Polarisation curve interpretation

A guide to polarisation curve interpretation:
deconstruction of experimental curves
typical of the Fe/H2O/H+
/O2 corrosion system
Harvey J. Flitt, D. Paul Schweinsberg *
School of Physical and Chemical Sciences, Queensland University of Technology,
G.P.O. Box 2434, Brisbane, Queensland 4001, Australia
Received 14 May 2003; accepted 26 October 2004
Available online 8 February 2005
Abstract
Experimental DC polarisation curves are the sum of the cathodic and anodic components
and can be diﬃcult to interpret. Schematic representations of ÔtypicalÕ curves (together with
their anodic and cathodic components) are available in the literature for comparison purposes.
A better approach to curve analysis is to generate mathematically the experimental curve
which is then deconstructed into its components. Unfortunately the appropriate computer
programmes are not readily available. We have considered it useful to revisit the collected
curve concept replacing schematic representations with experimental curves. Using an up-
dated programme we have accurately analysed curves representative of the Fe/H2O/H+
/O2
corrosion system.
Ó 2004 Elsevier Ltd. All rights reserved.
Keywords: Iron/low carbon steel corrosion; Computerised polarisation curve analysis; Curve deconstruc-
tion/deconvolution
0010-938X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.corsci.2004.10.002
*
Corresponding author. Tel.: +61 73 864 2111; fax: +61 73 864 1804.
E-mail address: p.schweinsberg@qut.edu.au (D.P. Schweinsberg).
Corrosion Science 47 (2005) 2125–2156
www.elsevier.com/locate/corsci

1. Introduction
The generation of polarisation curves continues to be important in aqueous cor-
rosion research. The time-consuming potentiostatic method has been largely re-
placed by the potentiodynamic approach where the potential (E) of the corroding
metal is automatically varied with time. The current (I) needed to maintain the metal
(working electrode (WE)) at each applied potential (Ew) is ascertained and the poten-
tial/current data is plotted to give the experimental polarisation curve. In corrosion
studies it is common practice for the curve to be displayed with the independent
variable (in this case the potential) rather than the dependent variable as ordinate.
Further, the logarithm10 of the current density (logi) is plotted in the positive x-
direction, notwithstanding the convention that anodic current is positive and catho-
dic current is negative.
The magnitude of Ew can be regarded as a measure of the oxidising power of the
corrodent [1], with the logi axis reflecting the rate of each reaction in the corrosion
process. Depending on the corrosion system under study it follows that from the
shape of the experimental curve it may be possible to obtain information on the
kinetics of the corrosion reactions, protectiveness of a passive film, ability of a com-
pound to act as a corrosion inhibitor, relative corrosivity of process streams and cor-
rosion rate (icorr) of the metal.
Unfortunately, extracting any of the above from the experimental curve may be
quite difficult. This is because at each applied potential the recorded current is the
sum of the anodic and cathodic components of the corrosion reaction and the exper-
imental curve (e.g., for the simple case of pure Fe in O2-free dilute H2SO4) will be the
sum of two true polarisation curves, one describing oxidation of Fe to Fe2+
and the
other reduction of H+
ion. This means that for potentials not greatly removed from
that of the freely corroding WE (corrosion potential (Ecorr)) the shape of the anodic
and of the cathodic portions of the experimental curve will differ from that exhibited
by each true curve. However, for potentials further from Ecorr the effect of the catho-
dic reaction on the anodic reaction and vice-versa is progressively lessened, and the
shape of the experimental curve eventually becomes an accurate representation of
the kinetics of the anodic and cathodic corrosion reactions. Of course, if an alloy
is involved or if the corrodent contains more than one oxidant (commonly H+
ion
and dissolved O2) the net experimental curve will more complex, and correspon-
dingly harder to interpret in terms of its components.
An example where failure to correctly analyse the experimental curve can lead to
error is when the curve is employed to evaluate corrosion rate. The Tafel extrapola-
tion method is well known but it is often forgotten that the metal is required to be
uniformly corroding and at the corrosion potential either the anodic or the cathodic
reaction needs to be under complete activation control. Further, for accurate estima-
tion of icorr the identified linear portion of the experimental curve should extend over
about one decade on the logi axis. Unfortunately, in practice these requirements are
not always met: the relevant cathodic reaction may be experiencing both activation
and concentration polarisation at Ecorr and extrapolation of what is perceived as a
ÔshortenedÕ Tafel portion is completely erroneous. Another example pertaining to
2126 H.J. Flitt, D.P. Schweinsberg / Corrosion Science 47 (2005) 2125–2156

corrosion rate evaluation is when corrosion-monitoring probes based on the polar-
isation resistance method are used. The reaction kinetics of the corrosion process
must be established before installation as these devices again assume that at Ecorr
the anodic and cathodic corrosion reactions are under activation control. A final
example where the experimental curve can be difficult to interpret is when the metal
spontaneously passivates/pits in the corrodent prior to polarisation. The anodic por-
tion of the experimental curve may now exhibit Ôstraight line behaviourÕ but, because
localised corrosion is involved, extrapolation of this portion of the curve does not
lead to a Ôcorrosion rateÕ. Also, in this case the cathodic portion of the experimental
curve may exhibit either a confusing Ôcathodic loop or dipÕ (negative peak).
In practice it is difficult, except for the simplest corrosion systems, to visualize an
experimental curve in terms of its anodic and cathodic components. Schematic rep-
resentations of experimental curves with their schematic ÔtrueÕ anodic and cathodic
curves have been published [1,2]. Thus Liening [1] discusses nine possible experimen-
tal curves for the reaction
M þ Hþ
! Mþ
þ 1=2H2
These examples may be useful in that it may be possible to associate features of an
experimental curve with one depicted in the collection. However, the best approach
for the interpretation of a polarisation curve is one based on electrochemical theory.
Here the appropriate thermodynamic and kinetic parameters are inserted into the
relevant mathematical functions to synthesise the approximate true cathodic and
true anodic curves for the corrosion system. These curves are then combined to give
the approximate synthesised experimental curve, which is then overlaid on the experi-
mental one. Values of the input parameters are now varied, and by trial and error
the shape of the synthesised experimental curve is altered until a good match is ob-
tained with the experimental one. (Note: literature and experimental values may be
used as a guide to the magnitude of the various parameters.) Finally, the matched
curve is deconstructed (deconvoluted) to show its true anodic and cathodic compo-
nents. Various computer-based programmes have been devised to effect the calcula-
tions and the results for a number of corrosion systems are described in the literature
[3–20]. We have also used this approach in SYMADEC, a programme for the syn-
thesis, matching and deconvolution of curves for the M/H2O/H+
/O2 system. Earlier
versions of the programme have been successfully used to study the corrosion kine-
tics of carbon steel and low-alloy steels in different aqueous environments [21–27].
Unfortunately computer programmes for curve interpretation are not readily
available. We have therefore considered it useful to revisit the collected curve con-
cept, but instead of employing schematic representations have selected for compar-
ison purposes actual experimental curves (in this case for the corrosion of iron and
carbon steels). Each curve has been synthesised, matched and then deconstructed to
reveal the nature of its components. Knowledge of the experimental conditions is
important in curve interpretation and this information is provided in detail. The role
of the Pourbaix diagram for the pure iron/pure water system at 25 °C in curve anal-
ysis is also emphasised. The experimental curves were obtained either from
experiments carried out in our laboratories, or from examples published in the
H.J. Flitt, D.P. Schweinsberg / Corrosion Science 47 (2005) 2125–2156 2127

literature. Printed curves were scanned and then digitised using a programme written
for this purpose. Filtering and sampling were applied to the digitised data to mini-
mise the current/voltage set and optimise graphical representation. The curves cho-
sen range in complexity, starting with the simple case indicative of one anodic and
one cathodic reaction and undergoing activation polarisation only at Ecorr to corro-
sion systems involving both activation and concentration polarisation and more than
one oxidant. The effect of non-passive surface films is also covered together with the
more usual case of an active/passive transition followed by pitting.
2. Mathematical basis of SYMADEC
The most common cathodic reactions driving the aqueous dissolution of a metal
are
2Hþ
ðaqÞ þ 2eÀ
! H2ðgÞ ðequivalent 2H2O þ 2eÀ
! H2ðgÞ þ 2OHÀ
ðaqÞÞ
and
O2ðgÞþ2H2Oþ4eÀ
!4OHÀ
ðaqÞ ðequivalent O2ðgÞþ4Hþ
ðaqÞþ4eÀ
!2H2OÞ
The relationship between the rate of each of the above reactions, expressed as catho-
dic current density, ic and high values of the activation overpotential, gact,c (>approx.
À0.03 V) at the metal/solution interface is
ic ¼ i0 expðÀanF gact;c=RTÞ ð1Þ
where a = transfer coefficient; n = number of electrons involved in the reaction;
F = FaradayÕs constant; gact,c = Ew À Ereversible; R = 8.314 J KÀ1
molÀ1
; T = abs.
temp. Rearranging gives the Tafel equation:
gact;c ¼ bc logðic=i0Þ ð2Þ
where bc = Tafel slope = À2.303RT/anF.
At higher reaction rates concentration polarisation is present (this is most often
seen for the oxygen reduction reaction) and the relationship between the cathodic
current density and the cathodic concentration overpotential, gconc,c, is
ic ¼ iLf1 À expðnF gconc;c=RTÞg ð3Þ
where iL = limiting current density.
Rearranging
gconc;c ¼ ð2:303RT=nF Þ logf1 À ðic=iLÞg ð4Þ
Charge transfer and concentration overpotentials are additive, and for a single
cathodic process Eqs. (2) and (4) can be added to give
gtotal;c ¼ Àð2:303RT =anF Þ logðic=i0Þ þ ð2:303RT =nF Þ logð1 À ðic=iLÞÞ ð5Þ

It follows [18,28] that the approximate value of the total cathodic current density is
given by
itotal;c ¼ ½i0 expðÀanF g=RTÞŠ=½1 þ fi0 expðÀanF g=RTÞg=iLŠ ð6Þ
or
itotal;c ¼ iLic=ðiL þ icÞ ð7Þ
Appropriate versions of Eq. (6) are used to model the curves for H+
and O2 reduc-
tion. The current densities at each potential are then summed.
The general anodic reaction for active metal dissolution is M ! Mn+
+ neÀ
. Con-
sider the corrosion of iron. This process is pH dependent, and reference to the well
known Pourbaix diagram [29] for the iron/water system at 25 °C (dissolved ion activ-
ity <10À6
M) shows the following:
1. For pH < $4.2 as the potential of the iron (Ew) is made more positive the reaction
is
FeðsÞ ! Fe2þ
ðaqÞ þ 2eÀ
ðactive corrosionÞ ð8Þ
2. For pH $ 4.2 to $9.4 as Ew is made more positive active corrosion (formation of
Fe2+
) is followed by passivation due to precipitation of hydrous oxide,
Fe2O3 Æ nH2O. (Note: the precipitate is usually represented as Fe(OH)3.)
3. For pH $ 9.4 to $12.2 as Ew is made more positive iron passivates to form
Fe(OH)2 then Fe(OH)3.
4. For pH > $ 12.2 as Ew is made more positive iron is transformed to soluble
HFeOÀ
2 ions followed by passivation due to Fe(OH)3.
For active dissolution of a metal, e.g., Fe (Eq. (8) above) the Tafel equation is used:
ia ¼ i0 expðf1 À agnF gact;a=RT Þ ð9Þ
or
gact;a ¼ ba logðia=i0Þ ð10Þ
where ba = Tafel slope = 2.303RT/(1 À a)nF.
In order to model the anodic curve for a transition from active to passive beha-
viour, i.e., from the potential where passivation commences (passivation potential,
Ep) to that value where passivation is complete (Ecp), Hines [9] assumed that the
metal surface consists of two independent regions—one where metal dissolution
MðsÞ ! Mnþ
ðaqÞ þ neÀ
occurs, and the other where a film deposits. Initially, metal dissolution is seen over
the entire surface, but as filming starts the area on which the anodic reaction pro-
ceeds unimpeded gradually decreases, reaching a minimum when the potential at
Ecp is reached. Suppose S is the fraction of metal area on which no film forms
and (1 À S) is the fraction filmed. The rate of the anodic reaction on the total surface
itotal,a can now be expressed in terms of the anodic current densities (i) on the un-
filmed and filmed regions. Thus

itotal;a ¼ iuS þ if ð1 À SÞ ð11Þ
where iu and if are the rates on the unfilmed and filmed regions, respectively. S will be
equal to unity at Ep and will reach a value of zero at Ecp.
Hines [9] suggested two physical models for the dependence of S on the applied
potential E. However, Eqs. (14) and (15) in his paper do not generate the S curve
depicted in his Fig. 3 [9]. We have corrected these equations and the variation of
S with applied potential according to HinesÕ second model is now given by
S ¼ 2½expðÀAðEw À EpÞ
p
ÞŠ=½1 þ expðÀAðEw À EpÞ
p
ÞŠ ð12Þ
where p = constant used to shape the passivation peak (2 symmetrical; 2–3 asymmet-
rical) and A = constant (10À3
–10À4
) that determines the width of the passivation
peak. Both p and A are obtained empirically and appear to have no physical signi-
ficance [11].
Substitution in (11) gives the following for itotal,a
itotal;a ¼ iuf2½expðÀAðEw À EpÞp
ÞŠ=½1 þ expðÀAðEw À EpÞp
ÞŠg
þ if f1 À 2½expðÀAðEw À EpÞ
p
ÞŠ=½1 þ expðÀAðEw À EpÞ
p
ÞŠg ð13Þ
In summary, when S = 1 (no film) (11) reduces to itotal,a = iu and the Tafel relation-
ship applies. When S = 0, itotal,a = if = icp.
In the presence of certain anions (e.g., ClÀ
) the film is attacked and at points
where the film is thin metal dissolution may proceed (localised or pitting corrosion).
That part of the anodic curve from the point where pitting commences (Ebr) to the
maximum potential reached (Em) is now modelled. It is assumed that the metal dis-
solution can be described by a linear logarithmic current density/potential relation-
ship. The following empirical expression is proposed for the dependence of the
anodic current density ia on the potential Ew
ia ¼ icpfðicp þ mÞ=mg ð14Þ
where
m ¼ expfln icp þ ð1=iron transpassive slopeÞ½Em À ðEw þ EbrÞŠg ð15Þ
with respect to (14) and (15) the following applies:
(1) when Ew is equal to or less than jEbrjm becomes large and ia = icp;
(2) when Ew > jEbrjm is small and ia > > icp.
At higher positive potentials film breakdown (in the absence of aggressive anions)
and oxygen evolution may be possibilities. Currently these aspects have not been fac-
tored into the programme.
Resistance polarisation due to the presence of the passive film will also be present
and the recorded anodic potentials must be corrected for the IR drop. Sometimes an
ionically conducting but non-passive porous film (e.g., graphitic carbon) may form
on a metal and the IR drop across this film must also be taken into account. If a cur-
rent I is passed across a film whose resistance is RX there will be a potential drop

given by gX = IRX. Resistance polarisation has the effect of making the electrode po-
tential (Ew) for a corrosion system larger than the ÔtrueÕ value (Etrue). Thus
Etrue ¼ Ew À IR ð16Þ
This type of polarisation can be responsible for the anodic portion of an experimen-
tal curve (e.g., for mild steel in oxygen-free 0.5M sulphuric acid) exhibiting curvature
instead of the expected straight line indicative of Tafel behaviour. SYMADEC
allows for the insertion of different values of film resistance and subsequent
calculation of the true potential.
3. Synthesising and plotting polarisation curves using SYMADEC
SYMADEC contains the following series of drop-down menus (Table 1) to allow
coordinated entry of parameters required for synthesising polarisation curves. Guid-
ance as to the magnitude of certain parameters (Tafel slopes and exchange current
densities) can be obtained from the literature (see Refs. [24,25]) whilst others (tem-
perature; [H+
] and [O2]) will be known either from the conditions of the experiment
or may be obtained directly from the experimental curve (Tafel slope; limiting cur-
rent density; primary and complete passivation potentials; pitting potential). Due
attention to the magnitude of parameters employed should minimise the possibility
of synthesising and matching a curve by the inclusion of inappropriate values.
4. Examples of analysed experimental polarisation curves
4.1. Case 1: Pure iron corroding in oxygen-free H2SO4 (active corrosion, no film
formation)
Data for the experimental polarisation curve shown in Fig. 1a was recorded
potentiostatically by one of the authors (DPS). Conditions for recording the experi-
mental curve were as follows: The working electrode (WE) was the cross-sectional
surface of a 5 mm diameter rod of 99.999% ÔspecpureÕ polycrystalline iron (Johnson
Matthey) embedded in Teflon. The corrodent was nitrogen purged 0.5M H2SO4 at
30 ± 0.5 °C. The electrode assembly, electrochemical cell and associated apparatus
were similar to those described by Schweinsberg and Ashworth [30]. The reference
and counter electrodes were saturated calomel and Pt foil (1 cm2
) respectively.
Two hundred and fifty millilitre of nitrogen purged (1 h) corrodent was heated in
a 1 L RB flask to boiling under reflux. (High purity nitrogen gas was further purified
by passing it through alkaline pyrogallol solution. Under these conditions the purged
corrodent was considered to be oxygen-free.) The contents, after cooling to ambient
temperature, were introduced into the N2-flushed cell under positive N2 pressure.
Gas was then passed continuously over the corrodent. The WE was abraded manu-
ally with 1200 grade SiC paper, polished on filter paper saturated with MgO slurry,
degreased with warm AR grade acetone, washed with water and immediately placed

whilst wet in the corrodent. The Luggin capillary was adjusted adjacent to (about
1 mm from) the WE. After 10 min immersion the WE was pre-polarised at
À756 mV (SHE) for 40 min to remove residual oxide film. The used corrodent was
then transferred from the cell under positive N2 pressure to a waste bottle and imme-
diately replaced under pressure with fresh corrodent. Gas was passed over the
solution.
The potential of the WE was monitored with a chart recorder and reached a
steady state after 90 min. This was selected as the corrosion potential (Ecorr). The
WE was then polarised cathodically (20 mV steps) to À576 mV (SHE) (current
was recorded after 1 min intervals). After cathodic polarisation the WE was allowed
to rest for 15 min. Over this period the potential of the WE either returned to its
Table 1
Menus incorporated in SYMADEC
Drop-down menus Parameters Notes
Menu 1: Redox inputs pH; [O2] (mg LÀ1
); T (K);
[Mn+
] (0.056 mg LÀ1
)
Parameters for calculation of Erev for reactions:
M(s) ! Mn+
(aq) + neÀ
2H+
(aq) + 2eÀ
! H2(g)
(or 2H2O + 2eÀ
! H2(g) + 2OHÀ
(aq))
O2(g) + 2H2O + 4eÀ
! 4OHÀ
(aq)
(or O2(g) + 4H+
(aq) + 4eÀ
! 2H2O)
Menu 2:
Hydrogen inputs
Tafel slope (V decadeÀ1
);
i0 (A cmÀ2
); iL (A cmÀ2
)
Parameters to synthesise cathodic curve for
H+
reduction
Menu 3:
Oxygen inputs
);
i0 (A cmÀ2
); iL (A cmÀ2
)
Parameters to synthesise cathodic curve for
O2 reduction
Menu 4:
Metal: active inputs
);
i0 (A cmÀ2
)
Parameters to synthesise anodic curve up to Ep
Menu 5:
Metal: passivation to
film breakdown inputs
icp (A cmÀ2
); Ep (V);
Ecp (V); Ebr (V); p; A;
Tafel slope after film
breakdown (V decadeÀ1
)
Parameters to synthesise anodic curve from
Ep to Em
Menu 6:
Plotting synthesised
curve
(a) Displays synthesised anodic curve
(b) Displays synthesised cathodic curve(s)
(c) Combines (a) and (b) to display
complete synthesised curve
Menu 7:
Matching and
deconvoluting
synthesised complete
polarisation curve
The experimental polarisation curve is plotted.
Alternatively a printed curve is scanned/
digitised and plotted. The synthesised
polarisation curve is overlaid on the
experimental one and the former is adjusted
(by varying parameters) until it matches the
experimental curve. The matched curve is then
deconvoluted into its anodic and cathodic
components. All curves are plotted with
potential (versus either SHE or SCE) as
ordinate and the logarithm of the current
density in the positive x-direction

(a)
(b)
PotentialvsSHE(mV)PotentialvsSHE(mV)
Fig. 1. Case 1. (a) Experimental and synthesised polarisation curves for pure iron in O2-free 0.5M H2SO4
at 25 °C. (b) Deconvolution of synthesised polarisation curve for pure iron in O2-free 0.5M H2SO4 at
25 °C.

previous steady state value or was within about 2 mV. Anodic polarisation was
commenced (10 mV steps) concluding at À206 mV (SHE).
Parameters and data required to synthesise and match the experimental polarisa-
tion curve (shown in Fig. 1a) are listed in Table 2. Case 1 represents a very simple
corrosion system in that a pure metal is employed and there is only one oxidant,
H+
ion. The pH of the solution is approximately 0 and the reversible potential (Erev)
for the H2/H+
system is accordingly zero. Since the corrodent was prepared using
pure water and AR grade acid, the concentration of dissolved iron (as Fe2+
) will
be negligible, and a value of 0.056 mg LÀ1
(10À6
M) may be used to calculate Erev
for the Fe/Fe2+
system (À621 mV (SHE)).
As a guide to the corrosion behaviour, reference can be made to the iron Pourbaix
diagram [29]. The diagram shows the corrosion susceptibility for the pure metal in
pure water at 25 °C. Thus, for Fe2+
ion activity = 10À6
M and pH $ 0, at the mean
Ecorr (here about À320 mV (SHE) from the experimental curve), and for the more
positive potentials applied in the experiment), Fe reacts to form Fe2+
ions.
By deconstructing the matched synthetic curve the anodic and cathodic compo-
nents are revealed (Fig. 1b). The shape of the experimental curve indicates that over
the potential range employed corrosion is uniform, and both the metal dissolution
and H2 evolution reactions experience activation polarisation only. The high acidity
delays the onset of concentration polarisation and also ensures that polarisation due
to solution IR drop is negligible. It should be noted that concentration polarisation
can be reduced also by stirring the solution. Tafel behaviour is well defined on both
portions of the experimental curve and extrapolation will give an accurate value of
icorr at Ecorr (0.16 mA cmÀ2
).
Uniform corrosion was confirmed by examination of the WE after polarisation.
In this case, due to the purity of the material, the nature of the corrodent and the
experimental conditions, there is little need to deconstruct the experimental curve
in order to understand the corrosion process. The ease of interpretation, together
with the ability to accurately evaluate corrosion rate by Tafel extrapolation from
the curve, makes this system suitable for studies on the inhibition efficiency of
organic compounds for iron corrosion [30].
4.2. Case 2: Carbon steel corroding in oxygen-free H2SO4 (active corrosion,
non-passive film formation)
The scanned experimental polarisation curve shown in Fig. 2a was originally re-
corded galvanostatically by Bandy and Jones [31] for 1080 carbon steel (nominal
comp. 0.75–0.88% C; 0.60–0.90% Mn; 0.04 max P; 0.05 max S) immersed in oxy-
gen-free 0.5M H2SO4. Conditions for recording the experimental curve were as fol-
lows: A glass cell was used for the electrochemical experiments and the laboratory
temperature was 25 ± 1 °C. The corrodent was placed in the cell, deaerated by bub-
bling oxygen-free H2 before, and then continually during the experiment. The WE
was fashioned from rod (exposed surface area 2.5 cm2
). The reference and counter
electrodes were saturated calomel and Pt foil respectively. The exposed face of the
WE was abraded with emery paper (final finish 00 grade), degreased with detergent,

Table 2
Parameters and other data relevant to the synthesis of polarisation curves
Parameters Case 1 Case 2
ﬁlmed
Case 2
unﬁlmed
Case 3a Case 3b Case 4a Case 4b Case 4c Case 4d Case 5 Case 6
pH 0 0 0 4.9 5.52 9 9 9 12.3 8.8 7
[O2] (mg LÀ1
) – – – 2 3 0.01 0.2 7.9 8 0.01 8
[Fe] (mg LÀ1
) 0.056 0.056 0.056 0.056 0.056 0.056 0.056 0.056 0.056 0.056 0.056
Temperature (K) 303 298 298 308 316 303 303 303 298 313 313
H2 TS (mV decÀ1
) 98 98 98 130 134 151 101 169 120 30 130
H2 ECD (A cmÀ2
) 8.48EÀ8 2.43EÀ6 2.43EÀ6 3.86EÀ8 1.00EÀ5 1.00EÀ8 1.00EÀ7 1.00EÀ7 1.31EÀ6 5.77EÀ6 4.18EÀ6
H2 LCD (A cmÀ2
) – 7.58EÀ2 7.58EÀ2 1.00EÀ4 2.27EÀ3 – 1.24EÀ5 – – – –
O2 TS (mV decÀ1
) – – – 133 127 151 171 152 153 159 180
O2 ECD (A cmÀ2
) – – – 3.00EÀ13 1.00EÀ11 4.58EÀ14 1.43EÀ10 3.92EÀ11 1.35EÀ11 5.06EÀ10 5.00EÀ11
O2 LCD (A cmÀ2
) – – – 9.61EÀ5 3.51EÀ4 2.14EÀ5 3.41EÀ5 5.09EÀ5 2.56EÀ5 8.55EÀ6 6.50EÀ5
Fe TS (mV decÀ1
) 48 39 39 64 39 142 98 168 159 144 30
Fe ECD (A cmÀ2
) 8.84EÀ11 1.67EÀ12 1.67EÀ12 1.94EÀ7 7.11EÀ9 3.49EÀ7 4.55EÀ7 5.56EÀ7 7.72EÀ7 8.38EÀ7 1.25EÀ7
Fe PPP (mV) – – – – – À525 À575 À420 À732 À480 À580
Fe CPP (mV) – – – – – À250 À164 À210 À494 À148 À410
Fe BP (mV) – – – – – 79 90 À100 – 138 À152
Fe TTS (mV decÀ1
) – – – – – 114 65 158 – 150 60
Fe CPC (A cmÀ2
) – – – – – 3.00EÀ6 2.92EÀ6 1.43EÀ6 1.41EÀ6 4.01EÀ6 3.01EÀ6
Res (X) 0 1 0 90 0 543 3228 4900 0 321 4000
Fe Exp – – – – – 2 2 2 2 2.01 2.14
Fe Lin – – – – – 1.00EÀ4 1.00EÀ4 1.70EÀ4 1.00EÀ4 1.00EÀ4 3.97EÀ4
Fe R (mV) À621 À618 À618 À623 À629 À621 À621 À621 À618 À627 À627
O2 R (mV) – – – 876 819 589 609 642 447 580 761
H2 R (mV) 0 0 0 À300 À346 À541 À541 À541 À727 À547 À435
Notes: TS = Tafel slope; ECD = exchange current density; LCD = limiting current density; PPP = primary passivation potential; CPP = complete passivation
potential; BP pitting potential; TTS = transpassive Tafel slope; CPC = complete passivation current density; Res = resistance; Exp = exponential constant p;
Lin = linear constant A; R = reversible potential.
H.J.Flitt,D.P.Schweinsberg/CorrosionScience47(2005)2125–21562135

rinsed in distilled water, dried and then placed in the test solution. The potential of
the WE was monitored, becoming steady after 4 h. This potential was selected as
Ecorr. The current was then adjusted to give six increments per decade on a logarith-
mic scale. The cathodic and then anodic potentials were recorded after 3-min inter-
vals. Both cathodic and anodic portions of the polarisation curve were obtained.
Bandy and Jones [31] found that for repeated experiments Ecorr varied between
À260 and À275 mV (SHE). For the diagram illustrated in their paper [31, Fig. 9]
the corrosion potential prior to cathodic polarisation was À268 mV (SHE). How-
ever, no mention is made of Ecorr before anodic polarisation and it is not possible
to establish this potential from their Fig. 9.
tion curve (synthesised curve shown in Fig. 2a) are listed in Table 2 (Case 2 filmed).
(a) (b)
(c) (d)
Fig. 2. Case 2. (a) Experimental and synthesised polarisation curves for carbon steel in O2-free 0.5M
H2SO4 at 25 °C assuming presence of non-passive surface film. (b) Deconvolution of synthesised
polarisation curve for carbon steel in O2-free 0.5M H2SO4 at 25 °C assuming presence of non-passive
surface film. (c) Experimental and synthesised polarisation curves for carbon steel in O2-free 0.5M H2SO4
at 25 °C together with synthesised curve assuming no surface film. (d) Deconvolution of synthesised
polarisation curve for carbon steel in O2-free 0.5M H2SO4 at 25 °C assuming no surface film.

Again the corrosion system is relatively simple in that the WE consists essentially of
one metal (Fe) and there is only one oxidant, H+
ion. The pH of the solution is
approximately zero and Erev for the H2/H+
system is 0 mV (SHE). As for Case 1
the [Fe2+
] was taken as 0.056 mg LÀ1
. Erev for the Fe/Fe2+
system is now
À618 mV (SHE).
The Pourbaix diagram for pure iron can be used also as a guide to the corrosion
behaviour of carbon steel. Thus at 25 °C (ion activity = 10À6
M and pH = 0) at Ecorr
(À270 mV (SHE)) and for more positive potentials applied in the experiment, it can
be assumed that the anodic reaction is principally the dissolution of Fe to form Fe2+
ions.
By deconstructing the matched synthetic curve shown in Fig. 2a the anodic and
cathodic components are revealed (Fig. 2b). The shape of Bandy and JonesÕ curve
indicates that on polarisation from Ecorr in the negative direction the hydrogen reac-
tion is experiencing activation polarisation and Tafel behaviour is seen over about
one decade [31]. At more negative potentials the onset of concentration polarisation
is observed. Extrapolation of the linear portion of the experimental curve to their
mean Ecorr will give an accurate value of the corrosion rate.
The anodic portion of the experimental curve in Fig. 2a might also be expected to
show Tafel behaviour. However, marked curvature is seen, and Bandy and Jones [31]
attribute this to a number of factors including a change in the nature of the metal
surface as liberated corrosion products deposit to form a non-passivating, conduct-
ing surface film. They show how the anodic current density ianodic can be calculated
from iapplied = ianodic À icathodic in the potential region near Ecorr where iapplied does
not equal icathodic. The extrapolated Tafel line gives icathodic and the data points give
iapplied. Substituting these values into the above expression gives corresponding val-
ues of ianodic at a number of potentials [31]. A straight line can now be drawn through
these values of ianodic which is now representative of reasonable anodic Tafel beha-
viour. In their paper both cathodic and anodic Tafel lines are seen to intersect approxi-
mately at the mean value of Ecorr.
The current authors (Flitt and Schweinsberg) have also observed anodic curvature
for carbon steel polarised in oxygen-free sulphuric acid. The WE was covered with a
black film, probably graphitic carbon which will impart a resistance to the WE. It
follows that the values of the recorded anodic potentials are greater than the true val-
ues. The effect of this resistance polarisation can be calculated using SYMADEC and
1 X was required to synthesise and match the anodic portion of Bandy and JonesÕs
experimental curve shown in Fig. 2a. The anodic portion of the experimental polar-
isation curve (assuming no film and therefore no ohmic resistance) can also be syn-
thesised and this, a straight line exhibiting Tafel behaviour, is shown together with
the matched cathodic portion in Fig. 2c. The corresponding deconvoluted anodic
and cathodic curves (assuming no surface film) exhibiting Tafel behaviour and inter-
secting at the mean value of Ecorr (approx. À270 mV (SHE)) are also shown in Fig.
2d.
Compensating for anodic curvature due to resistance polarisation using SYMA-
DEC is an alternative approach to that employed by Bandy and Jones. Their esti-
mated corrosion rate and cathodic and anodic Tafel slopes were 1.18 ·

10À3
A cmÀ2
, 98 mV decÀ1
and 38 mV decÀ1
, respectively, whilst the calculated cor-
rosion rate and corresponding parameters used by SYMADEC for curve synthesis
and matching were 1.39 · 10À3
A cmÀ2
, 98 mV decÀ1
and 39 mV decÀ1
, respectively.
In conclusion it should be noted that pure iron is expensive, and for studies on
iron corrosion the WE is often fabricated from carbon steel. It is often assumed that
these steels behave like the pure metal and that in strong acidic solution Tafel beha-
viour will be seen on both the cathodic and anodic portions of the experimental
polarisation curve. However, as discussed above, this is not necessarily so: the Tafel
region of the anodic portion can be obscured due to film resistance polarisation.
4.3. Case 3(a): Mild steel corroding in dormant, mixed cane sugar juice (active
corrosion, non-passive film formation)
The scanned experimental polarisation curve shown in Fig. 3a was originally re-
corded potentiodynamically by Cash [24] for a mild steel WE (typical of pipeline
steel used in a cane sugar mill) in dormant, mixed cane-sugar juice (MJ), open to
air at 35 °C. Mixed juice contains about 13% by weight of sucrose together with
Na $ 52 ppm, K $ 1300 ppm, Ca > 113 ppm, Mg $ 109 ppm, Al $ 25 ppm,
Fe $ 25 ppm, Si $ 73 ppm, ClÀ
$ 1200 ppm, sand and fine fibre from the crushed
cane. Preliminary experiments showed that mild steel, on exposure to MJ, becomes
coated with a grey/black, porous, non-passivating film which was found to consist
mainly of organic material [24]. In the sugar mill under flow conditions the thickness
of this film increases with both increasing flow rate and exposure time.
The working electrode employed by Cash [24] was the cross-sectional surface of
10 mm diameter rod of mild steel embedded in chemical resistant epoxy resin. The
reference and counter electrodes were saturated calomel and Pt foil (1 cm2
) respec-
tively. The WE was abraded with 600 grade SiC paper, degreased with AR grade ace-
tone and then immediately exposed to the MJ contained in a 1 L glass cell. The
dissolved oxygen concentration was 2 mg LÀ1
and Ecorr (À445 mV (SHE)) was steady
after approx. 30 min. After approx. 100 min exposure to the mixed juice the WE was
polarised anodically from Ecorr (in order to least disturb any film deposited on the WE
during the exposure period). This was followed by the cathodic scan when Ecorr had
returned to within ±5 mV from its previous value. The scan rate was 60 mV minÀ1
.
tion curve (shown in Fig. 3a) are listed in Table 2. For mild steel the dominant ano-
dic reaction is iron dissolution and this is driven by H+
ion and O2 reduction. The pH
of the MJ was 4.9, and calculated Erev for the H2/H+
system is À300 mV (SHE).
The dissolved O2 concentration was established as 2 mg LÀ1
and calculated Erev
for the O2/H2O system is +876 mV (SHE). The total dissolved iron (ICP analysis)
in the MJ was $25 mg LÀ1
but because of the possibility of complexing with
organics and other species in the MJ the concentration of dissolved iron as Fe2+
is
probably much less than 25 mg LÀ1
. The exact value of [Fe2+
] was not established
and for curve synthesis 0.056 mg LÀ1
(10À6
M) was used. This value can be justified
in that it is accompanied by an acceptable i0 of 1.94 · 10À7
A cmÀ2
for the Fe/Fe2+
system. Erev for the Fe/Fe2+
system is À623 mV (SHE)).

Although the Pourbaix diagram for pure iron corresponds to equilibria a 25 °C
and here a higher temperature (35 °C) and carbon steel is involved, the diagram
can be used as an approximate guide to the corrosion susceptibility of the mild steel.
Thus at pH = 4.9 and Ecorr = À445 mV (SHE) (and for more positive potentials ap-
plied in the experiment) the anodic reaction is principally the dissolution of Fe to
form Fe2+
ions.
The cathodic portion of the experimental curve (Fig. 3a) has some appearance of
linearity but this does not indicate a Tafel region. Tafel behaviour refers to one reac-
tion, and in this case the cathodic portion of the experimental curve is actually the
sum of two curves (oxygen reduction and hydrogen evolution). This is made clear
in Fig. 3b where the deconvoluted anodic and cathodic components of the synthes-
ised curve seen in Fig. 3a are shown. The deconvolution reveals that at Ecorr the
dominant cathodic reaction driving the corrosion is oxygen reduction.
(a) (b)
(c) (d)
Fig. 3. Case 3a. (a) Experimental and synthesised polarisation curves for mild steel in dormant MJ at
35 °C (100 min exposure; 2 mg LÀ1
O2) assuming presence of non-passive surface film. (b) Deconvolution
of synthesised polarisation curve for mild steel in dormant MJ at 35 °C assuming presence of non-passive
surface film. (c) Experimental and synthesised polarisation curves for mild steel in dormant MJ at 35 °C
(100 min exposure; 2 mg LÀ1
O2) assuming no surface film. (d) Deconvolution of synthesised polarisation
curve for mild steel in dormant MJ at 35 °C assuming no surface film.

The black film will impart a resistance to the WE and, as in Case 2, the values of
the recorded anodic potentials will be greater than the true values. The effect of this
resistance polarisation can be calculated using SYMADEC and 90 X were required
to synthesise and match the anodic portion of the experimental curve shown in Fig.
3a. The anodic portion (assuming no film and therefore no ohmic resistance) can
also be synthesised and this, a straight line exhibiting Tafel behaviour, is shown to-
gether with the matched cathodic portion (and also the experimental curve) in Fig.
3c. The corrosion rate can be estimated by extrapolating this line to the corrosion
potential. The corresponding deconvoluted anodic and cathodic curves are shown
in Fig. 3d.
4.4. Case 3(b): Mild steel corroding in flowing, mixed cane sugar juice (active
corrosion, non-passive film formation)
An experimental polarisation curve originally recorded potentiodynamically by
Cash [24] for a mild steel WE in mixed cane sugar juice (MJ), open to air at 43 °C
was scanned. In contrast to Case 3(a) the MJ was flowing through a laboratory
flow-rig and conditions for recording the experimental curve were as follows. The
flow-rig was constructed from black polyethylene tubing (18 mm i.d.). The WE
and CE were mild steel discs (0.95 cm2
) mounted in the electrode assembly (PVC tub-
ing) and ground so that they were flush with the internal wall of the tubing. A com-
mercial Ag/AgCl reference electrode with the tip mounted as close as possible to the
WE was used. In this experiment the MJ flow rate was 24 dm3
minÀ1
(2 m sÀ1
).
The WE was abraded with 600 grade SiC paper, degreased with AR grade acetone
and then immediately exposed to the flowing MJ. The juice temperature (43 °C), dis-
solved oxygen concentration (3 mg LÀ1
) and Ecorr (À440 mV (SHE)) were steady
after approx. 30 min. After approx. 100 min exposure to the flowing juice the WE
was polarised anodically from Ecorr (as in Case 3(a) to least disturb any film depos-
ited on the WE during the exposure period). This was followed by the cathodic scan
when Ecorr had returned to within ±5 mV of its previous value. The scan rate was
60 mV minÀ1
.
A black film was deposited on the WE during establishment of the corrosion po-
tential and its resistance acts to make the values of the recorded anodic potentials
greater than the true values. As in Case 3(a) the film resistance can be calculated using
SYMADEC and the curved anodic portion of the experimental curve can be
matched. The anodic portion can also be synthesised assuming no film (and therefore
no ohmic resistance) and the result is a straight line exhibiting Tafel behaviour. The
experimental curve in Fig. 4a is shown as it would appear if there was no film. The
matched cathodic portion of the experimental curve is also shown in Fig. 4a. The cor-
rosion rate can now be estimated by extrapolating the anodic Tafel line to the corro-
sion potential. Because of the increased temperature and movement of the corrodent
iL is now greater than that seen in Case 3(a). The corresponding deconvoluted anodic
and cathodic curves are shown in Fig. 4b.
tion curve are listed in Table 2. The dominant anodic reaction is again iron dissolu-

(a)
(b)
Fig. 4. Case 3b. (a) Experimental and synthesised polarisation curves for mild steel in flowing MJ at 43 °C
(100 min exposure; 3 mg LÀ1
O2) assuming no surface film. (b) Deconvolution of synthesised polarisation
curve for mild steel in flowing MJ at 43 °C (100 min exposure; 3 mg LÀ1
O2) assuming no surface film.

tion driven by H+
ion and O2 reduction. The pH of the MJ was 5.52, and the calcu-
lated Erev for the H2/H+
system is À346 mV (SHE). The O2 concentration was
3 mg LÀ1
and the calculated Erev for the O2/H2O system was +819 mV (SHE).
As in Case 3(a) an [Fe2+
] = 0.056 mg LÀ1
was used and the calculated Erev for the
Fe/Fe2+
system was À629 mV (SHE).
Using the Pourbaix diagram for pure iron at 25 °C as a guide, it is a reasonable
assumption that at 43 °C, pH = 5.52, and Ecorr = À440 mV (SHE) (and for more po-
sitive potentials applied in the experiment) the anodic reaction is principally the dis-
solution of Fe to form Fe2+
ions.
As in Case 3(a) the cathodic portion of experimental curve appears to show some
linearity, but again this does not indicate a Tafel region as the cathodic portion of
the experimental curve is the sum of two curves (oxygen reduction and hydrogen
evolution). The deconvoluted cathodic curve seen in Fig. 4b shows that at Ecorr
the dominant cathodic reaction driving the corrosion is again oxygen reduction.
Fig. 4b also clearly indicates that when the potential is made more negative than
Ecorr the hydrogen evolution reactionÕs contribution to the total cathodic current be-
comes increasingly important. At a sufficiently negative potential this curve will also
come under complete diffusion control.
Cases 3(a) and 3(b) are good examples of situations in which the experimental
curve does not provide a Tafel region which in turn can be used to estimate corrosion
rate. Although the anodic portions of the curves are indicative of active corrosion
they are curved due to deposition of non-passive films. As for the cathodic portions
they are the sum of two reactions. The presence of a straight-line region is simply
fortuitous.
4.5. Case 4(a): Low-alloy steel corroding in oxygen-containing, simulated steam
turbine condensate (active corrosion, induced passivation and pitting)
The experimental curve (Fig. 5a) was recorded potentiodynamically by Otieno-
Alego et al. [32] for A-470 turbine rotor disc steel (0.24% C, 1.8% Cr, 3.68% Ni,
0.46% Mo, 0.3% Mn, 0.12% V, 0.0004% S, 0.0004% P, 0.05% Si) immersed in a syn-
thetic steam turbine condensate containing 2 ppm NaCl, 2 ppm Na2SO4, 2 ppm
NaOH and 5 ppm SiO2.
A single compartment Perspex cell (800 cm3
) fitted with a Perspex lid was used.
The WE (10 mm dia.) and Pt counter electrode (1 cm2
) were mounted in chemical
resistant epoxy resin and immersed in the test solution using a Perspex holder. A sat-
urated calomel electrode (SCE) connected to a Luggin capillary was used as the re-
ference electrode. The temperature was 30 °C and the solution pH = 9.0. Bottled
nitrogen gas (containing traces of oxygen) was passed continuously through the cor-
rodent and this resulted in a dissolved oxygen level of approximately 0.01 mg LÀ1
.
The WE was abraded with 1200 grade SiC paper, degreased with AR grade acetone,
inserted in the solution and then immediately pre-polarised at À756 mV (SHE) for
10 min to remove any air-formed oxide film. After reaching a steady Ecorr (approx.
1 h) the corroding WE was polarised cathodically. This was followed by anodic

(a)
(b)
PotentialvsSHE(mV)PotentialvsSHE(mV)
Fig. 5. Case 4a. (a) Experimental and synthesised polarisation curves for low-alloy steel in synthetic
condensate at 30 °C (0.01 mg LÀ1
O2). (b) Deconvolution of synthesised polarisation curve for low-alloy
steel in synthetic condensate at 30 °C (0.01 mg LÀ1
O2).

polarisation when Ecorr had returned to within ± 5 mV of the previous value. The
polarisation scan rate was 10 mV minÀ1
.
In this case there are two oxidants (H+
ion and small amount of O2) driving cor-
rosion. Ecorr was À539 mV (SHE). Although low-alloy steel is corroding, the mate-
rial is approximately 93% Fe, and the Pourbaix diagram for pure iron is a reasonable
guide to corrosion behaviour and subsequent anodic polarisation. The diagram
shows that at pH = 9.0 and for an Ecorr $ À539 mV (SHE) pure iron is actively cor-
roding to form Fe2+
ions. Further, if the WE is made more positive iron passivates
with the formation of precipitated Fe2O3 Æ nH2O (or Fe(OH)3).
The shape of the experimental polarisation curve (Fig. 5a) supports the use of the
iron Pourbaix diagram to predict corrosion behaviour. The curve suggests active cor-
rosion at Ecorr and indicates that polarisation in the positive direction (by means of
the potentiostat) results in a classical active/passive transition. This is followed at
more positive potentials by a rapid increase in current density suggesting in the pres-
ence of ClÀ
pitting corrosion. Otieno-Alego et al. [32] reported that pits were ob-
served on the WE after anodic polarisation.
Parameters and data required to synthesise and match Otieno-Alego et al.Õs [32]
experimental polarisation curve based on the above assumptions (Fig. 5a) are listed
in Table 2. Erev for the H2/H+
system is À541 mV (SHE) with Erev for the O2/H2O
system +589 mV (SHE). Again, using a minimum value of [Fe2+
] = 0.056 mg LÀ1
,
Erev for the Fe/Fe2+
The deconvoluted anodic and cathodic components of the synthesised curve are
shown in Fig. 5b. This shows that at Ecorr the corrosion is driven mainly by reduction
of the small amount of oxygen in solution (the reduction of H+
ion contributes rel-
atively little to the total cathodic current density at this potential). Further, both
cathodic reactions are under complete activation control at the corrosion potential.
The cathodic portion of the experimental curve is a composite one and it is futile
searching for a linear ÔTafelÕ region to ascertain corrosion rate. The anodic portion
of the experimental curve before onset of passivation is also curved and cannot be
used to estimate corrosion rate.
An estimation of the corrosion current density may be ascertained (Fig. 5b) from
the intersection of Ecorr with the synthesised anodic and oxygen curves.
4.6. Case 4(b): Low-alloy steel corroding in oxygen-containing, simulated steam
turbine condensate (spontaneous passivation and induced pitting)
The experimental polarisation curve (Fig. 6a) was recorded by Otieno-Alego et al.
[32] as for Case 4(a) except that the oxygen concentration was increased from 0.01 to
0.20 mg LÀ1
by passing a nitrogen/air mixture through the corrodent. Again there
are two oxidants (O2 and H+
) driving the corrosion and the iron Pourbaix diagram
shows that at pH $ 9.0 and Ecorr = À141 mV (SHE), pure iron spontaneously pass-
ivates with the formation of Fe(OH)3.
The more positive Ecorr (À141 mV (SHE) versus À539 mV (SHE) for Case 4(a))
and the shape of the experimental curve (Fig. 6a) and suggests that the higher oxygen
level has been instrumental in passivating the low-alloy steel WE upon its immersion

(b)
(a)
Fig. 6. Case 4b. (a) Experimental and synthesised polarisation curves for low-alloy steel in synthetic
O2).

in the corrodent. The experimental curve also suggests that subsequent polarisation
with the potentiostat in the positive direction from Ecorr results in localised corrosion
at approximately 90 mV (SHE). This behaviour was supported by the existence of
pits seen on the WE after anodic polarisation to $+250 mV (SHE) [32].
Assuming spontaneous passivation in the corrodent followed by induced pitting,
the polarisation curve was synthesised and matched to the experimental one (Fig.
6a). Parameters and data required to synthesise and match the experimental curve
are listed in Table 2. In this case values of some parameters (e.g., the primary pas-
sivation potential for the active/passive transition) cannot be estimated from the
experimental curve. Erev for the H2/H+
system is À541 mV (SHE), with Erev for
the O2/H2O system +609 mV (SHE). Again, using a minimum value of [Fe2+
] =
0.056 mg LÀ1
, Erev for the Fe/Fe2+
The deconvoluted anodic and cathodic components of the synthesised curve are
shown in Fig. 6b and this shows that at Ecorr the corrosion is driven overwhelmingly
by oxygen reduction. The cathodic oxygen curve cuts the anodic one in the passive
region. This masks the active/passive portion of the anodic curve and, unlike in the
previous case, no estimate of the primary passivation potential and the complete pas-
sivation potential can be obtained from the experimental curve. The deconvolution
shows that at Ecorr oxygen reduction is under complete activation control.
Deconvolution clearly shows that there are no ÔTafel regionsÕ on the experimental
curve. The alloy, on exposure to a synthetic steam turbine condensate in which the
oxygen concentration is 0.20 mg LÀ1
spontaneously passivates, and the complete
passivation current density as seen in Fig. 6b may be taken as an estimate of the
its corrosion rate.
The small ÔstepÕ at approximately À450 mV (SHE) arises from the closeness of the
tip of the Ôpassivation peakÕ to the cathodic oxygen curve.
4.7. Case 4(c): Low-alloy steel corroding in oxygen-containing, simulated steam
turbine condensate (spontaneous passivation and spontaneous pitting)
The experimental polarisation curve (Fig. 7a) was recorded by Otieno-Alego et al.
[32] as for Cases 4(a) and 4(b) except that the oxygen concentration was further in-
creased to 7.9 mg LÀ1
. Ecorr was À80 mV (SHE) and the Pourbaix diagram shows
that at this potential at pH $ 9.0 pure iron spontaneously passivates with the forma-
tion of Fe(OH)3.
The shape of the anodic portion of the experimental curve (Fig. 7a), coupled with
the more positive corrosion potential (compared with the previous case), suggests
that the low-alloy steel on immersion in the corrodent may have undergone sponta-
neous passivation followed by localised (pitting) corrosion. Thus there is now suﬃ-
cient dissolved oxygen to drive Ecorr to a value either equal to, or more positive than
the pitting potential. Further work by Otieno-Alego et al. [32] showed that pits
formed a few minutes after immersion.
The synthesised/matched curve (assuming spontaneous passivation/pitting) is
shown in Fig. 7a. Parameters and data required to synthesise and match the exper-
imental curve are listed in Table 2. Erev for the H2/H+
system is À541 mV (SHE) with

(a)
(b)
Fig. 7. Case 4c. (a) Experimental and synthesised polarisation curves for low-alloy steel in synthetic
O2).

Erev for the O2/H2O system +642 mV (SHE). Using a minimum value of
[Fe2+
] = 0.056 mg LÀ1
system is À621 mV (SHE). Here it is
impossible to estimate the values of parameters for the active/passive transition
and pitting from the experimental curve. The iron breakdown/pitting potential
and Ecorr were taken as coincident. The deconvoluted anodic and cathodic portions
are shown in Fig. 7b and this reveals that at Ecorr the localised corrosion is domi-
nated by reduction of oxygen, and the cathodic reaction is under complete activation
control at the corrosion potential. Because the alloy is pitting the concept of corro-
sion rate (which applies to uniform corrosion) is meaningless.
The cathodic portion of the experimental curve exhibits (as in the previous case) a
small ÔstepÕ at approximately À350 mV (SHE).
Cases 4(a), 4(b) and 4(c) show how the oxidant concentration (here mainly oxy-
gen) can determine whether an alloy on immersion in the corrodent experiences ac-
tive corrosion, spontaneous passivation, or spontaneous passivation/pitting.
4.8. Case 4(d): Low carbon steel corroding in oxygenated pure water also
containing a basic detergent (spontaneous passivation)
The experimental curve shown in Fig. 8a was recorded potentiodynamically by
one of the current authors (HJF) [33] for 1020 carbon steel immersed in distilled
water containing a commercial detergent (25 mg LÀ1
; 25 °C). The solution was open
to air ([O2] $ 8 mg LÀ1
) and the pH = 12.3. The WE (abraded with 1200 grade SiC
paper and degreased with AR grade acetone) was immediately placed in the test solu-
tion and pre-polarised at À756 mV (SHE) for 5 min to remove residual oxide film.
The electrode was then polarised from this potential at 20 mV minÀ1
to approxi-
mately +740 mV (SHE).
In this case at pH = 12.3 and dissolved oxygen is the main oxidant driving the cor-
rosion. Ecorr is apparent from the experimental curve (À345 mV (SHE)). The Pour-
baix diagram for pure Fe shows that at pH = 12.3, and as the potential is made more
positive, the metal oxidises to form firstly soluble HFeOÀ
2 ions, followed by passiv-
ation due to deposition of protective Fe(OH)3.
The shape of the experimental curve and the value of Ecorr suggest that the high
oxygen level polarises and then spontaneously passivates the WE when it is im-
mersed in the corrodent. Fig. 8a also indicates that induced anodic polarisation from
Ecorr to $+740 mV (SHE) was insufficient to result in pitting. In addition HJF [33]
did not observe any pits on the WE after the experiment.
The synthesised and matched polarisation curve (assuming passivation) is also
shown in Fig. 8a and the deconvoluted anodic and cathodic components are shown
in Fig. 8b. Parameters and data required to synthesise and match the experimental
curve are listed in Table 2. Erev for the H2/H+
system is À727 mV (SHE) with Erev
for the O2/H2O system +447 mV (SHE). Again, using a minimum value of
[Fe2+
] = 0.056 mg LÀ1
system is À618 mV (SHE). In this case
estimating the values of parameters for the active/passive transition from the experi-
mental curve is less difficult than in the previous two cases. Fig. 8b shows that at
Ecorr the corrosion is driven entirely by reduction of oxygen. This example can be

(a)
(b)
Fig. 8. Case 4d. (a) Experimental and synthesised polarisation curves for mild steel in distilled water
containing commercial detergent at 25 °C (8 mg LÀ1
O2). (b) Deconvolution of synthesised polarisation
curve for mild steel in distilled water containing commercial detergent at 25 °C (8 mg LÀ1
O2).

compared with Case 4(b). The size of the passivation peak is markedly reduced be-
cause; at the higher pH (12.3 versus 9) fewer HFeOÀ
2 ions are required to precipitate
the hydrated oxide. Also, because the ÔnoseÕ is very small the cathodic portion of the
experimental curve does not exhibit a step as seen in Cases 4(b) and 4(c).
4.9. Case 5: Low carbon steel corroding in oxygen-containing water (induced
passivation and induced pitting)
The experimental curve shown in Fig. 9a was recorded potentiodynamically by
one of the current authors (HJF) [33] for 1020 carbon steel immersed in distilled
water (open to air) at 40 °C containing 25 mg LÀ1
NaCl and 150 mg LÀ1
of an oxy-
gen scavenger (activated hydrazine hydrate (LEVOXINTM
)). The oxygen concentra-
tion during polarisation was measured as $0.01 mg LÀ1
and the pH of the solution
was 8.8.
The WE (abraded with 1200 grade SiC paper and degreased with AR grade ace-
tone) was placed in the test solution and pre-polarised at À580 mV (SHE) for 5 min
to remove any residual oxide film. The electrode was then immediately polarised in
the positive direction (20 mV minÀ1
) through to approximately +300 mV (SHE).
The activated hydrazine hydrate, in addition to reducing the oxygen concentra-
tion, reacted with the water raising the pH of the solution to 8.8. Although the
amount of oxygen remaining in solution is small it will act in conjunction with the
H+
ions to drive the corrosion.
At this point it should be noted that the procedure adopted for recording an
experimental curve can add to difficulties in its interpretation. Here the corrosion po-
tential was not established by letting the WE stabilise after pre-polarisation, and as a
result it might be thought that the experimental curve shown in Fig. 9a exhibits three
such potentials, and perhaps two Ôactive/passive transitionsÕ. This dilemma can be
partly resolved by referring to LieningÕs schematic diagrams [1]. He shows that such
a curve will arise when the concentration (diffusion) controlled portion of the true
cathodic curve intersects the true anodic curve at two points on the active/passive
ÔnoseÕ, and the activation-controlled portion intersects the passive region. There is
only one active/passive transition, and what appears to be a second transition (at
more positive potentials) is actually a Ôcathodic loopÕ.
In the current example the corrosion potential was established in a separate experi-
ment (HJF [33]) and corresponded to the most negative of the Ôthree possibilitiesÕ
(À503 mV (SHE)) seen in Fig. 9a. From the Pourbaix diagram it can be assumed
that at this potential and for pH = 8.8 and in the presence of the dissolved oxygen
the low carbon steel is actively corroding to Fe2+
. It can also be assumed from the
shape of the curve that induced polarisation in the positive direction from Ecorr re-
sults in an active/passive transition followed by a cathodic loop. At even higher ap-
plied potentials film breakdown occurs at approximately +138 mV (SHE). (Note:
pits were observed by HJF [33] on the WE after induced polarisation to +300 mV
(SHE).)
Parameters and data required to synthesise and match the experimental curve
(Fig. 9a) are listed in Table 2. Erev for the H2/H+
system is À547 mV (SHE) with

(a)
(b)
Fig. 9. Case 5. (a) Experimental and synthesised polarisation curves for mild steel in NaCl salt solution
plus O2 scavenger (LEVOXIN) at 40 °C (0.01 mg LÀ1
O2). (b) Deconvolution of synthesised polarisation
curve for mild steel in NaCl salt solution plus O2 scavenger (LEVOXIN) at 40 °C (0.01 mg LÀ1
O2).

Erev for the O2/H2O system +580 mV (SHE). Using a minimum value of
[Fe2+
] = 0.056 mg LÀ1
system is À627 mV (SHE). In this case
it is again impossible to estimate values of parameters for the active/passive tran-
sition from the experimental curve. The deconvoluted anodic and cathodic com-
ponents of the synthesised/matched curve are shown in Fig. 9b. This clearly
shows how induced polarisation from the pre-polarisation potential (À580 mV
(SHE)) results in a diminution in both the rate of H2 evolution and oxygen
reduction. At Ecorr the corrosion is seen to be driven mainly by the oxygen reduc-
tion reaction. At more positive potentials there is sufficient Fe2+
ion in solution to
induce passivation and this is followed at higher potentials by pitting in the
aggressive ClÀ
solution. Fig. 9b also shows that the actual oxygen curve is under-
going combined activation and concentration polarisation when it intersects with
the actual anodic curve in the passive region (where a stable passive film has
formed) and at a more negative potential (where the film is unstable). These
points of intersection are responsible for the cathodic loop with the current den-
sity for oxygen reduction exceeding the anodic current density between the upper
two intersection points.
4.10. Case 6: Low carbon steel corroding in oxygen-containing water (spontaneous
passivation and pitting)
The experimental curve shown in Fig. 10a was recorded potentiodynamically
[21,22] for 1020 carbon steel immersed in distilled water at 40 °C containing
25 mg LÀ1
NaCl and 100 mg LÀ1
of a commercial inhibitor for iron (zinc phosphi-
nocarboxylic acid (ZnPCA)). An extra 15 mg LÀ1
of zinc was added (as zinc sul-
phate) and the pH was adjusted to 7.0 with dilute KOH solution. The test
solution was open to air and the oxygen concentration was measured at $8 mg LÀ1
.
The WE (abraded with 1200 grade SiC paper and degreased with AR grade acetone)
was placed in the test solution and pre-polarised at À600 mV (SHE) for 5 min to re-
move any residual oxide film. The electrode was then polarised in the positive direc-
tion (20 mV minÀ1
) to approximately +100 mV (SHE).
In this case at pH = 7.0 and [O2] = 8 mg LÀ1
the main oxidant driving the corro-
sion is dissolved oxygen. Although Ecorr was not measured after the cathodic, pre-
polarisation step, its value is obvious from the experimental curve (À142 mV
(SHE)). The low carbon steel can be expected to corrode similarly to pure iron
and from the Pourbaix diagram for Fe at pH = 7.0, and as the potential is made
more positive (from approximately À560 to +100 mV (SHE)), Fe is oxidised to
Fe2+
ions.
Phosphinocarboxylic acid (PCA), combining both the phosphino functional
group and the carboxylic functional group in one molecule, has been used as a
corrosion inhibitor for steel in cooling water and it is assumed that the molecule
is chemisorbed on the metal to act principally as an anodic inhibitor [34]. Inhi-
bition efficiency of PCA is markedly increased by the addition of zinc (optimum
inhibition in the range approximately 20–40% by weight of Zn). It has been pro-
posed that Zn(II) reacts with the PCA and the resulting zinc complex (ZnPCA)

(a)
(b)
Fig. 10. Case 6. (a) Experimental and synthesised polarisation curves for mild steel in NaCl salt solution
plus Zn-augmented ZnPCA at 40 °C (8 mg LÀ1
O2). (b) Deconvolution of synthesised polarisation curve
for mild steel in NaCl salt solution plus Zn-augmented ZnPCA at 40 °C (8 mg LÀ1
O2).

is also chemisorbed on the steel surface, reducing the rate of both the cathodic
and anodic corrosion reactions [34]. In the present case the cathodic portion of
the experimental curve (Fig. 10a) reveals a cathodic ÔdipÕ at approximately
À500 mV (SHE). Liening [1] notes that such a ÔdipÕ can arise when there is an
active/passive transition, and the current density of the concentration controlled
portion of the cathodic curve is just greater than that at the tip of the active/pas-
sive ÔnoseÕ. Evidence for the Zn-augmented ZnPCA promoting the formation of a
passive film is provided by the relatively positive value of Ecorr and by the shape
of the anodic portion of the experimental curve. The latter suggests active corro-
sion at Ecorr deriving from adsorption of aggressive ClÀ
ions and subsequent
localised corrosion. Under these conditions Ecorr is more positive than the pitting
potential. At the conclusion of the polarisation pits were observed on the steel
[21,22].
It can be assumed therefore that the low-alloy carbon steel on immersion in the
corrodent in the presence of inhibitor and chloride ions undergoes spontaneous pas-
sivation/pitting. On this basis the synthesised/matched curve is shown in Fig. 10a
and parameters and data required for synthesis are listed in Table 2. Erev for the
H2/H+
system is À435 mV (SHE) with Erev for the O2/H2O system +761 mV
(SHE). Using a minimum value of [Fe2+
] = 0.056 mg LÀ1
sys-
tem is À627 mV (SHE). It is again impossible to estimate the values of parameters
for the active/passive transition and film breakdown from the experimental curve.
Deconvolution (Fig. 10b) reveals the dominance of the oxygen reaction and shows
how as the potential is made more positive the rates of hydrogen evolution and
reduction of oxygen decrease. The figure also shows how the ÔdipÕ is generated with
the current density of the concentration controlled portion of the oxygen curve just
greater than that at the tip of the active/passive ÔnoseÕ. Finally, Fig. 10b also shows
the corrosion potential more positive than the pitting potential resulting in sponta-
neous pitting.
5. Conclusions
• Experimental polarisation curves for the corrosion system Fe/H2O/H2/O2 can be
synthesised using the appropriate mathematical relationships and kinetic and
thermodynamic data for the reactions involved in the corrosion process.
• Deconstruction of the synthesised, accurately matched curve reveals the true ano-
dic and cathodic components operative in the following corrosion systems: active
corrosion; active corrosion and non-passive film formation; active corrosion fol-
lowed by induced passivation and induced pitting; spontaneous passivation and
induced pitting; spontaneous passivation and spontaneous pitting. Curves exhi-
biting either a cathodic loop or a cathodic dip can also be analysed.
• The accurately analysed curves replace schematic representations and are a valu-
able reference source for the interpretation of experimental curves for the aqueous
corrosion of pure iron/carbon/low-alloy steels.

Acknowledgments
The authors wish to thank the School of Physical and Chemical Sciences for pro-
viding facilities for the writing of this paper. We would also like to acknowledge
those researchers whose results have been used in our analysis of experimental polar-
isation curves.
References
[1] E.L. Liening, in: B.J. Moniz, W.I. Pollock (Eds.), Process Industries Corrosion, NACE, 1986,
p. 85.
[2] O.W. Siebert, in: G.S. Haynes, R. Baboian (Eds.), Laboratory Corrosion Tests and Standards,
ASTM STP 866, ASTM, Philadelphia, 1985, p. 65.
[3] O.F. Devereux, Corrosion 35 (1979) 125.
[4] O.F. Devereux, K.Y. Kim, Corrosion 36 (1980) 262.
[5] J.G. Hines, J.H. Cleland, Proc. 8th Int. Congr. Metall. Corros. Mainz. 2 (1981) 1959.
[6] S.O. Berhardsson, R. Mellstrom, in: F. Mansfeld, U. Bertocci (Ed.), Electrochemical Corrosion
Testing, ASTM STP727, 1981, p. 352.
[7] R.S. Munn, Mater. Perform. 22 (August) (1982) 29.
[8] O.F. Devereux, K.Y. Kim, K.S. Yeum, Corros. Sci. 23 (1983) 205.
[9] J.G. Hines, Br. Corros. J. 18 (1983) 10.
[10] J.H. Cleland, C. Edeleanu, Br. Corros. J. 18 (1983) 15.
[11] P.A. Brook, J.S.L. Leach, B.R. Pearson, in: Proc. 166th Meeting of the Electrochem. Soc., Louisiana,
USA, 1984, p. 243.
[12] H.E.H. Bird, B.R. Pearson, P.A. Brook, Corros. Sci. 28 (1988) 81.
[13] K.S. Yeum, O.F. Devereux, Corrosion 45 (1989) 478.
[14] B.R. Pearson, P.A. Brook, Corros. Sci. 32 (1991) 387.
[15] K.R. Trethewey, J.S. Keenan, Corros. Prev. Control. 89 (August) (1991).
[16] K.R. Trethewey, J.S. Keenan, I. Wilson, Corros. Prev. Control. 115 (October) (1991).
[17] K.R. Trethewey, J.S. Keenan, in: R.S. Munn (Ed.), Microcomputer-based Corrosion Modelling
Applied to Polarisation Curves, ASTM STP 1154, ASTM, Philadelphia, USA, 1992, p. 113.
[18] S. Nesic, J. Postlethwaite, S. Olsen, Corrosion 52 (1996) 280.
[19] A. Anderko, P. McKenzie, R.D. Young, Corrosion 57 (2001) 202.
[20] D.W. Shoesmith, in: CorrosionMetals Handbook, vol. 13, ASM International, Metals Park, OH,
USA, 1987, p. 29.
[21] H.J. Flitt, in: Proc. 7th RACI Electrochemistry Conf., Australia, 1988, p. 287.
[22] H.J. Flitt, G.A. Cash, D.P. Schweinsberg, in: Proc. 7th European Symp. on Corrosion Inhibitors,
Ann. Univ. Ferrara, Italy, 1990, p. 1435.
[23] V. Otieno-Alego, G.A. Hope, H.J. Flitt, G.A. Cash, D.P. Schweinsberg, Australasian Corrosion
Association Conference No. 31, Sydney, paper F09, 1991.
[24] G.A. Cash, Ph.D. Thesis, Griﬃth University, Brisbane, Queensland, Australia.
[25] V. Otieno-Alego, G.A. Hope, H.J. Flitt, G.A. Cash, D.P. Schweinsberg, Corros. Sci. 33 (1992)
1719.
[26] V. Otieno-Alego, G.A. Hope, H.J. Flitt, D.P. Schweinsberg, Corros. Sci. 34 (1993) 1289.
[28] V.S. Bagotzky, Fundamentals of Electrochemistry, Plenum Press, New York, 1993.
[29] M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, NACE, Houston, 1974.
[30] D.P. Schweinsberg, V. Ashworth, Corros. Sci. 28 (1988) 539.
[31] R. Bandy, D.A. Jones, Corrosion 32 (1976) 126;
see alsoD.A. Jones, Principles and Prevention of Corrosion, Macmillan, New York, 1992.

[33] Personal Communication from Dr. H.J. Flitt.
[34] (a) A. Harris, A. Marshall, Corros. Prev. Control. (June) (1980) 18;
(b) A. Harris, A. Marshall, Corros. Prev. Control. (August) (1980) 17.

Polarisation curve interpretation

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Polarisation curve interpretation (20)

Recently uploaded (20)

Polarisation curve interpretation