Sie sind auf Seite 1von 9

Applied Surface Science 449 (2018) 445–453

Contents lists available at ScienceDirect

Applied Surface Science


journal homepage: www.elsevier.com/locate/apsusc

Full length article

Influence of designed electrode surfaces on double layer capacitance


in aqueous electrolyte: Insights from standard models
Md. Shihab Uddin a , Himadri Tanaya Das a , Thandavarayan Maiyalagan b ,
Perumal Elumalai a,∗
a
Electrochemical Energy and Sensors Lab, Department of Green Energy Technology, Madanjeet School of Green Energy Technologies, Pondicherry
University, Puducherry 605014, India
b
SRM Research Institute & Department of Chemistry, SRM University, Kattankulathur, Chennai 603203, India

a r t i c l e i n f o a b s t r a c t

Article history: Standard models, Helmholtz, Gouy-Chapman, Stern and Grahame models have been validated on the
Received 26 October 2017 electrical double layers (EDLs) formed from three different electrode surfaces, i.e., plain, semi-sphere
Received in revised form and cylindrical pore in aqueous electrolyte. The double layer potential (␾) and areal capacitance (Cdl )
24 November 2017
were estimated in the three models using MATLAB Simulink. It was found that the areal capacitance
Accepted 11 December 2017
Available online 13 December 2017
of the double layer strongly influenced by the nature/ type of the electrode surface in contact with the
electrolyte. The cylindrical-type pore surface exhibited the large areal capacitance. The charge-discharge
and impedance characteristics of the EDLs were also analyzed.
Keywords:
Supercapacitor © 2017 Elsevier B.V. All rights reserved.
Electrical double layer
Helmholtz
Gouy-Chapman
Stern
Grahame models
Electrochemical Impedance

1. Introduction ized. Electrical double layer capacitor (EDLC) is based on mere


charge separation at electrode/electrolyte interface. The electrodes
Batteries and supercapacitors are typical electrochemical for the EDLCs are typically made of high surface area carbon-based
energy storage devices. Among them, the batteries have the capa- materials because of their porous nature. Aqueous electrolytes are
bility to deliver energy over a long time at a relatively lower rate generally employed as an electrolyte in EDLC, where water dipole
of current (low power). The batteries fail to deliver high power acts as a dielectric material to provide high relative permittiv-
because, at high power discharge, the lifetime of the battery reduces ity. The presence of water dipole in the electrolyte significantly
drastically due to degradation of electrode material and diffusion reduces the distance between the two electrodes, to the order of
limitation of ions involved in the electrochemical reactions. Thus, angstrom (Å). Because of aforesaid aspects, the EDLC having high
current trends in the electrochemical energy storage technology capacity is different from the conventional parallel plate capaci-
focus on the development of storage device that can deliver high tor whose capacity is directly proportional to relative permittivity
power as well as high energy. In this regard, supercapacitors seem (␧) and surface area (A), but inversely proportional to the distance
attractive storage device that can deliver high power for the large between the two electrodes [1]. Modeling and simulations provide
cycle of as high as 105 cycles. In fact, the supercapacitors can bridge insights about designing novel materials for developing efficient
the power gap between the batteries and conventional capacitors, device by reducing manufacturing cost as well as technical losses.
especially for hybrid electric vehicles and electrical tools. Unfortu- Recently, Bang et al., has studied optical pore development of car-
nately, the supercapacitors have severe issues such as low energy bon models [2]. Similarly, Lu et al., has reported the porous carbon
density, high cost and high self-discharge. By addressing these as high ion transport channel for electrode materials in superca-
issues, the market demand for the supercapacitors can be real- pacitor [3]. Thus, to develop highly performing EDLC electrode for
energy storage device, a proper understanding of electrical double
layer (EDL) formed by different electrode surface in an electrolyte
∗ Corresponding Author. is indispensable.
E-mail addresses: drperumalelumalai@gmail.com, Wang et al., presented scaling law for the estimation of the
elumalai.get@pondiuni.edu.in (P. Elumalai). integral capacitance of EDLC supported by rigorous analysis and

https://doi.org/10.1016/j.apsusc.2017.12.088
0169-4332/© 2017 Elsevier B.V. All rights reserved.
446 Md.S. Uddin et al. / Applied Surface Science 449 (2018) 445–453

trostatic interaction between the electrode surfaces and counter


ions of the solution [13,14]. A uniform dipole distribution is present
at the interface where potential varies linearly from the electrode
surface to the bulk solution as schematically shown in Fig. 2(a).
However, Helmholtz model did not account for the influence of
thermal motion and ion absorption on the resultant double layer
capacitance (Cdl ). Gouy-Chapman described the EDL by considering
the thermal motion of ions near the charged surface and is driven
by diffusion and electrostatic forces, in particular at low concentra-
tion of electrolyte [15,16]. However, this model did not account for
the influence of ion absorption on the electrode/ electrolyte inter-
face. According to the Gouy-Chapman model, the double layer is
not rigid at the electrode/ electrolyte interface; rather there is ion
concentration gradient from electrode surface to the bulk solution
due to thermal motion of ions as schematically shown as in Fig. 2(b).
The Stern model recommends a combination of the concepts
of the Helmholtz and Gouy-Chapman models [17]. According to
the Stern model, the electrical potential varies linearly with the
distance from electrode surface up to Gouy-Chapman layer (rigid
layer) and after that decreases exponentially (i.e., diffused layer)
as schematically shown in Fig. 2(c). So, as per the Stern model, the
Fig. 1. Schematics of cross section of (a) Planar Surface, (b) Cylindrical pore surface, capacitance is regarded as two series connected capacitors as given
(c) Semi-sphere pore surface and (d) Front view of pore surfaces. in (Eq. 1):

1 1 1
experimental data on porous carbon-based electrodes [4]. Gav- = + (1)
C dl CH C GC
ish et al., examined the influence of relative permittivity on the
EDLC properties by changing electrolyte concentration [5]. It was where, Cdl , CH , and CGC are double layer, Helmholtz and Gouy-
reported that higher concentration having less number of dipoles Chapman capacitances, respectively [14].
in the solution resulted in low capacitance, whereas lower concen- The currently accepted Grahame model represents the real sit-
tration reduced the capacitance because ions extend to a longer uation of an EDL. Grahame introduced concepts of inner Helmholtz
distance from electrode surface due to thermal motion [6–9]. plane (IHP) where ion absorption occurs on the electrode surface
Ohshima et al., described a simple algorithm for the potential and outer Helmholtz plane (OHP). As shown in the Fig. 2(d), the
distribution in EDLC cylindrical pore surface [10]. He considered anions are tightly attached to the electrode surface and partially
Poisson-Boltzman equation to describe potential distribution in a desolated. This fact causes a reversal of electrostatic potential in
charged cylindrical pore filled with an aqueous electrolyte solu- the region between the IHP and OHP. The capacitance as per the
tion. Martin et al., estimated capacitance of EDLC based on carbon Grahame model is shown below [14]:
nanotube (CNT) using classical theory [11]. The curvature effect
on Gouy-Chapman limit has been established. Shuai Bana et al., 1 1 1 1 dqa
= +( + )(1 − ) (2)
worked on the electrochemical equivalent circuit (Randle circuit) Cdl CIHP COHP CGC dqM
to analyze the charging-discharging properties of the EDLC [12].
To the best of our knowledge, the influence of different types where qa and qM are absorption charge and metal surface charge,
of electrode surfaces on the capacitance of EDLC has not been well respectively.
examined. Thus, in this work, we re-visited the EDLC characteris-
tics of electrode surfaces having different types/ shapes of pores
(plain, semi-sphere and cylindrical) as shown in Fig. 1 by applying 3. Analysis
Helmholtz, Gouy-Chapman, Stern and Grahame models in aqueous
electrolyte. Alongside, we have also studied electrochemical charg- In this work, we have modeled EDLC for three different elec-
ing/discharging and impedance features of the aforesaid model trode surfaces namely plain layer, semi-sphere and cylindrical in
surfaces in aqueous electrolyte. the aqueous electrolyte. The schematic of three different surfaces
and the cross-section of the semi-sphere/ cylindrical surface are
shown in Fig. 1. The areal capacitances of the EDL resulted on
2. Various models for electrical double layer the three electrode surfaces were estimated based on Helmholtz,
Gouy-Chapman, Stern and Grahame models. We have assumed the
A metal in contact with an aqueous solution generates solvated surface potential of 0.5 V and uniform dipole distribution. To calcu-
charged ions on its surface, where positive and negative ions are late the potential change, in the parallel plane layer, linear Gauss
equally distributed throughout the surface. The free electrons will law equation has been applied and similarly in the bulk solution,
move away from or towards the interface depending on the direc- Poisson-Boltzmann equation has been applied. In addition, charg-
tion of the electric (E) field. Thus, a net charge is induced on the ing/ discharging and impedance characteristics of the EDLC have
metal, which will be equal in magnitude and opposite in sign to that been also examined.
on the solution side of the phase boundary. Therefore, charge sepa-
ration occurs in the electrode surface/ solution interface known as
an interfacial potential difference (␾). The properties of such elec- 3.1. Helmholtz capacitance of model surfaces
trified interface have been explained by Helmholtz (compact layer),
Gouy-Chapman (diffuse layer), Stern and Grahame models. According to Helmholtz, the electric double layer is the align-
As per Helmholtz model, two compact layers of charges are ment of charge on the electrode and electrolyte interface similar to
formed at the electrode/ electrolyte interface as a result of elec- parallel plate capacitor. Gauss law states that the surface integral of
Md.S. Uddin et al. / Applied Surface Science 449 (2018) 445–453 447

Fig. 2. Schematics of (a) Helmholtz, (b) Gouy-Chapman, (c) Stern and (d) Grahame model for the EDL.

the electric field in any closed surface is equal to the enclosed charge where ␳ is ionic charge density and ␸(x) is the electric potential
with the multiplication of inverse permittivity as shown below: at distance x [19]. The Boltzmann distribution gives the position of
 a particle in a system over the various possible states. In the bulk
Qenclose
E · dA = = Q · ␧0 −1 solution, the Boltzmann equation for ions can be written as Eq. (7):
␧0 zi qϕ(x) zi qϕ(x)
ni = n∞
i
e(− kB T
)
= C∞ NA e(− kB T
)
(∵ n∞
i
= C ∞ NA ) (7)
where E is the electric field, Qenclose is total enclose charge and
␧0 is free space permittivity, respectively [18]. So, the potential where zi , q, T, kB , C∞ and NA are the valence of ionic species i, unit
difference (␾) between the two rigid layers at distance x is charge (q = 1.6 × 10−19 C), temperature (T = 20 ◦ C) and Boltzmann
constant (kB = 1.63 × 10-23 m2 kg K-1 s-2 ), molar concentration of
Q
ϕ (x) = ϕ0 − xH (4) solution and Avogadro number, respectively [18]. So, the resultant
Aεr ε0 Poisson – Boltzmann equation is,
where ␸0 is the electric potential at xH = 0, A is area, xH is the z i qn∞

d 2 ϕ(x) z qϕ(x)
− i
thickness of Helmholtz layer and ␧r is the relative permittivity of =− i
e kB T (8)
dx 2 ε0 εr
the electrolyte. The first derivative of the potential with respect i
∂ϕ
to charge ( ) will give inverse capacitance (1/CH ). Therefore, the
∂Q For binary solute of unit charge (z1 = z2 = z) and same con-
Helmholtz model capacitance per unit area for the EDLC formed
centration of positive and negative ions (C∞- = C∞+ = C∞ ), the
from three different electrode surfaces derived from Eq. (4) are:
Poisson-Boltzmann Eq. (8) can be rewritten as
⎧εε


r 0
for plain surface (5a) d 2 ϕ(x)
zqϕ(x)

⎪ xH   =−
zqC ∞ N A −
zqϕ(x)
−e

⎨ εr ε0 dx2 ε0 εr
(e kB T kB T ) (9)
r2
CH = 1+ for semi-spherical pore surface (5b)
xH A Or




zqϕ(x)


⎩ εr ε0 1 + 2Lr for cylindrical pore surface (5c) d 2 ϕ(x) 2zqC ∞ N A
xH A =− sinh (10)
dx2 ε0 εr kB T
where r is pore radius and L is the length of the cylindrical pore. Multiplying 2
dϕ(x)
on the both sides of Eq. (10) yields Eq. (11).
dx

3.2. Gouy-chapman capacitance of model surfaces dϕ(x) d 2 ϕ(x) 2qzC ∞ N A dϕ(x)


zqϕ

(x)
2 =− 2 sinh (11)
dx dx2 ε0 εr dx kB T
The electric potential due to charge distribution in any medium
is given by the Poisson equation shown below: The Eq. (11) can be rearranged as
dϕ(x) 2
d 2 ϕ(x)  d( dx
) 2qzC ∞ N A zqϕ(x)
=− (6) dx = − sinh( )dϕ(x) (12)
dx2 ε0 εr dx ε0 εr kB T
448 Md.S. Uddin et al. / Applied Surface Science 449 (2018) 445–453

Fig. 3. (a) Charging equivalent circuit of EDLC, (b) Discharging equivalent circuit of EDLC, (c) Electrochemical equivalent (Randle) circuit, (d) Impedance diagram of Randle
circuit excluding Warburg coefficient and (e) Impedance diagram of Randle circuit including Warburg Coefficient.

Integrating Eq. (12) and applying the boundary conditions (13a, ∂ϕ


The surface charge density ( ) is the change of potential with
∂x
13b) respect to distance in any arbitrary medium as shown below:
At x = 0, ϕ (x) = ϕ (0) (13a) dϕ(x)
(x) = εo εr (17)
dϕ(x) dx
At x = ∞, ϕ (x) → 0, =0 (13b)
dx Thus, the capacitance is given by:
the solution for Eq. (12) is, d
 zqϕ(x)

CGC =

(18)
dϕ(x) 8C ∞ N A kB T
= sinh (14)
dx ε0 εr 2kB T So, the Gouy-Chapman capacitance for the electric double layer
 is,
dϕ 8C ∞ N A kB T 

= dx (15) 2z 2 q2 C ∞ N A ε0 εr zqϕ(x)
zqϕ ε0 εr C GC = cosh( ) (19)
sinh 2kB T kB T 2kB T
The solution for Eq. (15) is It is noted that the Gouy-Chapman capacitance is independent
⎧  of the surface area of the electrode; the change in capacitance for

⎪ 2C ∞ NA

⎪ (qzx ) three surfaces are almost negligible.

⎨ ϕ(0)e ε0 εr kB T
; where
zqϕ(x) zqϕ(0)
, ( →= )0 (16a)
ϕ(x) =  4kB T 4kB T 3.3. Stern model capacitance for model surfaces

⎪ 2C ∞ NA

⎪ 4kB T (qzx )

⎩ −1
tanh (e
ε0 εr kB T
tanh(
zqϕ(0)
)); Otherwise (16b) The Stern model is a combination of both Helmholtz and Gouy-
zq 4kB T Chapman models. So, the Stern model capacitances for the EDLC
formed in the three different surfaces are:




⎪ 1 1

⎪ εr ε0 +  2 2 for planar surface (20a)



⎪ 2z q C∞ A 0 εr
N ε zqϕ(x)


xH cosh( )


kB T 2kB T
1 1 1
= + for Semi-spherical pore surface (20b)
Cdl ⎪ εr ε0 r 2 2z 2 q2 C∞ NA ε0 εr

⎪ (1 + ) zqϕ(x)

⎪ xH A cosh( )


kB T 2kB T

⎪ 1 1

⎪ + for cylindrical pore surface (20c)

⎪ εr ε0 2Lr
2z 2 q2 C∞ NA ε0 εr
⎩ x (1 + A ) cosh(
zqϕ(x)
)
H kB T 2kB T

3.4. Grahame model capacitance for model Surfaces

Grahame considered the effect of adsorption of the ions at


the electrode surface. He divided the electric double layer into
three-parts; the first layer is electrode surface to IHP, second layer
Md.S. Uddin et al. / Applied Surface Science 449 (2018) 445–453 449

is IHP to outer Helmholtz OHP and third layer is OHP to bulk multiplying 1/Cdl Rs in both sides gives Eq. 26,
solution (Fig. 2(d)). The electric field (E) due to Coloumbic attrac-
dV c Rp + Rs Vs
tion/repulsion of unit charge on the dipole separated by a distance +( )V c = (26)
dt C dl R p R s C dl R s
x [20] is given below:
q The Eq. (26) is the first order partial differential equation and
E= (21) boundary conditions are:
4ε0 εr x2
The potential at three different layers, i.e., surface to inner att = 0, Vc = 0 (27a)
Helmholtz plane is:
att = ∞, Vc = Vs (27b)
q
ϕ (x) = ϕ0 − (22) So, the solution for the first order differential Eq. (26) is
4ε0 εr xIHP
R p +R s

where 4ε εqr x define the potential drop due to ion adsorption. V c (t) = V s [1 − e− C dl R p R s
t
] (28)
0 IHP
On the other hand, the potential at the inner Helmholtz plane Thus, the Eq. (28) gives the charging voltage for the EDLC.
to the outer Helmholtz plane is same as Helmholtz potential (Eq.
4). The potential variation from Outer Helmholtz plane to the bulk
3.5.2. Discharging
solution is similar to Gouy-Chapman potential (Eq. 6).
∂ϕ The Kirchhoff current law at node ‘a’ of Fig. 3(b) [22] is,
The first derivative ( ) of Eq. (22) give the inverse capacitance
∂Q
which is similar to that of Helmholtz model. As a result, the capac- dV c Vc Vc
C dl + + =0 (29)
itance variation as per the Grahame model is almost similar that dt Rp Rs + RL
of the Stern model, when the ion absorption at electrode surface is
Re-arranging the Eq. (29)
negligible. In the case of high ion absorption, the electrode would  
lose the EDLC property and behave as an ideal Pseudocapacitor. dV c Rp + Rs + RL
So, the areal capacitance for the three electrode surfaces accord- + Vc = 0 ((30))
dt C dl R p (R s + R L )
ing to Eq. (5) is written as:
⎧ ⎛ ⎞



⎪ ⎜ 1 ⎟


⎪ 1 1 ⎟ 1 − dqa

⎪ +⎜ ε ε +  (23a)
⎪ ε ε ⎝
⎠ dqM


r 0 r 0
2z 2 q2 C∞ NA ε0 εr zqϕ(x)

⎪ xIHP xOHP cosh

⎪ ⎛ kB T 2kB T ⎞




⎨ 1 ⎜ 1 1 ⎟

1
=   +⎜   + ⎟ 1 − dqa (23b)
Cdl ⎪ r 2 ⎝ εr ε0 r 2 zqϕ
⎠ dqM


εr ε0
1+ 1+ 2z 2 q2 C∞ NA ε0 εr (x)

⎪ x A x A cosh


IHP
⎛ OHP kB T 2kB T ⎞





⎪ ⎜ ⎟



1

+ ⎜
1

+ 
1 ⎟ 1 − dqa

⎪ εr ε0 2Lr ⎝ εr ε0 2Lr zqϕ(x)
⎠ dqM
(23c)

⎪ 2z 2 q2 C∞ NA ε0 εr
⎩ xIHP 1 + A xOHP
1+
A cosh
kB T 2kB T

The Eqs. (23a), (23b) and (23c) corresponds to planar surface,


semi-spherical pore surface and cylindrical pore surface, respec-
tively.
Initial conditions:

3.5. Charge/ discharge characteristics of electric double layer att = 0, Vc = Vs (31a)

att = ∞, Vc = 0 (31b)
3.5.1. Charging
For electrostatic capacitors, the current and voltage are related So, the solution
as given below:
for the Eq.

(30) is:
R p +R s +R L
V dc (t) = V s e− C dl (R s +R L )R p
t
(32)
dV c
iC = C dl (24)
dt Thus, Eq. (32) gives the discharging voltage of EDLC.
where Cdl is the double layer capacitance, Vc is the potential across
the capacitor and iC is the capacitor current [21]. 3.6. Electrochemical impedance characteristics of EDLC
The Kirchhoff Voltage law for the closed loop as shown in
Fig. 3(a) can be written as Eq. (25a) [22,23]. The Randle circuit for the EDLC can be best described as shown in
  Fig. 3(c–e), where Rs is solution resistance, Cdl is double layer capac-
Vc dV c itance and W is Warburg element. The Randles circuit consists of
Vc + + C dl Rs = V s (25a)
Rp dt an active electrolyte resistance (RS ) in series with the parallel com-
bination of the double-layer capacitance (Cdl ) and an impedance of
where Vs is the supply voltage. Re-arranging this Eq. (25a), we get Faradic reaction in series with Warburg element [24–26]. Solution
Eq. (25b) resistance is often a significant factor in the impedance of an elec-
  trochemical cell. The resistance of an ionic solution is a function of
Vc Rs dVc
Vc + + Cdl Rs = Vs (25b) ionic concentration, type of ions, temperature, and the geometry of
Rp dt
the area of the electrochemical cell.
450 Md.S. Uddin et al. / Applied Surface Science 449 (2018) 445–453

3.6.1. Randle circuit excluding Warburg element The R and F are gas constant (R = 8.13 J K−1 mole−1 ) and Faradic
If re-drawn the Randle circuit with respect to impedance where, constant (F = 96500 C), respectively, whereas Do , and Dr are oxidant
1
Z1 = Rs , Z2 = jωC , Z3 = Rp and ␻ is the angular velocity, the equiv- and reductant diffusion coefficients.
dl
alent impedance of the circuit shown in Fig. 3(d) is
Z 2 *Z 3 4. Results and discussions
Z eq = Z 1 + = Z re + jZ im (33)
Z2 + Z3
The variation of double layer potential and areal capacitance for
where real impedance three different electrode surfaces were estimated using MATLAB
Rp R2009a software on aforementioned models. The surface of double
Zre = Rs + 2 layer depends on the concentration of medium in which the double
1 + ωCdl Rp layer forms. Higher the concentration longer the length of the dou-
ble layer formed and vice-versa. For the bulk solution, the extent
and imaginary impedance,
of diffusion length can be calculated by using Debye length (xD ),
ωCdl Rp2 shown in Eq. (37).
Zim = 2 
1 + ωCdl Rp ␧r ␧0 kB T
xD = (37)
2q2 NA C∞
3.6.2. Randle circuit including Warburg element where ␧r, ␧0 B, T, q, NA , C∞ are relative permitivity, Boltzmann con-
If re-drawn the Randle circuit with respect to impedance where, stant, temperature, charge, avogadro number and concentration,
1
Z1 = Rs , Z2 = jωC , Z3 = Rp , Z4 = √
Aw
= √ Aw ; so the equivalent respectively. So, in all the cases the distance from the electrode
dl jω ej90 ω
impedance of the circuit (Fig. 3(e)) is given by: to the solution was kept constant at 20 nm. The capacitance was
estimated using 1 M aqueous electrolyte of relative permittivity ␧r
Z 2 *Z 5 = 78.5, temperature T = 20 ◦ C, valence z = 1 and electrode surface
Z eq = Z 1 + = Z re + jZ im (34)
Z2 + Z5 potential ␸(0) = 0.5 V. In case of concentrated electrolyte, more
1 Aw number of ions will be present in the solution. These ions require
Z re = R s + (R p + √ ) (35) high potential to accumulate/ integrate on the electrode-electrolyte
K ω2 C 2dl 2ω
interface. Thus, for all models, standard potential difference of 0.5 V

2 1

1 Aw Aw Aw was assumed for comparison.
Z im = Rp + √ +√ +√ (36)
K ωC dl 2ω 2ω ωC dl 2ω
4.1. Variation of potential on the model surfaces
where Aw is the Warburg coefficient.

2
2
√Aw 1 √Aw
Fig. 4(a) shows the variation of double layer potential with the
and, K = Rp + + ωCdl
+
2ω 2ω distance from the electrode surface to the bulk solution estimated
RT 1 1 as per the Helmholtz model using Eq. (4). The potential varies lin-
Aw = √ ( +√ ) early with the distance. It approaches to zero with the distance (x∼
AZ 2 F 2 2 Do Co Dr Cr
20 nm). As per the Gouy-Chapman model estimated from Eq. (16b),

Fig. 4. (a) Variation of potential as per the Helmholtz (a), Gouy-Chapman (b), Stern (c) and Grahame models (d).
Md.S. Uddin et al. / Applied Surface Science 449 (2018) 445–453 451

Fig. 5. Areal capacitance as per the Helmholtz (a), Gouy-Chapman (b), Stern (c) and Grahame models (d).

the potential variation is exponential with the distance as shown in pores) with the distance, ‘x’ (from the electrode surface to the bulk
Fig. 4(b). It is seen that the significant potential is extended even at of the solution) estimated based on aforementioned models (Eqs.
the distance of 20 nm. As is seen in Fig. 4(c), the double layer poten- 5(a–c), 20(a–c), 23(a–c)). It is seen that in all cases the Cdl decreases
tial as per the Stern model decreases sharply when we move from exponentially with the distance x’. As per the Eq. (19), the Gouy-
electrode surface to within 1–2 nm distance as predicted by Eqs. Chapmann capacitance is independent of nature of surface area as
(4 and 16b). Beyond 2 nm distance, the potential decays gradually all the three surfaces showed similar Cdl . In all other models, it can
and attain nearly zero at a sufficiently bulk solution (∼20 nm). The be clearly seen that the type of pore is strongly influencing the areal
sharp decrease in double layer potential (␸) at the vicinity of the capacitance. High capacitance is seen in the vicinity of the dou-
electrode is due to the accumulation of charges due to the formation ble layer and low in the bulk solution. Interestingly, the cylindrical
of IHP. On the other hand, the gradual decrease of ␸ beyond 2 nm pore electrode exhibits high capacitance compared to the plain and
distance is due to diffused layer (OHP) formed by thermal motion. semi-spherical surfaces. Such a high capacitance is due to the large
Fig. 4(d) shows the potential variation from electrode surface to the surface area associated with the cylindrical surface. It should be
bulk solution according to Grahame model (Eqs. 4, 16(b) and 22). noted that the larger surface area will lead to better interaction
The potential increases within a short distance of 1 nm and reach with the electrolyte leading effective charge separation and hence,
the maximum at 0.5 nm, then linearly decreases to ∼3 nm distance. resulted in high specific capacitance. The areal capacitance reduces
Beyond 3 nm, the potential gradually decreases with x. It is inter- sharply around 2–3 nm and beyond 3 nm the capacitance trends
esting to note that a significant difference in the potential as per to zero. This shows that all the charge storage occurs in between
the Stern and Grahame models within a short distance of 1 nm. It is the 2–3 nm forming a surface of charge layer on the electrode sur-
noted that as per the Grahame model, the ␸ at very near the elec- face within this range. The pore size should be good enough for ions
trode surface increases (within the distance) due to IHP formation penetration; otherwise, the electrode behaves like as plane surface.
as well as ion absorption. Such an increase of potential has not been Alongside, the dipole size would be as small as possible to ensure
observed in the Stern model (Fig. 4(c)). Thus, the Grahame model r1 → r2 . As a result, the capacitance increases due to the small dis-
accounts for the ion adsorption at the vicinity of electrode surface tance between charge planes according to Helmholtz model. It is
which significantly influences on the potential. Thus, larger sur- noteworthy that in this work, simple analytical models have been
face area leads to better electrode/ electrolyte interaction which developed for three difference electrode surfaces which accom-
has resulted in larger potential at the electrode surface. Interest- modated the standard EDL models. Unlike the works reported by
ingly, in all models, ␾ extent to the bulk in the cylindrical pore others which involved assumption and rigorous treatments, the
electrode compared to the plain and semi-spherical surfaces. Such analytical expressions derived for both potential and areal capaci-
a high potential is due to the large surface area associated with tance in all three different surfaces are simple which did not involve
the cylindrical surface. The larger surface area will lead to better assumption. This shows that cylindrical pore electrode can be the
interaction with the electrolyte leading reduction of resistance and best option to choose for developing high-performance practical
hence, resulted in high potential. EDLC material.

4.2. Variation of areal capacitance on the model surfaces 4.3. Charge/ discharge characteristics

Fig. 5 shows the variation of areal capacitances for three dif- Fig. 6(a) shows the obtained charge-discharge profiles for the
ferent electrode surfaces, (i.e. plain, semi-sphere and cylindrical EDLC formed by each of three different surfaces in the 1 M aqueous
452 Md.S. Uddin et al. / Applied Surface Science 449 (2018) 445–453

4.4. Electrochemical impedance characteristics

Fig. 6(b) shows the Nyquist plot of the EDLC simulated using
Eqs. (33 and 34). The Nyquist plot was simulated in 1 M aqueous
electrolyte of relative permittivity ␧r = 78.5, temperature 20 ◦ C, ion
diffusion coefficient Do = Dr = 1 × 10−9 m2 S-1 , whereas Faradic resis-
tance RP = 5 , solution resistance RS = 0.15 , and frequency f = 1 to
105 Hz. It can be seen that the Nyquist plot consists of the semicir-
cle in the high-frequency region and a spike at the low-frequency
region. The appearance of spike (Warburg element) confirms the
presence of ion diffusion influencing the EDLC property.

5. Conclusions

The double layer potential (␾) and the areal capacitance of


the EDLC have been examined on three different surfaces on the
basis of standard models. The potential and the Cdl depend on the
type of surface, cylindrical pore surface exhibiting larger poten-
tial as well as areal capacitance. The Grahame model accounted for
the total capacitance which involves Cdl and ion adsorption. The
charging/ discharging and impedance characteristics confirmed the
presence of ion diffusion contributing to Cdl . This study revealed
that a large surface area electrode can be utilized for developing
high-performance EDLC practically.

Acknowledgements

The authors acknowledge the Pondicherry University for com-


puter and laboratory facilities. PE thanks SERB, New Delhi, Govt.
of India for the research scheme (EMR/2016/001305). MSU thanks
South Asian Foundation for Fellowship.

References

[1] J.J. L-García, J. Horno, C. Grosse, Influence of the finite size and effective
permittivity of ions on the equilibrium double layer around colloidal particles
in aqueous electrolyte solution, J. Colloid Interface Sci. 428 (2014) 308–315.
[2] J.H. Banga, H.-M. Lee, K.-H. An, B.-J. Kim, A study on optimal pore
Fig. 6. (a) Simulated charge-discharge profiles of EDLC for three different surfaces development of modified commercial activated carbons for electrode
and (b) Nyquist plot of EDL with Warburg component. materials of supercapacitors, Appl. Surf. Sci. 415 (2017) 61–66.
[3] H. Lu, Q. Li, J. Guo, A. Song, C. Gong, J. Zhang, J. Zhang, Hierarchically porous
carbon with high-speed ion transport channels for high performance
supercapacitors, Appl. Surf. Sci. 427 (2018) 992–999.
[4] S. Bana, J. Zhang, L. Zhang, K. Tsay, D. Song, X. Zou, Charging and discharging
electrolyte (Eqs. (28) and (32)). The charge/ discharge characteris- electrochemical supercapacitors in the presence of both parallel leakage
tics of the EDLC was simulated using Vs = 0.45 V, ␧r = 78.5, RS = 0.15 process and electrochemical decomposition of solvent, Electrochim. Acta 90
(2013) 542–549.
 and Faradaic resistance 5  in the (1 M) aqueous electrolyte. The [5] N. Gavish, K. Promislow, Dependence of the dielectric constant of electrolyte
obtained charge/ discharge profiles are nearly mirror image in all solutions on ionic concentration, Phys. Rev. E 94 (2012) 1–5.
the cases. This means that the charging/ discharging process are [6] R.G. Babbadas, J.M. Sedlak, Dielectric coefficients in the interphase between
electrode and ionic solution, Electrochim. Acta 17 (1972) 683–693.
highly reversible. It is noted that the charging time is 8 s and the [7] J.B. Hasted, D.M. Ritson, C.H. Collie, Dielectric properties of aqueous ionic
discharge time is also 8 s, leading to coulombic efficiency of 100% solutions, J. Chem. Phys. 16 (1948) 1.
which is highly possible for an ideal EDLC. The semi-sphere and [8] D. Wei, G.N. Patey, Dielectric relaxation of electrolyte solutions, J. Chem. Phys.
94 (1991), 6795.
plane surface electrode also exhibits charging time of 7 s and 4 s
[9] B. Hess, C. Holm, N. van der Vegt, Modeling multibody effects in ionic
respectively, with 100% Coulombic efficiency. Significant IR drop solutions with a concentration dependent dielectric permittivity, Phys. Rev.
occurs in particularly during discharging, due to solution resistance Lett. 96 (2006) 147801.
offered by the electrolyte. The nature of electrode surface has a [10] H. Ohshima, A Simple algorithm for the calculation of the electric double layer
potential distribution in a charged cylindrical narrow pore, Colloid Polym. Sci.
sharp influence on the charging/ discharging process. The cylindri- 294 (2016) 1871–1875.
cal pore electrode exhibits large charging/ discharging time which [11] M.C. Henstridge, E.J.F. Dickinson, R.G. Compton, On the estimation of the
is substantiated by the Cdl data shown in Fig. 5. diffuse double layer of carbon nanotubes using classical theory: curvature
effects on the Gouy–Chapman limit, Chem. Phys. Lett. 485 (2010) 167–170.
Further, when ions lose their solvation shell, it undergoes the [12] H. Wang, J. Fang, L. Pilon, Scaling laws for carbon-based electric double layer
charge transfer reactions (i.e. the ions come in direct contact with capacitors, Electrochim. Acta 109 (2013) 316–321.
electrode surface) contributing to pseudocapacitance along with [13] H. Helmholtz, Studies of electrical boundary layers, Ann. Phys. 243 (1879)
337–382.
the EDLC. Generally, the presence of pseudocapacitance in the EDLC [14] S. Srinivashan, Electrode/Electrolyte Interfaces: Structure and Kinetics of
results in one or more plateau regions in charge/discharge profiles Charge Transfer, Fuel Cells (from Fundamentals to Application), Springer,
corresponding to the Faradaic reactions. It can be seen from Fig. 6(a) 2006, pp. 27–92.
[15] G. Gouy, Constitution of the electric charge at the surface of an electrolyte, J.
that there is no plateau regions in the charge/discharge profiles. So, Phys. 9 (1910) 457–467.
the simulated charge/discharge profile is entirely contributed by [16] D.L. Chapman, A contribution to the theory of Electrocapillarity, J. Sci. 25
the EDLC. (1913) 475–481.
Md.S. Uddin et al. / Applied Surface Science 449 (2018) 445–453 453

[17] H. Butt, L. Graf, M. Kappl, Physics and Chemistry of Interfaces, 2nd ed., [22] K.T. Oldham, Swain, The Doctrine of Description: Gustav Kirchhoff, Classical
Wiley-VCH, Weinheim, Germany, 2006. Physics, and the P̈urpose of All Scienceïn 19th-Century Germany’ Ph. D.
[18] O. Stern, The theory of the electrolytic double layer, Phys. Chem. 30 (1924) Thesis, University of California, Berkeley, USA, 2008, pp. 52.
508–516. [23] H.W. Jackson, Introduction to Electronic Circuits, Prentice-Hall, 1959.
[19] R.A. Serway, Physics for Scientists and Engineers With Modern Physics, 4th [24] J.E.B. Randle, Kinetics of rapid electrode reactions, Discuss. Faraday Soc. 1
ed., 1996, pp. 687. (1947) 11–19.
[20] Coulomb, Histoire de L’Académie Royale Des Sciences, Imprimerie Royale, [25] L.Earl Consoliver, I.Grover Mitchell, Automotive Ignition Systems,
1785, pp. 569–577. McGraw-Hill, New York, 1920, pp. 4.
[21] R. Boylestad, Introductory Circuit Analysis’, 10th edition, Prentice Hall, Ohio, [26] T. Erdey-Gruz, M. Volmer, Zur theorie der wasserst-offüberspannung, J. Phys.
USA, 2003, pp. 408. Chem. A 150 (1930) 203–213.