Professional Documents
Culture Documents
1952 Bockris HER
1952 Bockris HER
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.
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- difference between the metals. If AV is a potential
rent, directed attention away from the fact that difference measured by means of a potentiometer
the problems coflnected with overpotential were connected to wires of composition M, then,
purely kinetic, and that attention should be focused
A V = VM-- V p t + X .
upon the current. Second, considerable attention
has been paid to the determination of i0, the current Now, X ~ ~M -- qh,t, where q)M and ~p~ are the
work functions of the metals M and Pt respectively
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,
D I S S O L ~
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,
O LIMITINGCllfiRENT ~ it
~1 R[$I$1A~IC-EERROR i - T Y P E /
~
,~[ 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 April 1952
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 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,
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
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)
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
CRT ,
t - k/~0ff m [ 7, [ + const. (59) B = K•F ~ - a, exp , (67)
Equation (59) shows that the decay curve of over- and AH* and AS* are respectively the heat and
potential against time becomes asymptotic to the entropy of activation corresponding to (AG*)I.
time axis, the relation being exponential. The equa- Assuming AH* and AS* are not temperature de-
tion may be used to find the differential capacity pendent, and that over small ranges of temperature
(25). B is constant, the plot of In i0 against; 1/7' has a
(it) Relations in which dC/d7 = f(7). slope of - A H * / R .
During decay, i" = C dT'Jdt. Using this in (38) Further, when n is more negative than - 7 5 my,
and (39) gives (25) and (66) give
- c
d7t - c
d,/t (60) ic = B e • AH*;T~TF ). (68)
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
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
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)
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,
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- In both reaction schemes desorption of hydrogen
tion overpotential decreases numerically by 19 mv may occur by means of the reactions,
at 20~ upon decreasing the pH of the solution by HgH + HgH ~ H2 + 2Hg (f)
one unit.
(via) Effect of neutral salt at constant p H . - - F r o m H20 + HgH + e ~ H: + OH-. (g)
(91), (92), (93), and (126) and considering (a.+)B It is improbable that reaction (a) occurs because
as constant, evidence is available (7) that alkali metal deposition
occurs on Hg without being accompanied by an ap-
2RT
-- Const. - 3 F - In i, - ~'/3. (131) preciable overpotential. Reaction scheme cdef will,
therefore, be formulated. It can be shown that very
Hence, as ~" becomes more positive and approaches similar results are obtained for the scheme cdeg (6).
zero on addition of neutral salt, overpotential Let v5 be the velocity of reaction c, v~ that of d,
increases numerically. v7 that of e, and Vs that of f. Let Xl be the fraction
(iiib) Tafel line.--Special case (b). Conditions of the surface of mercury covered with alkali metal
(121) and (124). From (116), atoms in the steady state and x2 be the similar frac-
tion for hydrogen atoms. Then,
io = 2FkaaH+10 -9 exp
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 B. Numerical Calculations
alkali metal concerned at unit activity in mercury.
Hence, from (141) and (142) and (143), 1. General Remarks and Preliminary Calculations
The results of calculations presented here are to
ir = zFau~o (aM~+)' k7 exp 2~ [e0 -- A4] . (144) be regarded as of order accuracy only. This is due
to (i) the approximation involved both in the
A reasonably good value of e~ can be obtained from equations themselves and in the method of numer-
the emf of a reversible amalgam electrode of the ical application. (e.g., calculation of reversible elee-
type M amalgam/M ~+ ions at unit activity. If trode potential, see below), and (ii) the inaccuracy
it is assumed that the interaction energy between M of some of the experimental and theoretical data.
and Hg is included in e~, and that it is constant The following results are intended to be exemplify-
over a small change of activity, then the electrode ing rather than exhaustive (e.g., throughout, the
potential e, is given by concentration of hydroxonium ions in the electrical
double layer has been taken to be 10-1 gram equiv-
, 1
ea = e0 + - - In ~ , (145) alents per liter and no quantitative examination
has been made of the effect of variation of concen-
where C M ---- concentration of alkali metal M in tration).
Hg. From (145), knowing e, experimentally, e'0 is Evaluation of terms such as exp (AeorF)/RT has
found. been made by expressing A4: as the potential of the
reversible hydrogen electrode with respect to the
VII. DIscussioN OF MECHANISM AT Hg, Ag, Ni, eleetroeapillary maximum of the metal concerned.
AND Pt CATHODES The approximations involved in this procedure,
involving the neglect of the potential difference at
A. Table of Distinguishing Criteria the interface due to adsorbed solvent dipoles,
These are given in Table I which is based upon should be recognized; in comparison of rates on
the general results to be found in Sections IV and VI. 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
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 9. Conditions for Applicability of Slow Discharge
HC1. The two limiting slopes deduced in (103) Mechanism (Path C)
and (105) are shown (see Fig. 3). The condition to satisfy for (120) to apply is (118).
7. Minimum Coverage of Cathode by Hydrogen for the However, in this case it is not possible to calculate
Atomic Hydrogen Mechanism to be Desorptive sufficiently accurate values of k, or AG* for potential
conditions to be found. Limiting values for k, and
The required m i n i m u m coverage is obtained from
AG* when P a t h C is rate-determining have there-
( l l 1) as a function of current and is given in Table V.
fore been calculated and are given in Table V I I .
-020 ] J 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
-- 01 1 ~
rate-determining if the desorption were electro-
chemical.
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 185
(VII, B, 2) show that mechanism A cannot take perimental determination of B is prohibitively diffi-
place at overpotentials more negative than - 0 . 3 cult [see also Butler (19)]. In order that the proto-
volt, and it is interesting to note that the Tafel tropic transfer theory can yield a satisfactory
line begins to attain a limiting current at about theory of pH effects and of reversible electrode po-
this potential when measurements are made using a tentials an unusual structure of the double layer has
fresh cathode surface (10). to be assumed (20). Forecasts made upon this basis
(iN) In alkaline solution.--The coefficient a is of similarities between hydrogen and oxygen over-
89and )` is observed to be 1 (10), which indicates potentials have also not been confirmed (27). The
that mechanism E or G is operative. This is con- second theory (18) suffers the fundamental disad-
firmed by the interpretation of the observed pH vantage that it involves the formulation of the
effects. kinetic processes at working hydrogen electrodes
in terms of equations applicable only under equi-
4. Smooth Platinum in Acid Solution librium conditions; it is impotent in indicating ex-
The value of a at intermediate current densities pected pH and salt effects; and the explanation
indicates that the mechanism is B (16). The mecha- suggested for different values of b is based on ex-
nism at current densities above the limiting cur- perimental results which are not in accord with
rent density may be C or D, but is probably D those obtained using very pure solutions.
owing to the saturation of the surface with ad- Lastly, a quantitative, statistical mechanical
sorbed hydrogen at the limiting current density. formulation of the discharge reaction from both
hydroxonium ions and water has recently been
VIII. SUMMARY OF SOME RECENT ADVANCES
made (5). This work indicates rates of the discharge
The controversy concerning the value of b in reaction in good accord with those experimentally
the Tafel equation now seems resolved experi- observed on the two electrode materials considered
mentally as follows. If the solutions are pure, b (Hg and Ni). There also appears to be only a very
is not (except transiently at a limiting current) small probability that discharge can proceed from
greater than about 0.13, i.e.,/~ > 0.45/),. This fact water molecules in acid solution.
confirms that the energy barrier for the discharge
ACKNOWLEDGMENT
reaction is usually nearly symmetrical and is in
accordance with the equations formulated in Section Thanks are due to Dr. R. Parsons for critical dis-
(VI). Higher values of b are probably connected cussion and to Dr. R. G. H. Watson for verification
with a distortion of the energy barrier (the position of the numerical calculations.
of the transition state complex being displaced
Any discussion of this paper will appear in a Discussion
toward the electrode) due to the specific adsorption
Section, to be published in the December 1952 issue of the
of poisons on the electrode. JOURNAL.
The realization of the use of )` as a distinguishing
REFERENCES
criterion of mechanism, and its intensive statistical
application, has made available a criterion of greater 1. J. O'M. BOCKRIS AND B. E. CONWAY, Trans. Faraday
Soc., 45,989 (1949).
power than previously existed. 2. A. M. AZZAM, J. O'M. BOCKRIS, B. E. CONWAY, AND
Knowledge of AH0* and the potential of the elec- H. ROSENBERG, Trans. Faraday Soc., 46, 918 (1950).
trocapiUary maximum aid application of the quan- 3. J. O'M. BOCKRIS AND n . E. CONWAY, J. Sci. InsIru-
titative conditions established above. ments, 25,283 (1948).
In recent years, it has been suggested (17) that 4. J. O'M. BOCKmS, Chem. Revs., 3, 561 (1948).
5. R. PARSONS AND 3. O'M. BOCKmS, Trans. Faraday Soc.,
prototropic transfer of hydrogen from water in the
47,914 (1951).
solution to water adsorbed on the electrode, and 6. J. O'M. BOCKmS AND R. G. H. WATSON,J . chim. phys.,
also (18) that the atomic hydrogen desorption re- 49, 1 (1952).
placed at higher current densities by electrochemical 7. J. HEYROVSKY, Disc. Faraday Soc., l , 212 (1947).
desorption are the main rate-determining reactions 8. J. JOFA AND Z. B. PECHKOVSKAYA,Doklady Akad. Nauk.
S.S.S.R., 59,265 (1948).
to be considered for the hydrogen evolution reaction. 9. B. E. CONWAY,Thesis, University of London (1949).
Evidence in favor of the first was based mainly 10. J. O'M. BOCKRIS AND E. C. POTTER, J. Chem Phys.,
upon the constancy of the factor B for various elec- March (1952).
trolyte concentrations and for a few metals. Con- 11. A. N. FRVMKIN, Acta Physicochim. U.R.S.S., 18, 23
sideration of (67) shows t h a t the first requirement (1943).
12. F. P. BOWDEN ANn K. E. W. GaEW, Disc. Faraday ~oc.,
would be expected for any mechanism, however, 1, 86 (1947).
and it has in any case been shown in an earlier sec- 13. A. M. AZZAM AND J. O'M. BOCKRIS, Nature, 165, 403
tion of this paper that a sufficiently accurate ex- (t95o).
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).
186 JOURNAL OF THE ELECTROCHEMICAL SOCIETY April 1952
14. E. C. POTTER, Unpublished results. 21. J. HORIUTI AND M. IKUSIMA,Proc. Imp. Acad. Tokyo,
15. P. LUKOVTSEV,S. LEWINA, AND A. N. FRUMKIN,Acta 15, 39 (1939).
Physicochim. U.R.S.S., 11, 21 (1939). 22. B. KABANOV AND J. JOFA, Acta Physicochim. U.R.S.S.,
16. A. M. AZZAM, Thesis, University of London (1949). 10, 616 (1939).
17. H. EYRING, S. GLASSTONE, AND K. J. LAIDLER, J . 23. J. N. A6AR, Disc. Faraday Soc., 1, 81 (1947).
Chem. Phys., 7, 1053 (1939). 24. J. A. V. BUTLER, Trans. Faraday Soc., 28, 379 (1932).
18. A. HICKLING AND F. W. SALT, Trans. Faraday Soc., 25. P. LUKOVTSEVAND S. LEWINA, J. Phys. Chem. U.S.S.R.,
38, 474 (1942). 21,599 (1947).
19. J. A. V. BUTLER,J. Chem. Phys., 9,279 (1941). 26. R. PARSONS, Z. Elektrochem., 55, l l l (1951).
20. G. E. KIMBALL,S. GLASSTONE, AND A. GLASSNER,J . 27. A. HICKLING AND S. HILL, Disc. Faraday Soc., 1, 236
Chem. Phys., 9, 91 (1941). (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).