The Marcus-Hush model of electrode kinetics at a single nanoparticle

Martin C. Henstridge, Kristopher R. Ward, Richard G. Compton

Department of Chemistry, Physical & Theoretical Chemistry Laboratory, Oxford University, South Parks Road, Oxford OX1 3QZ, United Kingdom
a r t i c l e i n f o
Article history:
Received 29 August 2013
Received in revised form 13 October 2013
Accepted 15 October 2013
Available online 26 October 2013
Keywords:
Electrode kinetics
Electrochemical simulation
Nanoparticles
a b s t r a c t
We examine the effect of the Marcus-Hush model of electrode kinetics on electron transfer at the surface
of a single nanoparticle impacting an electode. Using numerical simulation we demonstrate the possibil-
ity of observing a kinetically limited steady state current which is smaller than the mass transport lim-
iting current for such a system.
2013 Elsevier B.V. All rights reserved.
1. Introduction
The differences between the Marcus-Hush model and the ubiq-
uitous ButlerVolmer model were rst highlighted for surface-
bound redox couples by Chidsey [1]. Where the ButlerVolmer
model gives rate constants which increase ad innitum, the Mar-
cus-Hush model gives rate constants which level off at large
overpotentials.
Later, Feldberg examined the conditions that would be neces-
sary to observe this difference in behaviour between the two mod-
els for solution-phase systems using steady state voltammetry at a
microdisc electrode [2]. Notably, this work demonstrated that the
Marcus-Hush model could, in principle, lead to a kinetically limited
steady state current which is smaller than the mass transport limit
provided that the limiting electron transfer rate is sufciently slow.
Since then several studies have been undertaken comparing the
two kinetic models in the tting of both cyclic and pulse voltam-
metry for several redox systems both under diffusion-only condi-
tions at a hemispherical microelectrode [3,4] and under
convective mass transport conditions at the channel ow electrode
[5]. These studies showed that the so-called symmetric Marcus-
Hush model, which assumes the potential energy curves of reac-
tant and product to be parabolae of equal curvature, was generally
unable to t practical experimental voltammetry while the Butler
Volmer model was consistently able to produce accurate tting.
Further work then introduced the asymmetric Marcus-Hush
model [6], which removes the assumption that the parabolic po-
tential energy curves are of equal curvature, which was seen to
be equal to the ButlerVolmer model in its ability to accurately
t both cyclic and pulse voltammetry.
While the limiting behaviour of the Marcus-Hush model has
been reported many times for surface-bound redox systems, it
has thus far not been possible to observe such behaviour for solu-
tion-phase systems. The conditions which Feldberg identied as
necessary to distinguish the models have not been realised exper-
imentally [2], and another study using the channel ow electrode
found that the two models would likely be indistinguishable for
real redox systems [7].
Nevertheless, the chances of observing a kinetically limited
steady state current are increased by decreasing the size of the
electrode. Mass transport to the electrode surface then becomes
more efcient and thus the rate of electron transfer can be higher
than at a larger electrode and yet still be current limiting.
One possible route to achieving such conditions is to study the
rate of electron transfer at nanoparticles impacting on an electrode
whereby electron transfer to and from the solution occurs only at
the surface of the impacting nanoparticle. Recent work has demon-
strated the possibility of extracting quantitative kinetic data from
such systems [8]. Such nanoparticle impacts can be modelled as
an isolated sphere supported on a planar, non-conducting surface
[9]. Using such a model, we aim to demonstrate the possibility of
experimentally observing a steady state current which is limited
by the rate of electron transfer, rather than mass transport, for a
solution-phase redox system.
2. Theory
2.1. Sphere on a plate
We model the impacting nanoparticle as a single conductive
sphere of radius r
e
conductively supported on a surface which
is innite in extent. Electrolysis is conned to the surface of the
nanoparticle, the plate simply provides an electrical connection.
1572-6657/$ - see front matter 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jelechem.2013.10.012

Corresponding author.
E-mail address: richard.compton@chem.ox.ac.uk (R.G. Compton).
Journal of Electroanalytical Chemistry 712 (2014) 1418
Contents lists available at ScienceDirect
Journal of Electroanalytical Chemistry
j our nal homepage: www. el sevi er . com/ l ocat e/ j el echem
Such a system can be described using cylindrical polar coordi-
nates (r, z, x) with the origin at the point of contact between
the sphere and the surface. This system is axisymmetric (i.e. all
properties of the system are invariant with respect to angle x)
and so we need consider only a 2-dimensional (r, z) slice through
the axis.
We assume the presence of a sufciently high concentration of
supporting electrolyte such that the effects of migrative mass
transport may be neglected. The mass transport in this system is
thus purely diffusional and may be described by Ficks 2nd law,
which takes the following form in this coordinate system:
@c
@t
D
@
2
c
@r
2

@
2
c
@z
2

1
r
@c
@r
_ _
1
where c is the concentration of the species in question, D is its dif-
fusion coefcient and t is time.
Throughout this study, we consider a one-electron reduction of
the form:
A e

B 2
for which the diffusion coefcients of both species are equal,
and only species A is initially present in bulk solution. In a linear
sweep voltammetry (LSV) experiment, the potential, E, is swept
at a constant rate, m, from some initial potential, E
i
, to some nal
potential such that at any time, t > 0:
E E
i
mt 3
The ux of species A normal to the electrode surface at a given
value of E is described by:
D
A
@c
A
@n
k
0
k
red
c
A;0
k
ox
c
B;0
4
where n is some coordinate normal to the electrode surface, c
A,0
and
c
B,0
are the concentrations of species A and B respectively at the
electrode surface, k
0
is the standard heterogeneous rate constant
and k
red
and k
ox
are the rate constants for reduction and oxidation
respectively.
2.2. Electrode kinetics
Within the ButlerVolmer model of electrode kinetics, k
red
and
k
ox
are dened as:
k
red
exp
aF
RT
E E
>
f
_ _
_ _
5
k
ox
exp
1 aF
RT
E E
>
f
_ _
_ _
6
where E
>
f
is the formal potential of the A/B couple, a is the so-called
transfer coefcient, F, R and T have their usual meanings.
Within the Marcus-Hush model of electrode kinetics, k
red
and
k
ox
are given by:
k
red
k
0
S
red
h; K
S
red
0; K
7
k
ox
k
0
S
ox
h; K
S
ox
0; K
8
where h is dimensionless potential and K is the dimensionless
analogue of the reorganisation energy (assuming units of eV):
h
F
RT
E E
>
f
9
K
F
RT
k 10
and S
red/ox
(h, K) is an integral of the form:
S
red=ox
h; K
_
1
1
exp DG
z
sym;red=ox
x=RT
_ _
1 exp x
dx 11
where
DG
z
asym;red=ox
x
RT

K
4
1
h x
K
_ _
2
c
h x
4
_ _
1
h x
K
_ _
2
_ _
c
2
K
16
12
x is a dimensionless integration variable:
x
F
RT
E 13
and the parameter c accounts for the differences between the
inner-shell force constants of oxidised and reduced species [10].
Note that when c = 0 the asymmetric Marcus-Hush model is re-
duced to its symmetric counterpart.
2.3. Simulation procedure
The simulation model is normalised according to the set of
dimensionless parameters shown in Table 1. Under this system,
Ficks second law becomes:
@C
@s

@
2
C
@R
2

@
2
C
@Z
2

1
R
@C
@R
14
and the ux equation becomes:
@C
A
@N
j k
red
C
A
k
ox
C
B
15
where N = n/r
e
is some dimensionless coordinate normal to the par-
ticle surface.
The outer boundary of the simulation space is placed such that
it is sufciently far from that nanoparticle surface that the diffu-
sion layer will be entirely contained within the simulation space
for the duration of the experiment. This condition is satised when
the boundary is placed a distance of 6

s
max
p
from the particle sur-
face in both the R and Z directions where s
max
is the dimensionless
experiment duration [11]. At this outer boundary we may there-
fore x the concentrations of both species A and B as their respec-
tive bulk value throughout the simulation.
The 2-dimensional simulation space is discretized by dividing it
up into a nite grid of spatial points. In the dimensionless (R, Z)
coordinate system, the surface of the spherical nanoparticle is de-
scribed by:
R
2
Z 1
2
1 16
where the dimensionless radius of the nanoparticle is (by deni-
tion) 1. The distribution of spatial points in the vicinity of the nano-
particle is determined by specifying an angular increment D/, and
calculating the values of R and Z at each successive value of /,
where / is the angle shown in Fig. 1. This allows grid points to t
precisely to the curved surface of the spherical particle. A similar
Table 1
Dimensionless parameters.
Parameter Normalization
Radial coordinate R = r/r
e
Axial coordinate Z = z/r
e
Time s Dt=r
2
e
Potential h F=RTE E
>
f

Concentration C
i
c
i
=c

A
Heterogeneous rate constant j = k
0
r
e
/D
M.C. Henstridge et al. / Journal of Electroanalytical Chemistry 712 (2014) 1418 15
method has been used previously to simulate conical electrodes
[12] and hemispherical electrodes [13]. The grid spacing expands
from the surface of the nanoparticle in both directions to the outer
boundary [9].
At the beginning of the simulation (s = 0), species A is present
throughout the simulation space at its bulk concentration while
species B is initially absent. As the system is symmetric about the
Z-axis (R = 0), there can be no material ux through this boundary
and so the boundary condition is simply one of zero ux
(@C/@R = 0), and likewise for the supporting surface (@C/@Z = 0 at
Z = 0).
To perform simulations, the mass transport equation (Eq. (14))
for each chemical species is discretized to the grid of (R, Z) spatial
points using the alternating direction implicit (ADI) method [14].
The resulting set of partial differential equations is non-linear,
and so it is solved for each species at each successive time step
using the iterative NewtonRaphson method [15], subject to the
boundary conditions described above.
The current is calculated at each time step by integrating the
ux across the entire nanoparticle surface as follows:
J 2p
_
p=2
p=2
J
/
cos /d/ 17
where J
/
is the ux given by:
J
/

@C
@Z
sin/
@C
@R
cos / 18
The dimensionless ux is then converted into current via:
I 4pFDc

A
r
e
J 19
Using this method simulations were carried out, using a scan
rate which was sufciently small to yield steady state behaviour,
for a wide range of simulation parameters, the results of which
are discussed below.
3. Results and discussion
3.1. Effect of varying asymmetry parameter, c
The asymmetry parameter c has previously been reported as
affecting voltammetry in a manner analogous to that of the But-
lerVolmer transfer coefcient a. This can be seen in Fig. 2, which
shows steady state voltammetry for a range of values of c, each
voltammogram reaches the same limiting current, but the steep-
ness of the current increase varies positive values of c give rise
to the steepest increases in current (analogous to a > 0.5) and
negative values of c give rise to more gradual increases in current
(analogous to a < 0.5).
3.2. Effect of varying reorganisation energy, k
More interesting, however, is the effect of the reorganisation
energy on the limiting current. It has previously been shown [2]
that the Marcus-Hush model can, theoretically, give rise to a kinet-
ically limited (rather than mass transport limited) steady state cur-
rent. Such conditions arise due to the levelling off behaviour of
rate constants with the Marcus-Hush model at large overpoten-
tials. Oldham has shown (for symmetric MH kinetics, i.e. c = 0) that
the rate constant reaches half of its limiting value when h = K [16].
This effect is therefore seen most clearly for small values of reorga-
nisation energy. Note that, for real systems in the solution phase,
values of the reorganisation energy typically lie in the range 1.0
3.0 eV [17,3].
Figs. 3 and 4 shows the effect of varying reoganisation energy
on the steady state voltammetry. As has been seen previously for
different electrode geometries, small values of k give rise to a
kinetically limited steady state current which is smaller than the
mass transport limit. For reorganisation energies of 1 eV and above
the mass transport limiting current is realised, however values of
0.75 eV and below give rise to limiting currents which are substan-
tially smaller than would be expected.
3.3. Effect of varying standard heterogeneous rate constant, k
0
Small values of reorganisation energy can give rise to a kineti-
cally limited steady state current, thus the limiting current should
also be sensitive to k
0
. This is conrmed in Figs. 4 and 5. The former
demonstrates that for a sufciently reversible system, even reorga-
nisation energies as small as 0.5 eV are not sufcient to observe a
kinetically limited current. The latter, however, demonstrates that
even values of k as large as 1 eV can, theoretically, yield signicant
deviations from the mass transport limited steady state current.
Also of note is that for the slowest kinetics (k
0
= 10
5
cm s
1
)
the initial current increase occurs at larger overpotential for the
Marcus-Hush model than for the ButlerVolmer model, even for
reorganisation energies as large as 5 eV. Unfortunately, however,
such an effect (effectively a shift in potential across the entire vol-
tammogram) could be the result of different values of either the
formal potential or the reorganisation energy, as well as the heter-
ogeneous rate constant.
r / nm
z
/

n
m
Fig. 1. The reduced 2D (r, z) coordinate system used to model a spherical
nanoparticle in simulations. Though not one of the coordinates, angle / is useful
for current calculations and for specifying the discretized spatial grid.
Fig. 2. Steady state voltammetry simulated for a single nanoparticle of radius
100 nm on a planar surface. In each case k
0
= 10
3
cm s
1
, k = 2 eV and
0.25 < c < +0.25 (in increments of 0.05). The dotted line highlights the case of c = 0.
16 M.C. Henstridge et al. / Journal of Electroanalytical Chemistry 712 (2014) 1418
In order to observe a kinetically limited steady state current we
require a small value of limiting rate constant, which is a function
of both k and k
0
. If we wish to detect this phenomenon experimen-
tally then we require a limiting rate constant which is as small as
possible, which in turn means small values of both k and k
0
. This,
however, is problematic since these two parameters are not inde-
pendent Marcus theory tells us that a small value of reorganisa-
tion energy leads to fast electron transfer, i.e. a large value of k
0
,
and vice versa.
Feldberg outlined the conditions necessary to distinguish Mar-
cus-Hush kinetics from ButlerVolmer kinetics at steady state for
microdisc electrodes [2]. In real systems, however, such condi-
tions are unlikely to be met when the above limitation is consid-
ered. A similar study using the channel ow electrode also
concluded that it was unlikely that it would be possible to distin-
guish the two models either from the shape of the voltammetry
or from the steady state current deviating from the mass transport
limit [7].
3.4. Effect of varying nanoparticle size, r
e
Next, the effect of varying nanoparticle radius was examined. As
can be seen in Fig. 6, for those kinetic parameters which yielded a
steady state current lower than the mass transport limit for the
100 nm nanoparticle, the same kinetic parameters yielded an even
smaller steady state current for smaller nanoparticles. A reorgani-
sation energy of 1 eV gave a steady state current which was indis-
tinguishable from the mass transport limit for a nanoparticle of
100 nm, but gave a steady state current which is 5% smaller than
the mass transport limit when the nanoparticle radius is decreased
to 25 nm.
Decreasing the nanoparticle radius has the effect of increasing
the efciency of mass transport to the nanoparticle surface. Under
such conditions the rate of electron transfer can be higher, in com-
parison to the situation for larger nanoparticles, and yet still be
small enough that the mass transport limit is not realised.
Decreasing the nanoparticle radius below 25 nm could result in
kinetically limited being observable for systems with even faster
electrode kinetics, further increasing the likelihood of observing
this phenomenon experimentally. For small (110 nm) nanoparti-
cles, however, the situation is complicated by the diffusion layer
around the nanoparticle becoming sufciently small so as to be
on the same scale as the double layer [18]. Thus migration can
no longer be neglected, regardless of the concentration of support-
ing electrolyte, and the present model can no longer be applied.
4. Conclusions
We have shown through numerical simulation that, as is the
case for microdisc electrodes and channel ow electrodes, it is the-
Fig. 3. Steady state voltammetry simulated for a single nanoparticle of radius
100 nm on a planar surface. For each of the solid lines k
0
= 10
3
cm s
1
, c = 0 and
k = 0.5, 0.6, 0.75, 1.0, 1.5, 2.5 and 5.0 eV. The dotted line shows the case of
ButlerVolmer kinetics with a = 0.5 for comparison.
Fig. 4. Steady state voltammetry simulated for a single nanoparticle of radius
100 nm on a planar surface. All parameters are as in Fig. 3, except k
0
= 10
1
cm s
1
.
Fig. 5. Steady state voltammetry simulated for a single nanoparticle of radius
100 nm on a planar surface. All parameters are as in Fig. 3, except k
0
= 10
5
cm s
1
.
Fig. 6. Steady state voltammetry simulated for a single nanoparticle on a planar
surface. All parameters are as in Fig. 3, except r
e
= 25 nm.
M.C. Henstridge et al. / Journal of Electroanalytical Chemistry 712 (2014) 1418 17
oretically possible to observe a kinetically limited steady state cur-
rent at a nanoparticle on a surface which is smaller than the mass
transport limit. We have also demonstrated that, in contrast to
other electrode systems, it may be possible to observe such a kinet-
ically limited steady state current experimentally provided that the
nanoparticles are sufciently small.
Acknowledgments
The research leading to these results received funding from the
European Research Council under the European Unions Seventh
Framework programme (FP/20072013)/ERC Grant Agreement
No. (320403). MCH is grateful for funding via an EPSRC Doctoral
Prize.
References
[1] C.E.D. Chidsey, Science 251 (1991) 919922.
[2] S.W. Feldberg, Anal. Chem. 82 (2010) 51765183.
[3] M.C. Henstridge, Y. Wang, J.G. Limon-Petersen, E. Laborda, R.G. Compton,
Chem. Phys. Lett. 517 (2011) 2935.
[4] E. Laborda, Y. Wang, M.C. Henstridge, F. Martnez-Ortiz, A. Molina, R.G.
Compton, Chem. Phys. Lett. 512 (2011) 133137.
[5] D. Suwatchara, M.C. Henstridge, N.V. Rees, R.G. Compton, J. Phys. Chem. C 115
(2011) 1487614882.
[6] E. Laborda, M.C. Henstridge, R.G. Compton, J. Electroanal. Chem. 667 (2012)
4853.
[7] M.C. Henstridge, N.V. Rees, R.G. Compton, J. Electroanal. Chem. 687 (2012) 79
83.
[8] J.M. Kahk, N.V. Rees, J. Pillay, R. Tshikhudo, S. Vilakazi, R.G. Compton, Nano
Today 7 (2012) 174179.
[9] K.R. Ward, N.S. Lawrence, R.S. Hartshorne, R.G. Compton, J. Electroanal. Chem.
683 (2012) 3742.
[10] E. Laborda, M.C. Henstridge, C. Batchelor-McAuley, R.G. Compton, Chem. Soc.
Rev. 42 (2013) 48944905.
[11] A. Einstein, Ann. Phys. 17 (1905) 549560.
[12] E.J.F. Dickinson, I. Streeter, R.G. Compton, J. Phys. Chem. C 112 (2008) 11637
11644.
[13] K.R. Ward, N.S. Lawrence, R.S. Hartshorne, R.G. Compton, J. Phys. Chem. C 115
(2011) 1120411215.
[14] D. Britz, Digital Simulation in Electrochemistry, third ed., Springer, 2005.
[15] W.H. Press, S.A. Teukolsky, W.T. Vetterling, B.P. Flannery, Numerical Recipes,
third ed., Cambridge University Press, Cambridge, 2007.
[16] K.B. Oldham, J.C. Myland, J. Electroanal. Chem. 655 (2011) 6572.
[17] J.O. Bockris, S.U.M. Kahn, D.B. Matthews, J. Res. Inst. Catal. Hokkaido Univ. 22
(1974) 121.
[18] E.J.F. Dickinson, R.G. Compton, J. Electroanal. Chem. 661 (2011) 198212.
18 M.C. Henstridge et al. / Journal of Electroanalytical Chemistry 712 (2014) 1418