You are on page 1of 8

56 JOURNAL OF T H E E L E C T R O C H E M I C A L SOCIETY January 1957

aqueous reaction to form the hydrated fluorides which are tion studies, to B. A. LaMont and co-workers for the
subsequently dehydrated at elevated temperatures in an chemical analyses, and to C. Lentz and associates for the
atmosphere of H F gas. Extensive studies were made of the spectrographic analyses.
conditions for the preparation of high-purity ZrF4, and
pilot plant equipment is described which was used to pre- Manuscript received February 23, 1956. Contribution
No. 480. Work was performed in the Ames Laboratory of
pare 100 lb batches of the fluorides.
the A.E.C.
The reduction step was investigated thoroughly, par- Any discussion of this paper will appear in a Discussion
ticularly for Zr, and those factors which affect metal Section to be published in the December 1957 JOURNAL.
quality and yield were determined. Reduction yields of
96 % were obtained with both Zr and Hf. After are-melting, REFERENCES
the sponge Zr had a hardness of 40-45 Rockwell A and 1. K. A. WALSrI, U. S. A.E.C. Report No. AECD-3640,
was readily cold-rolled into sheet. Zr metal thus prepared (1950).
had a purity of about 99.8%. 2. D. PETERSON AND H. A. WILI-IEL?r Unpublished work,
Hf metal, similarly prepared, had a hardness of 69 Ames Laboratory, Iowa State College, (1950).
Rockwell A and was hot-rolled but was too brittle to be 9SON, Oak Ridge National Lab. Report No. Y-595,
easily cold worked. The Hf was low in metallic impurities, (1950).
but contained considerable amounts of C, N, and oxygen. 4. J. L. WILLIAMS AND B. WEAVER, U. S. A.E.C. Report
No. AECD-3329, (1950).
ACXNOWLI~DGMENTS 5. M. U. ConEs, Rev. Sei. Instr., 6, 68, (1936).
6. R. M. TRECO, J. Metals, 5, 344, (1953).
The authors are especially grateful to J. W. Starbuek 7. R. A. F o o s AND H. A. WILtIELM, V. S. A.E.C. Report
for his valuable contribution in the experimental redue- ISC-693, (1956).

Electrochemical Polarization
I. A Theoretical Analysis of t h e Shape of Polarization Curves


Metals Research Laboratory, Electro Metallurgical Company, A Division of Union Carbide and Carbon Corporation,
Niagara Falls, New York

At low overvoltage values, deviations from Tafel behavior for a noncorroding elec-
trode are due primarily to the reverse reaction of the oxidation-reduction system, and
at high overvoltages to concentration and/or resistance polarization. It is
shown further that the practice of placing straight lines through a few experimental
points is extremely hazardous, while the indiscriminate introduction of "breaks" is
contrary to the electrode kinetics described.
Further complexities arising from a corroding electrode are described. In this in-
stance, the forward and reverse reactions of both of the oxidation-reduction systems
forming the corrosion couple must be considered. This representation of the local
polarization diagram of a corroding metal is more fundamental than that used pre-
viously in the literature, and thus provides a clearer picture of the various factors
which affect the corrosion rate and the shape of polarization curves.
A region of linear dependence of potential on applied current is described for a
corroding electrode by treating it in a manner analogous to that for a noncorroding
electrode. An equation is derived relating the slope of this linear region to the corrosion
rate and Tafel slopes. This relation provides an important new experimental approach
to the study of the electrochemistry of corroding metals since, in some instances, in-
terfering reactions prevent determination of T~fel slopes at higher current densities.

Polarization measurements are an important research data and the correct interpretation of "breaks" in polariza-
tool in investigations of a variety of electrochemical phe- tion curves. Abrupt changes in slope of overvoltage vs.
nomena. Such measurements pernfit studies of the reac- log current have been given considerable significance in the
tion mechanism and the kinetics of corrosion phenomena past few years. Logan (1) examined various methods of
and metal deposition. I n spite of their wide applicability plotting cathodic polarization measurements to evaluate
and extensive use, considerable uncertainty in the inter- the correspondence between current required for complete
pretation of polarization measurements still exists. Some cathodic protection of a system and current flow at the
of the uncertainties include the proper method of plotting potential break. He reported that the potential break

Downloaded on 2016-01-19 to IP address. Redistribution subject to ECS terms of use (see unless CC License in place (see abstract).

method indicated a different current requirement than * o . Z ~ o B t . ' E ..... i ........ J ........ , ','~;,,
that indicated by other criteria9 I n addition, breaks could N /
be obtained regardless of whether the potential was plot-
ted as a linear or logarithmic function of applied current.
However, examination of the data presented indicates that
there is no real experimental evidence pointing toward
existence of breaks. I n most cases, the plots would best
have been considered smooth curves rather than a series
of straight lines. Schwerdtfeger and McDorman (2) pre-
sented a theory which permitted calculation of the corro-
sion rate of a metal from its polarization characteristics.
The calculation was based on plotting both anodie and
cathodic polarization as linear functions of current and
using currents at the potential "breaks" in a formula de-
rived essentially from trigonometric consideratior/si Again,
however, experimental evidence for the existence of breaks
was not clear-cut. Still another discussion of potential
"breaks" was presented recently by Johnson and Babb (3)
who used irreversible thermodynamics to derive Pearson's
Z .I
....... I

FIG. 1. Relationship between overvoltage and current
I0 I00

(4) equation for the relation between the corrosion cur- for the anodic and cathodic reactions of a single electrode
rent and potential breaks. Data presented by these in-
vestigators for the corrosion of Fe in KC1 solutions indi-
of Z is equal to the rate of reduction of Z+(Z+ + e ~ Z).
eate that a great deal of imagination must be used when
The reaction rate and current flow are directly related ae-
drawing the curves in order to obtain "breaks" 9
cording to F a r a d a y s Law. If {, is defined as the current in
Changes in slope rather than a sharp break can be ob-
the forward (reduction) direction (Z+ + e -~ Z) and
tained in H activation overvoltage measurements as a
g~ as the current in the reverse (oxidation) direction
result of the H being discharged by two different rate-
(Z --* Z+ + e), the electrode equilibrium may be expressed
determining reactions. Parsons (5) discussed a dual mecha-
in the form
nism of H discharge and from energy considerations
calculated the potential ranges where more than one rate- ~z = ~ = io,z (I)
determining step might be expected. However, he did not
where io,, is commonly called the exchange current9 When
imply that a "break" should be found. Bockris and Con-
the reaction rate is controlled by a slow step requiring an
way (6) found the cathodic overvoltage vs. log i plots for
activation energy, the dependence of current on over-
Ag in 0.1-7.0N HC1 solutions showed a marked change in
slope at current densities which depended on acid concen- voltage may be expressed as 2
tration. They attribute the change in slope to a change in
symmetry of the energy barrier at the electrode interface
rather than to a dual discharge mechanism. Here again a
sharp break was not found, the two linear portions of the ~ = /o,~exp + ~ ,~, ) (III)
plot being connected by a curve over a short range of
overvoltage values.
where the overvoltage, 7, is the difference between the
The purpose of this discussion is to analyze the shape
potential of the working electrode and the equilibrium po-
of polarization curves in terms of modern concepts of elec-
tential of the reaction being studied, and/3'~ and fl~ are
trochemistry. I t is shown that many of the reported breaks
constants. For a system with a symmetrical energy barrier
in polarization curves are not real, and result either from 9 t
at the electrode interface, fl~ is equal to fl~P. Therefore, Eq.
attempts tc~.apply activation overvoltage theory to data
(II) and (III) may be written as
obtained under conditions where other types of overvolt-
age are included in the measurements, or from a combina-
~7 = -fl~ log 7 - 9 (IV)
tion of insufficient data and an erroneous assumption that ~o,z
a break must exist. I n such a presentation, it is convenient
to consider first the shapes of polarization curves for a r/ = ~
q-fl~ log -=- (v)
noncorroding electrode, then to extertd the analysis to in- Zo,z
clude further complexities which arise from local action
where r = 2.3 fl~ = 2.3 fl~'.
To aid in visualizing the relations given by Eqs. (IV)
i ~ONCORRODING ELECTRODE SYSTEM and (V), they have been plotted on Fig. 1 by arbitrarily
Consider a substance Z in a solution containing its ions giving fl~ a value of 0.100 v and io.~ a value of 1.0 #a. These
Z+2 I n such a system at equilibrium, the rate of oxidation Other investigators have expressed these equations in
This might be Cu in equilibrium with Cu +. The same more fundamental terms. They are simplified here in order
analysis also applies to an inert electrode in an oxidation- not to detract from the main purpose of the discussion.
reduction system such as Pt in a ferrous-ferric solution or Excellent literature reviews and discussions on the subject
in a reducing acid solution saturated with H gas. are presented by Bockris (7, 8) and Vetter (9).

Downloaded on 2016-01-19 to IP address. Redistribution subject to ECS terms of use (see unless CC License in place (see abstract).
( i i t 4Tl~
are of the same order of magnitude commonly found by
experiment; exchange current values generally vary from
I0-~ to 100/aa and ~ values from 0.03 to 0.30 v.
Fig. 1 illustrates the relationship between the oxidation ~ -OI
and reduction rates and the exchange current at the equi- o\
librium potential and the effect of overvoltage on these rates.
When the electrode equilibrium is disturbed by external tl.I
polarization, the reaction rates change in accord with the ~.J -'O2
curves in Fig. 1. Experimentally, however, the individual O
Bz= moo %o
(E Io,z = I.O.laAMP \o\
oxidation or reduction rates cannot be measured. The ex- hi
ternal current used for polarization actually is a measure o_0.5
of the difference between the two rates. For example, if \
the electrode under discussion (Fig. 1) is cathodically po-
larized from its equilibrium potential to an overvoltage ~CTIVE
-O.~ 84~, . . . . . . . . i , , ,,H,,I . . . . . . . . I , , , ,,,,,I ,
\o + ..... N

of - 0 . 1 v, the rate of reduction is equivalent to 10 /~a Ol I.O IO IOO I,OOO IQOOO

while the rate of oxidation is 0.1 #a. The external current APPLIED CATHODIC CURRENT, ~x (MICROAMPERES)
required to polarize to this potential is the difference be- FIG. 2. Relationship between overvoltage and applied
tween ~ and ~ , or 9.9 ga. If the absolute difference be- cathodic current for a single electrode system.
tween the forward and the reverse reaction currents is
defined as g, so that Ol1--"-,~.,k..~ ~,, i ~+. % i i i fiLl' I I I I'I*II I , I I IIIII[ I ' I .....

I hOBLE~oQ~+~-I z
- ~, = g~ (VI) I / 0%
I /<rz %,
then ~ is the external cathodic current when the electrode
is polarized to some overvoltage value, ~7.I t is evident that
$~ approaches $~ at overvoltage values sufficiently removed ~
from the reversible potential. I t is important to emphasize
again that experimentally the only factors measured direc- -0.2 oN
tly are y and ~ or ~ . The individual rate of oxidation or =
reduction cannot be measured. Although theory shows (.9

that there is a linear (Tafel) relationship between y and

>'O.3 7 ~ = ' ( 3 z L O G ~ + ; "z~ - - I o ," nF
log ~ or log ~ , only ~ vs. log ~ is measured and plotted.
Therefore, substituting Eq. (VI) in Eq. (IV), the theo- ~z = o.loo ",,o,,,
o Io, z =1.O~ AMP
retical relationship between overvoltage and the logarithm
IL =5OOO.~ AMP ""\x
of the external cathodic current is obtained.
-0.4 2.3 n ~ = O . 0 5 9

n =-~,mg : (VII)
-aft ]
Using the same arbitrary values for the constants ~ ~,CTIVE
, , ..... i .... ,,,
and io,.. as indicated previously, and knowing the relation 0. I
, , ,,J,,,i



, i LI ',,,I

, if

between ~ and L (Eq. V), a plot of the variation of y with
log ~ may be constructed. This is shown in Fig. 2. 3 Note
that deviation from a Tafel slope exists at the low values FIG. 3. Effect of concentration polarization on the re-
lationship between overvoltage and applied cathodic
of applied current. Only when the reverse (oxidation) cur- current for a single electrode system.
rent, L , becomes insignificant in comparison to the for-
ward or reduction current can a true Tafel relation be ex- The measurement of activation overvoltage may be
pected. Further, Tafel slopes cannot be obtained until complicated by two interfering phenomena--concentra-
applied currents reach magnitudes of several times io,~. tion polarization and resistance drop effects--when the
If it is assumed that experimental verification of a linear value of the exchange current is large. Concentration po-
relationship between ~ and log ~ requires linearity over a larization occurs when the reaction rate or the applied
range of about two logarithmic cycles of current, reliable external current is so large that the species being oxidized
estimates of the Tafel constants ~ and io.~4 require meas- or reduced cannot reach the surface at a sufficiently rapid
urements in the region of 1000 times i .... rate. The solution adjacent to the electrode surface be-
s Points shown on this figure and all subsequent figures comes depleted of the reacting ions, and the rate then is
are not experimental. They are calculated from the de- controlled by the rate at which the reacting species can
rived equations and are included to illustrate the need for diffuse to the surface. The electrode potential changes
considerable data to define accurately the shape of a po- sharply in this region until a potential is reached where a
larization curve. It is quite evident, on inspection of
several of the figures, that insufficient data might lead to new reaction proceeds. The change in potential caused by
a series of straight lines. ment of the slope of ,1 vs. log i~ in the Tafel region, while
4 These constants are derived experimentally from data io,= is found by extrapolation of the Tafel region to the
obtained in the Tafel region, f~= is obtained by measure- reversible potential.

Downloaded on 2016-01-19 to IP address. Redistribution subject to ECS terms of use (see unless CC License in place (see abstract).

concentration polarization m a y be represented in its sim~ c r----~:~,, , , : ..... , ........ , ........ , . . . . . . .

NOBLE// ~162 z
plest form ~ as
RT iL - ~ // Iz %..
~?..... = 2.3 ~ - log iL (VIII) -O.I "x\o\

where iL is the limiting diffusion current for the forward ~176

reaction, R the gas constant, T the absolute temperature,
and F Faraday's constant. When ~ approaches 0.i iL, ~-0.~
~ =-,oz LOG o-'T~,z- IxKr "\x
concentration polarization starts to become significant ex-
perimentally. The shape of a curve, including both activa- . ~z= moo ~
tion and concentration polarization, is illustrated in Fig. 3. o
9:t - 0 3
Io, z = 1.0.~ A M P ~o,~ > = \
For the purpose of this illustration, iL has arbitrarily been
Kr = I0 OHMS ~,\ \
selected as 5000 /za. The limiting diffusion current is a
function of the concentration of the reacting species, the
stirring rate of the solution, and all the other factors which -O.4 \
influence the maximum rate at which an ion can approach
a surface. ~ ACT IVE '~
Examination of Fig. 3 shows that deviation from Tafel -0.5 ......... l ........ I ........ I ........ *.......
behavior caused by concentration polarization is quite 0.1 1.0 I0 I00 1,000 IOpO0
marked and illustrates the difficulties in overvoltage meas~ APPLIED CATHODICCURRENT,Tx(MICROAMPERES)
urements when working near the limiting diffusion current. FIG. 4. Effect of resistance polarization on the relation-
Note that it would be quite convenient to separate an ship between overvoltage and applied cathodic current
for a single electrode system.
experimental curve of this type into at least three distinct
straight line regions and then to attempt to interpret the
ple matter to draw a break in this curve, but t h a t there
breaks theoretically. Such complicating interpretations are
is considerable latitude in choice of the break position.
obviously unnecessary. Stern (13) illustrated the inter-
ference which occurs when H overvoltage measurements CORRODING ELECTRODE SYSTEM
are conducted in the region where concentration polariza-
tion becomes significant. Both Stern (14) and King (15) One additional factor is responsible for deviations from
criticized the recent work of Schuldiner (16), pointing out Tafel linearity in the low-current region: corrosion or local
that the breaks in the reported H overvoltage curves on Pt action current. When shifting from a noncorroding to a
were caused by concentration polarization effects. In ad- corroding system, nlany complicating factors arise. Two
dition, King pointed out that corrections for concentra- co-existing electrochemical reactions now appear: the pre-
tion polarization cannot be made quantitatively without viously discussed oxidation reduction system, Z+ + e ~- Z,
accurate data for ion diffusion rates. Unfortunately, how- and the oxidation reduction system of the metal, M + + e
ever, concentration polarization appears to have been M. Each of these systems has its own exchange current
overlooked again in more recent work with Pd (17). It is and Tafel slope so that the steady-state potential of the
interesting to note that more than 25 years ago Bowden corroding metal occurs where the total rate of oxidation
(18, 19) recognized the possibility that concentration po- equals the total rate of reduction. Thus at the steady-
larization was the cause of breaks which he obtained dur- state corrosion potential,
ing H overvoltage measurements on 14g. ~z + ~,,, = ~ + T~ (IX)
Resistance between the reference electrode and the po-
larized electrode contributes still a third term to the total where Sm is the rate of reduction of M + and Fm is the rate
overvoltagc measured. This is a linear function of current of oxidation of metal M, and ~z and L are the rates of re-
and can be expressed as ~/~os = $xK~. If the resistance duction and oxidation of species Z, respectively. Since the
term K~ ~ is arbitrarily given a value of l0 ohms, the re- corrosion rate by definition is ~,,, - {m, it is evident that
sulting deviation from Tafel behavior appears as illus- the rate may also be defined as ~, - T~ at the corrosion
trated in Fig. 4. Note here again that it is not only a sire- potential. When the corrosion potential is sufficiently re-
moved from the equilibrium potentials of the reactions,
5 Tobias, Eisenberg, and Wilke (1O, 11) and Petrocelli
F,, and T~ become insignificant in comparison to {~ and T,~.
(12) have presented extensive discussions on concentration
polarization. Thus the corrosion rate becomes equal to 5 or ~ . This
6 The limiting diffusion current for H ion reduction in has been used directly by Stern (13) to calculate corrosion
HCI containing an indifferent electrolyte and stirred by rates of Ye from H overvoltage measurements, and indi-
natural convection is approximately 100 ga/cm ~ at pH 3.1, rectly by Elze and Fisher (22) to determine corrosion rates
1000 #a/cm ~ at pH 2.1, and 10,000 tta/cm z at pH 1.1 (13).
7 The resistance is a function of solution conductivity, in inhibited acid environments. Fig. 5 illustrates the po-
distance between the reference electrode and the sample, tential-current relationships for such a mixed electrode
and the geometry of the system. Barnartt (20) has pre- system. The metal oxidation reduction system has been
sented an analysis of the magnitude of the IR drop ex- drawn assuming an io,m of 0.1 #a, a~m value of 0.060 v,
pected as a function of both the current density and the and a reversible potential of -0.160 v. The Tafel con-
solution conductivity. K, may actually be a function of
applied current if the conductance of the solution adjacent stants for the Z oxidation reduction system are as pre-
to the electrode interface changes (21). viously described. Thus, equations for the various reaction

Downloaded on 2016-01-19 to IP address. Redistribution subject to ECS terms of use (see unless CC License in place (see abstract).

+0.1 derived in the following manner. The external applied ca-

thodic current, ~ , is equal to the difference between the
sum of the rates of all the reduction reactions and the sum
of the rates of all the oxidation reactions. Thus,

~ = G- + ~m) - G + ~ ) (XlV)

Since the local action current during cathodic polarization

~mO, I IA may be defined as
t- i~ = ~-.,- ~,. (xV)
rr ~ = ~ - L - i~ (XVI)
~ -02

n = -fl~ log 7 - (X)

~,CT I V E therefore,
0.01 0.1 1.0 IO I00
CURRENT (MICROAMPERES) n -- -fl~ log G + .[~ + iz~ (XVII)
FI~. 5. Relationship between overvoltage and curren~
for a corroding electrode system consisting of two co-
OI ~ L E .... 9 ' '- ..... I 'J- ..... I ........ I ........
existing electrochemical reactions.
/ /~Iz \'- 9 /// //

rates, using the equilibrium potential of the Z species re- -3 ....

action as a zero reference, are as follows:
...1~ \
Z Reduction i~,
o /..\ o%~

*t = - 3 ~ log -:---
~o,~ = -0.100 log 1-~ (X) ,,z \\
-J B Z = 0.100 \o
Z Oxidation Io~z * 1.0= AMPS

Metal Reduction
= +3z log __ = +0.I00 log - -
(XI) > -o~ \
-04' , , , ,,,,,I ..... ,,,I , ~ ...... I , , , ,,,,,I , , ,,,,
o.I I.O Io Ioo LOO0 IO000
n = -0.160 - fl~ log ~~'~ = -0.160 - ~'~
0.060 log b-~ (XlI)

Metal Oxidation Fzo. 6. Relationship between overvoltage and applied

cathodic current for a corroding electrode system.
= -0.160 +/3,, l o g / ~ - = -0.160 + 0.060 l o g ~ (XIII) 0 t . . . . . . j/T% ........ I ' '.~ ..... l ........ I . . . . . "
NO6LE // ",,(.--Iz ,,,'"

The constants are all arbitrary values. At any given poten-

tial, the rate of each reaction is indicated in Fig. 5. As
already discussed, the corrosion potential is closely ap-
o,..<.j-- -%,
proximated by the potential at which ~ = $,,. This cur- = \

rent is labeled i~o, in Fig. 5. I t is quite important to note } \\

both the similarities and differences between this diagram -o2
of the electrochemistry of a corroding metal and that com-
' to,z - .r k " "%
monly used in the earlier literature (23, 24). Since the
various oxidation reduction reactions occur at a finite rate Bz= o,1oo ~'.
=I O ~ A M P
ILO'z= 5 0 0 0 ~ AMP
\" xx
even at the equilibrium potentials, it is apparent that the
K r = I0 OHMS
2~ RT
~ " x x
usual simplified polarization diagrams which have a linear 9 nF = 0.059

current ordinate starting at zero are not strictly valid. ACTIVE ~ ~'x
....... d I I Illllll I i ,,i,,,l , , i1,iiii i ill,'
Although the diagram in Fig. 5 is more in accord with ol I.O I0 IOO 1,000

modern electrochemical principles, its greatest value lies
in the ease with which it perufits an understanding of the
FIG. 7. Effect of concentration polarization and re-
shape and nature of experimental polarization curves. An sistance polarization on the relationship between over-
expression describing the shape of the experimental ca- voltage and applied cathodic current for a corroding elec-
thodic polarization curve of a corroding electrode may be trode system.

Downloaded on 2016-01-19 to IP address. Redistribution subject to ECS terms of use (see unless CC License in place (see abstract).

I t
Since the variations of ~ and i ~ with 7 are known [Eq. i

(XI), (XII), and (XIII)], Eq. (XVIt) permits a calcula- ~. Ot '~ d)~\ Bz .
tion of the potential change as a function of applied ca-
thodic current. A similar equation m a y be derived for
anodic polarization and would be of the form:
~.o~o~- o~
7 = --0.160 + fl,, log ~ + i~ + $,, (XVIII)

Fig. 6 shows the expected curve for the overvoltage as

a function of the applied cathodic current, (~), for the
system described above. This should be compared with A~UED eArMoele CURRZNT,7~(MICRO~PERES)
Fig. 2 which shows the corresponding cathodic polariza-
tion curve for a noncorroding electrode with the same FI~. 8. Linear relationship between overvoltage and
applied current for a single electrode system at over-
Tafel constants. Note that deviation from Tafel behavior voltage values only slightly removed from the reversible
occurs at much highec polarizing currents for the corroding potential.
electrode. True Tafel behavior is not evident until po-
larizing currents of the order of several times the corrosion (IV) and (V) and substitution of these into Eq. (VI) yields
current are applied.
g~ = io,.[10-~/e= - 10 +~/~.] (XX)
If concentration polarization and resistance drop effects
are included in the measurements, Eq. (XVII) becomes For small values of 7//3~ where 10- ' / ~ may be approxi-
mated by 1 - 7/B~ (2.3) and 10+~/~~ m a y be approxi-
7 = -fl~ log mated by 1 + 7//3~ (2.3), Eq. (XX) reduces to
(xix) ~ = - (2) (2.3) (io,~)7/fl, and (XXI)
RT iz - "ix - iz~
+ 2.3 ~F- log ~K~
- io,.(2) (2.3) (XXII)
For the same values of iL and K , used previously, Fig. 7 Thus, measurements of overvoltage close to the reversible
shows the effect of external applied cathodic current, ~,, potential will yield results in accord with Eq. (XXII).
on the measured potential, 7. Fig. 7 contains only a very For the system Z+ + e ~ Z where B~ = 0.100 and io.~ =
short region which exhibits the Tafel slope, ~ , . This is,
of course, due to the choice of numbers used in this exam- 1.0 ~a (Fig. 1), Eq. (XXII) indicates that d-~,--,7+0 =
ple. An increase in the corrosion current or in K , and a
-0.0217 v/#a. Fig. 8 is a plot of 7 as a function of 7~
decrease in iz would completely eliminate any observable
calculated from Eq. (VII) for small values of 7. Note t h a t
Tafel behavior. I t is worth emphasizing here again t h a t
the linear relation predicted by Eq. (XXII) applies for ~
with a curve such as Fig. 7 a variety of straight line sec-
values up to about 20 my.
tions may be drawn with breaks placed at convenient
positions. Obviously, such a treatment would be quite in- Corroding Electrode
correct and contrary to the electrode kinetics described. The same analysis m a y be applied to a corroding elec-
trode where the corrosion potential is determined by the
LINEAR VS. TAFEL POLARIZATIONBEHAVIOR intersection of two logarithmic polarization curves. The
In addition to the problem of polarization breaks, the corrosion current in this ease is analogous to the exchange
question arises continually as to whether polarization current of a noncorroding electrode.
should be a linear or logarithmic function of applied cur- Thus 8
rent. For example, Straumanis, Shih, and SchIeehten (25,
~=-2.3i .... e (~%.+---R~)and (XXIII)
26) have found Tafel behavior for H overvoltage on Ti in
HC1, HBr, and H~SO~, but report a linear dependence of
overvoltage on applied current in HF. I n addition, these de ) /3, t3m
authors show that the linear relation is maintained if the ,+0 = - (2.3) (ir (/~, + fl~) (XXIV)
Ti dissolves, while the Tafel relation holds if the Ti ceases For the mixed electrode system described earlier and illus-
to dissolve because of fluoride additions. That these re-
trated in Fig. 5,
sults are in accord with the electrode kinetics described
above is illustrated here. d_e~ = -0.0040 v/t~a
az~/ e~O

A Noncorroding Electrode Fig. 9 is a plot of e as a hmction of ~ calculated from Eq.

Butler and Armstrong (27) have shown that the over- (XVII) for small values of e. Here again a linear relation
voltage of a reversible electrode is a linear function of ap~ between electrode potential and applied current is found.
plied current for values of overvoltage only slightly re- I t is important to note that the linear behavior extends to
moved from the reversible potential. Proof of this can be s e is the difference between the polarized potential and
found in the following derivation. Rearrangement of Eqs. the corrosion potential.

Downloaded on 2016-01-19 to IP address. Redistribution subject to ECS terms of use (see unless CC License in place (see abstract).

' ' ' ..... I ' ' ''''"1 ........

0 I--DI C
~g ~z Bm "
~ -o'- 2.3 i r z + ,~m )" ,0004(
j j " -.02 (
zz 9

o o
a.a. ~'. -0..'

-.050 Z "~ 4 5 6 7 8 9 I0 II 12 13 14 15 ~ -O.I


FIG. 9. Linear relationship between potential and ap-

plied current at potentials only slightly removed from the ........ a ........ ~ ........ ~ .... ",',',,J , ,,~,;'~,~'~"
corrosion potential. ~ is the difference between the po- -O'~O~ I0 ~ I0 ~ I0 ~ I0 7

larized potential and the corrosion potential. APPLIs CATHODIC CURRENT.~x(MICROAMPERES)

FIG. 10
applied current values higher than the corrosion current.
As stated previously, Tafel behavior would not be indi- served in a potential range where two or more reduction
cated until applied current values approximately ten reactions occur at similar rates.
times i .... are reached. A further extension of the concepts presented here shows
In the data reported by Straumanis (25, 26) for the that an improvement in the efficiency of reducing M + at
linear dependence of H overvoltage on applied current for any given potential may be achieved by increasing flz,
Ti in H F , the region where Tafel behavior would be ex- decreasing /~,~, increasing i . . . . decreasing i . . . . or by
pected to appear had not been reached experimentally be- reducing the difference between the reversible potentials
cause of the high corrosion current under the conditions of the two oxidation reduction systems. Thus, the analysis
of test. For example, the corrosion rate of Ti in 1N H F presented here is important in the study of metal plating
has been reported as equivalent to 85 m a / c m 2 (28), while efficiency as well as corrosion phenomena.
the overvoltage measurements were not carried farther
than 50 ma/cm 2. DISCUSSION
Eq. (XXIV) is valuable from both a theoretical and an The picture of a mixed electrode presented above could
experimental point of view, since it relates the corrosion be made more extensive by including even a third oxida-
rate and the Tafcl slopes to polarization measurements tion reduction system2 I t is doubtful whether a detailed
close to thc corrosion potential. Thus, low current po- discussion of such a system would be of benefit at the
larization measurements combined with corrosion rate present, but it will be included in a future publication.
data permit a calculation of one of the Tafel slopes if the I t is worth noting in passing, however, that the polariza-
other is known. This is of great value when concentration tion curves would be even more complex than those dis-
polarization or I R drop effects interfere with measure- cussed here.
ments at the higher currents. In addition, measurements This analysis has illustrated several important concepts
are made close to the corrosion potential, thus eliminating which arc worth listing for emphasis.
any surface changes which may result from high current 1. The representation of a corroding metal by polariza-
polarization. tion diagrams should be extended to include the reverse
reactions of the various oxidation reduction systems which
The Tafel constants for Fig. 5 have been picked con- 2. Deviations from Tafel behavior may be caused by
veniently to yield a single Tafel relation at the currents local action currents, concentration polarization, and I R
illustrated in Fig. 6. Inspection of Fig. 5, however, reveals drop effects, and by a change in the predominant electrode
that ~,~ Lecomes significant in respect to $~ at very high reaction.
current values. I n this current region, the Tafel slope of 3. An experimental polarization curve may show a linear
Fig. 6 will gradually change from ~ to f ~ . When ~ >> $~, dependence of potential on applied current for small
a new Tafel slope exists with a value of fl~. A pl0t of the amounts of polarization.
overvoltage curve for this situation is presented in Fig. 10. 4. An equation has been derived which relates the slope
Obviously, a different choice o f / ~ , fl,~, i . . . . and i .... could of the linear region of a polarization measurement to the
create this shift in Tafel slope at lower currents and smaller corrosion rate and the Tafel slopes. This equation will
values of overvoltage. I t is important to note that the prove valuable when interfering reactions prevcnt the de-
change in slope takes place gradually and results in a curve termination of the Tafel constants at higher currents.
which is only "Tafel-like" in nature over a considerable 5. The shape of an experimental electrochemical po-
range of current. I t would be a simple matter, although larization curve, either cathodic or anodic, can be analyzed
incorrect, to draw several Tafel lines through this curve. 9 An example of this would be Fe corroding in acid with
Thus, it is evident that real Tafel behavior will not be ob- a depolarizer such as oxygen or ferric ion.

Downloaded on 2016-01-19 to IP address. Redistribution subject to ECS terms of use (see unless CC License in place (see abstract).

if sufficient d a t a are obtained to permit an accurate de- K, = Resistance factor for calculating the IR drop
scription of the curve. Placing straight lines through four included in polarization measurements.
or five experimental points is hazardous, while the indis- This factor includes the solution conduc-
criminate introduction of " b r e a k s " is contrary to modern t i v i t y and the system geometry.
= The difference between the potential polarized
electrochemical concepts. Sufficient information concern-
by applied current and the corrosion poten-
ing the system should be available to estimate whether
concentration polarization or resistance drop effects h a v e k
been included in the measurements. d~) = The slope of the overvoltage vs. current curve for
Although this analysis contains only calculated polar- small overvoltage values.
ization curves, a subsequent discussion will illustrate how
experimental measurements comply with the electrode = The slope of the ~ vs. current curve for small
d~ ~-~o
kinetics described. vahles of e.
The picture of a corroding metal presented here is con-
sistent with the concept or theory of mixed potentials REFERENCES
originally treated by Wagner and T r a u d (29) and subse- 1. K. H. LOGAN, Corrosion, 10, 206 (1954).
quently discussed by Petrocelli (30). The concept of dis- 2. W. J. SCttWERDTFEGER AND 0 . N. McDoRMAN, This
crete anodie and cathodic areas in electrochemical corro-
Journal, 99, 407 (1952).
3. P. A. JOHNSON AND A. L. BABB, Ind. Eng. Chem., 46,
sion m a y be considered a special case of this theory. 518 (1954).
4. J. M. PEARSON, Trans. Electrochem. Soc., 81,483 (1942).
Manuscript received Feb. 24, 1956. 5. R. PARSONS, J. Chim. Phys., 49, C 82 (1952).
Any discussion of this paper will appear in a Discussion 6. J. O'M. BOCKRIS AND B. E. CONWAY, Trans. Faraday
Section to be published in the December 1957 JOURNAL. Soc., 48, 724 (1952).
7. J. O'M. BocKRIS AND E. C. POTTER, This Journal, 99,
169 (1952).
~ = Cathodic or reduction current of Z oxidation 8. J. O'M. Boc~ms, "Modern Aspects of Electrochemis-
reduction system. t r y , " Academic Press Inc., New York (1954).
~,~ = Cathodic or reduction current of M oxidation 9. K. J. VETTER, Z. Elektrochem., 50, 435 (1955).
reduction system. 10. C. W. TOBIAS, M. EISENBERG, AND C. R. WILXE, This
Journal, 99, 395C (1952).
~ = Anodic or oxidation current of Z oxidation 11. C. R. WlLKE, M. EISENBERG, AND C. W. TOBIAS, ibid.,
reduction system. 100, 513 (1953).
~,,~ = Anodic or oxidation current of M oxidation 12. J. V. PETROCELLI, ibid., 98, 187 (1951).
reduction system. 13. M. STERN, ibid., 102, 609, 663 (1955).
io.~ = The exchange current of the Z reaction. This 14. M. STERN, ibid., 102, 356 (1955).
is equal to the oxidation or reduction cur- 15. C. V. KING, ibid., 102, 193 (1955).
rent at equilibrium 16. S. SCHULDIN~R, ibid., 101, 426 (1954).
io.m = The exchange current of the M reaction. This 17. J. P. HOARE AND S. SCHULDINER, ibid., 109., 485 (1955).
is equal to the oxidation or reduction cur- 18. F. P. BOWDEN, Trans. Faraday Soc., 24, 473 (1928).
19. F. P. BOWDEN,Proc. Roy. Soc., 126A, 107 (1930).
rent at equilibrium
20. S. BARNARTT, This Journal, 99, 549 (1952).
dn 21. J. N. AGAR AND F. P. BOWDEN, Proc. Roy. Soc., 169A,
~ = The Tafel slope or ~ for the Z oxidation
206 (1938).
reduction system. 22. J. ELZE AND H. FISHER, This Journal, 99, 259 (1952)
dn 23. U. R. EVANS, "Metallic Corrosion, Passivity, and
t~,~ = The Tafel slope or d ~ o g i for the M oxidation Protection," p. 350, Longmans Green and Co., New
York (1948).
reduction system. 24. H. H. UnLIG, ed., "Corrosion Handbook," p. 436, J.
~ = The external applied cathodic current. Wiley & Sons, Inc., New York (1948).
~ = The external applied anodic current. 25. M. E. STRAUMANIS, S. T. SHIH, AND A. W. SCHLECH-
n = Overvoltage or difference in potential between TEN, J. Phys. Chem., 59, 317 (1955).
a polarized electrode and an electrode at 26. M. E. STBAUMANIS, S. T. SHIH, AND A. W. SCHLECH-
equilibrium for the same reaction. TEN, This Journal, 102, 573 (1955).
27. ft. A. V. BUTLER AND G. ARMSTRONG, J. Chem. Soc.,
~/.... = Concentration overvoltage.
1934, 743.
~s = Resistance overvottage. 28. M. E. STRAUMANIS AND P. C. CHEN, This Journal, 98,
iL = Limiting diffusion current. 234 (1951).
i~ = Local action current. 29. C. WAGNER AND W. TRAUD, Z. Elektrochem., 44, 391
Qo~ = Corrosion current or the local action current (1938).
at the corrosion potential. 30. J. V. PETROCELLI, This Journal, 97, 10 (1950).

Downloaded on 2016-01-19 to IP address. Redistribution subject to ECS terms of use (see unless CC License in place (see abstract).