Measurements of differential capacitance in room temperature

Measurements of differential capacitance in room temperature
ionic liquid at mercury, glassy carbon and gold electrode interfaces
Muhammad Tanzirul Alam, Md. Mominul Islam, Takeyoshi Okajima, Takeo Ohsaka *
Department of Electronic Chemistry, Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology,
Mail Box G1-5, 4259 Nagatsuta, Midor-ku, Yokohama 226-8502, Japan
Received 25 June 2007; received in revised form 3 July 2007; accepted 6 July 2007
Available online 17 July 2007
Abstract
Differential capacitances were measured in 1-propyl-3-methylimidazolium tetrafluoroborate (PMIBF4) ionic liquid at three different
electrode substrates (Hg, GC (glassy carbon) and Au) as a function of potential. Essentially different capacitance–potential curves were
obtained at different electrodes. From the parabolic electrocapillary curve measured at dropping Hg electrode in PMIBF4, the potential
of zero charge (PZC) was found to be À0.31 V vs. Ag/AgCl (wire). However, the capacitance–potential curve at Hg electrode was found
not to show any valley related to the PZC, whereas at GC and Au electrodes a minimum was observed at 0.29 and À0.51 V, respectively.
The results are in disagreement with the recent theoretical study which implies that the capacitance–potential curve should be of bell
shape with the maximum value of capacitance at PZC. The parabolic capacitance–potential curve similar to those obtained in inorganic
molten salts was also observed for the first time at GC electrode. Probable causes of the difference in their capacitance–potential curves
were also discussed.
Ó 2007 Elsevier B.V. All rights reserved.
Keywords: Ionic liquid; Capacitance; PZC; Double layer; Hump
1. Introduction
Over the last decade, there has been a surge in interest in
room temperature ionic liquids (RTILs) as ‘‘green’’ alter-
natives to the conventional solvents due to their unique
physiochemical properties such as low vapor pressure, non-
flammability and tunability as solvents by ready incorpora-
tion of one or more functional groups [1–3]. Large
electrochemical potential window and dual role as solvent
as well as electrolyte in electrochemistry have added
another avenue in this connection and various studies have
already been done using RTILs as electrolyte in capacitors
and in other investigations [4,5]. However, some funda-
mental aspects remain to be solved yet, including the struc-
ture of the electrical double layer (EDL) at the electrode/
RTILs interface on which our knowledge is still in the
primitive level.
The classical methods of evaluating the structure of the
EDL at electrode/electrolyte interfaces involve the mea-
surements of interfacial tensions and differential capaci-
tances as a function of potential [6]. The only work that
had been done in this regard in RTILs is the work of Koch
et al. [7] where they measured the differential capacitance at
Hg electrode for comparing it with that at carbon cloth
electrode aiming to use it in practical electrochemical
capacitor. Spectroscopic techniques have also been used
for studying the orientation of the ions at the electrode/
RTILs interface and it has been found to depend on the rel-
ative bulkiness of the ions, electrostatic attraction and local
structure of the RTILs [8–12].
All of the accepted previous models of EDL [6] and their
relevant theoretical works require the assumption of the
presence of a molecular solvent. However, those models
are barely applicable to RTILs due to the absence of any
1388-2481/$ - see front matter Ó 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.elecom.2007.07.009
*
Corresponding author. Tel.: +81 45 9245404; fax: +81 45 9245489.
E-mail address: ohsaka@echem.titech.ac.jp (T. Ohsaka).
www.elsevier.com/locate/elecom
Electrochemistry Communications 9 (2007) 2370–2374

molecular solvent in RTILs and their high concentration as
electrolytes. The only possible relevance can be found in
the work related to the interfacial structure in molten salts.
Unfortunately, the results in inorganic molten salts suffer
from lack of reliability due to the operational high temper-
ature, content of water as impurity and different extent of
thermal expansion resulting in the absence of well accepted
unanimous model of EDL and the corresponding explana-
tion [13,14]. On the other hand, more recent theoretical
study of Kornyshev [15] based on the local density approx-
imation dependent mean-field theory on double-layer at
ionic liquid/electrode interface has demonstrated that the
capacitance–potential curve should have the maximum
value at the PZC, rather than the usually observed mini-
mum and at more negative and positive potentials than
the PZC, the capacitance has been shown to decrease with
the square root of potential giving the capacitance–poten-
tial curve of a bell shape. However, experimental data is
in need to justify the above approach.
Unlike in aqueous or organic media containing electro-
lytes, ions are essentially in direct contact with the electrode
surface in RTILs. Thus, it is of real fundamental interest to
know how the pattern of the capacitance–potential curves
varies with the variation of electrode, potential or others
and the possible causes behind them. In the present paper,
we report for the first time the parabolic capacitance–
potential curve at GC/RTIL interface similar to those
obtained in inorganic molten salts. The essentially different
capacitance–potential curves are observed in PMIBF4 at
Hg, GC and Au electrode interfaces. Possible causes of
their different capacitance–potential curves are also
discussed.
2. Experimental
2.1. Reagents
1-Propyl-3-methylimidazolium tetrafluoroborate
(PMIBF4) with a purity of more than 99% and less than
30 ppm (i.e., ca. 2.1 mM) water was obtained from Stella
Chemifa Co. (Japan). All of the chemicals were of reagent
grade and used without further purification.
2.2. Electrochemical measurements
Electrochemical experiments were performed in a three-
electrode cell containing a hanging mercury drop (Hg)
[model CGME 900, Bioanalytical Systems, Inc. (BAS),
area: 0.018 cm2
], a glassy carbon (GC, diameter = 1 mm)
or a bare gold (Au, diameter = 1.6 mm) as a working elec-
trode, a spiral platinum wire as a counter electrode and a
homemade silver/silver chloride (Ag/AgCl (solid, wire))
as a reference electrode. The reference electrode was pre-
pared by coating silver chloride on a freshly polished silver
wire via an electrolysis at 0.35 V vs. Ag/AgCl/KCl (satu-
rated) in 1 M KCl solution. The electrolysis was run until
a clear brown film develops on the surface of silver wire.
The thus prepared electrode was used after washing with
Milli-Q water and drying at a moderate temperature
(50 °C). All potentials are reported against this Ag/AgCl
(solid, wire) reference electrode and were found to be stable
with a ±10 mV fluctuation of the potential. The formal
potential (E00
= 0.156 V vs. Ag/AgCl (wire)) of the ferro-
cene/ferricinium ion couple was used as an internal stan-
dard of electrode potential. Each experiment was carried
out at a freshly formed Hg drop or at a newly polished
clean working electrode. The solid working electrodes were
first polished on emery paper and then with aqueous slur-
ries of fine alumina powders (1 and 0.06 lm) on a polishing
cloth followed by rinsing with doubly distilled water and
acetone in an ultrasonic bath, each for 10 min, and were
finally rinsed with Milli-Q water. After that the Au elec-
trode was further treated electrochemically by successive
potential cycling in N2-saturated 0.05 M H2SO4 until the
cyclic voltammogram, characteristic of a clean Au elec-
trode, was obtained.
Solarton SI 1260 and SI 1287 were used as impedance/
gain phase analyzer and electrochemical interface, respec-
tively, for the measurement of capacitances. Impedance
measurements were done at a constant frequency
(200 Hz) by scanning the electrode potential (i.e., dc poten-
tial) from negative to positive direction at a scan rate of
5 mVsÀ1
and the ac potential with 5 mV peak to peak
amplitude was superimposed on dc potential. From the
impedance measurement data, the value of the capacitance
(C) was derived using the equation –Z0 0
= 1/(2pfC), where
Z0 0
is the imaginary component of impedance [16,17]. Elec-
trolytes were deoxygenated by purging pure N2 gas for
about 30 min, and the gas was kept flowing over the liquids
during the electrochemical measurements.
The drop-time measurement was done with a homemade
natural dropping Hg electrode in an electrochemical cell
with the same arrangement of reference and counter elec-
trodes as described above. In this experiment, the Hg drops
were allowed to fall through the capillary from a height of
56.5 cm to N2-saturated PMIBF4. The lifetime of a drop
(i.e., drop time) was estimated as an average by counting
the total time required for 10 drops to fall at a certain
applied potential. The measurement at each applied poten-
tial was repeated three times to confirm the reproducibility
of the obtained drop time. In such a way, the drop times
were measured in the potential range of À1.2 to 0.55 V at
a potential interval of 0.1 or 0.05 V. All the experiments
were carried out at room temperature (25 ± 2 °C).
3. Results and discussion
Fig. 1. illustrates the capacitance–potential curve at Hg
electrode in N2-saturated PMIBF4. The measured capaci-
tance curve, like the capacitance–potential curve at Hg
electrode in aqueous media, has a shallow minimum at
moderate negative potential (À0.9 V) with the capacitance
value of 11.2 lF cmÀ2
. The obtained curve can be divided
into two parts: The value of capacitance increases sharply
M.T. Alam et al. / Electrochemistry Communications 9 (2007) 2370–2374 2371

(up to 27.4 lF cmÀ2
) with increasing the potential in the
range of potential more positive than À0.9 V (to 0.2 V),
while a small elevation in the capacitance value was
observed at more negative potential than À0.9 V. The mea-
sured minimum capacitance value (11.2 lF cmÀ2
) even in
this highly concentrated PMIBF4 (5.85 M) RTIL is surpris-
ingly comparable with that in aprotic and aqueous solu-
tions containing relatively low concentrations of
electrolyte (typically 1–100 mM) and in the latter case such
a minimum was assumed as the capacity of the respective
solvent [18–21]. This valley and the capacitance inflations
on its both sides are characteristic of the capacitance–
potential curve at Hg electrode in the absence of a strong
adsorption of electrolyte ions. Several possibilities have
already been put forward in this regard, e.g., the orienta-
tion of the solvent molecules and pseudo-capacity associ-
ated with cation deposition [18,19,22], but till now none
of these is unanimously accepted. The observed minimum
at À0.9 V in no way represents the PZC [18,22], as men-
tioned below.
Fig. 2. shows the drop time vs. potential curve measured
at dropping Hg electrode in N2-saturated PMIBF4. The
curve is parabolic in shape in analogy with those in aque-
ous and aprotic solvent media with a maximum at
À0.31 V, which theoretically corresponds to the PZC.
Therefore, the valley observed at À0.9 V in Fig. 1. does
not correspond to the PZC. The shape of the electro-
capillary curve is quite similar to those reported by Koch
et al. [7] measured in a series of 1-ethyl-3-methylimidazo-
lium based RTILs where the potentials corresponding to
the maxima of the electrocapillary curves were designated
as the PZC. However, the capacitance–potential curve in
Fig. 1. does not show any valley related to the PZC at this
potential (À0.31 V).
According to the theory of Gouy–Chapman–Stern
model of electrical double layer [6], the capacitance–poten-
tial curve in highly concentrated media should not show
any distinguishable minimum associated to the PZC as
the capacitance elevation around it results from the com-
pactness of diffuse layer. Capacitance change in an ideal
system is usually accounted as the change of diffuse layer
thickness with potential and electrolyte concentration. At
a given concentration, the thickness of the diffuse layer is
the largest at the PZC due to the absence of any electro-
static attraction which results in the lowest value of capac-
itance at that point. However, in RTILs there is no
molecular solvent and the ions are reported to maintain a
solid like crystalline structure [23] even in their liquid state,
and thus the scope of the ions to move in response to the
applied potential is small and consequently the effect of
the diffuse layer will be negligible.
Fig. 3. illustrates the capacitance–potential curve at
GC electrode in N2-saturated PMIBF4. This curve shows
a minimum at 0.29 V with a capacitance value of 11.4 lF
cmÀ2
. Capacitance elevation on both sides of this mini-
mum is almost the same, giving the capacitance–potential
curve of a familiar parabolic shape similarly to those
observed in inorganic molten salts, where the potential
-1.2 -0.8 -0.4 0.0 0.4
12
16
20
24
28
C/μFcm-2
E/V vs. Ag/AgCl (wire)
Fig. 1. Capacitance vs. potential curve measured at N2-saturated Hg/
PMIBF4 interface.
-1.6 -1.2 -0.8 -0.4 0.0 0.4 0.8
3.6
3.8
4.0
4.2
4.4
4.6
4.8
5.0
Droptime/s
Fig. 2. Drop time vs. potential curve measured at dropping Hg electrode
in N2-saturated PMIBF4.
-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5
10
12
14
16
18
20
C/μFcm-2
Fig. 3. Capacitance vs. potential curve measured at N2-saturated GC/
PMIBF4 interface.
2372 M.T. Alam et al. / Electrochemistry Communications 9 (2007) 2370–2374

corresponding to the minimum of the capacitance–poten-
tial curve was ascribed to the PZC. This similarity could
be a demonstration of the presence of a similar interfacial
structure in both systems. The magnitude of capacitance
inflation with potential is quite low in PMIBF4 compared
to those in inorganic molten salts. Here, the capacitance
change is only 8.3 lF cmÀ2
in the examined potential
range of À1.3 to 1.3 V, whereas in inorganic molten salts
it is much larger, e.g., the capacitance was found to
change up to 65 lF cmÀ2
for a potential change of
0.3 V from the PZC at lead/LiCl-KCl melt interface
[13,14]. This relatively smaller capacitance change in
PMIBF4 is an indication of the absence of a strong spe-
cific adsorption of ions related to the charge transfer pro-
cess which was proposed as one of the factors of the large
capacitance change in molten salts [13,14]. Devanathan
[20] had shown that a strong specific adsorption of ions
should always be followed by large change of capacitance
with potential. Cyclic voltammogram taken at GC elec-
trode in N2-saturated PMIBF4 also did not show any
peak corresponding to charge transfer within the above
examined range of potential (data not shown here). How-
ever, due to the absence of any solvent sheath around the
ions, some weak electrostatic interaction of the ions with
the GC surface can not be ignored and can be considered
as one of the factors of capacitance change in PMIBF4.
Model of multilayer structure [13,14] can be ruled out eas-
ily as it would increase the interfacial distance with a con-
sequent decrease of capacitance, while a continuous
increase of capacitance is observed on both sides of the
PZC in Fig. 3.
The special structure of PMIBF4 at the interface may
be closely concerned with the observed change of the
capacitance curve. Baldelli et al. [8–11] and Nanbu et al.
[12] had shown this quite nicely using in situ sum fre-
quency generation and FTIR spectroscopic techniques.
They observed the orientational change of imidazolium
ring at Pt and Au electrode interface in response to the
applied potential. At negative potential, the imidazolium
ring tends to become parallel to the electrode surface,
whereas at positive potential it moves away to make a
room for the anion to approach to the electrode surface.
Thus, the orientational change along with the change of
the distance of the ions with respect to the electrode sur-
face could be a cause of the capacitance change with
potential. On the other hand, at Hg drop electrode, an
inherent streaming [24–26] of its surface may impede such
a potential-dependent orientational change of the imi-
dazolium ring.
Fig. 4. represents the capacitance–potential curve mea-
sured at Au electrode in N2-saturated PMIBF4. In analogy
with the capacitance–potential curve at GC electrode, this
curve is also parabolic in shape in the potential range of
À1.2 to 0 V with a minimum at À0.51 V and the same
explanation as that used in Fig. 3. can also be applied
for this case. Baldelli et al. [8,11] confirmed the minimum
of the capacitance–potential curve as the PZC using sum
frequency generation spectroscopic technique in two differ-
ent RTILs/Pt interfaces, and this finding is in disagree-
ment with the recently reported theoretical results which
predict that the capacitance should not be a minimum
but a maximum at the PZC [15]. The significant feature
of Fig. 4. is the capacitance rise on the anodic side of
the PZC in the form of a hump. Hump is a general aspect
of the capacitance–potential curves measured in aqueous
as well as organic media. Several causes have already been
proposed in this regard including the adsorption of elec-
trolyte ions or solvent at the electrode surface, orienta-
tional change of the solvent or ions, structure of the
electrolytes, etc. [18–20,22]. Our recent results show that
the hump in Fig. 4 is related to the adsorption pseudoca-
pacitance at Au surface. Further work regarding this is in
progress.
Capacitance–potential curves described above do not
show any maximum which can be designated as the
PZC according to the theoretical study of Kornyshev
[15]. Moreover, on the contrary to his prediction, at large
potentials our curves also did not decline with the square
root of potential. Particularly, the completely parabolic
capacitance–potential curve obtained at GC electrode
was found to differ significantly from the theoretical pre-
diction. Because of the absence of theory and poor
knowledge on the structure of electrical double layer at
the RTILs/electrode interfaces, it is difficult to designate
the PZC unambiguously, but by comparing the capaci-
tance–potential curves with the corresponding electro-
capillary curve and spectroelectrochemical data [8–12], it
is evident that PZC does not correspond to the maximum
value of capacitance. Additionally, the capacitance–
potential and electrocapillary curves measured at Hg
drop electrode in other RTILs (1-hexyl-3-methylimidazo-
lium tetrafluoroborate and 1-octyl-3-methylimidazolium
tetrafluoroborate) also do not support the prediction of
the above-mentioned theoretical study and will be
reported elsewhere.
-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5
10
12
14
16
18
20
C/μFcm-2
Fig. 4. Capacitance vs. potential curve measured at N2-saturated Au/
PMIBF4 interface.
M.T. Alam et al. / Electrochemistry Communications 9 (2007) 2370–2374 2373

4. Conclusions
Differential capacitance vs. potential curves were mea-
sured at PMIBF4 RTIL/Hg, GC and Au electrode inter-
faces using electrochemical impedance technique. Unlike
in aqueous or (conventional) organic solvents, capaci-
tance–potential curves were found to vary significantly
with the electrode substrates in this RTIL. It could be
due to the absence of inner Helmholtz layer of molecular
solvent between the electrode and ionic species which usu-
ally works as a dominant factor in shaping the capacitance
curves. The electrocapillary measurement at a dropping Hg
electrode showed the PZC at À0.31 V vs. Ag/AgCl (wire).
However, the capacitance–potential curve at Hg electrode
did not show any minimum or maximum (as predicted by
recent theory) at the PZC, whereas at GC and Au elec-
trodes a minimum was observed at 0.29 and À0.51 V,
respectively. The parabolic capacitance–potential curve
similar to those obtained in inorganic molten salts was also
observed for the first time at the GC/RTIL interface in
PMIBF4. None of the experimental capacitance–potential
curves is bell-shaped though the recent theoretical study
predicts that the curve should be of bell shape with a max-
imum value of capacitance at the PZC.
Acknowledgements
The present work was financially supported by Grant-
in-Aids for Scientific Research on Priority Areas (No.
417) and Scientific Research (A) (No. 19206079) to T.O.,
from the Ministry of Education, Culture, Sports, Science
and Technology (MEXT), Japan. M.T.A. gratefully
acknowledges the Government of Japan for a Monbu-
Kagakusho Scholarship.
References
[1] M. Galinski, A. Lewandowski, I. Stepniak, Eelctrochim. Acta 51
(2006) 5567.
[2] J.L. Anderson, D.W. Armstrong, G.-T. Wei, Anal. Chem. 78 (2006)
2892.
[3] H. Tokuda, K. Hayamizu, K. Ishii, Md.A.B.H. Susan, M. Watanabe,
J. Phys. Chem. B 109 (2005) 6103.
[4] D. Zhang, T. Okajima, F. Matsumoto, T. Ohsaka, J. Electrochem.
Soc. 151 (2004) D31.
[5] M.C. Buzzeo, R.G. Evans, R.G. Compton, ChemPhysChem. 5 (2004)
1106.
[6] A.J. Bard, L.R. Faulkner, Electrochemical Methods, Fundamental
and Applicationsecond ed., Wiley Inc., New York, 2001.
[7] C. Nanjundiah, S.F. McDevitt, V.R. Koch, J. Electrochem. Soc. 144
(1997) 3392.
[8] C. Aliaga, S. Baldelli, J. Phys. Chem. B 110 (2006) 18481.
[9] S. Rivera-Rubero, S. Baldelli, J. Phys. Chem. B 108 (2004) 15133.
[10] C. Romero, S. Baldelli, J. Phys. Chem. B 110 (2006) 6213.
[11] S. Baldelli, J. Phys. Chem. B 109 (2005) 13049.
[12] N. Nanbu, Y. Sasaki, F. Kitamura, Electrochem. Commun. 5 (2003)
383.
[13] A.D. Graves, D. Inman, J. Eelctroanal. Chem. 25 (1970) 357.
[14] A.D. Graves, J. Electroanal. Chem. 25 (1970) 349.
[15] A.A. Kornyshev, J. Phys. Chem. B 111 (2007) 5545.
[16] M.M. Islam, T. Okajima, T. Ohsaka, J. Phys. Chem. B 108 (2004)
19425.
[17] M.S. Saha, T. Ohsaka, Electrochim. Acta 50 (2006) 4746.
[18] D.C. Grahame, J. Am. Chem. Soc. 76 (1954) 4819.
[19] D.C. Grahame, J. Am. Chem. Soc. 79 (1957) 2093.
[20] M.A.V. Devanathan, B.V.K.S.R.A. Tilak, Chem. Rev. 65 (1965) 635.
[21] R. Payne, J. Am. Chem. Soc. 89 (1967) 489.
[22] B.E. Conway, J.O’M. Bockris, Modern Aspects of Electrochemistry,
vol. 13, Plenum Press, New York, 1979.
[23] S. Saha, S. Hayashi, A. Kobayashi, H. Hamaguchi, Chem. Lett. 32
(2003) 740.
[24] H.H. Bauer, in: A.J. Bard (Ed.), Electroanalytical Chemistry, vol. 8,
Marcel Dekker Inc., New York, 1975, p. 169.
[25] M.S. Saha, T. Okajima, T. Ohsaka, J. Phys. Chem. B 1006 (2002)
4457.
[26] M.M. Islam, T. Okajima, T. Ohsaka, J. Phys. Chem. B 110 (2006) 8619.
2374 M.T. Alam et al. / Electrochemistry Communications 9 (2007) 2370–2374

Measurements of differential capacitance in room temperature

More Related Content

What's hot (18)

Viewers also liked (13)

Similar to Measurements of differential capacitance in room temperature (20)

Recently uploaded (20)

Measurements of differential capacitance in room temperature