Electrochimica Acta, 1961. Vol. 4, pp. 325 to 361. Pergamon Pre*r Ltd.

Printed in Northern Ireland

THE ELECTRODE KINETICS OF THE DEPOSITION

AND DISSOLUTION OF IRON*
J. O’M. BOCKRIS, D. DRAZIC~ and A. R. DE~PIC~
John Harrison Laboratory of Chemistry, University of Pennsylvania
Philadelphia, Pa., U.S.A.

Abstract-The rate of the deposition and dissolution of Fe has been measured as a function of potential
for constant solution compositions : Fee+ concentration at constant pH; pH at constant Fe*+
concentration; and the presence of the anions SO1S-, Cl-, ClO&-, AC- and NOI- at constant Fe*+
concentration. The method of measurement was that of galvanostatic transients. The rate of the
hydrogen evolutidn reaction, both from HBO+ and H,O, has been examined during these transients, as
a function of potential and pH.
The ohmic correction to the total potential was the subject of separate measurements. Cathodic
transients manifested regions of Hf discharge from HsO +, Fez+ deposition and H+ discharge from
HaO. Anodic transients exhibited a characteristic maximum, removed by successive anodic pulses.
Capacitance values, calculated from initial gradients of potential/time transients, are some five times
higher than those expected. Transition times are consistent with Sand’s equation in respect of Fe*+
deposition.
Steady state anodic Tafel lines have slopes of about IRTIF; the slope becomes RT/2F in impure
solutions. The cathodic slope (after correction for the partial current density due to H discharge) is
apparently RT/F in (total) current density regions above those corresponding to the limiting current
density for in. When account is taken of the pH change at the cathode surface due to H discharge,
the cathodic Tafel slope becomes ZRT/F.

aecorr a log i00,, =

- = 0.06; - -0.48; at pH = 3 and cpe+s = 0.5 M, the velocity of
~PH ~PH
Fe;+ + 2eo- + Fe depends upon the anion present in the order CIO1- > Sod*- > Cl- >
AC- > NOa-. The deposition of H from Ho0 is slower than that from HIOf to an amount
equivalent to a heat of activation for proton discharge which is 9 kcal/mole greater for Ha0 than
from H,O+.
It is shown that the linear dependence of the reaction rate at constant potential and cpsa+ upon pH
cannot be due to dissolved H in the Fe. General kinetic expressions are deduced for an overall
electrode reaction Ma+ + 2e,- + M, the mechanism of which involves OH- ions: they relate Tafel
slopes of anodic and cathodic partial reactions to the reversible potential, corrosion potential and
corrosion rate. Application of them to data in the literature suggests that for Fes+ + 2e,- + Fe,
CL,= $ or 8 and a0 = g or f respectively. Seven paths and mechanisms for this reaction are proposed
and their kinetic consequences evaluated. Attention is particularly drawn to inconsistencies in one
previous suggestion. If the anodic Tafel slope is RT/2F, and a cathodic Tafel slope of RT/ZF, an order
of reaction 2 with respect to Fe e+ and 1 with respect to OH-, would also be expected, if amultielcctron
transfer reaction is not acceptable. A mechanism consistent with the results reported here is: Fe +
OH- + FeOH + e,-; FeOH -+ FeOH+ + et,-; FeOHf + Fe++ + OH-. This path is near in
energy requirements to another, in which b. = RT/2F. Anion effects, except that of NOs-, may be
consistent with a model in which specific adsorption of the anion changes the area free for ion-metal
exchange.
The difference in the heat of activation of hydrogen deposition on Fe from H*O+ and HI0 is
consistent with rate-determining proton discharge.

* Manuscript received 11 August 1960.

t Present address: Faculty of Technology, University of Belgrade, Belgrade, Yugoslavia.
325
326 J. O’M. BOCKRIS,D. DRAZIC and A. R. DESPIC

R&sume-On a mesure la vitesse de d&p& et de dissolution du fer en fonction du potentiel pour des
compositions constantes de la solution: concentration de Fee+ a pH constant; pH a concentration
constante de Fez+; et presence des anions SOp2-, Cl-, C104-, Al- et NO,- a concentration constante
de Fez+. La methode de mesure Btait la mtthode galvanostatique des courants transitoires. On a
examine, pendent ces courants transitoires, la vitesse de la reaction d’tvolution d’hydrogene a partir
de HsO+ et de H,O en fonction du potentiel et du pH.
La correction ohmique a apporter au potentiel total a fait l’objet de mesures separees. Les courants
transitoires cathodiques comportaient des regions de d&charge de H+ a partir de H30+, de depot de
F2+ et de d&charge de H+ a partir de H,O. Les courants transitoires anodiques comportaient un
maximum caracteristique, supprime par des pulsations anodiques successives.
Les valeurs de capacitance, calculees a partir des gradients initiaux des courbes transitoires
potentiel-temps, sont environ cinq fois plus &levees que celles que I’on s’attendair a obtenir. Les
temps de transition sont compatibles avec Equation de Sand en ce qui concerne le depot de Fez+.
Les droites de Tafel anodiques en Ctat stationnaire ont des pentes d’environ 3 RTIF; les pentes
deviennent RT/ZF dans les solutions impures. La pente cathodique (ap& correction pour la densite
de courant partielle due a decharge de H) est apparemment RT/F dans les regions de densite de
courant (totale) audessus de celles correspondant a la densite de courant limite pour in. Quand on
tient compte du changement de pH a la surface de la cathode dii a la d&charge de H, la pente de Tafel
cathodique devient 2RTIF.

aeeorr
--
ai0g icorrcorrosion
= 0,06; - = -0,48; ii pH = 3 et C&+ = 0,5 M, le vitesse de Fee+
8PH 8PH
+ 2 e,- Fe depend de I’anion present dans l’ordre ClO,- S0,2- Cl- AC- NO,-. Le depot de H a
partir de H,O est plus lent qu’a partir de HsO+ dune quantite equivalente a une chaleur d’activation
pour la d&harge du proton qui serait de g kcal mole-l plus Be&es pour Hz0 que pour H,O+.
On montre que la dependence lineaire de la vitesse reactionelle a potentiel et C&a+ constants par
rapport au pH ne peut pas &tre due a H dissous dans Fe. Des expressions cinetiques g&&ales sont
deduites pour une reaction d’electrode globale Mz+ + 2e,- + M, dont le mecanisme implique les
ions OH-: elles relient les pentes de Tafel des reactions partielles anodique et cathodique au potentiel
reversible, au potentiel de corrosion et a la vitesse de corrosion. Leur application aux don&es de la
littbature suggere que, pour Fe2+ + 2e,- + Fe, G(&= $ ou Q et a- = 3 ou $, respectivement. On
propose sept chemins et mdcanismes reactionnels pour cette reaction et on &value leurs consequences
cinetiques. On attire l’attention tout particulierement sur des contradictions dans une suggestion
pr&&dente. Si la pente de Tafel anodique est RTIZF, une pente de Tafel cathodique RTI2F et un
ordre de reaction 2 par rapport a Fe I+ et 1 par rapport a OH- devraient en resulter, si une reaction
de transfert comportant plusieurs electrons n’est pas acceptable. Un mecanisme compatible avec les
resultats obtenus ici est le suivant: Fe + OH- + FeOH + e,-; FeOH + FeOH+ + e,-; FeOH+ +
Fe”+ + OH-. Ce chemin est proche au point de vue des demandes Bnergetiques, dun autre pour
lequel Ca = RT/2F. Les effets dus aux anions, sauf pour NOs-, peuvant ttre compatibles avec un
mod&e pour lequel l’adsorption specifique de l’anion modifie l’aire disponible pour l’khange ion-
metal.
La difference de chaleur d’activation pour le depot d’hydrogene sur Fe 1 partir de H,O+ + He0
est compatible avec une dechange regulatrice du proton.

ZusammenfassunS-Die Abscheidungs- und Auflosungsgeschwindigkeit des Eisens wurde mittels

galvanostatischer Messmethoden in Funktion des Elektroden-Potentials und der folgenden System-
Variabeln untersucht: Fe++-Konzentration, pH, Gegenwart der Anionen SO,e-, Cl-, C104-, Ac-
und NO,-. Gleichzeitig wurde die Geschwindigkeit der Wasserstoff-Entwicklung aus H,O+ und
H,O in Ftmktion von Potential und pH bestimmt.
Man wertete die derart erhaltenen stationlren Tafel-Geraden aus und leitete aus allgemeinen
kinetischen Betrachtungen unter Berticksichtigung der Literaturdaten miigliche Reaktionsmechanis-
men ftt den Gesamtvorgang Fea+ + 2e,- + Fe her. Als wahrscheinlicher, und mit den gemessenen
Werten tiberemstimmender Reaktionsverlauf ist Fe + OH- + FeOH + e,-; FeOH + FeOH+ +
e,-; FeOH+ + Fez+ + OH- anzusehen.
Anionen-Effekte, ausgenommen NOa-, k6nnen dadurch erkhirt, werden, dass durch spezitische
Adsorptionsvorgiinge die freie Elektrodenfllche ftir den Austausch Metal1 + Ion verlndert wird.
 The electrode kineticsof the depositionand dissolutionof iron 327

A. INTRODUCTION
THE kinetics of deposition and dissolution of metals has usually been studied under
steady state conditions, often giving rise to complex results which do not aid insight
into the molecular mechanism. Recent studies1 show that the use of short time transi-
ents under galvanostatic conditions has certain advantages. (a) It allows direct evalua-
tion of ohmic contributions to overpotential. (b) Diffusional effects in solution may
be reduced; and the transient for hydrogen discharge from water separated from that
for metal deposition. (c) The small number of monolayers deposited or dissolved
minimizes change of surface at a given current density.
Rojtar, Juza and Polujat? examined the decay of the potential of an Fe electrode
in 1939, but in contact with Picein. Okamoto, Nagayama and Sato3 examined the
cathodic and anodic behaviour of Fe in H,SO, by applying pulses at successively
increasing current densities on the same electrode in the absence of added Fez+. Similar
experiments were done by Kusnezov and Iofa4. Heusle$ examined the dissolution
kinetics in acid Na,SO, and NaClO, solutions, but the mechanism proposed involved
a “catalysis,” the mechanism of which was left open. Hoar and Hurlen examined the
steady-state polarization and current efficiency in NaHSO, solution containing FeSO,,
and obtained Tafel slopes of RTI2F for cathode and anode lines. Kabanov and
Leikis’, and Kabanov, Burstein and Frumkin8 examined dissolution of iron in alkaline
solution.
Other than that of Hoar and Hurlen6, previous studies have concerned principally
dissolution. Evaluation of path and rate-determining step in the corrosion of iron
demands a knowledge also of the cathodic behaviour in order to distinguish between
alternative hypotheses for the mechanism of the anodic reaction. Because adsorbed
anions may take part in the rate-determining step, it is desirable with respect to inter-
pretation that each solution contains a minimum number of anionic species, i.e., that
buffer solutions are avoided.
Here, galvanostatic transient studies are reported on Fe deposition and dissolution
from simple salt solutions as a function of pH, Fe2+ concentration and nature of the
anion, using highly purified solutions and electrodes.

B. EXPERIMENTAL
Cell. The cells are shown in Fig. 1. That of Fig. la resembles cells described
elsewhere15. New features are : (a) Furnace A (Fig. 1a) allows treatment of electrodes
in hydrogen up to 95O“C*. Introduction of the electrode without contacting air can
be made through B. (b) A pre-electrolysis electrode passes through C, the pre-electro-
lysis current being broken by withdrawal of wire from solution. (c) A glass electrode
for pH measurement passes through D.
The cell of Fig. lb was used for distinction between current density for hydrogen
evolution, iH, and that for Fe deposition, ipe. Two Fe electrodes could be introduced
successively without contacting air. Measurement of iH was made by observation
of the displacement of solution in capillary C (Fig. 1b), the solution being
saturated with H,. Compartment B is separated from A by a closed stopcock. The
capillary has a diameter of 0.3 mm. (Smaller diameters cause an impracticable
capillary rise). Measurements of less than 10e3 ml (50 monolayers) of H, are possible.
* “‘Oxygen-free” hydrogen was purified in a manner described e1sewhere.O
328 J. O’M. BOCKRIS,D. DRAZIC and A. R. DEI.SPIC

Calibration was with hydrogen evolved from a Pt electrode in 1 M Na,SO, solution

(100 per cent current efficiency). The cell was thermostated (*O-05°C) with a water
jacket, W. A cause of inaccuracy was leaks through ungreased glass taps (X, Y and Z,
Fig. lb). These were replaced by those of Teflon.

FIG. l(a). Cell for galvanostatic investigation of Fe electrodes. A-Nichrome wire furnace;
B-Tap dividing the furnace from the electrode compartment. C-Slip joint for pre-electrolysis
electrode; D-Slip joint for the glass electrode; E-Reference electrode.

Reference electrode. A saturated calomel electrode (E, Fig. la and lb) was used.
Circuit. The circuit (Fig. 2) contains a vacuum-tube millivoltmeter (high imped-
ance), attached in parallel to the oscilloscope for recording the corrosion potential.
This potential fixed the reference point on the oscilloscope screen during measurement
of transients. A recorder (response time 0.1 set) was attached across RM (Fig. 2) for
measurement of current. An additional recorder was attached to the output amplifier
of the millivoltmeter to record the change in corrosion potential with time after a pulse.
Solution preparation. Solutes were recrystallized 2-3 times from conductance
water under hydrogen atmosphere. Concentrated solutions were made by distilling
 The electrode kinetics of the deposition and dissolution of iron 329

in conductance water in the same way. They were analysed (see below) and the final
solution made by mixing portions of the Fe-containing solution with that containing
neutral electrolyte, and diluting with comluctance water; they were impelled between
parts of the cell by H, pressure.

FIG. l(b). Cell for partial H,O+discharge curren&~)measurements. A-Anodic compartment,

B-Cathodic compartment, C-Capillary (diameter 0.3 mm). W-Water jacket, Fe-Iron Test
electrodes, E&Reference electrode, X, Y, Z, Teflon taps, G, Taps for connection with Nichrome
furnaces, same as B on Fig. la.

A. R. grade FeSO, and FeCl, were used respectively with A. R. Na,SO, and KCl.
During recrystallization of these salts, A. R. Fe powder was introduced, together
with small quantities of HeSO, and HCl respectively (to reduce traces of Fes+).
Ferrous acetate was produced under hydrogen atmosphere by adding 100 ml cont.
acetic acid and 50 ml of conductance water to excess Fe powder. The mixture was
heated on a water bath and left at room temperature until reaction was complete
(circa two days). All liquid was consumed, and green crystals produced. These were
recrystallized twice from conductance water. Recrystallized sodium acetate was
the inert electrolyte. Concentrated ferrous nitrate was produced by mixing con-
centrated solutions of recrystallized Ba(NO& (HNO, added) and FeSO, in stoichio-
metric quantities. Potassium nitrate was the inert electrolyte. FeClO, was formed by
dissolving Fe powder in 1: 1 HClO, (H, atmosphere). Sodium perchlorate was the
inert electrolyte.
330 J. O’M. BOCKRIS,D. DRAZIC and A. R. DESPIC

The solutions were pre-electrolysed for twenty hours at a current density of 5 mA

cm2. Pre-electrolyses were carried out at a potential below that corresponding to
appreciable evolution of H, to avoid significant change of pH in the solution. The
solution was protected by bubbling highly purified hydrogen.

3
CELL
FIG. 2. Galvanostatic circuit diagram. Rs-Current controlling series resistance of lO-IflOO
KG. R,,,-Current measuring standard resistance of l-100 G. &-Limited time pulse switch.
&-Fast switch for activating the relay. S1-Polarity reversing switches. C.R.O.--Cathode ray
oscilloscope. V-High impedance vacuum tube voltmeter. Pen Ret-Recording potentio-
meter. REL-Western Electric Relay 275B.

Electrodes. These were Fe wires, having: C = O-025% Mn = 0.028 % P =

0.001%; Se = 0.003 % and S = 0.024%. Spectroscopic analysis revealed: Cu, Mn
(O.Ol-O*l %); Si, Mg, P (0+lO1-0~1%).
The Fe wire (d = O-36 mm, 1 = O-5-3 cm) was sealed to the tungsten wire and
both coated with thin (As-free) Pyrex glass. The electrodes were mounted on slip
joints, introduced into the furnace, and heated fifteen minutes. (Very dry Ha, passed
through the traps containing charcoal at liquid air temperatures, had to be used to avoid
formation of oxide). Variation of the temperature of preparation between 600 and
900°C did not effect electrode behaviour.
Analysis of solutions. Fez+, Cl- and pH were determined by conventional means.
Other anion concentrations were determined by converting salts into the respective
hydroxides through an anion-exchange column (Amberlite, IRA 400) and titrating
with HCI.
Procedure. Each electrode was immersed in solution and a stable corrosion poten-
tial measured. The electrode was submitted to a series of anodic pulses at various
current densities and time. Thereafter, the electrode was subjected to a series of
cathodic pulses, each cathodic pulse being preceded by an anodic pulse of the same
 The electrode kinetics of the deposition and dissolution of iron 331

length and c.d. as the preceding cathodic pulse. After each pulse, care was taken to
assure that the electrode potential returned to the steady state corrosion potential
before a further pulse was applied. The V/t transients on the oscilloscope screen were
recorded photographically and enlarged fifty times.
For observations of transition times and the potentials for hydrogen evolution,
cathodic sweeps up to 10 set in duration were applied.

C. RESULTS
Measurements were carried out in solutions containing crest = 0*008-0~9 mole/l,
at pH 3.1-3.6. At cFe++ = 0.05 M and O-5 M, the pH range of measurements was
1.249. These measurements were carried out in FeSO, in the presence of Na,SO, at a
concentration sufficient to keep ionic strength at 0*5 (exceeded for solutions containing
O-5moles l-3. For cFe++= PO5 and pH in the range 3-6, measurements were carried
out in the presence of the corresponding alkali chloride, acetate, nitrate, and per-
chlorate.
The current-density range was 5.10-5-10-1 A/cm2. The time of sweep was deter-
mined by that necessary to reach the steady state. The range was 100 msec-IO sec.

(i) Ohmic overpotential

Separate transient measurements were used to evaluate ohmic overpotential.
The commencement of the transient at high c.d. was diffuse. Lower c.d.‘s on the
normal time scale gave ohmic overpotentials too small to be accurately observed.
Increase of sensitivity of the potential scale allowed more accurate measurement.
Decrease of the time scale to that of 50 psec produced a transient at a lower angle to
the horizontal, the commencement of which was sharply distinguishable. An q,hmic/i
line, deduced from transients in the c.d. range 104-10-2 A/cm2, allowed evaluation of
ohmic overpotential for any c.d. This was subtracted from the steady state value and
the potential thus obtained expressed with reference to the normal hydrogen scale.

(ii) Transients
Cathodic and anodic transients are exemplified in Fig. 3a. The second plateau on
the cathodic transient was observed only at higher current densities. At c.d.‘s 1O-2
A/cm2, a peak appeared on the cathodic transient (“superpolarization”).
A peak was observed, for anodic transients, at medium and high current densities
and increased with c.d. At high pH, the peak is less sharp and time to reach steady
state increases with increase of pH. If a series of anodic pulses is applied, e.g., every
50 msec., the peak is temporarily removed.
Decay transients are shown in Fig. 3b. The rate of decay in the lower section of the
cathodic V/t relation was increased by stirring. An anodic transient applied during
decay of a cathodic transient reduced the time needed to attain the steady state cor-
rosion potential.
A typical anodic V/trelation during decay shows more rapid change of the poten-
tial than that of the corresponding cathodic decay transient. Upon decay, it passes
through a minimum, which is smaller than the maximum observed on charging. Stir-
ring decreases this minimum.
332 J. O’M. B~CKRIS.,D. DRAZICand A. R. DESPIC

CATHODIC TRANSIENT
-2

time

ANODIC TRANSIENT

FIG. 3(a). Typical cathodic and anodic galvanostatic transients. vextrs. = extrapolated value
of overpotential for t = 0; r],,hdc = pseudo-ohmic overpotential; qai = activation over-
potential; 7 = transition time; 7Ss-time necessary to establish steady state.

-mV

CATHODIC DECAY
150

50 ANODIC DECAY

+mV I
(b). Typical cathodic and anodic decay transients.
In AC-- and NO,--containing solution, the potential/time relation on decay
exhibits a plateau (at circa 0-O V), length of which is proportional to the duration of
the previous anodic pulse.
(iii) Capacitance
Double layer capacitancevalues were calculated from c,, = iTotal(dt/dV) under the
condition that the faradaic current density iF < O-1 iTotal(Table I). Cathodic and
anodic transients were used.
 The electrode kinetics of the deposition and dissolution of iron 333

(iv) Transition times

Calculation with the equatiorP 7 = (.rm2FDc,2>/4i2 indicates that the minimum
current densities at which transition times were less than 10 set are (approx., A/cm2);
lOa for cr++ = 10-s M; 10-s for cFea+ = 1O-2 M; 10e2 for cFea+ = 10-l M and 5 x
1O-2 for cFes+ = O-5 M. In Table 2, calculated and observed transition times are given
for Fe2+ for current densities above those stated and calculated values are given for
H30+.
TABLE 1. CAPACITANCXDATAFOR Fe ELECTRODEIN 0.05M Fez+
SOLUTIONSIN THE PRESENCEOF 0.2 EQ. 1-l OF NEUTRAL SALT.

Capacitance @~Fcn-~)
Electrolyte
from cathodic from anodic
transients transients

FeSO&+ Na,SO, 235 f 15 223 * 30,

FeCI, + KC1 160 * 22 147 f 40
FeAc, + NaAc 109 & 7 133 f6
Fe(NO& + NaNO, - 15Ok60

TABLE 2. TRANSITION TIME FOR Fe e+ AND H,O+ DISCHARGEIN 0.062 M FeSO, + 0.1 M Na,SOp
SOLUTIONAT pH = 3.4
\ ,
Calculated 7x,6+ for
Current density Experimental QQ+ Calculated 7xez+
HsO+ discharge
(A/cm’) (set X lo-? (Xc x 10-S)
(set x 10-O)

0.35 9 6 0.9
0.32 10 8 1.1
0.30 10 9 1.23
0.26 13 12 1.63
0.24 14 14 I.75
0.20 24 20 2.5
0.15 40 38 4.5
0.08 140 125 17
0.06 290 222 31

In solutions containing NO,- and AC-, transition times were observed in anodic
transients, whereafter passivation occurred and 0, was evolved (Fig. 3b). The anodic
transition times were equal to those for the corresponding cathodic sweeps.

(v) Steady state

(a) Anodic. In Tables 3 and 4 are given b values of the anodic Tafel line and i,,
values calculated from the point at which the extrapolated log ian line intersects the
potential axis at the thermodynamically calculated value of the reversible Fe electrode*
for the Fe2+ concentration concerned.

* The activity coefficients for the corresponding solution of Ni*+ were used in the absence of data
for Fez+.
 TABLE3. TAFELPARAMETERS
FORDEPOSITION
ANDDISSOLUTTON OF IRON IN 05 M FeSO, + 05 M Na,SOI AT DEFERENT pH
-
Slopes Exchange current densities Corrosion Potential
I - I - Calculated
b ir~ter~~tion reversible
b b b i.
PH (anodic h3 L3 potential current of Tafel potential*
(anodic (cathodic (cathodic (cathodic
steady (anodic) (intersection) etorr ‘corr lines (fO.Olv)
peak) experimental), corrected) corrected)
state)
@W bv) (mv) WV) (A/cm”) (A/cm21 (A/cm’) 03 (A/cm? 09 (VI
-- _-
1.2 54 70 - - 3.0 x 10-g - - -0.278 3.2 x lo-& - -0.488
1.9 45 65 - - 4.5 x 10-S - - -0.323 1.7 x IO-4 - -0,488
2.0 48 6.5 65 123 1.6 x lo-’ 5.6 x 1O-6 3.0 x 10-1 -0.333 1.2 x 10-b -0.460 -0,488
3.1 38 56 63 113 6.8 x IO-’ 9.0 x 10-7 8.5 x 10-7 -0.398 1.2 x 10-d -0.483 -0.488
4.0 34 53 60 110 1.1 x 10-S 2.4 x 1O-8 2.0 x 10-G -0.438 1.0 x 10-S -0.480 -0.488
40 42 55 60 120 1.1 x 10-S 3.4 x 10-e 4.5 x 10-G --0458 7.0 x 10-S -0.510 -0.488
4,l 41 45-50 60 - 3.0 x 10-s - - -0.458 1.7 x 10-h - -0.488
4.9 40 55 30? - 7.0 x lo-’ - - -0448 1.0 x 10-s - -0.488
4.9 41 50 50 - 7.0 x lo-’ - - -0.453 4.0 x 10-G - -0.488
- - - -
* Reversible potential was calculated using activity coefficients for corresponding Ni2+ solutions. Change of the activity coefficient with pH can be
shown to influence the reversible potential for less than 1.5 mV, so it was neglected.
 TABLE 4. TAFELPARAMETERSFORDEPOS~ONANDDISSOLU~ONOFIRONINSOL~ONWITHD~FPERENTFERROUSIRON 2
CONCENTRATIONSAT pH = 3.1 0
8
.‘i
Fee+ Slopes I Exchange current densities Corrosion %
T Potential
at inter- (Xculated !?.
b b section reversible B
=.
b b i0
Concen anodic cathodic 10 10 Potential Current of Tafel G!
Activity anodic cathodic cathodic h 1Jotential*
tration steady experi- anodic intersection ecorr. ZCO,,. lines I~~O*Olv) s
peak corrected corrected
state mental
B
(M) 05) (mv) (mv> (mv) (mv) (A/cm’) (A/cm3 (A/cma) WI @/cm3 (VI &
-- cd
0.5 0.02 38 56 63 113 68 x lo-’ 1.1 x 10-B 8.5 x lo-’ -0.398 1.2 x 10-d -0.483 -0.488 8.
0.5 0.02 50 - 57 110 8.0 x lo-’ 1.3 x 10-S 9.0 x 10-1 -0.398 3.0 x 10-S -0.473 -0.488 c:
2.7 x 10-O 9.0 x lo-’ -0.411 8.8 x 10-S -0.482 -0.499 g
~
0.14 0.011 40 59 62 112 8.0 x lo-’
0.14 0.011 42 64 65 120 9.0 x 10-1 1.2 x 10-G 1.3 x 10-e -0.412 6.4 x lo-& -0.486 -0.499 E
0.036 0.003 39 57 65 130 1.2 x 10-1 2.1 x lo-’ 1.8 x lo-’ -0.413 4.7 x 10-S -0.508 -0.514 g
0.036 0.003 35 53 - - 3.0 x 10-1 - - -0.413 1.5 x 10-d - g
0.011 owO9 44 62 113 9.0 x 10-n 8.0 x lo-* 8.5 x 1OV -0.412 7.0 x 10-S -0.537 -0.530
-0530 E
0.011 0*0009 39 58 113 75 x 10-a 1.8 x lo-’ 1.4 x 10-1 -0.413 8.5 x 1O-5 -0.523 g
* - -
$
* Reversible potential was calculated using activity coefficients for corresponding Nil+ solutions. G
g
 eh v

-040 -

=corr. -

-035 -

-0.30

-02 5

i
I ! I I,,!,1 I I I111111 I I I ,I,,, I I ,l,,lL
10-4 10-3 IO-' i (A,,-2 )
FIG. 4. Anodic Tafel plot for solution of 0.5 M FeSO, + 05 M NasSO* which was prepared
without pre-electrolysis, pH = 2.6, b,’ = 30 mV, LJ.”= 60 mV.

t-1

eCOrl

V
(+I

FIG. 5. Schematic representation of polarization lines in solution containing Fe’+, at pH w 3.

Heavy lines-cathodic and anodic V/log i relation obtained from the steady state values of
the potentials. l-Anodic Tafel line. 2--H,O+ discharge limiting current density. 3--C&h-
odic Fe*+ deposition Tafel line obtained as V/log (iTOtal - iH). 4-Cathodic Fez+ deposition
Tafel line corrected to the pH of the bulk. AV’-concentration overpotential for HsO+ dis-
charge. Correction = Av’, i.e., BC = E
336
 The electrode kinetics of the deposition and dissolution of iron 337

In general, purified SO,” and Cl- solutions gave anodic Tafel slopes corresponding
to 2RT/3F. Anodic Tafel slopes differing from this were observed as follows:
(i) In solutions not rigorously purified, anodic Tafel slopes had two values, RT/2F,
at low c.d.‘s, RT/F at high c.d.‘s (Fig. 4). (ii) With AC- and NO,- the slope was
2RT/3F in the lower cd. ranges and 2RT/F in the higher ranges. (iii) If peak values
from the anodic transients (see Section C, iv) are plotted, the slope is O-07to 0.05 as the
pH is changed from l-2 to 4-9 at cFesO, = O-5M.
(b) Cathodic. Here, the hydrogen evolution reaction (h.e.r.) interferes. For a high
c.d. V/t line (Fig. 3a), the T for the first plateau is that calculable from a Sand-type
equation for Fe s+. At lower c.d.‘s, -r was too high to be observed, .so that no such
verification of lack of interference of iH with the potential of the first plateau could be
made for such values. The cathodic galvanostatic transient contains, in principle,
three plateaus. The first is due to deposition of H from H30+. At high c.d. this would
be of much smaller magnitude than that for Fe2+ (see Table 2). The second plateau is
due to Fe2+ discharge and a third plateau, observable only at high current densities,
represents the discharge of H from H,O (Fig. 3a).

FIG. 6. Plot of (a) anodic, (c) cathodic (steady state), (iH) partial HsOf discharge, (&B>partial
Fe deposition, and (Z&,,& corrected Fe deposition current densities as functions of
polarization.

eh V

-0.50

-0.40

-0.30-

/‘L
-0*20 -
i,,,,.
I I I I I
-6 -5 -4 -3 -2 -I
log i

(a). 0.5 M FeS04 + @5 M Na,SO, @H = 2.0) bC,exper = 65 mV, bc,correckd= 123 mV,
banodlc = 48 mV, i0 = 3 X lo-’ A/ems, e&,= = -0.333 V, iCorr = 1.2 x lo-& A/cma.
eh V

-060

-0.50

-040

-0*30-

-o.zo-

.-7 -6 -5 -4 -3 -2 -I
log i

(b). @5 M FeSO, -I- 0.5 M Na,S04 (pH = 3.1), bc,exper = 63 mV, bC,C,,rr&ed = 113 mV,
bano& = 38 mV, i. = 8.5 x lo-’ A/cm*, e,& = -@398 V, iCorr = 1.2 x 10-d A/cm*.

eh CI

-04 IO-

‘O-

10 -

io -s
rev.
,ecorr.

IO-

50 -

20-
I I
-7 -6 A5
log i

(c). 0.5 M FeSO, + 0.5 M Na,S04 (pH = 4*0), bc,exper = 60 mV, bc,cw.&ed = 120 mV,
banotic = 42 mV, i. = 45 X 10-O A/cma, et,,, = -0.458 V, icorr = 7 x 10-S A/cm’.
338
 The electrode kinetics of the deposition and dissolution of iron 339

-0.20 -
I I I t I I
-7 -6 -5 -4 -3 -2 -I
log i

(d). 0.5 M FeSO, + 0.5 M Na,SO, (pH = 47), bc,exr,er = 60 mV, ban,,& = 30 mV. No data
from hydrogen evolution measurements.

Cathodic Tafel plots (ire B+obtained as described in Section B) are in Figs. 6 and 7.
Anodic Tafel lines intersect with cathodic ones thus derived approximately at the
thermodynamically calculated value for the reversible potential for the reaction Fez+
+ 2e; + Fe (and clearly not at the corrosion potential), thus verifying that the
cathodic line is initially free of effects of the codeposition of H.
(c) Correction of cathodic Tafel IineforpH changes at the electrode surface. During
the deposition of iron, H,O+ discharge is occurring and pH change at the electrode
surface must occur. It is, therefore, not valid to accept the cathodic Tafel lines, cal-
culated directly as discussed above, to represent Tafel lines for Fe deposition at the
bulk pH, for at each current density, the pH at the electrode surface differs from that
for given bulk PH. These differences are large, because under the current densities for
appreciable Fe deposition, the limiting diffusion current density for H,O+ discharge is
already exceeded (see Fig. 6a).
It has been experimentally found (see later, eq. 12) from the dependence of iO,,odic
a ln iO.an
on the bulk pH, that w 1; and consequently the cathodic deposition reaction
a In aOH-
also involves one OH-, so that the pH changes at the electrode surface will affect the
velocity of the cathodic deposition reaction. Consequently, the cathodic Tafel line
with the slope of -60 mV represents the net effect of potential upon current density
for a constant pH in the double layer, together with the dependence of this pH upon
 340 J. O’M. BOCKRIS,D. DRAZIC and A. R. DESPIC

potential. To obtain Tafel slopes which can be used to indicate the mechanism of the
reaction, a correction must be made to the cathodic slope so that it refers to constant
bulk pH at all potentials, the conditions to which kinetic equations apply.
From the experimental results (Tables 3, 4) one can write

’
I&Fe = kc,Fe (~~~-)~e-(~~~~‘~*),

ahicFe F
L a vF, = - R-T’
(1)

where (aon-) is the activity of OH- ions in the bulk.

To allow for the effect of the potential dependence of (ao& (activity of
OH- ions at the electrode surface) on the reaction rate, it is necessary to deduce
aln a(a~~-)~
v (cf. the dependence of (i&t on pH15).

FIG. 7. Plot of (a) anodic, (c) cathodic (steady state), (in) partial HIlO+ discharge, (ids) partial
Fe deposition, and (iI&,&corrwted Fe deposition current densities as a function of
polarization

eh V

-0BC I-

-0.7(

-06( I-

-o.!x )-

-0*4(

-&I

-0.2( II-

-7 -6 -5 -4 -3 -2 -I
lo9 i
(a). 0.14 M FeSO, + @13 M Na,SO, (pH = 3.1). bc,exper= 65 mV, bC,$om&ed = 120 mV,
banodic = 42 mV, i0 = 1.3 x lo-@ A/cmS, e,& = -@412 V, iCorr = 6.4 x lO-6 A/cm*.
Ccorr L
-G.sot-
I
)

-0.30 -
I I
-5 -4 -3 -2 -I
6 FC -6
109 i

(b). 0.036 M FeSO, + 0.13 M Na,SO,. (pH = 3.1), b c,exper = 65 mV, &corrected = 130 mV,
banodiC= 39 mV, i. = 1,8 X IO-’ A/cm%, en,,, = -0.413 V, icOn = 4.7 X 1O-5 A/cm*.

I
(i,),

I I I I I I
-7 -6 -5 -4 -3 -2
109 i

(G). 0.011 M FeSQ + 0.15 M Na,SO, @H = 3.1), bc,exper = 57 mV, bc,wmc~d = 113 mv,
= -0.412 V, iCon = 7 x 10-E A/ems.
&odic = 44 mV, C = 8.5 X lo-*, e&,rorr
341
342 J. O’M. Bocms, D. DRAZIC and A. R. DESPIC

Concentration overpotential for HsO+ discharge can be written as

(%,o+)e RT @OH-)B
% = ‘Fe - vE*‘k=o =
sin
/j&F (aH30+)B = GF’n (aoH-)e ’ (2)

where V,, is the actual potential of the Fe electrode at a certain current density, and
Vn,V,=ois the potential which the electrode would have if only the hydrogen evolution
occurred upon it at this given current density and in the absence of concentration
overpotential (qc = 0).
Equation (1) can be rewritten as

a(=_IIn icFe FP, Fe + a ln (%H-)e

(3)
a vFe RT 8 vFe ’
and using (2)
--- (VFe - vH,qo=O) +$ > (4)
a vFe = - Ri? a vFe

where /?e,Peis the symmetry factor for the deposition reaction, at a certain definite pH
value (i.e., the bulk pH). /&Fe can be found from (1) and (4),

&Fe= 1 - g-
Fe
(v,, - V&@)& = 1 - ,& + BEf aFi”=’- Fe
(5)

7. can be expressed also by the equation

(6)
Differentiating (6) we get

avH.vo=O = 1 _ (64
a vFe

Because in = io~-(B~FIRT)V~s~o=O
2

a (-E) a (_ ~,-6‘&FiRTP’~.v.-o)
av,, = avFe
lo -----_PEP _ (B,NRTU’,,rl,=O avH,%=O ;
= --
( i
‘>( RT avFe
therefore

iE PHF~VH,~~=, (6b)
avFe =<RTx’
Substituting (6b) into (6a)
1
avFe ia . (6~)
If---
iL - iH
 The electrode kinetics of the deposition and dissolution of iron 343

Introducing (6~) into (5)

r6H
Bc,Fe =1-/%X+ .
(7)
1+Y-K-&-

For the condition when iH -+ iL, i.e., when the iron deposition actually occurs
with high current efficiency, the last term in eq. (7) tends to zero and

pc,lTe = 1 - PH. (8)

From the present measurements on the h.e.r. on Fe (see Fig. 14), /& = O-51

Bc,Fe= 0.49. (9)

This shows that the cathodic Tafel line would have slope of 2RT/F, if the pH in the
vicinity of the electrode were constant with change of current density and this is, then,
the value which must be compared with expectations of the various mechanisms.
A graphical correction is obviously possible. Schematic representation of the
anodic and cathodic Tafel lines, with the correction of the experimental cathodic Tafel
line to the bulk pH, is given in Fig. 5. Because PC& = /Ic,H, the correction for each
point on the experimental V/log iFeline is equal to the corresponding numerical value
of hydrogen concentration overpotential qC= V,, - VH,9+,, which causes this pH
change. The corresponding experimental and corrected Tafel lines are shown on
Fig. 6 and 7, and the Tafel parameters are tabulated in Tables 3 and 4.
The point of intersection of the cathodic and anodic Tafel lines, which, if the
measurements and method of correction is correct, must equal eRev.,Fe,is plotted in
Fig. 8 as a function log aFeat. From it, a eRev.,/alOg%,so= O-03, and the extrapolated
-0.43 V, in substantial support of the method bf evaluation (thermodynamic
-044V).

(vi) Anodic Tafel lines from peaks of transients

Theb values for the Tafel lines at various pH’s for the peaksof the anodic transients
varied from 0.050 in pH 49 to 0.070 in pH 1.2 (see Table 3).

(vii) Efict ofFe2+ concentration on exchange c.d. at constantpH

A plot of log (iJanodiCas a function of log aBes+(aB,s+
= 04009402 M) is given in
Fig. 9 for pH = 3.1. This line shows that

A plot of log (i,), against log a,,r+ is shown in the same Fig. 9. It gives:
 344 J. O’M. Fkmms, D. DRAZIC and A. R. DJBPIC

(viii) E$ect of pH on log i,, at constant aF,a+

Fe2f concentrations used were O-05and 0.5 M. The plots of log (i,,),, as a function
of the pH in the bulk of the solution is in Fig. 10. It gives, in the linear section:

ah (i,),, (12)
ahaon-= ’
--I

eh v

-0.55 -

-oflo-

-0.45-

I I I lllll I I II1111 I 1 I111111 I I I IlllL

-0.40
0.001 001 0.1 activity

FIG. 8. Plot of potentials of intersection of anodic and corrected cathodic Tafel line as a
aebsv
function of log a~~~+, = -0.03.
ai0gme2+

(ix) Effect of pH and a,,%+ on corrosion potential

A plot of corrosion potential as a function of pH is shown in Fig. 11. It gives

a%0rr o

a= * *
o6

(13)

This value is in excellent agreement with values reported by Bonhoeffer and Jena,lO
D’Ans and Breckheimer,l’ and Stern.12
A plot of ecorr as a function of log aFeat at constant pH is shown on Fig. 12. The
variation with aFeS+is less than 3 mV per decade of a,+.
 The electrode kinetics of the deposition and dissolution of iron 345

I
Oon.

IO-s

10-7
4
c

0- i, onodic

l - i,, cathodic

A- i,, intersection

I I I,11111 I I111,111
0~001 0.01 0.1
activity

FIG.9. Plot of log i. as a function of log a~~‘+. i,,,,~O-current density of intersection of

anodic Tafel line with calculated reversiblepotential. z&sthodje-the same, but for the corrected
cathodic Tafel line. i. intersection-current density at the intersection of anodic and corrected
cathodic Tafel line.
Composition of the solution: FeSOI and Na$O( in concentrations to give obtained ux,z+
and constant ionic strength (0.5), except for the most concentrated solution, which had
0.5 M FeSOI + 0.5 M NaBSO+ pH = 3.1. e = 0.8.
FB

(x) E#ect of pH on corrosion current

A plot of corrosion current obtained from the intersection of anodic Tafel line
with the measured corrosion potential is shown in Fig. 13.
It gives

(14)
(xi) Eflect of anions
This is shown in Table 5.
 PH

6- I ’
’ ,’
7- / /’ ,’

/’ I ,I //

6- /, /
/ /
/ /
’ 1’
/’
S- / / 1,’
/ ’ /’
’ /’

4- A’

3-

2-

I-

I I I I I I I I
-9 -6 -7 -6 -5 -4 -3 -2
log i.

FIG. 10. Plot of log i,,sno~C as a function of pH in the bulk of the solution. alnio an
) = 1.
a In aoH-
O-In 0.5 M FeSOd + 0.5 M Na,SOI solutions. O.-In O-05 M FeSO, + 0.15 M NaBSO
solutions. Dotted lines-calculated lines with Lindstrand’s18 value for the hydrolysis constant.

h
ecor1

-05

-0.4

-0.:

I I
2 3 4 5 PH
FIG. 11. Plot of the corrosion potential eL
cormsion as a function of pH in the bulk of the solution.
aeCo,,
0.5 M FeS04 + 0.5 M Na,SO,, - a PH = 0.06.

346
 I , , ,l,l, I I I111111 , I I111111
-@30
0.001 0.01 0.1 Fe2+ activity

FIG. 12. Plot of the corrosion potential eLmston as a function of log aBe”+.

PH

I I , ,,,,, I f I,11111 I I I IllIll

lO-s 16' lO-3 i CO,r.
FIG. 13 Plot of the corrosion current densities, obtained for the intersection of anodic Tafel
line with ehcorrosjonas a function of PH. 0.5 M FeS04 -I- 0.5 M Na,SO,. a logicon: - __0.5.
a hi pHoa
 348 J. O’M. BOCKRIS, D. DRAZICand A. R. DESPIC

(xii> Kinetics of the deposition of H from HsO+ and Hz0

Typical Tafel lines for H,O+ discharge (from extrapolated values qextrp. of cathodic
transients (Fig. 3a)), and H,O discharge (steady state after the transition time was
exceeded for Fez+ deposition) are shown in Fig. 14. In Fig. 15, iO,HS,,+and iO,H,O
are plotted as a function of pH in the solution of O-5M of Na,SO, without addition of
TABLE 5. EXCHANGECURRENTDENSITYASAFUNMON
OF ANIONS FOR pH = 3 (CRes+= 0.05 M)

Solution i. Xlo*
(A/cm’)

Fe(ClO& + NaClO, 50
FeSOp + NaBSOl 30
FeCl, + KCl 8
FeAcl 4 NaAc 1
Fe(NO3, + KNOr 0.02

Fe salt, and in the presence of O-5M of FeSO,. No effect on i0,H80+ was observed.
The dependence of i,,H80+ on pH is

(15)

which is in agreement with the calculated value for H,O+ discharge as rate-determining
step.
D. DISCUSSION
(1) General Kinetic Relations for the Deposition and
Dissolution of Iron, Involving OH-
The following relations are independent of path and rate determining step, except
for the assumption that OH- takes part in the mechanism,* and that in steady
state corrosion, the limiting current for H,Of is not exceeded.
ia,re = ka,Fe a:;!!- e%FevFIRT, (16)
ic,Fe = ke,Fe a::!- a;:;: e-%FeVFIRT, (17)
iC,H= kc,Haq+ e-‘%HVFtRT. (18)
If t&Fe = hFe9 V = VR, and from (16) and (17):
nFea+ RT
vR,Fe = vR.0 + y h aF$+, (19)
‘%,Fe + %,Fe >

Or,

%,Fe + aa& = 2nFeB+. (20)

* It might be thought that the pH dependence of the reaction could be caused by hydrogen adsorbed
on, or coming out of, the metal. This hypothesis is, however, inconsistent with the intersection of
cathodic and anodic Tafel lines at the reversible potential for Fee+ + 2e0- + Fe, and with the result
log (iJan Cc aBea+. Were the anodic transients concerned with H emanating from the interior of the
electrode, it can be. shown that the surface concentration of H with this origin would have to be
related to that in the bulk of the metal by an equation of the type 0x,, = KerRI”/“‘,and thus imply an
electron exchange between the two types of H. A pH effect due to blocking of sites on the electrode
surface requires the law 0x = 1 - Kacx-, to agree with the observed pH dependence, and this relation
is clearly unlikely.
-0*60 -

-0.70 -

/
-0.60 - /

-0.50-
/
/
/ /
-0.40 - / ,/g
/
/*\ /i3
A’ ‘*
-0*30- / , a
/ /
/ /
/
-0.20 - %ev.
io, Hz0 i,, H,O+

I I I I I I
-6 -7 -6 -5 -4 -3

log i,
FIG. 14. Plot of the polarization data for hydrogen evolution reaction on Fe electrode in 1 M
Na,SO, at pH = 3.6. O-Cathodic steady state line showing limiting ditfusion current for
H*O+ discharge and at higher c.d.‘s Tafel line for Hz0 discharge. + -Tafel line for H*O+
discharge without concentration overpotential obtained from the extrapolated value to t = 0
from cathodic transients (c.f Fig. 3a).

0- io for H,O+dischorge in the presence of Fe++

l - lo for HsO+dischorge in the absence of Fe++

A- io for lie0 discharge

PH

I I I I I
-6 -7 -6 -5 -4 log io

FIG. 15. Plot of &,uro+ and i,,,HtO as a function of pH.

ah h,Hso+
= -0.5.
ahaoIi-
2AQ
350 J. O’M. Beams, D. DRAZIC and A. R. DESPIC

For steady state corrosion at a potential V,,,, > 50 mV from vn,,+,

ia.re = Lu, (21)
i.e.,

vcDlT= Vwrr,ll- (22)

Hence,
(23)

Similarly, the steady state corrosion rate is:

(24)

(25)

For net anodic currents such that V,, is at least 50 mV from V,,, eqn. (16) gives

= aoe FIRT, (26)

q)H++,=+

and for net cathodic currents such that VFeis at least 50 mV from V,, eqn. (17) gives

= -acF, AfiT. (27)

aOE-hes+

Further, for i = 0, io,Fe= i,,Fe = 6, (16) and (17) give with (19)

(28)

(30)

Experimental knowledge of the coefficients of eqn. (20), (23), (25) and any one of
(26), (27), (2% (29) or (30) gives all parameters needed for complete determination of
mechanisms. It is noteworthy that the experimental dB?culties of studying the
cathodic reaction may be avoided by choosing (26) (together with (20), (23), (25)).
In the absence of detailed data on the Tafel slopes, some mechanism-indicating
results can be obtained by application of equations (23) and (25), utilizing the thermo-
dynamic equation (20), and considering that the only physically reasonable values
which can be @WI t0 noH- and n&a+ are 1 or 2. Some conclusions drawn on such a
basis are shown in Table 6.
Inspection of Table 6 immediately indicates a number of “impossible” cases, e.g.,
those in which a transfer coefficient is 0 or negative or 24, and this eliminates a number
of combinations of parameters, if it is assumed that multielectron transfer stages do not
occur in the rate-determining step. Further selection of likely parameters can
 The electrode kinetics of the deposition and dissolution of iron 351

TABLE 6. TRANSFER COEFFICIENIS CONSISTENT WITH SOME REPOWED DEPENDENCIES OF CORROSION

POTENTIAL AND CORROSION RATE ON pH.

avoorr: -- RT
a In UI-JH- F

aIn ieorr
-: -&EB; 0-E
a In aoH_

Observed Transfer Coejicients Consistent

Result (Calculated from equations 20, 23 & 24)

IIOIX- nFe’+ ac,Fe Set No.

(Assumed) (Assumed)

AA
A.E- -

2 1 2/l
0 7

\2 2 8

0 9

/ 1 8 2/l12 2 10

%
P/l 14 13

L2 3) 14
D,E-
-# 15
2 14
l& 16

Note; Reported: A-by Bonhoeffer and JenalO, D’Ans and Breckheimer’l and Stem’*; B-our results;
D-by Heusleld; E-by Heusler’ and StemI*.

be made on the basis that ue,H is usually about one half, sometimes one and possibly
two thirds (though this value has not been reported for Fe).
With these “selection rules,” Table 6, therefore, indicates the following expected
combinations of parameters (Table 7) :
18
352 J. O’M. BOCKRIS,D. DRAZIC and A. R. DESPIC

Because of the expected PC,=condition, the most likely parameters for the assump-
tion stated are
3 1
%,Fe = - with u,,re = - ,
2 2
5 3
%,Fe = - with CC,,*~
= -.
2 2
TABLE 7. PARAMETERS EXPECTED IN IRON DISSOLUTION AND DEPOSITION
BASEDON pH DEPENDENCE OFCORROSION RATEANDPOTENTIAL

now nFe'+ %I %&Fe %FB Implication

1 1 4 4 A,B
2 2 t : s A,B
1 1 1 ?f 2 A&
2 2 1 2 2 ) A3

(2) Some Paths and Rate-determining Steps in the

Deposition and Dissolution of Active Iron
The unexpected fact that the velocity of deposition and dissolution of Fe in acid
solutions is a function of pH, as it has for long been known to be in alkaline solutions,s
eliminates consideration of simple mechanisms earlier proposed, such as
Fez+ + 2e,,- q? Fe
or
Fez+ + e,- z? Fe+
Fe+ + e,,- z$ Fe.
Two other general facts must be taken into account. (i) In the cathodic reaction, the
co-deposition of H+ causes pH changes in the vicinity of the interface, and because of
the pH dependence of the mechanism, this fact complicates a comparison between
experimental t?mlings and predictions of various mechanisms; (ii) the little discussed
existence, due to hydrolysis, of the FeOH+ ion, present at significant concentrations
even at low pH values, eliminates the physical difficulty of having to assume direct OH-
effects.
The available experimental results5*6indicate a variety of sharply different kinetic
behaviours, suggesting that the energy differences between various possible paths are
unusually small and that one can go from one path to another with only a small
change of conditions. It is, therefore, necessary to examine several mechanisms.

(i) Mechanism A
Fe + OH- + FeOH + e,- (31)
FeOH s FeOH+ + e,- (32)
FeOH+ + Fez+ + OH-. (33)
Or
Fez+ + H,O + FeOH+ + H+
FeOH+ + e,,- 5 FeOH
FeOH + H+ + e,- s Fe + H,O.
 The electrode kinetics of the deposition and dissolution of iron 353

(a) Tafel slopes

i, = k,,FeaoH_e(‘+B)vFJRT, (34)
aV 1 RT
(35)
aIn=--’ l+BF
-_BVFIRT
i, = k c.&+e~+aO~-e , (36)

av
-=---*
IRT
(37)
8 In i, ,i3 F

(b) Velocity of dissolution as a function of pH and aFea+, From (34),

(aFE.i:H)
v,aF,r+= += ‘p
Bea
(38)

for V = V, from (36),

(39)

(c) Corrosion potential as a function of PH. If V,,,, is at least 50 mV from VRSFe,

and VRSH, from (34) and (18),
k a.Fe’aOH-e (1+B)VmrrFIRT = k -tVcoorrFlRT
c.H,O+G-e ,

2 RT
= ---. (41)

(d) Velocity of corrosion as a function of PH. It is assumed that during change of

pH the change of the ratio of anodic to cathodic area is negligible. The assumption is
justified by the agreement with results of the present experiments, obtained by its
application.

(1 +B)-_T
v,,o--- 2 RTlna OH -
(42)
i\ Z+pF
L
)
I
’

zcorr = kz.FeaOH-e

(43)

(ii) Mechanism B
The kinetics of most reactions at electrodes are regarded as consecutive sequences
in which all particles entering the system undergo permanent change. It was suggested
by Heusleti that in the case of the deposition and dissolution of Fe a “catalyst” takes
part in the reaction, i.e., the rate is affected by species which undergo no net change
during one act of the overall reaction. This concept was formulated by this author in a
somewhat intuitive way. It can be examined in terms of electrode kinetics as follows.
354 J. O’M. F%XXRIS,
D. DRAZICand A. R. DE.WIC

Suppose that the path is that of mechanism A and that the catalyst (C) forms an
activated complex with Fe and OH-, the rate of decomposition of which is the rate-
determining step of the overall reaction of dissolution.
AfC+AC-+B+-C (44)
Hence,
0, = kKa,ac. (45)
Suppose the net path is that of mechanism A, (45) becomes

i,,p, = k,,FeK~aoH-ace2(1-~)vP’RT. (46)

Suppose C be identified with FeOH, and that this radical is in an independent
equilibrium with OH- at the interface, according to reaction:
Fe + OH- + FeOH + e+,-, (47)
I his reaction not taking part in the overall reaction.
Then
ac = aFeOH= KaaoH-eVFIRT, (48)
ia,Fe= k,,FeKlK2a&e2(3/2-~)VF’RT. (49)
To obtain the cathodic current, one has, by reasoning analogous to that which
leads to (46) (cf. (32)),
’
lc,Fe = k,,seaFeoH+Kl’ace-2~VF~RT. (50)
As the potential becomes more negative, ac formed by mechanism (47) diminishes
and were the same catalyst present for backward and forward reactions, the cathodic
current density would be invariant with potential. Let it be assumed (Heuslers) that
the catalyst now forms according to
Fee+ + OH- + e,- + FeOH. (51)
If this independent reaction is in equilibrium,

G,r-e = k,,peK~a~eoH+aoH-a,2+e-(1+28)VF~RT. (52)

Hence, using (33),

&,Fe = kc,FeK~K2’a&8+a&-e-(l+28)VFIRT. (53)

(a) TafeZslopes. From (49) and (53),

(54)

8V 1 RT
-a In i, = - (1 + 2/3) I; * (55)

(b) Velocity of deposition and dissolution as a function Of pH and aWPC. From (49),

(56)
 The electrode kinetics of the deposition and dissolution of iron 355

From (49) and (53),

a ln ke
2, (57)
8 ln aF@+“on-,v =

aIn ia.Fe 0, (58)

aln +es+aOH-,v =

a In iO
(59)
a In +@+aOH- = I.

(c) Corrosion potential as a function of PH. For the conditions when i&Fe = ic,=
the equation similar to (40) gives

aho,, 3 RT
(60)
aha,,-= -

(d) Velocity of corrosion as a function of PH.

(e) “Catalytic” nature of mechanism B. From (47) and (51),

ao,o = ~oH-aFe~+e-(V-Vo,sl)F/RT, (63)
(64)
If ~c,~ (activity of the catalyst during the cathodic transient) is not to effect the
position of the reversible potential, for V = VBev,Be,

ac,c = %7,a. (65)

With this condition (63) and (64) give
VO,ltev,Fe
= ii(vOA, + v0.51h (66)

But, considering (47) and (51) in the form

FeOH + e,- + Fe + OH-, (67)
Fez+ + OH- + e; + FeOH, (68)
one has

A&, + A&s = AGO,,, + AGO,, + RTln +,a+ = AGo,F.e-tRTh aFeS+ (69)

.’. A&w + AGO,,, = bGo,~e (70)

and this is equivalent to (66), i.e., (ac,JRev = &JR,. The catalytic nature of the
mechanism is thereby established.
356 J. O’M. E~OCKRIS,
D. DRAZIC and A. R. DESPIC

(iii) Mechanism C
OH,, + O&j, + e,,- (71)

OHad, + 2Fe + Fe,OH (72)

Fe,OH a.D.s. Fe” + FeOH + 2eo- (73)

FeOH Z+ Fe2+ + OH;, + e,-. (74)

or, H20 + OHad, + H+ + eo- (W

OHad, + 2Fe + Fe,OH (72a)

Fe,OH a.D.s. Fe2+ + FeOH + 2eo- (73a)

FeOH + Hf + Fe2+ + H,O + eo-. (74a)
These reaction steps may be treated in a manner analogous to that of the foregoing
A and B and the results of such a treatment are shown in Table 8.

(iv) Mechanism D
Fe + H,O + FeOH + H+ + e,- (75)
FeOH + FeOH+ + e,- (76)

FeOHf + Fe R.D.S. Fe,OHf (77)

Fe,(OH)+ + Fe2+ + FeOH + e,- (78)
FeOH + H+ + Fez+ + H,O + e,-. (7%
The associated magnitudes of mechanism D are shown in Table 8.

(v) Mechanism E
Fe + OH- R.D.S. Fe(OH)+ + 2eoW (80)
Fe(OH)+ + Fe2+ + OH-. (81)
(vi) Mechanism F
Fe + 20H- a Fe(OH), + 2eo- (82)
Fe(OH), + Fe2+ + 2OH-. (83)

(vii) Mechanism G (Kabanov, Burstein and Frumkin)8

Fe + OH- + FeOH + e,- (84)

FeOH + OH- R.D.S. Fe0 + H,O + e,- (85)

Fe0 + OH- + HFeO,- (86)
HFeO,- + H,O C+ Fe(OH), + OH- (87)
Fe(OH), + Fe2+ + 20H-. (88)
 TABLE 8. C~MPARLWNOPEXPERIMENTALBEHAVIOURS WITHTHATEXPECTED FOR VARIOUS MECHANISMS@ = O-5)

Mechanism
T Present results
Quantity (coefficients x 2.303)
A B C D E F G
_-
8 vFe 2RT RT RT RT RT RT 2RT 0.042 f OGO8
--
aIni. 3F G 2F 2F 7 T 5

avpe 2RT RT RT RT RT RT 2RT

-- -- -- -- -- -- -- -0.116 f OGO6
a In iCom F 2F 2F 2F F F F

1 2 2 2 1 1 1 0.8

1 2 1 1 1 2 2 0.9 f @05

2RT -- RT -- RT RT RT 2RT 4RT

--- -- -- -- -_-
3F F 2F 2F F F 3F

t 1 1 1 t 3 3 0.8 f 0.1

RT 6RT 4RT 4RT 4RT 2RT 3RT

-- --- --- --- --- -- --- -0.060 f oaO3
F 5F 5F SF 3F F 2F

-g -g -f 0 -f -0.5 f 0.01
358 J. O’M. BCEKRIS,D. DRAZIC and A. R. DESPIC

(3) Discussion of Mechanisms

(i) Catalytic theory (Mechanism B, Table 8)

This theory involves the following difficulties.*
(i) A physical picture of the catalytic action of FeOH is absent. Indeed, in the
cathodic reaction, FeOH increases with increasingly negative potential and this might
have been thought to block increasingly the amount of surface area free for the dis-
charge of Fe(OH)+. Similarly, it is a part of the theory that although FeOH is in
equilibrium with the cathode steady state, it is “irreversible” and its concentration
independent of potential, during the transients.
(ii) The theory involves the following internal inconsistencies.
(a) It indicates that
a VW,, 6RT
= ---9
alna,,- 5 F

whereas the values experimentally reported5 are significantly less than this (Heusler’s
method of obtaining the order of reaction with respect to non- involves the assumption
that:
~vcor, = --*RT
8 h aOH- F
(b) It indicates that
8 In icotr 2
8 In aoa- = -5
whereas Heusler reported no variation of icorr with pH.
(c) The observation that
aln ieorr
= 0 implies that, from (25),
a ln aon-
‘% Fe
noH- = i
%H

whence, because Heusler reported ‘$ge = 2 and non- = 2, c(,,n = 1. This value has
not been reported for hydrogen evolution on pure iron.

(ii) Mechanisms giving b, = f$.

From (20),
%Fe = 2nFer+ - &,Fe.

If &a+ = 1, “e,Fe = 0, which is impossible. Hence, Ra,Fe= 2, iII$ieS nFe8+ = 2,

RT
and thereby implies E,,Fe = - .
2F
Similarly, from (23) and a,,,, = 2,

nOH- + 1
<l
2.5
.*. noH-<I&
* The difficulties of this work are not lessened by the absence from the publication of presentation
of data except for Tafel lines at two pH values.
 The electrode kinetics of the deposition and dissolutionof iron 359

.‘. If ua,Fe is 2, nOH- is probably 1 (to avoid this, a,,= would have to be the
unreported value of 1).
Hence, a&Fe = 2, makes it likely that the mechanism involved has noH- = 1,
?Zgee+= 2 and &,Fe = 2. Among the theories discussed above, mechanism C and D
are consistent with these requirements. Of these two, D is preferable because it
involves only one-electron transfer per individual step, the existence of two-electron
transfers per individual step in a consecutive series being a process not usually postu-
lated, e.g., in non-electrochemical systems.

(iii) Mechanism consistent with the present work

The present results are clearly consistent with mechanism A (see Table 8). The
only other scheme involving OH- ions to which they approach is scheme F, which is of
importance primarily in high pH ranges. The present results are inconsistent with
i3V
mechanism Fin respect to of variation of the corrosion potential,
( alna OH-i,ap,ac >’
and corrosion rate with pH.
A further test can be applied to the proposed mechanism.
Consider the equilibrium constant of Fez+ + OH- + Fe(OH)+.
K~COH-(~Fes+hOtaI
CFeOH+ = (89)
1 + @OH-

and
(CFe2+)Total
(CFe++holution =
1 + KCOH-’
(90)
Hence, for mechanism A of Table 8,
kK, c~,~-(cFe*+)&$~’
i. = (1 + K, COH_)(W12)’ (91)
Thus, at low pH, one would expect
a In iO
= 1, (92)
( 8 In COH-W+hotcrl

as experimentally reported (Fig. 10).

At higher pH’s,
a In iO
= p/2. (93)
( ah COH-(cFe++hotsl >

A tendency to decrease of a rnlz lo _ with increase of pH is indeed noticed in Fig. 10.

The dotted lines of Fig. 10 show $0 the result of a numerical calculation using the KI
value from Lindstrand.13
(4) Instability of Mechanisms
The results reported here showed instability in the following respects: (i) About
10 per cent of the measurements gave b, as about g2F (11)
” Change from purified to
2RT to RT
unpurified solutions caused b, to change from - - . This abrupt change is
3F 2F
360 J. O’M. %XKRIS, D. DRAZIC and A. R. DESPIC

different in character from, e.g., that of the gradual effect of trace impurities in solution
on b, for the h.e.r., the poisoned value not being that characteristic of an alternate
mechanism. It suggests that the energy difference between mechanism A and a path
RT.
characterized by 2F is not great.
This is consistent with the fact that an extrapolation of the present i,, values to a
value of pH = 1, namely the approximate pH of NaHSO,, gives a value of about
lo-lo A/cm2 whereas that as reported by Hoar and Hurlen for aFea+= @05 M shows
a value i. ti lo-l1 A/cm2.
(5) E@ct of Anions
The effect of anions on i. (Table 5) is in the order ClO,- > SOa2- > Cl- > Ac-
> NO, and this is the order of the specific adsorbabilities of these ions on Hg, except
for NO,-. The electro capillary maximum for Fe is at -0.3 to -0.4V on the H scale,
i.e., the potential of the Fe surface during the experiments is not far from that of the
e.c.m. No knowledge exists concerning the heat of adsorption of these anions upon
iron. In the absence of such knowledge, it is only possible to state that the order of
effect of the anions is consistent with an increasing coverage effect in the direction
of decreasing io, due to specific adsorption (activity and pH effects having been
accounted for the values of Table 5).

(6) Hydrogen Evolution from H30+ and H,O.

The difference in heat of activation for the discharge of protons from H30+,
(A Hr’g), and H,O, (A HIIoX), has been discussed (Parsons and Bockris14), and it has
been shown that a rough calculation indicates that that for the discharge from water
should be some 7 kcal/mole more than that for discharge from H,O+.
Now
(io)E,O+= KHaO+ag$,+ e--h%?RT (94)
-II2 -AHIIo'IRT
(~o)H,o= KH,O;~H,O%,O+~ (9%
Therefore
(io>E,O+ = KH O+@H o+>
~~e-(AH~-AHI~X)IRT
(96)
(i0)H20 KH*O aHBO

if the i,‘s are compared at the same pH.

Thus for pH = 3.6,
6.8 10-7 _ KHso+ lo-=
KHaO m e(-An~-Anl~*I)Rr (97)
1.2 10-S
(A H$” - A H,““) = $9.6 kcal/mole. (98)
AcknowZedgements-The authors wishto acknowledgegratefullyfinancialsupport from the Corrosion
Research Council, and also, for the latter stage of the work, support from the United States Steel
Corporation. Two of us (D. D and A. R. D) wish to thank the Council of the University of
Belgrade for leave of absence.
REFERENCES
1. W. MEHL and J. O’M. BOCKRIS, J. Chem. Phys. 27,818 (1957); C’unad.J. Chem. 37,190 (1959);
H. GEREXHER, .I.Electrochem. Sot. 62, 256 (1958).
2. W. A. ROJTER, W. A. JUZA and E. S. POLUJAN, Acta Physicochim. 10, 389 (1939).
 The electrode kinetics of the deposition and dissolution of iron 361

3. G. OKAMOTO,M. NAGAYAMAand N. SATO,Proc. 8th Meeting CITCE, Madrid, 1956, p. 12. Butter-
worth, London (1958).
4. V. A. Kusmzov and Z. A. IOFA, Z. Fiz. Khim. 21,201 (1947).
5. K. E. HEUSLER,Z. Elektrochem. 62, 582 (1958).
6. T. P. HOAR and T. HURLEN,Proc. 8th Meeting CZTCE, Madrid, 1956,~. 445. Butterworth, London
(1958).
7. B. KABANOVand D. LETKIS,Dokl. Akad. Nauk. S.S.S.R. %I,1865 (1947).
8. B. KABANOV,R. BURSTEINand A. FRUMKIN, Disc. Far&y Sot. 1,259 (1947).
9. A. M. AZZAM, J. O’M. B~CKRIS,B. E, CONWAYand H. ROSENBERG,Trans. Far&y Sot. 46,918
(1950).
10. K. B~NHOE~ZR and W. JENA, Z. Elektrochem. 55, 151 (1951).
11. J. D’ANs and W. BRECKHEmER,Z. Elektrochem. 56, 585 (1952).
12. M. STERN,J. Electrochem. Sot. 102, 609 (1955).
13. F. LINIXTRAND,Svenska Teknologforewenger. Tekn. Tidskr. 56,282 (1944).
14. R. PARSONSand J. G’M. BOCKRL?,Trans. Faraday Sot. 47,914 (1951).
15. J. O’M. BOCKRIS,Modern Aspects of Electrochemistry Chap. 4. Butterworth, London (1954).
16. P. DELAHAY,NewZnstrumentalMethods in Electrochemistry, p. 184, Interscience N.Y. (1954).