3262 J. Phys. Chem.

B 2006, 110, 3262-3270

Dynamic Diffuse Double-Layer Model for the Electrochemistry of Nanometer-Sized

Electrodes

Rui He, Shengli Chen,* Fan Yang, and Bingliang Wu

Department of Chemistry, Wuhan UniVersity, Wuhan 430072, P. R. China
ReceiVed: January 5, 2006

A dynamic diffuse double-layer model is developed for describing the electrode/electrolyte interface bearing
a redox reaction. It overcomes the dilemma of the traditional voltammetric theories based on the depletion
layer and Frumkin’s model for double-layer effects in predicating the voltammetric behavior of nanometer-
sized electrodes. Starting from the Nernst-Planck equation, a dynamic interfacial concentration distribution
is derived, which has a similar form to the Boltzmann distribution equation but contains the influence of
current density. Incorporation of the dynamic concentration distribution into the Poisson and Butler-Volmer
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

equations, respectively, produces a dynamic potential distribution equation containing the influence of current
and a voltammetric equation containing the double-layer effects. Computation based on these two equations
Downloaded via UNIV OF NEW SOUTH WALES on July 3, 2024 at 06:01:56 (UTC).

gives both the interfacial structure (potential and concentration profiles) and voltammetric behavior. The
results show that the electrochemical interface at electrodes of nanometer scales is more like an electric-
double-layer, whereas the interface at electrodes larger than 100 nm can be treated as a concentration depletion
layer. The double-layer nature of the electrode/electrolyte interface of nanometer scale causes the voltammetric
responses to vary with electrode size, reactant charge, the value of formal redox potential, and the dielectric
properties of the compact double-layer. These voltammetric features are novel in comparison to the traditional
voltammetric theory based on the transport of redox molecules in the depletion layer.

1. Introduction effect, which is also called the Frumkin effect since Frumkin
Electrochemical processes confined into interfacial domains initially addressed the basic concept, deals with the influence
of nanometer scales can be of importance in many areas of potential drop in the diffuse double-layer on the concentration
including fabrication of nanomaterials and nanodevices, energy of redox species at the plane of closest approach and the
conversion, chemical sensing, and scanning probe microscopy. effective driving force of the interfacial charge transfer. The
Since the late 1980s, electrochemistry at nanometer-sized major problem in Frumkin’s approach is that it is based on
electrodes has been the subject of extensive studies.1-11 As well Gouy-Chapman’s equilibrium double-layer model and that the
as applications in scanning probe microscopies12-14 and mea- transport of redox molecules is totally ignored. The conventional
surements of fast electron-transfer rates,6,9-11,15 electrodes of electromigration theories go to the other extremity in dealing
nanometer size offers novel opportunities to investigate the with the interfacial electric filed effects. Most of these theories
interfacial electrochemical reactions confined into spatial do- only consider the transport of redox species in the depletion
mains of nanoscales. Recently, results presented in a number layer, and the effect of the double-layer is totally ignored.
of studies have shown that the voltammetric behaviors of Furthermore, most of the electromigration theories are based
nanometer-sized electrodes deviate significantly from the pre- on the assumption of electroneutrality in the depletion layer.17-19
dictions of classic electrochemical theories.1-5,7,8,16 For example, The problems in Frumkin’s approach for the double-layer effects
the voltammetric behavior of electrodes of submicrometer and the conventional electromigration theories would be magni-
dimension upon removal of the supporting electrolyte cannot fied at the electrode/electrolyte interface of nanometer scale.
be interpreted in terms of either the conventional electromigra- This is because the mass transport rates to/from nanometer-
tion theories or Frumkin’s double-layer effects.1,2,4 Even in the sized electrodes are very high, which, on one hand, can induce
presence of excess supporting electrolyte, the limiting currents significant charge separation in the depletion layer so that the
of the steady-state polarization curves obtained at electrodes of electroneutrality assumption becomes invalid and, on the other
nanometer scales for one-electron redox reactions were found hand, can significantly break the equilibrium distribution of ions
to vary with redox molecules,2 which obviously disagrees with in the double-layer.
the diffusion-based voltammetric theories. These results imply Recently, a few studies on the voltammetric theory have
that the effects of electric fields at electrode/electrolyte interfaces attempted to tackle these problems, especially the electroneu-
on an electrochemical reaction have to be reconsidered as long trality assumption.7,16,20-22 The basic approach employed in
as nanometer-sized electrodes are used.
these studies is to replace the electroneutrality approximation
The influences of the interfacial electric field on an electrode
by the more rigorous Poisson equation. The studies by Smith
reaction are mainly described in terms of either the double-
and White7 and Oldham and Bond16 are particularly devoted to
layer effect or the electromigration effect.17 The double-layer
the voltammetric responses of nanometer electrodes through
* Corresponding author. E-mail: [email protected]. Phone: +86 27- simultaneous solution of the Poisson and Nernst-Planck
68754693. Fax: +86 27-68754067. equations. Smith and White’s numerical solution of the Nernst-
10.1021/jp060084j CCC: $33.50 © 2006 American Chemical Society
Published on Web 02/02/2006
Model for the Electrochemistry of Nanoelectrodes J. Phys. Chem. B, Vol. 110, No. 7, 2006 3263

Planck-Poisson equations demonstrated that the diffusion-based solving the PNP equation directly with a numerical method,
transport theories become no longer valid for predicting the we have derived a so-called dynamic concentration distribution
voltammetric responses of electrodes of radii less than 100 nm equation from the Nernst-Planck equation and the Faraday law.
even in the presence of an excess of a supporting electrolyte. The electrode/electrolyte interface is then formulated in a way
Oldham and Bond solved the Nernst-Planck-Poisson equations similar to that which Gouy and Chapman used to deal with the
in a relatively analytical way. They concluded that detectable equilibrium double-layer. The difference is that the dynamic
charge separation in the depletion layer only occurs when the concentration distribution equation is used instead of the
electrode diameter is less than 10 nm even in the absence of a Boltzmann law. A final interfacial potential distribution equation,
supporting electrolyte. This is obviously inconsistent with the which reduces to the traditional Poisson-Boltzmann equation
findings in recent experimental studies.1-4 By carefully review- when current flowing through the interface is zero, is obtained.
ing the derivation processes in Oldham and Bond’s study, we In addition, combination of the dynamic concentration distribu-
found that the double-layer is actually ignored in their model. tion equation and the Butler-Volmer equation produces a
They only considered the electromigration effect although the voltammetric equation containing the double-layer effects. The
electroneutrality approximation is avoided. Neglecting the effect computation results based on this model show that the electro-
of the double-layer is rather unreasonable either when a chemical interface becomes more like an electric-double-layer
nanometer-sized electrode is used or the supporting electrolyte as the electrode approaches nanometer dimensions, making the
is absent. Smith and White’s model considered both the traditional voltammetric theory based on the depletion layer
electromigration and double-layer effects. According to their model inappropriate. It is also shown that the structure of the
calculation, peak-shaped instead of sigmoid-shaped steady-state compact double-layer has a crucial impact on the voltammetric
polarization curves would be obtained for the oxidation of a responses of nanometer electrodes.
positively charged reactant (or the reduction of a negatively The term “dynamic diffuse double-layer” was initially used
charged reactant) at submicrometer electrodes. So far, no by Levich in the 1940s to describe how the diffuse double-
experimental evidence has ever been observed to support such layer may be disturbed by an electrode reaction.23 Since he
a prediction. Instead, sigmoid-shaped polarization curves with considered interfaces formed at large flat electrodes, it was
limiting current plateaus are usually obtained for the oxidation shown that the influence of the double-layer on an electrode
of cations (e.g., ferrocenylmethyltrimethylammonium cation, reaction is negligible in the presence of an excess of a supporting
TMAFc+)1,4,6 or the reduction of anions (e.g., Fe(CN)63- or electrolyte. As will be shown in this paper, the situation at
IrCl62-)2,3 even at very small electrodes. electrode/electrolyte interfaces of nanometer scales is much
The Poisson-Nernst-Planck (PNP) equations are probably different.
the most rigorous approach on the basis of mean-field continuum
2. Dynamic Diffuse Double-Layer Model
theory to formulate interfacial electrochemical reactions that
involve transport of redox ions in the electric field. The varied
results produced by solving PNP equations for similar problems Oz + e- ) Rz-1 (1)
are mainly due to the diverse ways to define the boundary of In the present study, we consider an outer-sphere one-electron
an electrochemical interface in different studies. Establishing reduction reaction (eq 1) at a spherical (or hemispherical)
proper boundary conditions is of crucial importance in electro- electrode. Choosing the spherical geometry simplifies the
chemical theoretical analysis. It is the boundary conditions that theoretical derivation without losing the generality. As the
define the physical nature of a specific electrochemical system, electrode size is reduced to a nanometer scale, the interfacial
including the electrode geometries, properties of the redox field would become more or less spherical despite the electrode
reactions, and the applied electrochemical perturbations.17 shapes. It is assumed that the electrode is made of conductive
Inappropriate boundary conditions would lead to false results materials so that the interfacial double-layer only extends toward
and conclusions even if exactly the same mathematical equations the solution side and that the electrolyte solution initially
were employed. For instance, Oldham and Bond have treated contains the reactant Oz and a 1:1 ratio of a supporting
the double-layer in a very different way from that in Smith and electrolyte of A+B- with a concentration of 0.5 mol L-1, which
White’s paper. It should also be pointed out that both of these is large enough so that the conventional Frumkin double-layer
studies have assumed that the electrode processes are transport- effect can be avoided and a further increase of the supporting
limited, i.e., electrochemically reversible. This allows utilization electrolyte concentration would result in little change on the
of the Nernst relationship between the electrode potential and double-layer structure. For simplicity, we also make the
concentrations of electroactive species as a boundary condition following assumptions: the counterion accompanying the
for solving the PNP equations. A fact that has been ignored by reactant Oz is either A+ or B- depending on the sign of the
these researchers is that most of the electrochemical reactions reactant charge z; the initial concentration ratio between the
would become rather irreversible at electrodes of nanometer reactant AzO (or OBz) and the supporting electrolyte is 0.01,
scales due to very high mass transport rates. The Nernst relation which ensures a negligible contribution of electromigration of
is therefore no longer applicable. redox species to the voltammetric responses according to
In the face of the increasing importance of the electrochem- traditional electrochemical theories; none of the ions in the
istry at interfaces of nanometer dimensions and the wide solution specifically adsorbs on the electrode surface and all
applications of nanometer electrodes, it seems imperative to have the ions have the same plane of closest approach (PCA, which
a more accurate theoretic description of the electrochemistry is also called the outer Helmholz plane and is denoted as OHP)
of nanometer-sized electrodes. In this paper, we introduce a so that no charge resides in the compact part of the double-
dynamic diffuse double-layer model to describe the electrode/ layer; the plane of electron transfer (PET) coincides with the
electrolyte interface bearing an electrochemical reaction. In this PCA, i.e., the electron transfer occurs at r ) r0 + µ, where r is
model, the entire electrochemical interface is described as a the radial coordinate originating from the center of the spherical
dynamic electric double-layer; therefore, no distinction between electrode, r0 is the electrode radius, and µ is the thickness of
the double-layer and the depletion layer is required. Rather than the compact double-layer; the potentials are measured with
3264 J. Phys. Chem. B, Vol. 110, No. 7, 2006 He et al.

respect of that in the bulk solution, which is assumed to be We may solve eq 10 to obtain current-density dependent
equivalent to the potential of zero charge (PZC). concentration profiles for species j. Before doing so, it will be
The way we model the dynamic electrode/electrolyte interface useful to make the spatial coordinate r dimensionless by letting
is similar to that which Gouy and Chapman used to treat the p ) (r0 + µ)/r, which enables the problem of solving equations
equilibrium electrochemical interface. Gouy-Chapman’s equi- in the infinite range of (r0 + µ) e r e ∞ to be replaced by that
librium double-layer model is based on the Boltzmann distribu- in the finite interval of 1 g p g 0. Some other dimensionless
tion of ions (eq 2) and the Poisson equation (eq 3). variables will lead to further notational economy; they are the
following: the dimensionless potential, φ ) Fæ/RT; the
cj ) cjbe-(zjF/RT)æ (2) dimensionless concentration, cj* ) cj/cOb; the dimensionless
current density, i* ) i/idL, in which idL ) FDOcOb/(r0+µ)
F corresponds to the limiting current density for the reduction of
∇ 2æ ) - (3)
0 Oz to Rz-1 under pure diffusion control. It should be noticed
that idL thus defined is different from the limiting diffusion
In eqs 2 and 3, cj and cjb are the local and bulk concentration current density in conventional voltammetric theories. The latter
of ion j, respectively; zj is its charge; æ is the local potential;  is defined as FDOcOb/r0, in which the thickness of compact
and 0 are the local dielectric constant and the permittivity of a double-layer µ is ignored in the reaction coordinate. For
vacuum; F ) ∑zjcj is the local charge density; and F, R, and T simplicity, we also assume that O and R have the same diffusion
have their usual meanings. Combining eqs 2 and 3 produces coefficient D. Thus, eq 10 reduces to the following dimension-
the well-known Poisson-Boltzmann equation, that is less form,

1 dcj*
∇2æ ) - ∑
 j
cjbe-(z F/RT)æ
j
(4)
dp
+ zjcj*
dφ
dp
) -mji* (11)
0

Solving eq 4 gives a gradient potential profile in the diffuse Solving eq 11 with boundary conditions that cO* ) 1 and
double-layer. cR* ) cRb/cOb at p ) 0 gives
The major difference of a dynamic electrochemical interface
from the equilibrium one is that there is net movement of ions cj* ) e-zjφ(cjb* - i*Ωj) (12)
through the interface due to interfacial reactions. Thus, the
Boltzmann law that describes the equilibrium distribution of where
ions becomes inappropriate. The transport of ions in an electric
field is generally described by the Nernst-Planck equation, Ωj ) mj ∫0pez φ dp
j
(13)
according to which the flux J (mol cm-2 s-1) of the ion j is
The derivation details of eq 12 are given in Appendix A.
zj F This equation represents a dynamic distribution of ionic species
Jj ) -Dj∇cj - Djcj∇æ (5)
RT at an interface where a redox reaction is proceeding since the
influence of current density (i.e., the reaction rate) is included.
where Dj refers to the diffusion coefficient of ion j. As the steady Equation 12 will becomes the dimensionless Boltzmann equation
state has been reached, when mj ) 0. This says that the local electrostatic equilibrium
distribution of inert electrolyte species is maintained even though
∂cj
) -∇Jj ) 0 (6) there is a current flowing through the electrode/electrolyte
∂t interface. However, the distribution of redox species of Oz and
Rz-1 is altered due to the occurrence of interfacial reactions.
In spherical coordinates,
The deviation from the equilibrium profiles depends on the
dJj 2 current density. The larger the current density, the more the
1 d 2
∇Jj ) + J ) (r Jj) (7) redox species deviate from a Boltzmann distribution. As the
dr r j r2 dr current density goes to zero, the Boltzmann distribution is
recovered.
It is easily recognized from eqs 6 and 7 that
Similar to Gouy and Chapman’s approach for the equilibrium
double-layer, we can combine the derived dynamic distribution
r2Jj ) const (8)
equation and the Poisson equation to obtain a dynamic interfacial
According to the mass conservation and Faraday laws, we have equation. The Poisson equation has the following dimensionless
the following equation at PET form,

d2φ
(Jj)r)r0+µ ) -mj
F
i
(9) p4
dp2
) -(r0 + µ)2κ2 ∑zici* (14)

where i is the current density flowing through the electrode/ where

( )
electrolyte interface, mj is a integer constant and takes values
of -1, +1, and 0 for species Rz-1, Oz, and inert electrolyte 0RT -1

species (A+ and B-), respectively, if the reduction current is κ2 ) 2 b

(15)
F 2cO
defined as positive. Combining eqs 5, 8, and 9 gives

dcj zjF (r0 + µ)2i The reciprocal of κ has the dimension of thickness and is called
dæ
Dj + Djcj ) mj (10) the Debye-Hückel length. Substitution of eq 12 into eq 14
dr RT dr r2F yields
Model for the Electrochemistry of Nanoelectrodes J. Phys. Chem. B, Vol. 110, No. 7, 2006 3265

d2φ (r0 + µ)2κ2

dp2
) -
p4
[ ∑(zjcjb*e-z φ) + i*∑Ωjzje-z φ]
j j
(16)

It can be seen that eq 16 will reduce to the dimensionless

Poisson-Boltzmann (PB) equation when the current density is
zero. Similar to the PB equation, that is, the basic interfacial
equation under equilibrium, eq 16 is a characteristic equation
to describe the dynamic electrochemical interface where an
interfacial reaction is taking place. Unlike the PB equation, eq
16 alone is insufficient to reveal the interfacial properties and
behaviors since it contains an additional unknown variable, the
current density. It is well-known that the current-potential
dependence can be described by the Butler-Volmer equation,
which has the following dimensionless form

k0(r0 + µ) s -R(E-φPET-E0′)
- cRs*eâ(E-φPET-E ′)] (17)
0
i* ) [cO *e
D Figure 1. Schematic diagram of the compact double-layer (a) and the
dielectric variation in the interfacial region (b).
k0
In eq 17, is the standard rate constant of redox reaction of
eq 1 at formal potential E0′, and R and â are the transfer The potential at the boundary of the PET (p ) 1) depends
coefficients. E is the potential value at the electrode surface. on the electrode potential E at the surface and the potential drop
φPET, cOs*, and cRs* are the potential, concentration of Oz, and in the compact double-layer. To obtain the potential drop in
concentration of Rz-1 at PET, respectively. All the concentra- the compact double-layer, a fairly detailed model of the double-
tions and potentials are in their dimensionless forms. The layer and its dielectric properties must be specified. It is
difference between eq 17 and the conventional Butler-Volmer generally recognized that the dielectric permittivity changes with
equation is that the potential drop in the diffuse double-layer distance from the electrode surface in the interfacial region.25-27
(φPET) is incorporated in eq 17. This is similar to Frumkin’s For simplicity, we assume that the compact double-layer can
approach to correct the double-layer effects. The main problem be divided into two regions (see Figure 1a): the inner part that
in the Frumkin model for double-layer effects is that the is composed of the first layer of the solvent molecules bound
concentrations and diffuse double-layer potential are treated to the electrode surface and the outer part that spans from the
according to Gouy-Chapman’s equilibrium double-layer theory.17 outer edge (dotted line 1 in Figure 1a) of the inner part to the
For instance, cOs* and cRs* are determined using the Boltzmann plane of closest approach of ions (dotted line 2 in Figure
equation. This is obviously problematic as an interfacial reaction 1a).17,25,27 We also assume that the dielectric permittivity is
is taking place. Instead of using the Boltzmann distribution law, uniform within each region of the double-layer but that it has
we can substitute cOs* and cRs* with eq 12 by setting p ) 1. different values in different regions (Figure 1b). Such assump-
Considering that cOb* ) 1 and cRb* ) 0, we have tions enable us to correlate the potential at the PET, φPET, to
1 the electrode potential E as
i* ) (18)

( )( )
1 R(E-E0′) (z-R)φPET
+ ΛO + ΛRe(E-E ′)
0
e e (r0 + µ)2 µid (µ - µi)d dφ
λ φPET ) E - + (21)
(r0 + µi) r0ic (r0 + µ)oc dr PET
in which λ ) k0(r0 + µ)/D and Λj ) ezjφ dp. Equation 18 is
∫10
a voltammetric equation containing the double-layer effects. The In eq 21, ic, oc, and d are the effective dielectric constants of
traditional steady-state voltammetric equation without consider- the inner part of the compact double-layer, the outer part of the
ing the double-layer effects is compact double-layer, and the diffuse double-layer, respectively
and µi is the thickness of the inner part of the compact double-
1
i* ) (19) layer. The derivation details of eq 21 are shown in Appendix
1 R(E-E0′)
+ 1 + e(E-E ′)
0
e B.
λ Equations 16 and 18 are solved numerically with the boundary
which can be derived by combining the Butler-Volmer equation conditions of eqs 20 and 21 according to the method and
with Fick’s diffusion law.17,24 parameters given in Appendix C.
Thus, eqs 16 and 18 are the final equations of the dynamic
3. Results and Discussion
double-layer model. Simultaneously solving these two equations
with appropriate boundary conditions will be able to give the 3.1. Electric-Double-Layer Nature of Electrochemical
potential profiles at and the current density flowing through the Interfaces at Electrodes of Nanometer Scales. The dimension-
electrode/electrolyte interface. less current density i* given by simultaneous solution of eqs
Since eq 16 is a second-order differential equation about the 16 and 18 is actually the normalized value of the real current
potential, two boundary conditions related to the potential are density by the limiting current density due to diffusion, idL. Thus,
required to solve it. We have assumed that the plane of closest the limiting value of i* should be equal to one if the electrode
approach for all the redox molecules is at the PET; thus, the reaction is controlled purely by the diffusion of redox species.
two boundaries are the bulk solution (p ) 0) and the PET (p ) Figure 2 shows the computed steady-state i*-E curves corre-
1), respectively. The boundary condition in the bulk solution is sponding to the reaction of eq 1 with z ) -1 at electrodes of
different sizes. The numbers inserted in the figure refer to the
φp)0 ) 0 (20) electrode radii. It can be seen that the i*-E curves have sigmoid
3266 J. Phys. Chem. B, Vol. 110, No. 7, 2006 He et al.

Figure 4. Potential distribution inside the compact double-layer at

electrodes of different sizes. The electrode potential E ) 0.1 V. The
numbers inserted in the figure refer to the radii of electrodes.

Figure 2. Theoretical steady-state voltammetric curves corresponding concentration gradient is seen at an electrode of 1 nm in radius;
to the reduction of 5 × 10-3 mol L-1 Oz with z ) -1 at electrodes of therefore, pronounced effects of the electric field on the
a variety of sizes. The parameters are the following: k0 ) 1.0 cm s-1, voltammetric responses are expected. Thus, the potential and
D ) 1 × 10-5 cm s-2, R ) â ) 0.5, E0′ ) 0, (the concentration of concentration profiles in Figure 3 can explain the deviation of
supporting electrolyte) cAB ) 0.5 mol L-1. the limiting current density from idL at electrodes smaller than
100 nm shown in Figure 2 well, i.e., the pronounced penetration
of the electric field into the concentration field makes the
transport of redox molecules deviate from pure diffusion
behavior as the electrode size approaches nanometer scale. At
large electrodes, the interface is more like a concentration
depletion layer, whereas it is more close to an electric-double-
layer at an electrode of nanometer scale. The electric-double-
layer nature of the interface at nanometer electrodes may raise
some voltammetric features that are unseen in the voltammetric
behaviors of conventional large electrodes.
It should be pointed out that the voltammetric behaviors of
electrodes having radii of 1 nm or less would not be affected
Figure 3. Potential and reactant concentration profiles in the interfacial by the electrostatic double-layer effects addressed above only.
region. The conditions and parameters are the same as those for Figure The finite molecule size of ions may play a significant role in
2. the voltammetric responses of such small electrodes. As
discussed by Smith and White,7 the point-charge assumption
shapes regardless of the electrode size. At electrodes larger than of ions would unrealistically overestimate the screening of the
100 nm, the limiting current density is almost the same as the electrode charge by the inert ions, which tends to diminish the
diffusion-controlled limiting current density idL, i.e., i* ) 1. effects of the interfacial electric field. Thus, considering the
This indicates that the influence of the electric field on the finite molecule size of ions, the double-layer effects should be
voltammetric behavior at electrodes larger than 100 nm is more pronounced than those predicted by the present model.
negligible. However, the value of the limiting current density 3.2. Effect of Electrode Size. The double-layer nature of the
deviates gradually from idL as the electrode radius goes below interface makes the intrinsic interfacial properties and behavior
100 nm. The smaller the electrode size, the more the limiting become size-dependent. This has been depicted in Figures 2
current density deviates from idL. and 3, which show that the voltammetric behaviors and
Figure 3 shows the distribution of the potential and the interfacial structures are electrode-size dependent as r0 < 100
reactant concentration in the interfacial region at electrodes of nm. The electrode size effects are actually implied by eq 21
100 nm (dash line) and 1 nm (solid line) in radii, respectively, that defines the boundary of the diffuse double-layer. As
as E - E0′ ) -0.5 V where the limiting current is reached. mentioned in the Introduction, the boundary conditions have a
The potential distribution can be obtained simultaneously with crucial impact on the computation results. It is seen from eq 21
the current density by solution of eqs 16 and 18. The concentra- that the electrode size becomes significant only when r0 is close
tion distribution is then obtained according to eq 12 with the to µ. Otherwise, as r0 . µ, eq 21 becomes
calculated potential distribution and the current density. The

( )( )
x-axis in Figure 3 represents the distance from the PET µid (µ - µi)d dφ
normalized by a thickness parameter δ, which is measured from φPET ) E - + (22)
the PET to the location where the concentration of reactant Oz ic oc dr PET
reaches 95% of its bulk value. We may consider δ as the
thickness of the entire interfacial region. It can be seen that the in which the electrode size effect disappears. As shown in Figure
interfacial potential distribution at an electrode of 100 nm occurs 4, the structure of the compact double-layer also depends on
in a very narrow region that is negligible in the entire interface. the electrode size. For an electrode having a radius larger than
The penetration of the electric field into the concentration field 100 nm, the potential distributes itself linearly in the compact
is almost unseen (less than 0.01%). This implies that the double-layer, which has been generally recognized for conven-
influence of the interfacial electric field on the voltammetric tional electrodes of large scales.17,27 As the electrode becomes
responses of electrodes larger than 100 nm should be negligible. smaller than 100 nm, however, the potential distribution
Whereas, significant overlap between potential gradient and gradually becomes curved and varies with electrode size.
Model for the Electrochemistry of Nanoelectrodes J. Phys. Chem. B, Vol. 110, No. 7, 2006 3267

Figure 5. Steady-state voltammetric curves for the reduction of

differently charged reactants at an electrode of 10 nm. The other Figure 7. Theoretical voltammetric curves for k0 ) 1000 cm s-1. The
conditions and parameters are the same as those for Figure 2. other conditions and parameters are the same as those for Figure 2.

tions. According to Smith and White’s calculation, the limiting

current density for the reduction of reactants with a charge of
+1 will be enhanced by 50% at an electrode of 1 nm in radius.
The results in this study show an enhancement of about 5%.
Even for the reduction of reactants with a charge of +3, the
enhancement of iL is less than 20%. In addition, Smith and
White’s study indicated that the reduction of a negatively
charged reactant (or oxidation of a positively charged reactant)
at electrodes of submicrometer radii would be greatly inhibited
so that a peak-shaped rather than sigmoid-shaped i-E curve
would be obtained. The results in the present study have shown
Figure 6. Ratio between the limiting current density predicated by that the reduction of negatively charged reactants on very small
the dynamic diffuse double-layer model and that calculated according electrodes still exhibits sigmoid-shaped polarization curves.
to eq 19 for reduction of variously charged reactants at electrodes having
different sizes. The other conditions and parameters are the same as
There are mainly two differences between the theoretical
those for Figure 2. treatment in the present study and that by Smith and White.
First, we use the Butler-Volmer equation rather than the Nernst
3.3. Effect of Reactant Charge. As indicated by eq 18, the relationship to correlate the electrode potential and concentra-
effects of the double-layer cause the current-potential relation tions of redox molecules at the PET. This causes the voltam-
at nanometer electrodes to depend on the charge z carried by metric behavior for a redox reaction with k0 ) 1 cm s-1 to
the reactant. This is manifested by the voltammetric curves for become totally irreversible at nanometer electrodes. Figure 7
the reduction of differently charged reactants at an electrode of gives the voltammetric responses for a one-electron reduction
10 nm in Figure 5. The deviation from the diffusion behavior reaction with k0 ) 1000 cm s-1 so that k0r0/D ) 10, which
becomes more pronounced for a reactant with a higher charge. means that the reaction is totally reversible. In such a case, the
The reduction of a negatively charged reactant and a positively Butler-Volmer equation is actually equivalent to the Nernst
charged reactant are inhibited and enhanced, respectively. For equation. It can be seen that the limiting current density on i*-E
the reduction of a neutral reactant, the limiting current is almost curves for electrodes of the same size in Figure 7 shows less
the same as idL. However, the whole voltammetric curve does deviation from the diffusion behavior than that for k0 ) 1 cm
not necessarily overlap with the diffusion-controlled one s-1, as shown in Figure 2. This says that the assumption of
calculated according to eq 19 (dotted line). Although the reversibility would not enhance the deviation of voltammetric
transport of the neutral reactant is not affected by the interfacial responses from diffusion-controlled behavior.
electric field, the transport and profile of the product is still A second difference between the current study and that by
affected. In addition, the potential drop in the diffuse part of Smith and White is the compact double-layer structure that is
the double-layer modifies the effective driving force for the used to establish the boundary condition at the PET. Smith and
charge-transfer step, as suggested by Frumkin’s double-layer White treated the compact double-layer as a uniform dielectric
effects.17 of 0.2 nm in thickness with the same dielectric constant as that
in bulk solution, e.g., 78. Such a compact double-layer model
Figure 6 summarizes the effects of the electrode size and
is unreasonable in a few aspects. It is generally recognized that
reactant charge on the ratio between the real limiting current
the dielectric permittivity is not uniform in the double-layer due
density iL and the limiting diffusion current density idL. It can
to the interfacial electric field.26,27 In the absence of specific
be seen that the difference between iL and idL is negligible when
adsorption, the inner part of the compact double-layer is made
the electrode radius is larger than 100 nm, but it becomes larger
up of a layer of highly oriented solvent molecules with a
and larger with decreased electrode size and increased reactant
dielectric constant of about 6, which is much lower than the
charge.
value in bulk solution.27 The nonspecific adsorbed ions usually
3.4. Effect of the Compact Double-Layer Structure. It is approach the electrode with their solvent shell or other ligands
noticed that the differences between iL and idL calculated in this and cannot penetrate into the first solvent layer (see Figure 1a).
study are much less than that reported by Smith and White in The dielectric constant of this layer is between that of the inner
a recent study on voltammetric theories of nanometer elec- solvent layer and the bulk solution. The thickness of the compact
trodes,7 although we have chosen the same studying system, a double-layer may be approximately estimated using27,28
one-electron redox process of 5 mM reactant in the presence of
a 1:1 supporting electrolyte of 0.5 M, and have employed the
same theoretical approach, i.e., Nernst-Planck-Poisson equa- µ ) 2rw + x3rw + ri (23)
3268 J. Phys. Chem. B, Vol. 110, No. 7, 2006 He et al.

Figure 8. Theoretical voltammetric responses of an electrode of 1 nm

in radius calculated based on the assumption of different dielectric Figure 10. Theoretical voltammetric responses of an electrode of 100
properties in the compact double-layer. The other conditions and nm in radius for reactions with different redox formal potentials (shown
parameters are the same as those for Figure 7. by the numbers inserted). The other conditions and parameters are the
same as those for Figure 2.

1 cm s-1 would become reversible and the influence of the

double-layer on the interfacial reaction kinetics is negligible in
the presence of an excess of a supporting electrolyte. For redox
reactions with lower k0 values or in the absence of a supporting
electrolyte, a dependence of voltammetric response on E0′ would
also occur on large electrodes.
As well as the limiting current density, the double-layer nature
of the interface at nanometer electrodes also affects the shape
of the voltammetric curves, e.g., the half wave potentials. We
will give a detailed discussion on this issue in a separate paper.

Figure 9. Theoretical voltammetric responses of an electrode of 1 nm 4. Conclusion

in radius for reactions with different redox formal potentials (shown
by the numbers inserted). The other conditions and parameters are the A dynamic diffuse double-layer model has been established
same as those for Figure 2. to describe electrochemical interfaces bearing a redox reaction.
The model consists of a dynamic potential distribution equation
where rw is the radius of water molecules and ri is the radius of containing the influence of current and a voltammetric equation
the ions that carry the charges to compensate the electrode containing the influence of the double-layer. It is shown that
charge. The thickness of the inner region of the compact double- this model can well predicate the voltammetric responses of
layer can be considered to equal the hard sphere diameter of electrodes of various sizes. The computed results according to
water, 0.29 nm.27,28 The outer part of compact double-layer can this model indicate that the steady-state polarization curves
be considered to consist of the thickness of the solvent shell corresponding to outer-sphere redox reactions have sigmoid
and half of the ion diameter. As shown in Figure 8, voltammetric shapes even at nanometer electrodes. However, the limiting
curves calculated for the reduction of an anion with a charge current and the half wave potential on the voltammetric curves
of -1 using eqs 16 and 18 with boundary condition based on change with electrode sizes, reactant charges, dielectric proper-
the assumption that the compact double-layer has a uniform ties of the double-layer, and the redox formal potentials. These
dielectric constant of 78 does show curved shapes. The low particular voltammetric behaviors seen on nanometer electrodes
dielectric constant inside the compact double-layer causes the disappear from the voltammetric responses at electrodes larger
potential to mainly drop in the compact double-layer. If one than 100 nm calculated with the same model. It is shown that
ignores it, the diffuse double-layer potential will be overesti- the anomalous voltammetric features of nanometer-sized elec-
mated; therefore, the double-layer effects will be overestimated. trodes are due to the electric-double-layer nature of the electrode/
3.5. Effect of the Redox Formal Potential. The double- electrolyte interfaces.
layer nature of the interface at an electrode of nanometer scale
would result in another particular behavior: the votammetric Acknowledgment. This work is financially supported by the
response would depend on the value of the redox formal National Natural Science Foundation of China (NSFC No.
potential. This is depicted by the polarization curves at electrodes 20443009), the Research Fund for the Doctoral Program of
of 1 nm for redox reactions having different formal potentials Higher Education, State Education Ministry of China (RFDP
shown in Figure 9. For the reduction of a negatively charged No. 20050486025), and the Scientific Research Foundation for
reactant, a positive shift of E0′ enhances the reaction rate, as the Returned Overseas Chinese Scholars, State Education
indicated by the increased limiting current density. Such a Ministry of China (ROCS No. 2005-383).
dependence of voltammetric responses on E0′ does not occur
Appendix A: Derivation of Equation 12
on large electrodes, as indicated by the voltmetric curves for
different E0′ at an electrode of 100 nm shown in Figure 10. Equation 11 can be transformed into the following form by
The influence of E0′ on the voltammetric responses of nano- setting cj* ) Q1Q2,
electrodes originates from the variation of the charge density
on the electrode surface with the departure of the electrode
potential from the PZC, which affects the distribution of the Q1 (
dQ2
dp
+ zjQ2
dφ
dp )
+ Q2
dQ1
dp
) -mji* (A1)
potential and concentration of ions in the double-layer. At
electrodes larger than 100 nm, the electrode reaction with k0 ) where Q1 and Q2 are functions of p. We can chose Q2 so that
Model for the Electrochemistry of Nanoelectrodes J. Phys. Chem. B, Vol. 110, No. 7, 2006 3269

dQ2 dφ diffuse double-layer. For such an enclosure, eq B1 becomes

+ zjQ2 ) 0 (A2)
dp dp
This requires
-4π(r0 + µ)2d0 (dφdr ) ohp
)q (B6)

Replacing q in eq B5 with eq B6 leads to

Q2 ) e-zjφ (A3)

( )( )
(r0 + µ)2d 1 1 dφ
Combining eqs A1, A2, and A3 leads to φohp ) E - - -
ic r0 r0 + µi dr ohp
dQ1
( )( )
e-zjφ ) -mji* (A4) (r0 + µ)2d 1 1 dφ
dp - (B7)
oc r0 + µi r0 + µ dr ohp
We then have
Since we have assumed that the PET coincides with the OHP,
Q1 ) Q1b - mji* ∫0pez φ dp
j
(A5) eq 21 is actually equivalent to eq B7 by replacing the subscript
ohp with PET.
where Q1b refers to the value of Q1 when p ) 0 (bulk solution).
Appendix C: Numerical Method for Solving Equations
Thus, cj* can be expressed as
16 and 18
cj* ) e-zjφ[Q1b - mji* ∫0pez φ dp]
j
(A6) Equation 16 is solved numerically in the p-domain using the
finite-difference method. Second-order, center-finite difference
Considering that φ ) 0 and cj* ) cjb* when p ) 0, we have formulas are employed to discretize eq 16 and the boundary
conditions. To do so, the entire p-domain [0, 1] was divided
cj* ) e-zjφ[cjb* - mji* ∫0pez φ dp]
j
(A7) into 2 subdomains: [0, (r0 + µ)/(r0 + µ + A/κ1)] and [(r0 +
µ)/(r0 + µ + A/κ1), 1], where A is a constant that is chosen to
make sure that the potential at p ) (r0 + µ)/(r0 + µ + A/κ1) is
Appendix B: Derivation of Equation 21
negligibly small compared to the electrode potential, e.g., 10-5
According to the Gaussian law, the total charge within any V. The term κ1 is the reciprocal of the Debye length calculated
enclosure equals the outward normal electric field strength with the total concentration of ions in the solution, that is

( )
integrated over the boundary surface of the enclosure multiplied
d0RT -1/2
by the local permittivity. Due to the fact that there is no charge
κ1 ) - (C1)
in the compact double-layer, the total charge within any 2F2c
enclosures that have radii of r0 e r e. r0 + µ equals the charge
amount at the electrode surface q. Thus, where
dφ c ) cOb + cRb + cAb + zcBb (C2)
-4π r20 )q (B1)
dr
The subdomain of [0, (r0 + µ)/(r0 + µ + A/κ1)] is equally
Since we have assumed that  is constant within a specific region divided by N1 nodes in the p-domain with node 1 corresponding
of the double-layer, eq B1 may be integrated in the inner region to p ) 0 (bulk solution) and node N1 located at p ) (r0 + µ)/(r0
of the compact double-layer as + µ + A/κ1). The subdomain of [(r0 + µ)/(r0 + µ + A/κ1), 1]
is unevenly divided by N2 nodes where the node N1 + m is
-
q
4π0ic
∫rr + µ r12 dr ) ∫Eφ
0
0 i ihp
dφ (B2) located at
r0 + µ
pN1+m )
( )
so, (C3)
A m

( )
r0 + µ + 1 -
q 1 1 κ1 N2
φihp ) E - - (B3)
4πic0 r0 r0 + µi
This will result in an even node spacing in the real r-domain.
Similarly, Such a dividing strategy is used for better computation accuracy.

( )
According to the definition of dimensionless distance p, we may
q 1 1 have
φohp ) φihp - - (B4)
4πoc0 r0 + µi r0 + µ
r0 + µ
where φihp and φohp refer to the potential at the inner Helmholz ∆p ≈ - 2 ∆r (C4)
r
plane (IHP) and OHP, respectively. Replacing φihp in eq B4
with eq B3 gives It can be seen that the ∆r value becomes larger at smaller r for

( )
a specific ∆p value. Thus, evenly dividing the p-domain will
q 1 1
φohp ) E - - - render the real spacing in the r-domain near the PET very large,
4πic0 r0 r0 + µi therefore resulting in large computation inaccuracy. To gain the
q 1
-
1
4πoc0 r0 + µi r0 + µ (
(B5) ) necessary accuracy, the node spacing ∆p has to be very small;
this greatly increases the computation time. If the inner
subdomain is divided according to eq C3, both the computation
We can imagine a spherical Gaussian enclosure that is located accuracy and computation time requirements can be well
just outside of the OHP but does not include any charge in the satisfied. The computing procedure is as follows: First, eq 16
3270 J. Phys. Chem. B, Vol. 110, No. 7, 2006 He et al.

is solved by setting i* ) 0. This will produce the equilibrium (9) Slevin, C. J.; Gray, N. J.; Macpherson, J. V.; Webb, M. A.; Unwin,
potential distribution at a specific electrode potential. The P. R. Electrochem. Commun. 1999, 1, 282-288.
(10) Shao, Y. H.; Mirkin, M. V.; Fish, G.; Kokotov, S.; Palanker, D.;
resulting potential profile can be used to obtain a current density Lewis, A. Anal. Chem. 1997, 69, 1627-1634.
and concentration profiles according to eqs 18 and 12, respec- (11) Penner, R. M.; Heben, M. J.; Longin, T. L.; Lewis, N. S. Science
tively. The thus-obtained current density and concentration 1990, 250, 1118.
profiles are then used to solve eq 16 to produce a potential (12) Macpherson, J. V.; Unwin, P. R. Anal. Chem. 2000, 72, 276.
(13) Bach, C. E.; Nichols, R. J.; Meyer, H.; Besenhard, J. Surf. Coat
profile, which is used to obtain new current density and Technol. 1994, 67, 139.
concentration profiles according to eqs 18 and 12. Then, a new (14) Bach, C. E.; Nichols, R. J.; Beckmann, W.; Meyer, H.; Schulte,
potential profile is obtained by solving eq 16. Such a computa- A.; Besenhard, J. O.; Jannkoudakis, P. D. J. Electrochem. Soc. 1993, 140,
tion cycle is repeated until the current density and potential 1281.
(15) Sun, P.; Zhang, Z. Q.; Guo, J. D.; Shao, Y. H. Anal. Chem. 2001,
profile show negligible differences from the last calculation 73, 5346-5351.
cycle. In the computation, we have set N1 ) 1000 and N2 ) (16) Oldham, K. B.; Bond, A. M. J. Electroanal. Chem. 2001, 508, 28-
5000. The computation is performed with home-written program 40.
codes in Matlab. The parameters used in the computation are: (17) Bard, A. J.; Faulker, L. R. Electrochemical Methods, 1st ed.; John
Wiley: New York, 1980.
R ) â ) 0.5, µi ) 2rw ) 0.29 nm, ic ) 6, oc ) 40, d ) 78,
(18) Oldham, K. B. J. Electroanal. Chem. 1988, 250, 1.
T ) 298 K, E0′ ) 0 unless stated, D ) 1 × 10-5 cm s-2. (19) Amotore, C.; Fosset, B.; Bartlet, J.; Deakin, M. R.; Wightman, R.
M. J. Electroanal. Chem. 1988, 256, 255.
References and Notes (20) Bonnefont, A.; Argoul, F.; Bazant, M. Z. J. Electroanal. Chem.
2001, 500, 52.
(1) Watkins, J. J.; White, H. S. Langmuir 2004, 20, 5474-5483. (21) Murphy, W. D.; Manzanares, J. A.; Mafé, S.; Reiss, H. J. Phys.
(2) Chen, S. L.; Kucernak, A. J. Phys. Chem. B 2002, 106, 9396- Chem. 1992, 96, 9983.
9404. (22) Smyrl, W. H.; Eewman, J. S. Trans. Faraday Soc. 1966, 62, 207.
(3) Chen, S. L.; Kucernak, A. Electrochem. Commun. 2002, 4, 80-
85. (23) Levich, B. Dokl. Akad. Nauk SSSR 1949, 67, 309.
(4) Conyers, J. L.; White, H. S. Anal. Chem. 2000, 72, 4441-4446. (24) Mirkin, M. V.; Bard, A. J. Anal. Chem. 1992, 64, 2293-2302.
(5) Morris, R. B.; Franta, D. J.; White, H. S. J. Phys. Chem. 1987, 91, (25) Fawcett, W. R., Double Layer Effects in the Electrode Kinetics of
3559-3564. Electron and Ion Transfer Reactions. In Electrocatalysis; Lipkowski, J.,
(6) Watkins, J. J.; Chen, J. Y.; White, H. S.; Abruna, H. D.; Ross, P. N., Eds.; Wiley-VCH: New York, 1998; p 322.
Maisonhaute, E.; Amatore, C. Anal. Chem. 2003, 75, 3962-3971. (26) Fawcett, W. R. Can. J. Chem. 1981, 59, 1844.
(7) Smith, C. P.; White, H. S. Anal. Chem. 1993, 65, 3343-3353. (27) Bockris, J. O. M.; Reddy, A. K. N. In Modern Electrochemistry;
(8) Norton, J. D.; White, H. S.; Feldberg, S. W. J. Phys. Chem 1990, Plenum: New York, 1970; Vol. 2.
94, 6772-6780. (28) Fawcett, W. R. J. Electroanal. Chem. 2001, 500, 264.