can be of three 
types , individually or in combination : 

ii a potential drop across an ohmic resistance . This 

polarization is characterised by a V-IR response , and is 
Virtually time independent - i.e. instantaneous 
polarization with application of current . Capacitance or 
interfacial charging processes are included in this 
category as is the voltage drop across the electrolyte 
between the working and counter electrodes . 
ii . polarization due to the resistance to charge transfer 
acrOSs the electrified interface - i.e. so-called 
H Tafel" overvoltage . This kind of polarization is 
characterized by a 1 ogarithmic dependence upon current 
flow - the "Taf el Equation": 



10 



where n ; = degree of polarization produced by current I 
and B- = charge transfer resistance, 
iii. the potential drop across an insoluble reaction product 
or film which forms at the metal/electrolyte interface. 
The -degree of polarization is a function of the 
resistivity of the reaction product, the polarization 
rate is a function of the nucl eation/growth kinetics of 
the deposition process. It is this polarization process 
with which we will be most interested. 

A schematic representation of the polarization extremes - 
polarization resistance during active metal dissolution versus IR 
resistance across an insoluble corrosion product - is shown in 
Figure 3. 



Under conditions of 
repetitive current 
reversal , any change in 
either degree or in 
rate of polarization 
signifies changes in 
the one of the three 
processes enumerated 
above. Of the three, 
only the third should 
provide any significant 
contribution. Thus, by 
evaluating such 
changes, we should be 
able to establish 
criteria for evaluating 
the environmental 
stability of a 
particular material in 
a given environment. 

As "protective" 
films grow on bare or 
air-formed film covered 

metal substrates, there should be a regular increase in degree of 
polarization with each consecutive anodic cycle. Furthermore, the 
degree of polarization should progressively decrease as well, if 
the protective film is becoming more and more protective. Any 
change in this trend will be interpreted as an indication of 
development of instability in the system - a loss in ability of 
the system to resist the corrosive actions of the environment. 


limiting cases in polarization 
behavior - metal dissolution v, 
insoluble corrosion product. 


X.1.2 EXAMPLES of HOW CRITERIA. ARE APPLIED 


NASA/KSC is interested in predictive capability with respect 
to a large number of environmental degradation of materials 
problems. Specifically, accelerated test methods for predicting 
long term resistance of coated and uncoated metals to atmospheric 


ll 




corrosion are needed for rapid, early screening of candidate 
materials. Susceptibility to crevice corrosion, pitting and 
stress corrosion cracking - all localized forms of corrosion 
attack •* is another area to which this approach can be applied in 
developing accelerated test methodology. 

X. 1.2.X ATMOSPHER I C CORROSION 

Atmospheric corrosion can occur in two forms - at elevated 
temperatures by direct reaction between material and corrosive 
gasses, and through electrochemical means by interaction between 
material and condensed aqueous liquid films. KSC is considered 
one of the most, if not the most, severe locations for 
atmospheric corrosion in the world. This is due to its proximity 
to the ocean and to the corrosive nature of combustion products 
from Space Shuttle solid fuel rocket boosters [SRBs] . 

During atmospheric corrosion under condensed moisture films, 
the actual corrosion process naturally takes place only when the 
film is present upon the surface - long term exposure testing as, 
for instance, is performed at NASA-KSC's seaside test site, 
provides an integrated measure of atmospheric corrosion behavior 
under those conditions where condensation is present. Ocean 
spray, rain, condensation of dew, etc. provide the natural 
environment, The process of Condensation and evaporation actually 
increases the severity of the attack over what it would be under 
constant exposure conditions - and, as such, provides a clue as 
to an appropriate acceleration procedure. 

As Figure 4 demonstrates, although initial, as-condensed 
electrolyte concentrations may fall below a critical level to 
initiate corrosive attack [<5xlO' J M Cl’ 1 for carbon Steels, e . g . ] , 
a critical level can be achieved during evaporation. During test 
site exposures, this condition is only reached during some 
fraction of the total exposure times. Rainwater and dew 
condensation tend to return the process back to square One. 

It is therefore proposed that one possible acceleration test 
would be to increase chloride [or any Other aggressive species] 
by aliquot additions to a stock solution while the metal 
undergoes repetitive current reversal voltammetry. The time or 
number of cycles to a point where polarization behavior is 
affected would be a criteria for comparison of material 
susceptibi 1 ity . 


12 



Assume some critical [minimum] 
concentration of chemical species X 
causes resistivity of film to decrease. 

If C(X) In environment Is 0.5C(*), 
no problem, but.. 


if C (X) increased, or 



if solvent Is evaporated, 

we have a problem - 
film begins to thin 



Figure 4 - Schematic representation of how increased solute 
concentrations can be used to simulate evaporation of 
condensed moisture layers. 


1.1. 2. 2 CREVICE CORROSION 

Metals with the capacity to passivate themselves with barrier 
surface films [aluminum, stainless steels, nickel and titanium 
alloys] require a finite hydroxide ion concentration [from the 
cathodic reduction of oxygen] to retain their passivity. 

Occluding any area of a passive metal surface - with a crevice, 
for example, restricts oxygen transport to the creviced area, and 
leads to breakdown of passivity and active corrosion [Figure 5]. 
Degree of susceptibility of a material to this form of localized 
corrosion might also be rapidly characterized by an accelerated 
test combining Repetitive Current Reversal Voltammetry [RCRV] 
with oxygen depletion from the electrolyte. 

It is proposed that an accelerated method for determination of 
susceptibility to crevice corrosion would involve the effect of 
oxygen depletion on RCRV while deaerating the electrolyte. 


13 



1. Uniform "general" corrosion 
- "passive" film on surface 


t h mm 




2. Add crevice - 

Impede replenishment of 
cathodically reducible species - 
decrease cathodic current 
Inside the crevice, 
some decrease In anodic current 
outside crevice. 


3. With lower [OH ], passive film 1 

under crevice dissolves - WBBS&Bk 
anodic current Inside crevice Increases - 
anodic current outside crevice decreases. 


Figure 5 - Schematic representation of how a crevice leads to 
localized breakdown of passivity. 


1 . 2 PROPOSED EXPERIMENTAL PROTOCOL. 

a. Reversal frequency and current amplitude experimental 
operational variables. 

b. Determination of the effect of oxygen depletion on RCRV 
kinetics . 

c. Determination of the effect of increased chloride ion 
concentration on RCRV kinetics. 

d. Comparison of RCRV kinetics for two different materials for 
the purposes of distinguishing behavior - procedural 
feasibility determination. 

e. Determination of effects of other atmospheric contaminants 
on RCRV kinetics. 

f. Development of a "standard" test medium composition and 
methodology for characterization of atmospheric corrosion 
resistance . 


14 




2.0 EXPERIMENTAL 


MATERIALS and EQUIPMENT 


Feasibility studies were performed using two relatively common 
ferrous-based alloys in use at NASA/KSC - A36 [basically a carbon 
steel] and 4130 [a low alloy steel]. Their compositions are given 
in Table I . 

Based upon these 
compositions, the A36 
alloy would be expected 
to show the lesser 
degree of resistance 
toward atmospheric 
corrosion attack. 

Specimens of the two 
materials were machined 
into rectangular 
configurations, then 
assembled into the 
electrode configuration 
shown in Figure 6. This 
particular assembly 
configuration was 

selected for a> number of reasons - immersion of the tip end of 
the assemoly beneath the electrolyte precludes the need for 
isolation coating and their inherent tendency of providing 
crevices for crevice corrosion attack; multiple material 


Table I - Elemental Compositions of Iron 
Alloys Selected for this Study. 


A36 : 0.29 C(max), 0.8-1.20 Mn, 

0 .20Cu 

4130: 0.28-0.33 C, 0.40-0.60 Mn, 
0.15-0.30 Si, 0.8-1.10 Cr, 
0.15-.25 Mo 



Figure 6 - Photograph of machined samples 
assembled into the bi-specimen electrode 
configuration. 


experimental runs can be made consecutively in the same 







electrolyte without concern for changing, cleaning and 
repositioning electrodes between experiments. 

The electrochemical cell used in these studies was designed 
and constructed prior to my arrival at KSC - it is shown in 
Figure 7. It features, in addition to conventional items 
L 

IfK 

V i r ; 


Figure 7 - Photograph of Electrochemical Cell used 
in these studies. 


[reference electrode well/luggin capillary, isolated counter 
electrode well, working electrode access and gas dispersion tube 
input], two separate vessels for addition of test solutions. 

Electrochemical measurements were obtained using a Tacussel 
BiPad potentiostat/gal vanostat coupled with a signal generator. 
The waveform selected for these experiments was a square wave 
function in which the input variables were current amplitude and 
cycle frequency [Figure 8]. 

Current amplitudes used here were as high as +/- 2 
milliamperes with reversal frequencies from 10 milliseconds to 10 
seconds. Polarization kinetics were measured using a Tektronics 
storage oscilloscope. Data was obtained by analysis of 
photographs taken of the retained screen image. 

The standard or stock solution selected was 0 . 1M KH^PO^ 
adjusted to pH=7 with 0 . 1M NaOH. The basis for this selection was 
that phosphate solutions represent in innocuous environment 
insofar as steels are concerned. 0 . 1M NaCl was added when effect 
of aggressive ion concentration was to be evaluated. Unless 
otherwise noted, all solutions used were air-saturated and 
stagnant . 

Prior to running any experiment, specimens were mechanically 
polished using 100 grit silicon carbide paper, rinsed with 
distilled water, then dried with absolute ethanol rinse and a 
blast of warm air before assembling the electrode and placing the 


16 



assembly in the 
electrochemical cell. 
Solution was added from 


one of the reservoirs 
to a level just above 
the bottom of the 
electrode assembly. 
Rest or corrosion 
potentials were 
recorded once they had 
reached some steady 
state value. 


Appfed Current, mAnps 

2 
1 
0 
*1 
-2 

0 



- 

cycte 

frequency 


' ourrert 
ampttude 

f 












i 


Figure 8 - Schematic of Experimental 
Procedure with respect to the RCRV 
control variables. 


3.0 EXPERIMENTAL RESULTS and DISCUSSION 


In Figure 9, for 
those who derive 
comfort from 
conventional 
polarization curves, 
appears a pair of 
polarization curves 
describing 
potentiostatic 
polarization behavior 
for A36 steel in air- 
saturated and deaerated 
[with N 2 ] KHjPOi 
solution [pH 7j. The 
only feature of 
interest is the graphic 
demonstration of the 
ability of the A36 
alloy to passivate in 
this solution, whether 
0 2 is present or not. 



Figure 9 - Conventional Polarization 
Plot [I v. V] showing the Effects of 
Oxygen Depletion on Corrosion Kinetics 
of A36 Alloy. 


Notice that the 
potentiostatic method 

requires a critical anodic current of better than 500 mA to 
achieve passivation at potentials around -600 mV v SCE. Also note 
that the critical current requirements are higher in the 


17 





deaerated solution. 

When the same experiment is performed using the low frequency 
RCRV method, with a +/- 10 second frequency and a +/- 1.25 mA 
current amplitude, passivation is produced by the 5th cycle with 
one four hundredth [1/400] of the anodic current required for the 
potentiostatic polarization method [Figure 10]. 



solution. 

These results, as contained within Figures 9 and 10, serve to 
endorse my argument criticizing the reliance on potentiostatic 
polarization procedures for prediction of corrosion behavior. 
Their use, particularly in determining , from Tafel slope 
measurements, corrosion current densities are misleading to say 
the least. 

RCRV measurements of the 4130 alloy in the stock solution 
using low frequency current reversals were similar to the results 
displayed in Figure 10, only the number of cycles to produce 
passivity were noticeably less - by the 2nd cycle. However, 
experimental results for both alloys in this stock solution were 
extremely irreproducabl e . The problem lies with the 
electrochemical history of the process. Length of time at open 


18 




circuit was of prime consideration - after passivation was 
achieved with either alloy, it became impossible to distinguish 
behavior between the two alloys. Specimen pretreatment prior to 
the experiments was also an influencing factor - surface 
preparation and finish would affect the results. 

Addition of solutions containing chloride ion resulted in 
inability of both alloys to achieve passivation - the film 
resistance appeared, however, slightly greater for the 4130 alloy 
than for the A36. These results also were difficult to reproduce. 
The point is, unless a definitive criteria can be developed which 
is independent of operator procedure, the technique will never be 
applicable as a standard test method. 

One problem with the electrochemical cell was that, with its 
present design, aliquots of solutions containing the aggressive 
species could not be rapidly mixed with the stock solution. This 
suggests a design change such that second solution aliquots be 
added through the gas dispersion valve circuit - in this way, 
solutions can be rapidly mixed during RCRV sequences. 

3.1 HIGH FREQUENCY CURRENT REVERSAL CYCLING 

Increasing the cycle frequency increases the stability of the 
system - at least with respect to reproducibility in measurement. 
In Figure 11 is plotted polarization kinetics for a single 
typical cycle when a 
±10 mA amplitude 
current is impressed 
for 10 mS cycles . A 
number of relevant 
observations are worth 
noting - When forward 
and reverse current 
amplitudes are equal 
[current balanced], the 
median potential read 
from a slow response 
high impedance 
voltmeter is remarkably 
constant. So constant, 
in fact, this procedure 
might be considered as 
a standard test method 
for measurement of 
corrosion potentials. 

Changes in current 
amplitude results in a 
shift in median 
potential to higher or lower values, respectively. 

When current amplitudes are unbalanced, the median potential 
shifts in the direction of the current imbalance. This procedure 
should also be studied further in that the rate of polarization 



A36 alloy in the stock solution; 
experimental parameters shown on graph. 


19 




shift appears to reflect the degree of control of the 
contributing anodic and cathodic electrochemical kinetics on the 
overall corrosion process. 



Replotting the data of Figure H using the negative natural 
logarithm of the cell voltage, produces the graph shown in Figure 
12. Only the anodic polarization portion of the overall cycle is 
shown. Although the plot is not a linear function of the plotting 
variables over the whole range, it fits the later portion of the 
cycle quite well . Neither an exponential or parabolic plot would 
do as well. Without drawing any conclusions as to the mechanistic 
implications to be derived from this, experimental results will 
be presented in this format, since it is easier to see variations 
when comparing different alloys and when chloride ion solutions 
are introduced into the stock solution. 


20 




3.2.1 POLARIZATION BEHAVIOR IN O.IM KHjPO,^ 


- In (Cell Voltage) [v. SCE] stock solution 



Elapsed Time, msec 


Figure 13 - A comparison of polarization kinetics for the A36 and 
4130 alloy steels in the stock solution [0.1 KHjPO^ , pH 7]. 

Current amplitude was ±2 mA, cycle reverse frequency was 120 mS. 

In Figure 13 are compared polarization kinetics for our two 
reference materials, the A36 carbon steel and the 4130 low alloy 
steel. The operational variables were selected by trial and error 
to produce the optimum conditions for producing the greatest 
difference in behavior between the two - 20 mS current reversal 
frequency with a 2 mA current amplitude. What is interesting to 
note is that the A36 alloy evidences a larger degree of 
polarization than does the 4130 steel. Remember, now, the 4130 is 
more resistant to atmospheric corrosion than the A36. 

Furthermore, the IR drop [film electrical resistance] is 
substantially greater for the A36 material. Intuitively, you 
might think that this is contradictory in terms of the known 
corrosion resistance of the low alloy steel. It must be 
emphasized that corrosion resistance lies not necessarily with 
electrical resistivity, but with passivity - with the resistance 


21 




to ion transport through the barrier layer . The proof of the 
pudding, so to speak, Will lie with ability of a barrier layer to 
retain its ionic resistivity in the presence of aggressive or 
corrosive chemical species. Whatever, we can say at this point 
that polarization behavior of the A36 alloy is distinguished by a 
larger electrical resistance which, in the presence of impressed 
anodic currents , results in larger [more positive] degrees of 
polarization - larger electrical fields or potential drops across 
the barrier corrosion product layer. 


- lri(C&lj Voftage) [V. 


Stock solution +0. 00626 M Cl 


2.200 


1. 

1. 

1. 
1.200 
1.000 


41 iC polarization 


j i. -A " 

et> 


. * A 

" 


. * .. & ' 



6 


i 




16 


i msec 


A comparison of polarization kinetics for A36 and 
in the stdck solution to ,whiOh has been sufficient 
to Bfitii the overall Cl -1 concentration to 6.25x10' 


lire 14 - 
4i30 Steels 
^hi bride 


Immediately upon addition bf chloride ions^ here producing a 
concentration in the stock solution of 6 . 25xlO’ 3 H; just above the 
minimum concentration required to produce pitting in low carbon 
steels , there are changes in polarization behavior between the 
two metals. These results are shown in Figure 14. There is little 
apparent difference between polarization kinetics for the two 
materials . However , there are differences in polarization degree . 
in order to better visualize what the differences are , the 


22 




data contained in Figures 13 and 14 are compared by replotting 
and comparing the polarization behavior of each alloy with, and 

without, chloride ions being present. 

Notice that there is 
little change in the 
kinetics of 
polarization, but a 
noticeable decrease in 
the IR drop across the 
film - the film has 
become more 

electrically conductive 
[Figure 15]. Similarly, 
when the behavior of 
the 4130 alloy is 
compared in stock and 
chloride containing 
solution, there is also 
a decrease in IR drop 
across the corrosion 
product, but the 
relative amount of the 
decrease is far less 
[Figure 16] . 


There is little other 
than speculation as to 
the interpretation of 
this behavior. The fact 
that the degree of 
polarization is so much 
greater for the A36 
alloy indicates that 
there is a greater 
driving force for 
migration of negatively 
charged anionic species 
into the film toward 
the metal/film 
interface. This, by 
itself, does not 
necessarily result in 
increased electronic 
conductance - unless, 
the chloride ion 
somehow leads to the 
within the corrosion 
product. Perhaps, just perhaps, chloride ions within the 
corrosion product lead to an increased concentration of negative 
charge carriers, increasing the n-type extrinsic semiconductivity 



production of electronic charge carriers 



behavior for A36 with, and without, 
chloride ions being present [from 
Figures 13 and 14]. 





of the barrier layer. 

Whatever the explanation, within a few minutes, the IR drop 
decrease trend continues, until, after 5 minutes, the A36 has 
completely lost its capacity for producing a electrically 
resistive barrier layer completely [Figure 17]. 


- fn(Cell Voltage) [v. SCE] 5 minutes after Cl addition 

2.200 

zooo 

1.8Q0 

1.600 

| . .. 

1.400 

• 

1.200 
1.000 
0.800 

2 4 6 8 10 12 14 16 18 

Elapsed Time, msec 


Figure 17 After 5 minutes of RCRV in the presence of chloride 
ion, A36 has lost its capacity for production of electrically 
resistive corrosion products while 4130 is hardly affected at 
all. 

Notice that the 4130 is hardly affected at all. 

Bear in mind that this behavior has been measured without the 
occurrence of noticeable corrosion attack on the A36 alloy - only 
the tendency has been measured. 



24 




4.0 CONCLUSIONS 

a. Gal vanostatic measurements represent a more natural 
approach to the modelling and development of accelerated 
test methods than do conventional potentiostatic methods. 

b. Periodic or cyclic current reversal tends to simulate 
concurrent anodic and cathodic reactions on a freely 
corroding metal surface better than do potentiostatic 
methods where anodic and cathodic processes occur on 
separate electrode surfaces. 

c. Development of highly resistive barrier layers which result 
in large degrees of polarization during anodic current 
cycles appear to be directly associated with susceptibility 
to atmospheric corrosion. 

d. Repetitive Current Reversal Voltammetry [RCRV] has been 
demonstrated to be a highly sensitive means with which to 
differentiate relative susceptibility to localized 
corrosion attack. 


25 



Experimentation must J?e continued with attention focussed 
on variations in polarization behavior as a function of 
solution composition. This would require modification to 
the existing electrochemical cell used in these studies . 

Development of a standard test solution and procedure to be 
applied to a larger list of candidate materials. Emphasis 
most be placed on establishing experimental methodology 
wbi ch is not limited by the experience or skill of the 
operator - ’’idiot proof". 

For scientific purposes , parallel studies of mechanistic 
implications of polarization kinetics should be 
accomplished . This will require "up-grading" pf the 
equipment used - a digital oscilloscope with computer 
interfacing would speed up the data acquisition and 
interpretation process . 

The relationship between polarization behavior and both 
current amplitude and frequency needs to be established in 
more detai 1 than has been done for these feasibility 
studies . It is possible that the results of these 
experiments may be integrated with experimentation using AC 
impedance techniques thus permitting more meaningful 
interpretation of alternating current methods than has been 
achieved to date .