Double Layer

W Schmickler, University of Ulm, Ulm, Germany

& 2009 Elsevier B.V. All rights reserved.

Introduction absence of specific adsorption, the electrode surface is

covered by a layer of water molecules; the ions can enter
Electric double layers form whenever two conducting only into the second layer, whose center is called the
phases meet at an interface. Generally, one of the phases outer Helmholtz plane (Figure 1). Helmholtz assumed
acquires a positive excess charge on its surface, which is that all of the excess charge of the solution resides in this
balanced by a countercharge of the same magnitude and plane. This should be qualitatively correct at high ionic
opposite sign on the other phase. Well-known examples concentrations, where the conductivity of the solution is
are the interfaces between two metals with different work high, and its excess charge should hence be located at the
functions, of two semiconductors with n- and p-type surface. Thus, in this model, the interface acts like a
conductivities, and of a metal electrode in contact with parallel plate capacitor, and the effective plate separation
an electrolyte solution. Here, we consider only metal–
solution interfaces, because they are of greatest relevance
to fuel cells, supercapacitors, and batteries. In practical
applications, carbon is sometimes used as an electrode
material. The solution side of the double layer at carbon
electrodes can be described by the same models as for
Solvated
metals. However, because of the lower electronic density, anion
a space-charge layer exists on the surface of carbon. This
is similar to the space-charge layers of semiconductors,
but much shorter in extent, of the order of a fraction of a
nanometer. This space-charge region acts like a capacitor
Solvated
in series with the solution. Therefore, the double-layer cation
capacity of a flat carbon electrode is smaller than that of a
metal electrode in contact with the same solution.
Traditionally, electrochemical double-layer theory has
been concerned with so-called ideally polarizable inter-
faces, at which by definition no electrochemical reaction
takes place, and hence no direct current passes through the
interface. However, this condition can be relaxed: For
double-layer studies, it is sufficient that any electro-
chemical reaction does not disturb the distribution of the
excess charges at the interface, and this is the assumption
made throughout this article. In this case, a potential dif-
ference can be applied to the two sides of the interface,
which entails a change in the excess charges. Hence, the
interface resembles a capacitor, and, indeed, much research
has been focused on the capacity of the double layer.
Specific adsorption can have a large effect on the
charge distribution near the interface. Because each kind of
ion has its specific adsorption behavior, it is not possible to
make a general, nontrivial theory for the double layer in
the presence of adsorption. Hence, in the major part of this
article, assume that specific adsorption is absent, or has Metal Solution

only a negligible effect on the double-layer properties. Outer Helmholtz

The Helmholtz Model
The oldest model for the structure of the double layer at
metal electrodes was put forward by Helmholtz. In the
plane
Figure 1 Structure of the double layer in the Helmholtz model
without specific adsorption; the excess charge in the outer
Helmholtz plane is balanced by a countercharge on the metal,
which is not shown.

8
 Electrochemical Theory | Double Layer 9

is given by the distance of the outer Helmholtz plane to Inner Outer

Helmholtz Helmholtz
the metal surface, which is of the order of 0.4 nm. The layer layer
intervening water layer acts as a dielectric, because it
consists of a monolayer of water only – its dielectric
constant e will be appreciably less than that of bulk water.
Assuming an effective value of e ¼ 20–30, this gives
an interfacial capacity per unit area of the order of
0.4–0.5 C m  2, independent of the electrode potential.
For higher charge densities, dielectric saturation of the
water layer may be expected, because the field is so high
that the solvent dipoles will be largely oriented. So the Solvated
capacity should decrease at both sides of the potential of cation
zero charge (pzc). The resulting picture is, in fact, not a
bad average description of the interfacial capacity at high
ionic concentrations, as is evidenced by the capacity of
Ag(111) in contact with a 0.1 mol L  1 solution of po-
Solvated
tassium perchlorate (see upper curve in Figure 4). anion
In the presence of specific adsorption, this model must
be modified. The adsorbed ions are in contact with the
electrode; their centers form the so-called inner Helm-
holtz plane. Figure 2 depicts the case in which anions are Adsorbed
specifically adsorbed. The total charge at the interface is anion
still zero, so the charge on the adsorbed ions must be
balanced by the ions in the outer Helmholtz plane and by
the charge on the metal. Specific adsorption increases the
capacity, because the distance that separates the charges
of opposite sign is shorter.

Gouy–Chapman and Gouy–Chapman–

Stern Models
Metal Solution
Although the Helmholtz model gives a reasonable de-
Figure 2 Structure of the double layer in the Helmholtz model
scription at high ionic concentrations, it does not explain with adsorption; the charge on the metal is not shown.
the behavior at low concentrations. This was achieved by
the first statistical double-layer theory presented by
Gouy and Chapman at the beginning of the last century. where s is the charge per unit area of the electrode and f
It has much in common with the Debye–Hückel theory, the electrode potential. The Gouy–Chapman theory
which was developed several decades later. gives the following expression for capacity:
The Gouy–Chapman theory considers a planar metal
electrode in contact with an ionic solution, and starts ds ee0 ze0 ðf  fpzc Þ
CGC ¼ ¼ cosh ½2
from the following assumptions: (1) The metal is a per- df LD 2kT
fect conductor, and its excess charge is distributed evenly
on the surface. (2) The solvent is a dielectric continuum, where f is the electrode potential, fpzc the pzc, at which
characterized by a dielectric constant e. (3) The ions are the charge density s on the electrode vanishes, k is the
point particles, whose distribution is determined by the Boltzmann constant, and T the temperature. The Debye
Poisson–Boltzmann equation. For the case of an elec- length LD , which is familiar from the Debye–Hückel
trolyte containing cations and anions with the same theory for electrolyte solutions, is given by
charge number z, explicit expressions for the differential !1=2
capacity and for the profile of the electrostatic potential ee0 kT
can be derived. The quantity that is commonly measured LD ¼ ½3
2ðze0 Þ2 n0
is the differential capacity per unit area:
where n0 is the concentration of the ions in the bulk of
@s the solution. The Gouy–Chapman capacity has a pro-
C¼ ½1
@f nounced minimum at the pzc, and rises rapidly on both
10 Electrochemical Theory | Double Layer

300
150
10−1 M
10−2 M
10−3 M

C (µC cm−2)
200 100
C (µF cm−2)

100 50

0
0 −0.5 0.0 0.5
−0.1 0.0 0.1
( − pzc) V  (V vs SCE)

Figure 5 Capacity of a Au(210) electrode in contact with an

Figure 3 Interfacial capacity for an aqueous solution of a 1:1
aqueous solution of 0.1 mol L  1 KClO4 (blue squares) and of
electrolyte according to the Gouy–Chapman equation.
0.1 mol L  1 KClO4 þ 0.1 mmol L  1 KBr (red points).
Reproduced from Schmickler W. Interfacial Electrochemistry,
Reproduced from Schmickler W. Interfacial Electrochemistry,
copyright with author.
copyright with author.

60 Ag(111) charges, and for high concentrations, it significantly

x mM KClO4 overestimates the interfacial capacity. This is not sur-
55 5
10
prising, because under these circumstances, this theory
50
20 predicts an extension of the space-charge region that is of
45 50
C (µF cm−2)

100 the same order of magnitude as the diameters of the

40 solvent or of the ions. The simplest model that accounts
35 for the finite size of the solvent molecules is the intro-
30 duction of an ion-free layer, or Stern layer, as in the
25 Helmholtz model. The following resulting equation for
20
the capacity resembles two capacities in series:
15
−1.4 −1.2 −1.0 −0.8 −0.6 −0.4 −0.2 0.0 1 1 1
¼ þ ½4
ESCE (V) C CGC CH

Figure 4 Capacity of a Ag(111) electrode in contact with an

aqueous solution of KClO4 at various concentrations.
Reproduced from Schmickler W. Interfacial Electrochemistry,
copyright with author.
sides (Figure 3), and decreases strongly with the ionic
concentration of the solution.
For comparison, Figure 4 shows experimental values
for the interfacial capacity of a Ag(111) electrode in
contact with an aqueous solution of potassium per-
chlorate, an electrolyte that adsorbs only weakly. For
lower concentrations, the minimum at the pzc predicted
by the Gouy–Chapman theory is clearly visible; in fact,
for solid electrodes, the observation of this minimum is
practically the only way to determine the pzc. For a
concentration of 0.1 mol L  1, the minimum has dis-
appeared – a vestige can be seen as a shoulder. At po-
tentials far from the pzc, the capacity drops, whereas the
Gouy–Chapman theory predicts a continuous rise to
unreasonably large values.
Thus, for low concentrations and near the pzc, the
Gouy–Chapman theory works quite well. At high
where CH is known as the Helmholtz or inner-layer
capacity. If it is caused by the finite size of the ions, it
should be independent of the ionic concentration.
Equation [4] is best seen as a definition of the Helmholtz
capacity; it has little predictive value, but is a convenient
tool for describing the corrections to the Gouy–Chap-
man theory.
Specific adsorption has a large effect on the double-
layer capacity. As an example, Figure 5 shows the effect
that the addition of a small amount of Br  ions, which
adsorb quite strongly, has on the capacity of a Au(210)
electrode. With the exception of very high potentials, the
capacity is greatly enhanced. The magnitude of the en-
hancement and the shape of the capacity–potential
characteristics depend strongly on the type of ions, and
there are no general rules.
Often, it is difficult to decide whether or not a weak
specific adsorption occurs. For this purpose, a Parsons
and Zobel plot is a good method: According to eqn [4], a
plot of the inverse measured capacity 1/C versus the
inverse Gouy–Chapman capacity 1/CGC, performed at
constant charge density s, should result in a straight line
 Electrochemical Theory | Double Layer 11

with unit slope – the Gouy–Chapman capacity is varied
by changing the concentration of the electrolyte.
Figure 6 shows such a plot for a mercury electrode in
an aqueous solution of sodium phosphate. Actually, this
is one of the very few examples where a unit slope has
been obtained, indicating that there is no specific ad-
sorption. Generally, on solid metals, nearly all anions
show some specific adsorption near the pzc and at
higher potentials.
In any case, extrapolating the Parsons and Zobel plot
to the intercept corresponds to the limit of infinite ionic
concentrations and yields the Helmholtz capacity.
Figure 7 shows the corresponding curves for mercury
and Ag(111). Notice the large difference between the two
curves, which indicates that the metal itself has a con-
siderable influence on the capacity.
80
 = 0 / µC cm−2
 = 4 / µC cm−2
 = −4 / µC cm−2
1/C (cm2 µF−1)
Hard Sphere Electrolyte and Jellium
The Helmholtz capacity denotes the deviations from the
Gouy–Chapman model at high concentrations and
charges. It seems that there are two effects: The finite size
of the solvent and the ions, which explains why the real
capacity is smaller than the Gouy–Chapman theory
predicts, and a contribution of the metal that would ex-
plain why the capacity on Ag(111) is so much higher than
on mercury.
A simple model to describe finite size effects is to
regard the solution as an ensemble of hard spheres. There
are three types of hard spheres: Two types of ions with
appropriate charges at the center, and spheres with a
dipole moment at the center, which model the solvent.
The statistical mechanics even of this simple model is
quite complicated, and it has been solved only approxi-
mately – in the so-called mean spherical approximation –
and only for small excess charges. For this case, it gives an
explicit formula for the Helmholtz capacity as the first-
order correction to the Gouy–Chapman theory:
 
1 1 e1

60 ¼ si þ ss ½5
CH 2ee0 l

where si is the diameter of the ions, assumed to be the

same for cations and anions, and ss is that of the solvent
40 molecules. The parameter l describes the dielectric
properties at the interface and is obtained from the di-
electric constant using the formula l2(1 þ l)4 ¼ e. The
0 10 20 30 40 contribution of the solvent is significantly greater than
1/CGC (cm2 µF−1) that of the ions. This explains an experimental finding:
Figure 6 Parsons and Zobel plot for a mercury electrode in an The values of the Helmholtz capacity depend only very
aqueous solution of NaH2PO4. The line indicates unit slope. little on the nature of the ions, as long as they are not
Reproduced from Schmickler W. Interfacial Electrochemistry, specifically adsorbed.
copyright with author. For aqueous solutions, eqn [5] predicts Helmholtz
capacities of the order of 10–20 mF cm  2 at the pzc. This
is at least of the correct order of magnitude. The next
80 order correction to the Gouy–Chapman theory can also
70
be calculated explicitly although the formula is not so
Ag(111) simple. It has been successfully employed to explain the
60 small but systematic deviations from the straight line,
which can be seen in Figure 6 for small values of 1/CGC.
CH (µFcm−2)

50
Experimental values for the Helmholtz capacity lie in
40
the range of 30–80 mF cm  2 and, as discussed above,
30 depend on the nature of the metal, in the case of single
20
Hg crystals even on the orientation of the surface. This fact
has been explained through the response of the metal
10 electrons to the electric field in the double layer. At the
0 surface, the electronic density decays over a distance of
−12.0 −8.0 −4.0 0.0 4.0 8.0 12.0 about 0.1 nm. The response of this electronic tail to the
 (µC cm−2) double-layer field can be viewed as an electronic polar-
Figure 7 Helmholtz capacities for mercury and Ag(111) in
izability of the surface, which increases the capacity.
aqueous solution of a nonadsorbing electrolyte. Reproduced from Explicit calculations for this effect have mostly been
Schmickler W. Interfacial Electrochemistry, copyright with author. based on the jellium model, in which the electrons are
12 Electrochemical Theory | Double Layer
4.00
Hg
3.00 Tl
1/CH (m2 F −1)
Cd
2.00
In
Zn
1.00
Ga
0.00
8.0 12.0 16.0 20.0 24.0 28.0
Electronic density × 1000 (au)
Figure 8 The inverse Helmholtz capacity of sp metals at the
potential at zero charge (pzc) as a function of the electronic
density. The latter is given in atomic units (au); where 1 au of
density ¼ 6.76  1024 cm  3. The dashed line is the prediction of
jellium in contact with an ensemble of hard spheres. Reproduced
from Schmickler W. Interfacial Electrochemistry, copyright with
author.
described as an electronic plasma, and the metal ions are
represented by a constant positive background charge.
This model predicts that the interfacial capacity should
increase with the electronic density of the metal, a trend
that the capacities of the sp metals follow at the pzc
(Figure 8). Metals with d bands near the Fermi level are
not well described by jellium, and do not follow this
trend.
There have been various attempts to extend the hard
sphere model to higher charges, but without a major
success.
Computer Simulations
The statistical mechanics of more realistic models than
hard sphere is intractable by analytical methods; it has
been therefore treated by computer simulations, mostly
by molecular dynamics. Such simulations require inter-
action potentials between all the particles concerned,
solvent, ions, and electrode. Usually, these potentials are
assumed to have a plausible form with parameters that
are determined by ab initio calculations. For pure water
in contact with a metal, there have been a few simu-
lations based entirely on density functional theory.
Practically all simulations for pure water in contact
with a metal agree that water is structured at the elec-
trode surface. The distribution functions for the water
molecules show a distinct peak at the surface, and a
weaker secondary peak (Figure 9); obviously, this is a
packing effect. At uncharged electrodes, the dipole mo-
ment of water is mostly parallel to the surface, but with a
small preference for the oxygen end to be closer to the
metal. The magnitude of the latter effect depends on the
8 1.00
0.75
6
0.50
 (V)
norm
4
0.25
2
0.00
0 −0.25
0 3 6 9 12
x (Å)
32.0
Figure 9 Particle densities r, for hydrogen (solid line), oxygen
(dashed line), and the electrostatic potential f (dotted line), for
water in contact with a silver electrode. The densities for oxygen
and hydrogen have been normalized to unity and two,
respectively, in the bulk. Reproduced from Frank S, Hartnig C,
GroX A, and Schmickler W. ChemPhysChem 2008, available
online; copyright with Wiley-VCH.
metal, and on the details of the interaction potentials that
are used. It gives rise to a dipole potential drop at the
interface, which is typically of the order of several hun-
dred millivolts (see Figure 9).
The detailed structure of water at the surface is often
discussed in terms of the bilayer model proposed by
Doering and Madey to explain water adsorption on metal
surfaces in the vacuum at low temperatures. As the name
suggests, this model proposes two layers of water that are
hydrogen-bonded to each other. Obviously, in the liquid
state and at ambient temperatures, such a bilayer must be
strongly disturbed, although a vestige can be seen in
some simulations.
Simulations with ions have mostly shown unsurprising
results: Small cations such as Li þ and Na þ are not
adsorbed, whereas anions such as Cl  are adsorbed at
the pzc and at positive potentials. So far, these simu-
lations have not shed any light on the magnitude of the
interfacial capacity.
Nonaqueous Solutions and Molten Salts
The double-layer capacities of nonaqueous solutions
follow the same principles as those of aqueous solutions.
Because the dielectric constants are usually lower than
those of water, and the size of the solvent molecules
larger, capacities are generally lower than in aqueous
solutions. Unfortunately, there are few systematic stud-
ies, and those that do exist are mostly from the last
century.
The capacities of interfaces between metal electrodes
and molten salts are of the same order of magnitude as
those for aqueous solutions. Capacity curves for simple
molten salts such as the alkali halides usually have a

80
60
C (µF cm−2)
40
20 KI
KCl
0 e
−0.8 −0.6 −0.4
 (V)
Figure 10 Interfacial capacity of a Pb electrode in contact with
molten alkali halides at 800 1C.
U-type shape, with the minimum near the pzc; an ex-
ample is shown in Figure 10. Various models have been
proposed for the interfacial structure, but no agreement
has been reached yet. Computer simulations for alkali
halide melts are difficult because of the strong Coulomb
forces, which make it difficult to reach the equilibrium
states.
At present, ionic liquids, which have much larger ions
than the alkali halides, and which are liquid at ambient
temperatures, are a very active area of research, and they
are also amenable to computer simulations.
Conclusions
Qualitatively, the double layer at the interface between a
metal and an aqueous solution is quite well understood,
but a quantitative theory is lacking. Thus, after decades of
research, it is still not possible to calculate the double-
layer capacity of such a basic system like a Ag(111)
electrode in contact with a 0.1 mol L  1 aqueous solution
of potassium fluoride. Further progress in understanding
can be expected from a combination of computer simu-
lations and ab initio calculations. For molten salts, the
situation is worse, because there is not even a basic
theory on the Gouy–Chapman level. Nevertheless, ex-
perimentally all electrochemical double layers are well
investigated.
Because at the interface between a metal and a con-
centrated electrolyte solution the effective range of the
charge distribution is exceedingly small, of the order of a
few tenths of a nanometer, the capacity is very high. This
makes such interfaces interesting for technological ap-
plications as so-called supercapacitors: A high charge
density can be stored at these interfaces and discharged
during a very short time. They can therefore react faster
Electrochemical Theory | Double Layer 13
to a change in power demand than either batteries or fuel
cells, and can provide short boosts of power when needed.
Nomenclature
Symbols and Units
C capacity
CGC Gouy–Chapman capacity
CH Helmholtz or inner-layer capacity
Euler’s constant
e0 unit of charge
ESCE potential with respect to saturated
calomel electrode
k Boltzmann constant
LD Debye length
n0 concentration of the ions in the bulk
T temperature
qnorm normalized particle density
z charge number
e dielectric constant
e0 permittivity of free space
k dielectric properties at the interface
r charge per unit area
ri diameter of ions
rs diameter of solvent molecules
/ electrode potential
/pzc potential of zero charge
See also: Batteries: Capacity; Capacitors:
Electrochemical Capacitors: Ionic Liquid Electrolytes
Electrochemical Double-Layer Capacitors; Overview;
Electrodes: Porous Electrodes. History: Electrochemical
Capacitors; Measurement Methods: Electrochemical:
Impedance Spectroscopy.
Further Reading
Badiali JP and Amokrane S (1991) Capitance of an ideally polarizable
electrode: Role of the metal, temperature and solvent effects. In: Tosi
MP and Kornyshev AA (eds.) Condensed Matter Aspects of
Electrochemistry, pp. 157--175. Singapore: World Scientific.
Bockris JO’M, Conway BE, and Yeager E (1980) Comprehensive
Treatise of Electrochemistry, vol. 1, New York: Plenum Press.
Grahame DC (1947) The electrical double layer and the theory of
electrocapillarity. Chemical Reviews 41: 441.
Parsons R (1990) The electrical double layer: Recent experimental and
theoretical development. Chemical Reviews 90: 813.
Schmickler W and Henderson DJ (1986) New models for the
electrochemical interface. Progress in Surface Science 22: 323–420.
Spohr E (1999) Computer simulations of electrochemical interfaces.
In: Alkire RC and Kolb D (eds.) Advances in Electrochemical Science
and Engineering, vol. 6, pp. 1--75. Weinheim: Wiley.