The Mechanism of the Cathodic Hydrogen

Evolution Reaction'
J. O'M. BOCKRIS AND E. C. POTTER
Imperial College of Science, London, England
ABSTRACT
Some of the o u t s t a n d i n g problems of concept and m e c h a n i s m in the field of cathodic
h y d r o g e n evolution kinetics are discussed a n d clarified. A full d e r i v a t i o n and correla-
tion of kinetic e q u a t i o n s which assume no m e c h a n i s m reveals expressions for several
p a r a n m t e r s which take wflues specific to one or more mechanisins. The use of sta-
tistical methods of t r e a t m e n t of d a t a proves indispensible in e s t i m a t i n g the values of
these parameters. A f u r t h e r m e t h o d of ascertaining the m e c h a n i s m of the hydrogen ev-
olution reaction is to examine the kinetics of the individual reaction paths, so t h a t the
expected values of p a r a m e t e r s conmmn to all p a t h s m a y be deduced. A n u m b e r of mecha-
nisms i m p o r t a n t in acid and alkaline solulions are thus t r e a t e d , and are shown to be dis-
t i n g u i s h a b l e experimentally. Using already published data, the actual conditions under
which various reaction p a t h s take place at mercury, silvcr, nickel, and smooth p l a t i n u m
cathodes are calculated. I t is not only possible to compare these deduced d a t a with ob-
s e r v a t i o n , a n d t h e r e b y verify the occurrence of a p a r t i c u l a r reaction p a t h , b u t also to
d e n m n s t r a t e the impossibility of some m e c h a n i s m s in specific cases. The recent ad-
vances which the foregoing m e t h o d s have made possible are discussed in r e l a t i o n to
d a t a which have r e c e n t l y beconm available.

I. INTRODUCTION 10-1~ gram moles/liter of As203, CS~, CO, KCN,

The electrolytic evolution of hydrogen is gener-
ally chosen as the main subject of experimental and
theoretical research in electrode kinetics because it
was originally thought to be one of the simplest
electrode reactions. This expectation has not been
confirmed. However, work already done has em-
phasized the importance of the reaction, and its
practical significance to studies in corrosion has
stimulated further attempts to solve the problems
discovered. In the following, some general equations
of electrode kinetics are formulated and indicate the
experimental methods best suited to mechanism
determinations.
II. SOME DIFFICULTIES OF TECHNIQUE AND CONCEPT
It is notoriously difficult to measure the velocity
of an electrode reaction and the corresponding
electrode potential with reproducibility. The best
agreement obtained between workers in different
laboratories is + 8 my on Hg and =t=20 my on solid
(e.g., Ni) cathodes.
Many different ~ypes of Tafel lines are reported
in the literature. Some of these are collected in Fig.
1, from which it is seen that only the first and the
last two represent forms which are free from vitiat-
ing influences. The principal vitiating factor in ex-
perimental work arises from mimlte traces of im-
purities in solution. For example, addition of only
M a n u s c r i p t received M a r c h 26, 1951. This paper pre-
pared for delivery before the W a s h i n g t o n Meeting, April 8
to 12, 1951.
169
etc., detectably affect the electrode potential at a
given current density at nickel cathodes (1). Re-
producible results can only be obtained in solutions
which have been purified by pre-electrolysis on to an
auxiliary cathode (2). Use of this method involves
laborious work to determine the amount of pre-
electrolysis necessary for a given electrode and solu-
tion. The criterion by which the optimum conditions
for pre-electrolysis are found is that passage of
further coulombs at higher potentials makes no
further difference to the experimental results.
Even then, the initial state of the solution before
purification may be difficult to reproduce so that
sometimes less pre-electrolysis is necessary than
at other times.
Many difficulties in preparing the surface of solid
metals in a reproducible, clean state have caused
most studies to t)e made on mercury. In one method
(3) the electrode in wire form is heated in a stream
of hydrogen to remove oxide films and sealed into a
thin glass bulb containing pure hydrogen. After
pre-electrolysis of the solution the bulb is broken by a
glass probe thereby immersing the electrode in the
purified and oxygen-free solution. A further possible
method would be to heat the wire electrode electri-
cally, e.g., inductively inside the cell, containing
hydrogen, before placing it in solution.
In addition, a few difficulties connected with the
general attitude of much past work to studies of the
hydrogen evolution reaction may be mentioned.
First, the difficulty of measuring the electrode poten-

Downloaded on 2014-05-28 to IP 192.236.36.29 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 170 JOURNAL OF THE ELECTROCHEMICAL SOCIETY April 1952

tial, and the comparative ease of measuring the cur-
rent, directed attention away from the fact that
the problems coflnected with overpotential were
purely kinetic, and that attention should be focused
upon the current. Second, considerable attention
has been paid to the determination of i0, the current
difference between the metals. If AV is a potential
difference measured by means of a potentiometer
connected to wires of composition M, then,
A V = VM-- V p t + X .
Now, X ~ ~M -- qh,t, where q)M and ~p~ are the
work functions of the metals M and Pt respectively

D I S S O L ~
O LIMITINGCllfiRENT ~
~1 R[$I$1A~IC-EERROR i - T Y P E /
2
DEPOLARIZED (neglecting the Peltier heat in the usual way).
Hence,
VM = A V + V p t - - ~ M + ~ p t .
Or, if the experimental conditions are appropriate,
so that AV represents a hydrogen overpotential,
VM = ~ + V r t - - r + r (1)
where K is a constant potential difference for all
cells in which hydrogen overpotential, 7, is meas-
O ured at the same pH and temperature. But also,
it

~
,~[ OXIOE FILM TYPE the relation between current density, ic, and poten-
tial, for a kinetic electrode process in whieh the
reverse reaction can be neglected, is given (see
below) by an equation of the form,

I ELECTROCHEMICAL~. TYPE
SILVER J
where a and K1 are constants and the pH is assumed
to be constant. Thus,

[--a(~- r (2)
LOG CURRENT DENSITY----~- LOG CURRENT DENSITY ic = Kx exp RT "
FI~. 1. Some types of Tafel line for hydrogen deposition
16
Therefore, if when ~ = 0, ic = i0, then,
I i i i i i i t

= K ~ e x P k[ | R T J ' (3)
14

t2

I0
where K2 is a new constant.
,_o
A relation of type (3) is indeed observed, and is
shown in Fig. 2. It follows, then, that the most ap-
i
6
propriate potential at which to compare the rate of
MOFeW ~
the hydrogen evolution reaction at various electrode
4
materials is not the reversible hydrogen potential
2 but the absolute zero of potential. In the absence of
reliable knowledge of this, the potential of the
0 . ,I I I I I I I I
3.0 3.5 4.0 4.5 5,0 5.5 6.0 6.5 7.0 7,5 charge-free surface appears to be the best reference
dp IN s
potential.
FIG. 2. Relationships between exchange current and Similar conclusions apply to the heat of activa-
thermionie work function for a number of cathodes.
tion, AH*, which is usually quoted at the reversible
hydrogen potential. However, measurements of i0
passing in either direction at the electrode at the
and AH* retain considerable relevance because they
reversible potential, i.e., at zero overpotential.
assist in distinguishing, between various possible
Consider a complete galvanic cell at constant
mechanisms of the electrode reactions.
temperature consisting of a metal M (the cathode
Lastly, the rate of the evolution reaction depends
in a cell in which overpotential is measured) and a
upon the difference of two inner potentials $, i.e.,
Pt electrode in the same solution as the metal M.
it depends upon a Galvani potential, where 4~ is
There are three potential differences in the cell:
defined by
VM, that at the boundary M-solution; Vpt, at the
boundary Pt-solution; and X~ the contact potential ~b-- ~b+ x, (4)

Downloaded on 2014-05-28 to IP 192.236.36.29 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 Vol. 99, No. ~ CATHODIC H Y D R O G E N EVOLUTION R E A C T I O N 171

~b being the Volta potential of a phase and x the In these circumstances the required variances are
surface potential. The latter potential is strongly calculated from equations of the type,
dependent upon the surface condition of a phase,
and hence the overpotential at a given rate of reac- V(/5)
_ E(pi-
nr(nr-
p)2
1)'
(10)
tion would depend sharply upon trace impurities.
where V(p) = variance of p,
I I I . STATISTICAL METHODS IN MEASUREMENTS OF pi = the value of the parameter p of the
HYDROGEN ELECTRODE KINETICS -th
j Tafel line,
In order to observe with reasonable and known p = (Zp~)/n,
accuracy those of the theoretically important param- and nr = number of Tafel lines in the popula-
eters in overpotential studies (e.g., temperature tion.
and concentration effects) which may vary little It can be shown that the relation between log i0
more than the reproducibility of the measurements, and absolute temperature, T, is
it is essential to replicate the observations and to
apply statistical methods. log i0 = log B AH* (11)
2.303RT'
In applying the method of least squares to esti-
mate the parameters of a Tafel line, those formulae where log B contains a temperature term, but is
should be employed which correspond to the cur- usually considered to be temperature independent
rent as the independent variable and the potential over the ranges of temperature often used experi-
as the dependent variable. This means that the sums mentally. It is evident from (11) that since log B
of squares of the vertical deviations of the experi- is formally the value of log i0 at infinite temperature,
mental points from the computed Tafel line are the value of log B is obtained by a long extrapola-
minimized, and that, consequently, a prediction of tion of the observed relation. Also, since log i0
potential at a stated current has the minimum is itself obtained by lengthy extrapolation of a
possible error based on the original observations. Tafel line, it is not surprising that published values
Thus we have of log B show lack of agreement (4). In order to
obtain the most probable values of AH0* and log
a = ~ - b2, (5)
B from the data, it is clear that by treating 1 / T as
and the independent variable and log i0 as the dependent
variable, equations analogous to (5) to (8) can be
b - nZxy - Zx2y _ N (6)
n~x ~ - IF,x) 2 D '
applied.
The experimental accuracy which must be
where x and y take the corresponding experimental ac.hieved to attain any desired limits of error in an
values of log i~ and ~, respectively, n is the number estimate of log B may be calculated in the following
of pairs of observations, 9 is ( X y ) / n , and ~ is ( ~ x ) / n . way. The variance of log i0, V(log i0), is given with
The variances of the Tafel line parameters may be sufficient accuracy by
calculated from
V(m)D' (12)
V(log i0) - n' '
V(b) = [n~Y2 - (~Y)2]D - N2
(n -- 2 ) D ~- ' (7)
where V ( m ) is the variance of m,
V i a ) = V(b) [ D / n + ~2], (8) m is H*/2.303R,
D' is the appropriate denominator corre-
and sponding to D, in (6),
and n' is the number of pairs of observations
V(log i0) -- V ( b ) [ D / n + (log i0 - 2) 2] 9 (9)
be of log i0 and 1 / T .
Also, the variance of log B, V(log B), is obtained
where V(b), V i a ) , and V(log i0) are the respective from an equation analogous to (8) as,
variances of b, a, and log i0, each with n - 2 degrees
V(log B) = V ( m ) [ D ' / n ' + ~2], (13)
of freedom.
When replication of observations is carried out, where ~ = Y r / n ' = ( F . 1 / T ) / n ' . From (12) and (13)
it is often desired to replace the population of Tafel and noting by analogy to (6) that D ' / n ' = Z(r - e)~,
lines so obtained by one mean line. The parameters
of this line may be calculated by bulking the ob- V(log B ) Z ( r -- ~.)2 (14)
V(log i0) = + _
servations and applying (5) and (6), but the vari-
ances of these parameters cannot be found by In order to use (14) ntlmerically suppose that the
applying equations (7) to (9) to such bulked data. values of r are derived from the experimental

Downloaded on 2014-05-28 to IP 192.236.36.29 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 172 JOURNAL OF T H E E L E C T R O C H E M I C A L SOCIETY
temperatures 0 ~ 10 ~ 20 ~ 30 ~ and 40~ and fur-
thermore, let it be arbitrarily decided that the maxi-
mum inaccuracy tolerable in log B corresponds to
95 per cent confidence limits of 4-0.5. Hence, as-
suming that the errors in log B are normally dis-
tributed, we have 3.18 x/V-(log B) = 0.5, where 3.18
is the value of Student's t for 3 degrees of freedom
and the 0.05 probability level. Hence from (14)
V(log i0) is 0.000286. Now, from (6) and (9) it is
seen,
V(log io)
V(b)[(logio - log~)2 + Z(logi~ - logic) 2]
52
where log i~ = 2. This equation may be applied to
typical experimental data for the mercury cathode,
which (reference shows) gives the most reproducible
overpotential measurements. Let log i~ take the
nine values, - 7 , - 6 . 5 , - 6 , . - . - 3 , so that the
value of log i~ is - 5 . 0 . Taking b and log i0 for the
mercury cathode in aqueous acid solution as 0.115
volt and -12.0, respectively, and using the value
for V(log i0) calculated above, it is found that V(b)
is 0.0592 9 10-6 (corresponding to 95% confidence
limits in b of 4-0.0006 volt). Such an accuracy in
estimating the slope of the Tafel line is unattain-
able even with modern techniques. This is evident
since an error of only 1 mv in estimating the over-
potential at 10 7 amp/era 2 and a similar error in the
opposite direction at 10 3 amp/era 2 is sufficient to
cause an inaccuracy of 4-0.0005 volt in b.
In a recent investigation using nickel cathodes
(10) one twentieth of the experimental Tafel lines
attained less than the above inaccuracy in slope,
but the slope of the mean line of a population of
Tafel lines could not be determined with an accur-
acy (95% confidence limits) greater than 4-0.004
volt. The 95 per cent confidence limits for log B
were on the average 4-2.2, each limit being based
on an average of 25 pairs of observations. It, there-
fore, appears that the quantity B is not a useful
distinguishing criterion of reaction mechanism.
While variances (which should be quoted with
their number of degrees of freedom) are indepen-
dent of distribution of errors they are not readily
assimilable as measures of error, and it is customary
to assume (in the absence of evidence otherwise)
that the distribution of errors is normal so that
confidence limits may be quoted. Some caution is
necessary here; for if it is assumed that the param-
eters b and log i0 of the Tafel line have normal dis-
tribution of errors, then the errors in i0 itself are not
normally distributed. Consequently, the distribu-
tion of errors of X which is calculable from i0 (see
later) is not normal, although it is unlikely that seri-
ous error would arise by assuming normality.
Downloaded on 2014-05-28 to IP 192.236.36.29 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
April 1952
IV. GENERAL EQUATIONS
A. Current and Potential
Suppose that ki is the specific rate constant of
an unspecified rate-determining step in the forward
direction of a reaction (ions depositing). Then,
kT ( (AG*)I~
k~ = K-h- exp ~ /, (16)
where (AG*)~ is the standard free energy of the
activated complex of the rate-determining step
with respect to the initial state (here assumed to be
(15) X gram ions of hydrogen ions constituting part of
the monolayer adjacent to the electrode surface,
often termed the Helmholtz double layer) of the
reaction, where ~ is the transmission coefficient,
and where ]c and h have their usual meanings. Let
a, be the activity of H30 + ions in the initial state.
Then the forward velocity ~ is
: al]r (17)
Also, the forward current is
; = ~XF, (18)
where X is the number of electrons necessary so
that one act of the rate-determining step can occur.
Suppose a potential difference AO~ is applied be-
tween the electrode and the initial state of the reac-
tion. If this potential difference is positive it retards
the flow of reactants over the energy barrier of the
rate-determining step, i.e., it makes (AG*)I more
positive. The potential difference increases (AG*)I
by /~XFA~b~ where /~2~q~cis the potential difference
through which the electrons pass before they reach
the transition state. (Work done on the system
after it has passed the transition state does not
affect the velocity of the reaction.) Hence 0 < f~ < 1.
varies with the reaction mechanism, and is a
complex quantity, except in certain simple mecha-
nisms. Thus, in the discharge reaction H~O + ~- e -+
M H -t- H20, ~ = 1 if the energy barrier is sym-
metrical. General and limiting values of ~ are evalu-
ated below for common mechanisms. From (16),
(17), and (18),
= KXF -ff a~ exp RT . (19)
By a similar argument,
i -~ K},F aF exp
(20)
.( (zxa*h-(1- ~)x~+~F)
~7~
where (AG*)2 is the standard free energy of activa-
tion of the reverse reaction referred to the initial
 VoL 99, No. 4 CATHODIC H Y D R O G E N EVOLUTION R E A C T I O N 173

state of this reaction, and aF is the activity of the By expanding exponentials, equation (24) be-
entities in this initial state. comes,
If 5 ~ = A~b,, the reversible potential,
i~ = Xi0~F (32)
-- i ' = 0, where ~ = i0 = ~. (21) RT
From (19), (20), and (21) and since v = A~b~-- &b~, Thus, i~ and ~ are linearly related at sufficiently
low overpotentials. Also, from (32) and formally
z = io exp \ - ~ ), (22) allowing for the possibility of Oi~//O~?depending on 7,

= io exp (-~ (1 -RT~)XvF~], (23) X = -- Fi0 \ ~ / , - . 0 " (33)

X (see Ref. 21) can be termed the electron number

= - io L e x p k , - RT ) of the reaction and is estimated experimentally by
(24) applying equation (33). As shown below, X is a valu-
-- exp ((1 - ~ X v F ) I " able diagnostic criterion of reaction mechanisms.
Special Case 3.--Relations connected with non-
linearity between ~ and log ic.
Equation (24) is the most general expression for the
From (24),
cathodic current, and represents a more fruitful form
of the relationship than has been stated hitherto. ~XnF~
Special Case 1. Relation between i~ and y at i~ = i0exp R T ] [1 - f], (34)

appreciable overpotentials (e.g., v more negative where

than about - 7 5 my, see special case 3).
Equation (24) becomes f = exp \ R T ] " (35)

i~= i0[exp( ~XT/F~/j, (25) When the departure of the Tafel line [see equation
(27)] from linearity due to the reverse current in
or (20) is just detectable experimentally, then f is
RT RT just distinguishable from zero, and v = ~. The
v = ~ In i0 -- p~,~,
^~
in i~. (26) smallest detectable and significant value of f, i.e.,
f.~, depends on the experimental design, but, pro-
Equation (26) is Tafel's equation, vided a minimum number of about five pail's of ob-
servations fall within a region about 0.025 volt on
= a - b log io (27)
either side of ~.~, it may be assumed that f~ is 0.05.
or Therefore, from (35),
RT
= a -- a F In i~ (28) f~ = 0.05 = exp \ R T ] , (36)

so that, comparing (26) and (27) and (28), it follows or at 20~ with ~ expressed numerically in volts,
that: 0.075
RT X = -- (37)
a = ~ In io, (2~)
Equation (37) gives an alternative method of es-
2.303RT timating X experimentally and has not been pre-
b - - - , (30) viously described.
~hF
a = fiX. (31) B. Potential and Time

Equation (31) shows that a is a composite quantity, 1. During Build-up of Overpotenlial

and need not be between 0 and 1 since fl and X For a completely polarisable electrode, the net
can have maximum values of 1 and 2, respectively. constant cathodic current is the sum of the con-
Thus, the experimentally observed values of a of denser (i') and faradaic (i') currents:
2, which have been difficult to explain hitherto, i~ = i' + i'. (38)
are a special case of equation (31).
Special Case 2.--Relation between ir and v at Let C be the differential capacity of the electrode-
low overpotentials (e.g., n less negative than - 2 0 solution interface. Then,
mv). - C d~t = i' dt (39)

Downloaded on 2014-05-28 to IP 192.236.36.29 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 174 JOURNAL OF T H E E L E C T R O C H E M I C A L SOCIETY April 1952

where dyt and dt are infinitesimal changes in m Or, integrating between h and t2 and the correspond-
and time, t, respectively. Let ~t and % be values of ing values of A~,
at times t = t and t = or respectively. Since i"
is given by (24), then from (24), (38), and (39),
h - t~ -
CRT [
io~XF e•
x,.F 7
RT-/In (An~ - A~2). (48)
Cd~ = i~ - io exp
at
[ ( ~ ]
(40)
Equation (48) indicates that during the initial
build-up of overpotential, ~t relatively rapidly
-- exp ((i -R~XntF)]. reaches practically constant values. For example,
with 2.303RT/B~F = 0.1, the time required for
Special Case 1. dC/d~ = 0; ~t more negative overpotential to rise from within 10 mv to within
than - 7 5 my. 1 mv of ~ is the same as that required for the in-
Equation (40) becomes terval 1 my to 0.1 mv from %.

2. During Decay
-Cd'-~t=i~-i~ fl~_~F) ] . (41)
(i) Relations in which it is assumed that dC/d~ = O.
Let From (41) with i~ = 0, and *t~ more negative
than - 7 5 my,
i" _ f, (42)
io
dt - C exp , (49)
and suppose.f' < 0.05 so that i" is negligible in (38),
then from (38) and (39), where n't is the overpotential l seconds after the corn-
mencement of decay. Integrating,
io = Cdm (43)
dt "
t -- io~XF exp \ R T ] -F eonst. (50)
Hence, from (42) and if f~ < 0.05, d~t/dt is a con-
stant. To find the region of overpotential in which At t = 0, 7t~ = ~ . Hence,
this result is valid, we apply (38) which becomes
i~ = i" at t = ~ , so that from (25),
Const. = -i0/3X~ exp . (51)

i~ = i 0 [ e x p ( /~F)]. (44) From (50) and (51),

Also from (25),

t - i0~XF e x p \ R T ] exp . (52)
i" = io [exp (fl~F)]. (45)
Special Case 1. ~'t = % - A{, where Av' is less
negative than - 2 0 my.
Thus, from (42), (44), and (45), Using this condition in (52) yields

f'= exp ( f l X ( ~ % re)F). (46)

exp RT ]
(53)
For example, let f ' = 0.03, then from (46), taking a flXFi~t
typical value of 2.303RT/~XF from experiment as CRT
0.12, % - w = -0.15, i.e., the relation of poten-
Having regard to the condition for A~', (53) becomes
tial to time is linear [(43) is valid] to within 0.15
volt of the constant value n~.
, iot exp . (54)
Special Case 2. dC/dy = 0; ~ more negative
than - 7 5 mv; (v~ - vt) less negative than - 2 0
mg. Equation (54) shows that the initial decay of over-
Let ~ - ~t = An; at t = h, A7 = A ~ I and at t = potential is linear with time. Hence, linear extra-
t~, A~? = AV:. Hence, from (40) and (44), polations to zero time in the commutator method of
measuring overpotential are valid under the above
d,,_dt c.i~ ( /~F)] defined conditions.
Special Case 2. A~' more negative than approxi-
(47)
mately - 5 9 mv (normal decay); nt more negative
9 [1- exp \--R-T/J" than - 20 mv.

Downloaded on 2014-05-28 to IP 192.236.36.29 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 Vol. 99, No. ~ CATHODIC H Y D R O G E N EVOLUTION REACTION 175

Applying this condition to (52), and thus neglect- Equation (61) enables estimates of differential ca-
ing exp(~X7~F)/RT gives, pacity obtained from the build-up of overpotential
to be corrected for faradaic current from observa-
t - flXFio
exp (\ xT;F
RT ]" (55)
tions of the corresponding decay curve (22). The
differential capacity may be obtained from the de-
, RT RT CRT cay curve by application of (63), it being also neces-
9". 7t = ~ In t -- ~--~ In BXFi0--" (56) sary to know the values of b and log i0 of the cor-
responding Tafel line.
Hence from (56) and (30),

dTi 2.303RT
C. Current and Temperature 2
- - b. (57)
d(log t) /~XF From (19) and (21),
Equation (57) shows that overpotential decays t~T ( (AG*)~
logarithmically with time under the above condi- io = ~XF ~ - al exp - ~-7~ /, (64)
tions and that the slope of the logarithmic decay
curve is the same as that of the corresponding Tafel
line [cf. Butler (24)].
io = KXF ~ - ar exp ( ], (65)
Special Case 3. 7't less negative than - 2 0 my.
where tAG*)1 is the standard free energy of activa-
From this condition, the faradaic current i",
tion for the forward direction of the rate-determin-
is given by (32), also from (38), in which during
ing step at the reversible potential. Hence,
decay i~ = 0, and using (39) it follows that

dT't ~io 7t F i0 = B e x p ( AH:

RT/' (66)
dt - CRT " (58)
Integrating (58), where

CRT ,
t - k/~0ff m [ 7, [ + const. (59) B = K•F ~ - a, exp , (67)

Equation (59) shows that the decay curve of over-
potential against time becomes asymptotic to the
time axis, the relation being exponential. The equa-
tion may be used to find the differential capacity
(25).
(it) Relations in which dC/d7 = f(7).
During decay, i" = C dT'Jdt. Using this in (38)
and (39) gives
and AH* and AS* are respectively the heat and
entropy of activation corresponding to (AG*)I.
Assuming AH* and AS* are not temperature de-
pendent, and that over small ranges of temperature
B is constant, the plot of In i0 against; 1/7' has a
slope of - A H * / R .
Further, when n is more negative than - 7 5 my,
(25) and (66) give

- c
d7t - c
d,/t (60) ic = B e • AH*;T~TF ). (68)

or Hence, taking B and ~3~ as temperature inde-

ir pendent,
J
C = ( d( d' )T_t ~ _ d\ d[ ,t h] (61)
( 0(ln ic)~ _ AH* + fl~TF
-oT~] , R T2
(69)

where the values of C, dT't/dt, and dTt/dt refer to the

From (26), and with conditions assumed as for (69),
same overpotential 7.
Further, if (54) is applied to infinitesimal changes
during decay,
(0,) = - (70)

From (69) and (70) it follows that

= - - ~ exp \ (62)
\d-i/,, RT /
(07) AH0* + ~ T F (71)
so that ~o = flXFT
2 O t h e r relations between c u r r e n t a n d t e m p e r a t u r e have
In C = Inio RT \ dt/,h" (63) been given by Agar (23).

Downloaded on 2014-05-28 to IP 192.236.36.29 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 176 ,JOURNAL OF T H E E L E C T R O C H E M I C A L SOCIETY April 1952

V. POSSIBLE REACTION PATHS OF THE HYDROGEN M~+ + ze slow Ml

------+
EVOLUTION REACTION 3
M1 + zH,O fast M~+ + zH + zOH- (L)
-----)
A. Source of Proton is H~O+
H20 + M ~ H + e fast H2 +OH- +M1
The possible paths are (M being a metal atom):
slow M~+ + ze fast M~
H30 + + e _____~ MH + H20 -----)
(A)
fast Ml + zH20 slow M~+ + zH + zOH- (M)
MH + MH ____, 2M + H2 ----.+

H20 + M 1 H + e fast H2 +OH- +M1

fast --_~
H~0 + + e ___~ MH + H~O
(B)
slow M~+ + ze fast Mx
MH + MH ___, 2M + H2 ------->

slow ~1 "31- ZH20 fast M~+ -4- zH -4- zOH- (N)

)
H~0 + + e ___~. MH -~" H 2 0
(C) Ho.O + MtH + e slow H2 + OH- + M1
fast ------+
HaO + + MH + e ___, I.i~ + H20

fast The kinetics may be complicated by the presence

H~O+ + e - - - ~ MH + H20 (D) of: (a) simultaneous reactions in which, say, two
slow desorptive reactions proceed at the same velocity
H~0 + + MH + e ) H2 -~- H 2 0
in the steady state; (b) dual reactions in which two
B. Source of Proton is H 2 0 stages ill the evolution have virtually the same
slow energy barrier (to within, say, 3 kcal); (c) linked
H20 + e , MH + OH- reactions, where the energy barriers of various steps
(E)
fast are of appreciably different heights, but the kinetics
MH + MH ____. t{2 -~- 2M
of the overall process depend upon the heights
fast of two or more energy barriers.
HoO + e _____~ MH + OH-
(F) It has been shown recently that discharge from
slow
MH + MI~ __---~ H2 + 2M water molecules in acid solutions is improbable
slow (5). Hence, reactions E to N are likely only in al-
H20 + e ~ MH + O H - (a) kaline solutions. The reactions A to D are the best
fast known and the most reliable material exists by which
H20 + MH + e ____> H2 + O H - + M
their kinetics, developed below, may be compared
fast with experiment. A, B, and D are commonly termed:
H~.O + e ____~ M H + 0H-
(H) The Slow Discharge, Atomic Hydrogen, and Elec-
H20 -4- MH + e slow§ HHo
_ -4- OH- + M trochemical Mechanisms of the electrolytic evolu-
In alkaline solution it is also possible that the tion of hydrogen, respectively.
metal M ~+ ion discharge is an intermediate step in
V I . THE KINETICS OF SPECIFIC REACTION PATHS
the evolution reaction.
slow The most important reaction paths will be con-
M~+ + ze ~ M1 sidered here in some detail. It will be assumed that
M~ + zH20 f a s t M~+ + zH +zOH- (I) the adsorption of hydrogen upon the electrode is
governed by a Langmuir isotherm and that the
MH + MH fast+- H~ + 2M
velocity of the reverse desorption reaction may be
fast neglected at potentials not near the reversible hy-
M] + + ze M~ drogen potential.
M~ + zH20 slow M~+ + zH +zOH- (J)
A. The Source of the Protons is H~O+ and the Atomic
MH + M H f a s-t H.~ + 2M
Hydrogen Reaction is Desorptive (Paths A and B)
fast
M~+ + ze M1 1. General Kinetic Equations
M1 + zH20 _____~fastM] + + zH + zOH- (K) The reactions and velocities Vl, v2, va to be consid-
slow ered a r e :
MH + MH ____~ H2 + 2M
a The reaction paths suggested are by no means exhaus- H30 + + e v~ MH,
___. (72)
tive, e.g., it is possible t h a t the discharge of hydroxonium
ions may occur with an electrochemical desorption involv-
ing water, i.e., H~O + MH + e ~ H2 + O H - + M. MH - -~)2
4 H~O+ + e, (73)

Downloaded on 2014-05-28 to IP 192.236.36.29 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 Vol. 99, No. 4 CATHODIC HYDROGEN EVOLUTION REACTION 177

and 2. Kinetics when Discharge is Rate-Determining

V3
MH + MH _~ H2 + 2M
Let al~+ be the activity of hydrogen ions in the
electrical double layer in gram equivalents per liter4;
a , be the activity of hydrogen atoms on the elec-
trode surface in gram atoms per cm2; and x be the
fraction of the surface covered with adsorbed hydro-
gen (x is more rigorously defined as the fraction of
the available surface covered). Then an = l0 -9 x
if there are 101~ free spaces for adsorption per cm e,
of the electrode surface, and assuming the activity
coefficient of the adsorbed hydrogen is unity it
follows,
v~ = k~ all+(1 -- X) exp ( -- 2 R T / '
v~ = /~ 10-~ x exp \ 2 ~ ] '
(Path A)
(74) (i) General condition for slow discharge mechanism.
- - T h e condition is
10(al -]- a2) < a3, (85)
where the factor 10 is an arbitrary limit of signifi-
cance. The use of the terms "a" in the condition
rather than the "k" terms provides for the possibil-
ity of changes of mechanism occurring on varying
potential or hydrogen ion concentration. The
linking of a~ a M a2 by a positive sign in the condition
is clear, since, by the nature of the reactions con-
cerned, a decrease in a2 increases x, thereby de-
creasing a~ and increasing a3: that is, a decrease in a2
(75) increases the probability of the slow discharge
mechanism, as shown by the condition (85) above.
(ii) Coverage of surfaee.--From (85) and provided
(76) (a~ + a2) 2 > 10a2(a~ + a~), i.e.,

and al > 9a2, (86)

va = ka 10-1Sx2, (77) it is easy to show that 8a~a3 > (a~ + a2) 2. Hence
using this condition in (79) it follows
where k~, k~, k3 are specific reaction velocities, and
the energy barrier of the discharge reaction is as- x = 9 (87)
sumed to be symmetrical. 3r = (Ar -- ~-) where
Ar is the inner potential difference of cathode and (iii) Tafel line.--From (80), (83), and (87) it
solution and ~- is the potential difference between follows
the bulk of the solution and the plane passing
through the center of the ions adjacent to the
cathode surface.
ic = FklaH+ exp ( -- 2 R T / " (88)

In the steady state, Hence from (27), (29), and (30) for the slow dis-
v~ - v2 - 2v3 = 0. (78) charge mechanism,

Solving (75), (76), (77), and (78) for x gives the a = 0.5; (89)
real solution: and, since X = 1,

X ~-
--(a, + a2) -4- g/(al + a2) ~- + 8a, a3 (79) = 0.5. (90)
4a3
(iv) Effect of pH in pure dilute acid solution.-
where Assuming that the Stern model of the electrical
ACF '~ double layer prevails at the metal-solution inter-
al = /qan+ exp 2~]' (80) face, that the solution is dilute, that specific ad-
sorption of ions is absent, and that the interracial
potential conditions are more than about 0.3 volts
a2 = 10-9k~ exp \ 2 ~ / ' (81)
from those of the elcctrocapillary maximum, the fol-
lowing equations may be validly used,
and
a3 = 10-1Sk3. (82) an+ = (an+)B e x p ( - ~ T F ) , (91)

The current is, therefore, (92)

ACe= A C t + n ,
i~ = 2Fv3 = 2F10-~Sk~x ~ = 2Fa3x ~ (83)
Ar = R T In (all+)., (93)
= F [--(al "-~ a2) -4- %/(al -4- a2)2 -4- 8ala.~]2. (84)
8a3 and
4 For ease of representation the suffix H30+ is replaced t" = Const. -t- ~ ' In (aH+)B, (94)
by H +.

Downloaded on 2014-05-28 to IP 192.236.36.29 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 178 J O U R N A L OF T H E E L E C T R O C H E M I C A L S O C I E T Y April 1952

where (an+)B is the activity of hydrogen ions in the Hence values of a and, therefore, of fl are complex
bulk of the solution. Using these equations in (88) functions of potential, and can be calculated from
2RT (101) by the use of (29) and (30), taking ~ as 2.
it is found that n = Const. In i~. There-
F Two important limiting conditions arise from (101).
fore, overpotential is independent of the pH of the
solution with the above mechanism under the con- Limiting condition 1 :
ditions (85) and (86).
If k210 -9 exp > 10, (102)
(v) Effect of pH in the presence of excess neutral
salt.--In the presence of excess neutral salt (e.g.,
LaC13) ~ decreases to a constant value near zero. b - 2.303RT _ 0.029 at 20~ (103)
Under these conditions (91), (92), (93), and (94) 2F
may be combined to Therefore,
a = 2 a n d ~ = 1, as}, = 2 (104)
77 = Const. - 2R_
TF lni~ -t- %T In (a,,+)B. (95)
Limiting condition 2:
Equation (95) shows that under the above conditions If A0 tends to - ~ ,
(85) and (86), overpotential should decrease nu-
merically by 58 mv at 20~ upon decreasing the b~ ~, (105)
pH of the solution by one t~nit. so that
(vi) Effect of neutral salt at constant p H . - - F r o m a ~ 0, and ~ ~ 0. (106)
(88), (92), (93), and (94), and considering (all+),
as constant, The forms of (100) corresponding to (103) and (105)
are, respectively,
,7 = Const. 2RT In ic -- ~-. (96)
Y [klaIt+12 ( 2A~bF~
ic = 2 F k s [ _ ~ - 2 J exp RT]' (107)
Since on addition of neutral salt to the electrolyte
~" becomes more positive and approaches zero, and
overpotential increases numerically. ic = 2Fkal0 -18. (108)
3. Kinetics when Atomic Combination is Rate-Deter- Equation (108) clearly indicates the limiting cur-
mining (Path B) rent caused by the atomic combination reaction.
(i) General condition for atomic hydrogen mecha- (iv) Effect of pH in pure dilute acid solution.-
nism.--Comparing with (85), the condition is Using (91), (92), and (93) in (107) it follows
10ca < al ~ a~. (97) RT
= Const. - 2F- In ic. (109)
(ii) Coverage of surface.--Using condition (97)
and also the condition
Equation (109) indicates that with the above mech-
9a2 < al, (98) anism overpotential is independent of the p H of
and also since %/1 ~- n = 1 + 89 where n is small, the solution. It is evident from (108) that the mag-
it follows nitude of the limiting current is independent of pH.
a, (v) Effect of pH in the presence of excess neutral
x - . (99) salt.--Equation (109) shows that ~ is independent
al + 62
of ~" and, therefore, is independent of pH in the
(iii) Tafel line.--From (80), (81), (82), (83), and presence of excess neutral salt. Equation (108)
(99), shows that the limiting current is independent of the
i ~ - - 2F10-18ks al presence of neutral salt.
a l -~- 62 (vi) Effect of neutral salt at constant pH. ~ is
unaffected by the presence of neutral salt (see 109).
---- 2Fksl0 -1~ k~lO- i A4~F 9 4. Minimum Coverage of Cathode by Hydrogen for
1 -I- ~a~H+ exp ~
Atomic Combination to be Desorptive
d(A~)
Evaluation of d(ln ic) yields From (16), (77), and (83),

ic 2F ~T 10-18x2, (110)
d(A~b) _ b _ 1 + klk2aH--------~
10-9 exp \~][A~bF~
(101) if (AG*)I is taken as zero to give a maximum veloc-
d(ln i~) 2.303 2Fk~ 10-9 (A4~F~ "
RTklaa§ exp \ R T ] ity of combination, ~ = 1, and kl in (16) is identified

Downloaded on 2014-05-28 to IP 192.236.36.29 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 Vol. 99, No. 4 CATHODIC HYDROGEN EVOLUTION REACTION 179

with ka in (83). Hence, the least possible coverage (it) Coverage of surface.--Utilizing (118) in (114),
for the atomic hydrogen reaction to maintain a and also from (80) and (115),
current density of i~' a m p / c m ~ is given by x = a~/a4 = k~/lO-gk4. (119)
(iii) Tafel line.--Condition (118)utilized in (116)
x = 10~ . (111)
gives,

ic = 2FklaH+ exp ACF~

2RT]' (120)
B. The Source of Protons is H~O + and the Electro-
chemical Reaction is Desorptive (Paths C and D)
which relation is very similar to that found for a
1. General Kinetic Equations rate-determining discharge reaction followed by an
The reactions and velocities to be considered are atomic hydrogen desorption step. Hence, a = 0.5
andfl = } a s h = 2.
HaO + q- e
Vl
- - - ~ MH, (72) (iv) Effect of pH in pure dilute acid solution, (v)
effect of pH in presence of excess neutral salt, and
MH __v~_. HaO+ + e, (73) (vi) effect of neutral salt at constant pH.--These
three effects are as already given in Section A, 2.
and
3. Kinetics when Electrochemical Step is Rate-Deter-
HaO + +
V4
M H -b e ----* H~. (112) mining (Path D)
(i) General condition for electrochemical mecha-
v, and v2 are given by (75) and (76). v~ is given by nism.--The condition is
10a4 < al -b a2. (121)
v~ = tc~lo-ga~+x exp -- 2RT]" (113)
(it) Coverage of surface.--Special case (a)
The steady state value of x yielded by reasoning 10al < a2. (122)
analogous to that leading to (79) is
From (114), (121), and (122),
x - al (114) X = at~a2. (123)
al -~- a2 -~- a4' Special case (b)
where 10a2 < at. (124)
From (114), (121), and (124),
a4 = k410-ga~* exp -- 2RT]" (115)
x -~ 1. (125)
The current is, therefore, (iii) Tafel line.--Special case (a). Conditions
(121) and (122). From (116),
2Fk~k4 10-9(all+) 2 exp (

(k~ -b k410-~)a.+ q- k210 9 exp \ ~ ]

z~4~F'~2RT]
9 (116) ic = 2F k~ k~(a~+)2 exp ( . (126)

From (27), (29), and (30) it follows that

From (116),
a = 3/2 (127)
d(a~)
as }, = 2, fl -- 3/4. (128)
d(ln ic)
k21o-~ (~F~ (iva) Effect of pH in pure dilute acid solution.-
1+ (/c~ ~ k410-9)a~+ exp \ R T ] Using (91), (92), (93), and (94) in (126) it follows
(117)
2RT
= Const. - - - In G. (129)
2RT 1 + (k~ ~- k4 10-9)an+ exp \ ~ - ] _ ] 3F
Hence with the above mechanism under the above
- (k~ + k, IO-9)aH+RT exp \-~/3 conditions overpotential is independent of the pH
of the solution.
2. Kinetics when Discharge is Rate-Determining
(Path C) (va) Effect of pH in the presence of excess neutral
salt.--From (91), (92), (93), and (126) and as ~" -~ 0
(i) General condition for slow discharge mechanism. it follows that:
- - T h e condition is
= Const. - 2RT
3-F- In ic q- ~
R-T In (aH+)s. (130)
10(a~ + a2) < at. (118)

Downloaded on 2014-05-28 to IP 192.236.36.29 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 180 J O U R N A L OF T H E E L E C T R O C H E M I C A L SOCIETY April 1952

Equation (130) shows that under the above condi-
tion overpotential decreases numerically by 19 mv
at 20~ upon decreasing the pH of the solution by
one unit.
(via) Effect of neutral salt at constant p H . - - F r o m
(91), (92), (93), and (126) and considering (a.+)B
as constant,
2RT
-- Const. - 3 F - In i, - ~'/3.
Hence, as ~" becomes more positive and approaches
zero on addition of neutral salt, overpotential
increases numerically.
(iiib) Tafel line.--Special case (b). Conditions
(121) and (124). From (116),
io = 2FkaaH+10 -9 exp
In both reaction schemes desorption of hydrogen
may occur by means of the reactions,
HgH + HgH ~ H2 + 2Hg (f)
H20 + HgH + e ~ H: + OH-. (g)
It is improbable that reaction (a) occurs because
evidence is available (7) that alkali metal deposition
occurs on Hg without being accompanied by an ap-
(131) preciable overpotential. Reaction scheme cdef will,
therefore, be formulated. It can be shown that very
similar results are obtained for the scheme cdeg (6).
Let v5 be the velocity of reaction c, v~ that of d,
v7 that of e, and Vs that of f. Let Xl be the fraction
of the surface of mercury covered with alkali metal
atoms in the steady state and x2 be the similar frac-
tion for hydrogen atoms. Then,

From (27), (29), and (30) it follows that

2RT]"
v5 = ksan+(1 -- X l - x2) exp ( 2~]' (135)

a = 0.5 (133) v6 = k610 -9 xl exp \ 2 R T - ] ' (136)

as h = 2, /3 = 1/4. (134) vr = k7 ( 1 0 - 9 Xl) t a H 2 0 , (137)

(ivb) Effect of pH in pure dilute acid solution, where xi is raised to the power 89in accordance with
(vb) effect of pH in the presence of excess neutral the experimental data of Jofa and Pechkovskaya
salt, (vib) effect of neutral salt at constant p H . - - (8) on the velocity of reaction between potassium
Consideration of (91), (92), (93), and (132) shows amalgam and water, aH~o is the activity of water,
that the above three effects are as already given in and ks-kr are velocity constants.
Section A, 2. At low and medium current densities, l >> xl, x2.
Also, in the steady state,
C. The Source of the Protons is Water and a Metal
v5 - - v6 - - vr = 0, (138)
Cation Takes Part in the Reaction: A Special
Case v7 -- Vs = 0. (139)
Reference to the reaction paths E to N indicates According to the mechanism suggested,
that many more possibilities exist for the mecha-
nism of the evolution of hydrogen when the source k~all+ exp 2RT ] >> k710- 9/2 aH2o
of the protons is water than when it is the hydroxo-
(140)
nium ion. Since little data as yet exist with which com-
parison m a y be made, the kinetics of these schemes <~ kG10-9 exp \ 2 R T - ] "
will be formulated here only in one special case.
It has been recently shown by Bockris and Watson From (140) by an argument similar to that used in
(6) that the evolution of hydrogen at mercury cath- the derivation of (88), one obtains
odes from aqueous alkaline solutions involves the (k4 89 ~ (z~F~ (141)
alkali or alkaline earth metal cation. Hence, the io = Fk7 \k6] aH20(aM*+)~exp 2RT-]"
possibilities in this case are:
But,
slow
M ~+ + ze ~ M/Hg (a) k5 _ (aG*)~- (aG*)~)
ke exp R-T
(142)

or,
M / H g + zH20 f_~ast M,+ + z H g H + z 0 H - (b)
= exp ( -- R T ] '
fast where AG5 is the standard free energy change in
M ~+ + ze ---+M/Hg (c)
reaction (c). Because the departure from irrevers-
M/Hg f a s t M~+ +ze (d) ibility in (135) and (136) is small it follows that
approximately,
M/Hg+ zH20 sl~ +zHgH+zOH- (e) AG5 = -- ze~oF, (143)

Downloaded on 2014-05-28 to IP 192.236.36.29 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 Vol. 99, No. 4 CATHODIC HYDROGEN EVOLUTION REACTION 181

where e'0 is the reversible electrode potential for the
alkali metal concerned at unit activity in mercury.
Hence, from (141) and (142) and (143),
ir = zFau~o (aM~+)' k7 exp 2~
A reasonably good value of e~ can be obtained from
the emf of a reversible amalgam electrode of the
type M amalgam/M ~+ ions at unit activity. If
it is assumed that the interaction energy between M
and Hg is included in e~, and that it is constant
B. Numerical Calculations
1. General Remarks and Preliminary Calculations
The results of calculations presented here are to
[e0 -- A4] . (144) be regarded as of order accuracy only. This is due
to (i) the approximation involved both in the
equations themselves and in the method of numer-
ical application. (e.g., calculation of reversible elee-
trode potential, see below), and (ii) the inaccuracy
of some of the experimental and theoretical data.
The following results are intended to be exemplify-

TABLE I. Characteristics of various mechanisms of cathodic hydrogen evolution

Wi
$yl
metr cat[ LO(~nn
enel gY I Neutral salt
aH+)BAr, i effect at
Characteristic of mechanism Conditions for application X ban mr I
constant pH

Pure solo. + Excess

ol neutral salt
--i
Acid solution mech- Slow discharge 10(aL + a2)< at Nil 1 --v RT ~ more neg.
anisms
F a s t atmnie H (A) 9a2 < at I
m

Slow atomic H 10as < al + a~ 2 2.1 1 ; Nil Nil Nil

F a s t discharge (B) 9a~ < al

Slow discharge 10(a~ + a ~ ) < a~ Nil 1 --~ RT ~ more neg.

F
F a s t electrochemical (C)
--i
Slow electrochemical a, + a 2 > 10a4[a~ > 10m 2 I I1 Nil 1 -T RT ~ more neg.
I l+v F

F a s t discharge (D) 'al > 10a2 2 t I1 Nil 1 -'y RT ~ more ncg.

F
--i
Alkaline solution Slow discharge from Htl3 Analogous to m e c h a n i s m s --2RT --RT ~ more pos.
meehaIlisms w i t h either fast atomic n with discharge from | h O + Y Y
(Eb or fast electrochemical
from H~O* (G)

Slow M / H g ) reaction as > 10aT < a e z 2 --RT RT 0 [n(a~u z+) B --RT

~ - 47 ~ - 9 0 ln(a,+) B -Y- Nil
F a s t M z+ discharge
F a s t atomic H ( J ) If ~ = 0.5

~, = Bttrrier d i s y m m e t r y factor = Fraction of potential assisting forward direction of rate-determining step.

* Assuming u n i v a l e n t alkali metal ion.
~: A s s u m i n g t h a t relation ~- = const 4- ~ In (an+)B is valid.

over a small change of activity, then the electrode
potential e, is given by
, 1
ea = e0 + - -
where C M ---- concentration of alkali metal M in
Hg. From (145), knowing e, experimentally, e'0 is
found.
VII. DIscussioN OF MECHANISM AT Hg, Ag, Ni,
AND Pt CATHODES
A. Table of Distinguishing Criteria
These are given in Table I which is based upon
the general results to be found in Sections IV and VI.
ing rather than exhaustive (e.g., throughout, the
concentration of hydroxonium ions in the electrical
double layer has been taken to be 10-1 gram equiv-
In ~ , (145) alents per liter and no quantitative examination
has been made of the effect of variation of concen-
tration).
Evaluation of terms such as exp (AeorF)/RT has
been made by expressing A4: as the potential of the
reversible hydrogen electrode with respect to the
eleetroeapillary maximum of the metal concerned.
The approximations involved in this procedure,
involving the neglect of the potential difference at
the interface due to adsorbed solvent dipoles,
should be recognized; in comparison of rates on
various metals, this error is less serious because it is

Downloaded on 2014-05-28 to IP 192.236.36.29 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 182 JOURNAL OF THE ELECTROCHEMICAL SOCIETY April 1952

reasonable to assume that the dipole contributions Table II gives numerical values of some of the
are approximately the same at different metals. quantities used in the following calculations.
The terms AG~*, AG2*, AG*, and AG* (where AG*
is a standard free energy of activation; 1 and 2 2. Potential Condilions for Applicability of Slow
refer respectively to the forward and reverse direc- Discharge Mechanism (Path A)
tions of the discharge of hydroxonium ions; ~nd 3 The two conditions to be satisfied for (88) to
and 4 refer respectively to the forward atomic hy- apply are (85) and (86). Numerical application of
drogen and electrochemical desorption steps) have these conditions gives the results of Table III.
been obtained as follows. AG* and AG* have been
3. Coverage of Surface (x) for Slow Discharge Mecha-
taken from calculations by Parsons (26) of the ener-
nism (Path A)
getics of the discharge reaction 5. AG* (for the atomic
Table IV gives the surface coverages at various
T A B L E II. Certain numerical uantities at 20~ overpotentials obtained by applying equation (87)
Reversible I
at 20~
H poten- I
Log i0 ]tiai referred AG~*
Ca- (aa+ = to electro- kcal Log k~ AG~*
kcal
Log k~ k~/k2 4. The Value of AG* for Atomic Hydrogen Desorp-
thode 0.1) capillary
maximum lion
volts
For smooth Pt cathodes the observed experimen-
Hg -12 +0.20 38 -15.5 23.2 -4.5 --11.0 tal slope of the Tafel line at current densities of
Ag -7 --0.046 31 -10.3 23.9 --5.0 - 5.3
about 10 2 _ l0 ~ amp/cm 2 indicates that the
Ni -6 --0.30t 31 -10.3 23.6 -4.7 - 5.6
atomic hydrogen mechanism is rate-determining.
Value o b t a i n e d from i n t e r p o l a t i o n of an observed rela- The limiting current density for this mechanism is
tion between 9 a n d the p o t e n t i a l of the electrocapillary found to be 30 amp/era 2 in 1N aqueous acid solu-
m a x i m u m for a n u m b e r of metals.
tion (13). Substituting in (146) it follows that,
T A B L E I I I . Potential conditions for slow discharge mech- AG* = --2.5 kcal approx.
anism (Path A) at 20~ k3 = 1014'2
Cathode Condition (85) : Condition (86) :
Overpotential in volts Overpotential in volts The correct value of AG* can therefore be taken as
about 0 kcal, the corresponding value of k3 being
Hg More positive t h a n - 1.4 More negative t h a n
10'2's at ordinary temperature. With the approach
-0.43
Ag More positive t h a n - 0 . 5 4 All v a l u e s ~ s a t i s f y of saturation appreciable departures from ideality
condition among the hydrogen atoms on the surface are
Ni More positive t h a n - 0 . 2 9 All values satisfy likely to occur. Hence, the react'ion may be some-
condition what slower than indicated by the value of AG*.
A low value of AG* would be expected because the
T A B L E IV. Surface coverage with hydrogen at various over- reaction concerned is a surface radical reaction. To
potentials (Path A) a first approximation it is assumed that AG* (and
Overpotentia]
also k3) are the same for all cathodes.
Cathode (volts) --0A --0.3 -0.5 --0.7 --1.0 --1s

5. The Value of lcl/k2 for Pt Cathodes

Hg Coverage x l0 -6"2 10-a'4 10-4.5 10-3.4 10-2"4 Sat, d.
Utilizing the ahove determined value of AG*
Ag Coverage x 10-2"6 10-L~ 10-~ Satd. surface.
(or k3) in (100), (80), and (81) referred to the ape-
Ni Coverage x 10-1'5 10-~ Saturated surface cial case of the exchange current i0 so that '1 = 0,
it follows that
hydrogen desorption) will be shown below to have a kl/k2 = 10 -14"4.
value of approximately zero. AG* (for the electro-
chemical desorption) cannot yet be calculated even
6. Calculation of the Course of the Tafel Line for the
Pt Cathode
with sufficient accuracy for the present purposes;
k~, where n = 1, 2, or 3 depending on the reaction Substituting the above obtained values 8 for k3
concerned, has been calculated from AG* by means and kl/k2 in (100) gives the complete Tafel equa-
of the equation, 6 In this calculation the experimental value of ks i.e.
10 L4-2,has been used in order t h a t the l i m i t i n g c u r r e n t shall
k, = ~ - exp - RT]" (146) appear at 30 a m p / c m 2. If the value of k3 = k T / h = 10~*
is used the l i m i t i n g c u r r e n t appears a t 1 a m p / e r a 2. In
All numerical values q u o t e d i n following sections hence e i t h e r case the form of t h e calculated Tafel line remains
depend upon the accuracy of these very a p p r o x i m a t e cal- the same. T h e values of log i0 and the reversible h y d r o g e n
culations, except for values calculated here concerning p o t e n t i a l referred to the electrocapillary m a x i m u m h a v e
platinum. been t a k e n as - 3 , a n d +0.28 volts, respectively.

Downloaded on 2014-05-28 to IP 192.236.36.29 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 VoL 99, No. 4 CATHODIC HYDROGEN EVOLUTION REACTION 183

tion for hydrogen evolution from aqueous 0.1N
HC1. The two limiting slopes deduced in (103)
and (105) are shown (see Fig. 3).
7. Minimum Coverage of Cathode by Hydrogen for the
Atomic Hydrogen Mechanism to be Desorptive
The required m i n i m u m coverage is obtained from
( l l 1) as a function of current and is given in Table V.
-020 ] J
-- 01 1 ~
9. Conditions for Applicability of Slow Discharge
Mechanism (Path C)
The condition to satisfy for (120) to apply is (118).
However, in this case it is not possible to calculate
sufficiently accurate values of k, or AG* for potential
conditions to be found. Limiting values for k, and
AG* when P a t h C is rate-determining have there-
fore been calculated and are given in Table V I I .
l r
Calculations of AG* b y statistical mechanical
means are complex and have only been a t t e m p t e d
by Conway (9) in the case of Ag. He finds (hG*)^~
9 kcal, in which case, therefore, discharge would be
rate-determining if the desorption were electro-
chemical.

-0,1( TABLE VII. Limiting values of k4 and AG*~for Path C to be

rate-determinin at 20~
Cathode k4 limit for Path C, AG* limit for Path C,
Rate-determining Rate-determining
+005

Hg When ,1 more negative When ,1 more negative

~ "~ , , , than -0.43 volt k, than --0.43 volt aG~
must be > 10 -5.~ must be < 24 kcal
-3.0 -2 0 - IO 0 1.0 2.0
log Ic Ag k4 must be > 1O-~ at ~G~ must be < 17 kcal
FIG. 3. Calculated course of Tafel line for the smooth all overpotentials at all overpotentials
platinum cathode in 0.1N HC1 at 20~ (assuming path
B is rate determining). Ni As for Ag As for Ag

TABLE V. Minimum coveragefor atomic hydrogen desorption

TABLE VIII. Maximum values of surface coveragefor Path
ic amp/cm~ x C to be rate-determining
Cathode Maximum value of x
10-t~ 10-8
lO-g 10-4.5
Hg 10-~ at overpotentials more negative than -0.43
10-6 10-3
10-3 volt
lO-i .6
1 1 (satd. surface) 10-' at all overpotentials
Ag
10 Saturated surface
Ni As for Ag

TABLE VI. Potential conditions for atomic hydrogen mech-

anism (Path B) at 20~ 10. Coverage of Surface for Slow Discharge Mecha-
nism (Path C)
Cathode Condition (97) [ Condition (98)
.Overpotential
. in volts
. . Overpotential
. in volts
The coverage x is given b y (119), and by using
Hg More negative than' More positive than the limiting values of k4 given in Table V I I a maxi-
-1.65 [ -0.43 m u m value of x can be calculated (see Table V I I I ) .
Ag More negative than I All values satisfy con-
-0.78 than dition 11. Conditions for Applicability of Electrochemical
Ni More negative All values satisfy con- Mechanism (Path D)
-0.64 dition
The condition to satisfy for P a t h D to be rate-
determining is (121), b u t since k4 and AG* are not
8. Potential Conditions for Applicability of Atomic known accurately, Table I X gives limiting values
Hydrogen Mechanism (Path B) of these quantities (cf. Subsection 9).
The two conditions to be satisfied for (100) to
apply are (97) and (98). Numerical application of 12. Potential Conditions for Applicability of a =
these conditions gives the results of Table VI. and a = 89when Path D is Rate-Determining
For smooth platinum cathodes, the use of the ex- The condition to be satisfied for a to have values
perimental values of k,/k2 in condition (97) shows of ~ and 89 are, respectively, (122) and (124). The
t h a t for the atomic hydrogen mechanism to apply potential conditions appropriate to both values of
a t ~ = 0, kl m u s t be > 1 0 -6"s. are given in Table X.

Downloaded on 2014-05-28 to IP 192.236.36.29 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 184 JOURNAL OF T H E E L E C T R O C H E M I C A L SOCIETY April 1952

13. Coverage o.f Surface for Electrochemical Mecha- Measurements in this current density region have
nism (Path D) been made (12) and do indeed show an increase
The required coverage is found from (114) neglect- in the value of a. Here, however, the continuous
ing a4 and is as follows: increase of a with decreasing current density prob-
Hg: Surface saturated at overpotentials more ably indicates the presence of slight trace impurities
negative than - 0 . 3 7 volt, in the mercury used (to which traces the kinetics of
the electrode reaction would be very sensitive at the
Ag:~ Surface saturated at all overpotentials. very low current densities involved).
Ni: )
The mechanism is hence probably A or C. Ac-
C. Estimate of Rate-Determining Reactions at Hg, cording to the results of Sections (VII, B, 4 and 8),
Ag, Ni, and Pt Cathodes According to Evidence the mechanism cannot be A above a current density
at present Available of about 30 amp/cm 2 and below about 10-s amp/era 2,
1. Mercury outside which limits it must therefore be C. It is not
possible to distinguish at present between A and C
(i) In Acid Solution.--The value of a -- 89shows at intermediate current densities although for both
that the mechanism must be A, C, or D. At low the rate-determining reaction is the discharge step.
current densities the velocity of the atomic com- (ii) In alkaline sohdion.--As shown in the special
case worked out in Section (VI, 0), the rate-deter-
T A B L E I X . Limiting values of k4 and 5G*4for Path D to be
mining reaction here is tile combination of alkali
rate-deterndning at 20~
metal atoms with water. An implication of this
Cathode k4 limit for Path D, AG~'limit for Path D, mechanism is that it would give rise to a limiting
Rate-determining Rate-determining
current density at high current densities, but this
Hg When n more negative When n more negative aspect is as yet unexamined.
t h a n - 0 . 4 3 v o l t k4 t h a n - 0 . 4 3 v o l t AG~
m u s t be < 10-7-~ m u s t be > 27 kcal 2. Sih, cr in Acid Solution
Ag k4 m u s t be < 10 - 2 3 a t AG~ m u s t be > 20 kcal The value of a (9, 13) shows that the mechanism
all o v e r p o t e n t i a l s a t all o v e r p o t e n t i a l s must be A, C, or D. The results of Section (VII,
B, 2) show that mechanism A cannot occur at over-
Ni As for Ag As for Ag potentials more negative than - 0 . 5 4 volt, ~nd fur-
thermore is unlikely at high current densities since
TABLE X no limiting current is detectable up to current densi-
Cathode __C~176 for a = 3/2_ Condition for a = 1/2
ties of 150 amp/cm 2 (13). If the mechanism is D,
Section (VII, B, 12) shows that a must equal 1 as is
Hg n more positive than I ~ more negative than observed. A determination of h (14) yields the value
--0.32 v o l t [ -0.43 volt 1 which favors mechanism A ; and statistical mechan-
Ag N o v a l u e of ~ ] All v a l u e s of
ical calculations by Conway (9) have shown that
Ni N o v a l u e of n
II
All v a l u e s of
5G~ is about 9 kcal, whereas for mechanism D to
occur AG~ must be >20 kcal. Effect of pH is com-
bination is too great for mechanism D to occur plex but can be brought into better accord with A
(see Section VII, B, 11). I t is therefore highly or C if Stern's equations for the dependence of elec-
improbable that this mechanism occurs at all on trokinetic potential on concentration are developed
mercury because a change to it at higher current fully.
densities would involve a change in the value of i0
and this is not observed. Further, according to the 3. Nickel
conclusion of Section (VII, B, 13), if D were opera- (i) In acid solution.--The value of a showsthat
tive, the surface of the mercury would be saturated the mechanism must be A, C, or D (9, 10, 15).
with adsorbed hydrogen at all potentials accessible Determination of ~, is difficult since the Tafel line
to experiment, and this is contrary to evidence con- is of the dissolution type (see Fig. 1) and at inter-
cerning the variation of the interracial tension of mediate current densities in strongly acid solutions
mercury-solution interfaces with pH (11). Accord- ( > N / 2 ) the dissolution of the metal partially ob-
ing to the results of Section (VII, B, 12), mechanism scures pH and salt effects. Azzam and Bockris have
D with a value of a = ~ would occur up to an over- shown (13) that a limiting current occurs at about
potential of - 0 . 3 volt (corresponding to a current 1 amp/cm 2 which supports mechanism A. The ab-
density of about 10 9 amp/era 2) if electrochemical sence of pH effects at concentrations below N/100
desorption were the rate-determining reaction. also supports this mechanism. The results of Section

Downloaded on 2014-05-28 to IP 192.236.36.29 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 Vol. 99, No. ~ CATHODIC HYDROGEN EVOLUTION REACTION
(VII, B, 2) show that mechanism A cannot take
place at overpotentials more negative than - 0 . 3
volt, and it is interesting to note that the Tafel
line begins to attain a limiting current at about
this potential when measurements are made using a
fresh cathode surface (10).
(iN) In alkaline solution.--The coefficient a is
89and )` is observed to be 1 (10), which indicates
that mechanism E or G is operative. This is con-
firmed by the interpretation of the observed pH
effects.
4. Smooth Platinum in Acid Solution
The value of a at intermediate current densities
indicates that the mechanism is B (16). The mecha-
nism at current densities above the limiting cur-
rent density may be C or D, but is probably D
owing to the saturation of the surface with ad-
sorbed hydrogen at the limiting current density.
VIII. SUMMARY OF SOME RECENT ADVANCES
The controversy concerning the value of b in
the Tafel equation now seems resolved experi-
mentally as follows. If the solutions are pure, b
is not (except transiently at a limiting current)
greater than about 0.13, i.e.,/~ > 0.45/),. This fact
confirms that the energy barrier for the discharge
reaction is usually nearly symmetrical and is in
accordance with the equations formulated in Section
(VI). Higher values of b are probably connected
with a distortion of the energy barrier (the position
of the transition state complex being displaced
toward the electrode) due to the specific adsorption
of poisons on the electrode.
The realization of the use of )` as a distinguishing
criterion of mechanism, and its intensive statistical
application, has made available a criterion of greater
power than previously existed.
Knowledge of AH0* and the potential of the elec-
trocapiUary maximum aid application of the quan-
titative conditions established above.
In recent years, it has been suggested (17) that
prototropic transfer of hydrogen from water in the
solution to water adsorbed on the electrode, and
also (18) that the atomic hydrogen desorption re-
placed at higher current densities by electrochemical
desorption are the main rate-determining reactions
to be considered for the hydrogen evolution reaction.
Evidence in favor of the first was based mainly
upon the constancy of the factor B for various elec-
trolyte concentrations and for a few metals. Con-
sideration of (67) shows t h a t the first requirement
would be expected for any mechanism, however,
and it has in any case been shown in an earlier sec-
tion of this paper that a sufficiently accurate ex-
Downloaded on 2014-05-28 to IP 192.236.36.29 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
185
perimental determination of B is prohibitively diffi-
cult [see also Butler (19)]. In order that the proto-
tropic transfer theory can yield a satisfactory
theory of pH effects and of reversible electrode po-
tentials an unusual structure of the double layer has
to be assumed (20). Forecasts made upon this basis
of similarities between hydrogen and oxygen over-
potentials have also not been confirmed (27). The
second theory (18) suffers the fundamental disad-
vantage that it involves the formulation of the
kinetic processes at working hydrogen electrodes
in terms of equations applicable only under equi-
librium conditions; it is impotent in indicating ex-
pected pH and salt effects; and the explanation
suggested for different values of b is based on ex-
perimental results which are not in accord with
those obtained using very pure solutions.
Lastly, a quantitative, statistical mechanical
formulation of the discharge reaction from both
hydroxonium ions and water has recently been
made (5). This work indicates rates of the discharge
reaction in good accord with those experimentally
observed on the two electrode materials considered
(Hg and Ni). There also appears to be only a very
small probability that discharge can proceed from
water molecules in acid solution.
ACKNOWLEDGMENT
Thanks are due to Dr. R. Parsons for critical dis-
cussion and to Dr. R. G. H. Watson for verification
of the numerical calculations.
Any discussion of this paper will appear in a Discussion
Section, to be published in the December 1952 issue of the
JOURNAL.
REFERENCES
1. J. O'M. BOCKRIS AND B. E. CONWAY, Trans. Faraday
Soc., 45,989 (1949).
2. A. M. AZZAM, J. O'M. BOCKRIS, B. E. CONWAY, AND
H. ROSENBERG, Trans. Faraday Soc., 46, 918 (1950).
3. J. O'M. BOCKRIS AND n . E. CONWAY, J. Sci. InsIru-
ments, 25,283 (1948).
4. J. O'M. BOCKmS, Chem. Revs., 3, 561 (1948).
5. R. PARSONS AND 3. O'M. BOCKmS, Trans. Faraday Soc.,
47,914 (1951).
6. J. O'M. BOCKmS AND R. G. H. WATSON,J . chim. phys.,
49, 1 (1952).
7. J. HEYROVSKY, Disc. Faraday Soc., l , 212 (1947).
8. J. JOFA AND Z. B. PECHKOVSKAYA,Doklady Akad. Nauk.
S.S.S.R., 59,265 (1948).
9. B. E. CONWAY,Thesis, University of London (1949).
10. J. O'M. BOCKRIS AND E. C. POTTER, J. Chem Phys.,
March (1952).
11. A. N. FRVMKIN, Acta Physicochim. U.R.S.S., 18, 23
(1943).
12. F. P. BOWDEN ANn K. E. W. GaEW, Disc. Faraday ~oc.,
1, 86 (1947).
13. A. M. AZZAM AND J. O'M. BOCKRIS, Nature, 165, 403
(t95o).
 186 JOURNAL OF THE ELECTROCHEMICAL SOCIETY April 1952

14. E. C. POTTER, Unpublished results.
15. P. LUKOVTSEV,S. LEWINA, AND A. N. FRUMKIN,Acta
Physicochim. U.R.S.S., 11, 21 (1939).
16. A. M. AZZAM, Thesis, University of London (1949).
17. H. EYRING, S. GLASSTONE, AND K. J. LAIDLER, J .
Chem. Phys., 7, 1053 (1939).
18. A. HICKLING AND F. W. SALT, Trans. Faraday Soc.,
38, 474 (1942).
19. J. A. V. BUTLER,J. Chem. Phys., 9,279 (1941).
20. G. E. KIMBALL,S. GLASSTONE, AND A. GLASSNER,J .
Chem. Phys., 9, 91 (1941).
21. J. HORIUTI AND M. IKUSIMA,Proc. Imp. Acad. Tokyo,
15, 39 (1939).
22. B. KABANOV AND J. JOFA, Acta Physicochim. U.R.S.S.,
10, 616 (1939).
23. J. N. A6AR, Disc. Faraday Soc., 1, 81 (1947).
24. J. A. V. BUTLER, Trans. Faraday Soc., 28, 379 (1932).
25. P. LUKOVTSEVAND S. LEWINA, J. Phys. Chem. U.S.S.R.,
21,599 (1947).
26. R. PARSONS, Z. Elektrochem., 55, l l l (1951).
27. A. HICKLING AND S. HILL, Disc. Faraday Soc., 1, 236
(1947).

Downloaded on 2014-05-28 to IP 192.236.36.29 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).