Journal of Colloid and Interface Science 300 (2006) 655–662

www.elsevier.com/locate/jcis

An electrical impedance spectroscopic (EIS) study

on transport characteristics of ion-exchange membrane systems
Jin-Soo Park a , Jae-Hwan Choi b , Jung-Je Woo c , Seung-Hyeon Moon c,∗
a Polymer Electrolyte Fuel Cell Research Group, New Energy Research Department, Korea Institute of Energy Research (KIER),
71-2, Jang-dong, Yusong-gu, Daejeon 305-343, Republic of Korea
b Department of Chemical Engineering, Kongju National University, 182, Shinkwan-dong, Gongju, Chungnam 314-701, Republic of Korea
c Department of Environmental Science and Engineering, Gwangju Institute of Science and Technology (GIST), 1 Oryong-dong, Buk-gu,
Gwangju 500-712, Republic of Korea
Received 13 March 2006; accepted 14 April 2006
Available online 24 May 2006

Abstract
This study aimed at investigating ion-exchange membrane systems using impedance spectroscopy. Nyquist plots showed that the impedance
obtained in this study described the ion-exchange membrane system well, as consisting of (i) an ion-exchange membrane immersed in solution,
(ii) electrical double layers at the membrane surface, and (iii) diffusion boundary layers arising from the interface between the ion-exchange
membrane and the electrolyte solutions. Taking into account the physical and electrochemical understanding of the ion-exchange membrane
system, an equivalent circuit was suggested to quantitatively analyze each component of the ion-exchange membrane system. To confirm the
reliability of the proposed equivalent circuit, the resistance and capacitance were estimated from the impedance data and the values were compared
with other experimental results (e.g., I –V curves). The comparison showed good agreement and validated the equivalent circuit. Moreover, the
impedance measurements made it possible to confirm the electroconvective effects in the over LCD region.
© 2006 Elsevier Inc. All rights reserved.

Keywords: Ion-exchange membrane; Electrical impedance spectroscopy; Nyquist plot; Equivalent circuit; Electroconvective effects

1. Introduction Donnan exclusion. Moreover, the different transport number of

Ion-exchange membranes (IEMs) have been widely used
in various applications such as diffusion dialysis, electrolysis,
eletrodialysis, and fuel cells [1–10]. At the same time, much at-
tention has been paid to electrochemical characteristics of IEM
systems in order to understand complex transport phenomena
in IEM systems.
IEMs consist of highly swollen gel-type polymer struc-
tures carrying fixed positive or negative charges. The functional
groups, as fixed charges have a significant effect on the ion-
exchange behavior of the membranes [11]. When an electric
current passes through the membrane system, the current is
carried by both positive and negative ions in the bulk solu-
tion phase, but mainly by the counterions in the IEM due to
* Corresponding author. Fax: +82 62 970 2434.
E-mail address: shmoon@gist.ac.kr (S.-H. Moon).
0021-9797/$ – see front matter © 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcis.2006.04.040
the counterions in the bulk solution phase and the membrane
leads to depletion in electrolyte concentration on the diluate
side and to enrichment on the concentrate side [12,13]. As a re-
sult, concentration polarization occurs at the interface between
the IEM and the electrolyte solutions, and polarized layers, i.e.,
the diffusion boundary layer (DBL), are finally formed on the
membrane surface. In the case of an IEM system where two
equal 1:1 electrolytes are separated by an IEM, it is generally
believed that several electric fields may be superimposed: lo-
cal fields around the membrane, diffusion boundary layers, and
externally applied fields [14].
Until now, known transport phenomena of the IEM systems
have mainly been the fruit of experimental work using current–
voltage curves. However, unknown transport phenomena are
out of the field that can be investigated using current–voltage
curves. Such complex transport phenomena (i.e., electrocon-
vection, water dissociation, fraction of conducting, nonconduct-
656 J.-S. Park et al. / Journal of Colloid and Interface Science 300 (2006) 655–662
ing region of IEMs, etc.) in the field of IEM systems are re-
quired to be understood by new methods.
Previous researchers have proven that electrical impedance
spectroscopy (EIS) is a noninvasive technique for character-
izing (i) the electrical properties of materials, electrochemical
phenomena, and structural observations of bipolar membranes
[15–19], (ii) composite membranes [20], and (iii) synthetic
ion-exchange membranes [21–23]. Especially, the impedance
studies of the bipolar membranes have provided good electro-
chemical information on the bipolar junction structure and the
intermediate region between the two ion-exchange membranes
as well as related electrochemical phenomena (i.e., the superim-
posed electric fields). Especially, Smith et al. [16] predicted the
approximate distance (1 nm) between the two ion-exchange
membranes using the obtained impedance data. Park et al.
[24,25] reported that impedance spectra well explained fouling
effects on IEM systems that were hundreds-of-micrometers do-
mains. Taking account of the space resolution, the impedance
measurements can be a good tool for investigating unknown
transport phenomena in IEM systems.
In this study, we examined the analogy between current–
voltage curves and impedance spectra of IEM systems in terms
of conductive properties (i.e., resistance). Based on the im-
pedance results, each conductive and capacitive property in the
equivalent circuit as reported in [24] was quantitatively ana-
lyzed to describe the structures of IEM systems. Moreover,
microscopic information on sublayers (i.e., electrical double
layers and microfluid flows (so-called electroconvection phe-
nomena)) was provided at the phase boundaries between mem-
brane and solution.
2. Experimental
2.1. Electrolytes
Potassium chloride (Sigma, St. Louis, Missouri) in reagent
grade was used as a supporting electrolyte. KCl solutions with
different concentrations were used for measurement of the elec-
trical impedance spectra and the I –V curves of the IEM sys-
tems used in this study. Deionized water (18 M cm) was used
in preparing all the electrolytes.
2.2. Ion-exchange membranes
The Neosepta AMX (Tokuyama Soda Co., Japan) is a re-
inforced standard grade homogeneous ion-exchange membrane
containing quaternary ammonium groups as fixed charges for
use in general concentration or desalination [11,26]. Table 1
shows their main properties as given by the manufacturer [26].
Prior to all the experiments, the test membranes were immersed
in the desired electrolyte solution for 24 h to reach equilibrium.
2.3. Measurement of current–voltage curves and electrical
impedance spectra
The current–voltage curves and electrical impedance spec-
troscopy were obtained using an electrodialytic cell, which
Table 1
Main properties of the AMX anion-exchange membrane used in this study [24]
Name Neosepta AMX
Backing Poly(vinyl chloride)
Electric resistance ( cm2 ) 2.5–3.5
Burst strength (kg/cm2 ) 4.5–5.5
Exchange capacity (meq/g dry membrane) 1.4–1.7
Thickness (mm) 0.16–0.18
Transport number (total cations or anions) 0.98<
Water content 0.25–0.30
Reinforcing Yes
consisted of two compartments of equal volume (150 ml) sep-
arated by a membrane that was supported by rubber gaskets
across a circular hole of 0.785 cm2 . The details are found else-
where [27].
Direct (dc) and sinusoidal alternating currents (ac) were sup-
plied to a pair of Ag/AgCl planar electrodes at a current scan-
ning rate of 1 µA/s for I –V curves and in the frequency range
from 106 to 10−3 Hz for measurements of impedance spectra
by a potentiostat/galvanostat and a frequency response analyzer
(AutoLab, Model PGSTAT 30, Netherlands), respectively. The
Ag/AgCl planar electrodes minimized the water dissociation re-
actions at the electrodes because proton/hydroxyl ions may af-
fect the composition of electrolytes in the compartment for both
experiments. The potential difference and impedance across the
membrane system were measured using two Ag/AgCl elec-
trodes immersed into Luggin capillaries. All the experiments
were conducted at 25 ± 1 ◦ C.
3. Results and discussion
3.1. Impedance arising from the IEM systems
The sinusoidal ac responded to the several superimposed
electric fields in IEM systems as discussed earlier. The im-
pedance spectra can be evaluated qualitatively and quantita-
tively. Fig. 1 shows a typical Nyquist plot of the IEM systems
used in this study. The diameter of the capacitive loop decreased
in the presence of stirring. It is believed that the decrease in
thickness of DBLs reduces impedance by ion diffusion [12,13].
This behavior can be explained by the concept of complex con-
ductivity, which illustrates the characteristics of charge move-
ment and storage properties with frequency [22,28],
σ ∗ = σ + j ωε, (1)
where σ is the conductivity and ε the dielectric permittivity of
the electrical system. At a high frequency (ωε  σ ), the dielec-
tric permittivity term dominates in the complex conductivity.
This means that the dielectric permittivity of the interface is
greater than that of the membrane, and current can easily pass
through the membrane without significant electrical hindrance.
In fact, it was observed that the impedance of the IEM sys-
tem was constant at high frequencies (not shown here), corre-
sponding to the conductivity of the membrane. However, at low
frequencies (ωε  σ ), the conductivity term dominates in the
 J.-S. Park et al. / Journal of Colloid and Interface Science 300 (2006) 655–662
Fig. 1. Stirring effect on impedance spectra (Nyquist plots) of the AMX mem-
brane in 0.01 M KCl using the two-compartment electrolytic cell [25] when
solutions in each compartment were stirred and not stirred.
Fig. 2. Membrane effect on impedance spectra (Nyquist plots) of the AMX
membrane in 0.01 M KCl, using the two-compartment electrolytic cell under
no stirring.
complex conductivity. This implies that the conductivity of the
DBLs is much lower than that of the ion-exchange membrane.
Thus, this layer was apt to insulate the ion-exchange membrane
to some extent, resulting in high impedance to current passing
through the ion-exchange membrane. It makes a loop as shown
in Fig. 1. No capacitive loop was observed for the impedance
spectrum of the IEM system in the absence of the ion-exchange
membrane, as shown in Fig. 2. It can be inferred that at low fre-
quencies the capacitive loop depicts the electrical responses to
interface phenomena (e.g., DBL, heterogeneous transport [24])
originating from the ion-exchange membrane.
3.2. Concentration effects on the impedance spectra
The loop was also affected by the concentration of the elec-
trolyte and the distance between the two reference Ag/AgCl
657
Fig. 3. Equivalent circuit for the IEM systems used in this study, which consists
of effects of the membrane immersed in solution (SM), heterogeneous transport
(HT), and diffusion boundary layer (DBL) and is represented by a resistance
and two parallel combinations of a resistance and a constant phase element.
electrodes. Assuming that an IEM system is described as the
equivalent circuit with a resistance (R1 ) in series and the com-
bination between a resistance (R2 ) and a capacitor (C2 ) in
parallel, the total impedance of the electrochemical system is
expressed as
 
R2 ωC2 R22
Zsys = R1 + − j , (2)
1 + ω2 R22 C22 1 + ω2 R22 C22
where ω = 2πf (f is the frequency). At a high frequency (1 
ω2 R22 C22 ), the total impedance became
 
∼ 1 1 ∼
Zsys = R1 + 2
−j = R1 , (3)
2
ω R2 C2 ωC 2
and at a low frequency (1  ω2 R22 C22 )
Zsys ∼
= (R1 + R2 ) − j ωC2 R22 ∼
= R1 + R2 . (4)
Previous researchers have used this type of the equivalent
circuit to describe a simple membrane system [16,19–21,29].
The series resistance and the parallel combination generally
correspond to the impedances arising from the solution and the
membrane, respectively. Thus, the membrane results in the for-
mation of a loop. However, IEM systems should be different. It
should be noted that the membrane is highly conductive in the
IEM systems (0.263–0.555 S/m2 according to the manufactur-
er’s estimate of the ion-exchange membrane conductance [26]).
Accordingly, the converged impedance in Eq. (3) includes the
resistance of the solution (R1 ) and membrane (R2 ) in the IEM
system at high frequencies. It is necessary to introduce other
electrical components into this equivalent circuit to fully de-
scribe the impedance arising from the IEM system. Park et
al. [24] reported the equivalent circuit for IEM systems. In this
study, the IEM system was analyzed using the equivalent circuit
shown in Fig. 3.
The beginning part of the capacitive loop in Fig. 4 is the
pure resistance of the solutions and the membrane at zero phase
angle. The shift of the capacitive loops toward lower real im-
pedance is due to the decrease in solution resistance. It is well
known that resistance of ion-exchange membranes should be
independent of solution concentration due to the electroneutral-
ity between the counterions and the fixed exchangeable sites. It
658 J.-S. Park et al. / Journal of Colloid and Interface Science 300 (2006) 655–662
Fig. 4. Concentration effect on impedance spectra (Nyquist plots) of the AMX
membrane in different concentrations, 0.005, 0.01, and 0.05 M KCl, using the
two-compartment electrolytic cell under no stirring.
Fig. 5. Distance effect of the reference electrode on impedance spectra (Nyquist
plots) of the AMX membrane in 0.01 M KCl using the two-compartment elec-
trolytic cell under no stirring where the two reference electrodes are positioned
at different distances from the central membrane: 0.25, 0.50, 0.75, and 1.00 cm.
is noted in Fig. 4 that a change in the diameter of loops also
occurs. The diameter decreases as the concentration of the elec-
trolyte increases. It is inferred that there is another impedance
to restrict ionic transport from the solution to the membrane
(i.e., heterogeneous transport). Fig. 5 shows that increasing the
distance between two reference electrodes causes a shift of the
capacitive loop similar to the effect of electrolyte concentration,
but the diameter of the loops does not change. This implies that
the heterogeneous transport is related to transport phenomena
that are dependent on electrolyte concentration and occur in the
heterogeneous phase (i.e., at the interfaces).
3.3. Equivalent circuit of the IEM systems
Capacitive or inductive loops are often generalized as de-
pressed semicircles by parallel R (resistance)–Q (constant
(a)
(b)
Fig. 6. Change of the imaginary impedance part of the IEM system with re-
spect to the reciprocal of frequency in the 0.01 M KCl solution: (a) the linear
relationship between ω−1 and −Im(Z) showing the capacitance characteristic
due to the dispersion of the membrane from 139 to 16.8 Hz and (b) the non-
linear relationship between ω−0.5 and −Im(Z) showing the fractional power
characteristic due to the dispersion of the DBL from 11 to 0.002 Hz.
phase element, CPE) circuits, expressing different relaxation
times in a complex electrochemical system. The CPE im-
pedance is obtained using the following equation in the im-
pedance representation,
−Im(Z) = −Q(ω) = Y0 (j ω)−n , (5)
where Y0 is the admittance (/sn ) and n the empirical parame-
ter (0  n  1). When n = 0.5, 1, and −1, the CPE is called a
Warburg impedance, which is related to a diffusion process ac-
cording to Fick’s law, a reactance by capacitance, and one by
inductance, respectively [21,28]. The power dependency of the
constant phase element with frequency can be observed in the
plot of 1/ω vs −Im(Z), in which the slope is the reciprocal of
total admittance by Eq. (5).
Figs. 6a and 6b show two different relaxations with fre-
quency. The linear relationship with ω−1 (see Fig. 6a) at fre-
quencies above 16.8 Hz indicates the pure capacitance charac-
teristic, according to Eq. (6). On the other hand, at frequencies
 J.-S. Park et al. / Journal of Colloid and Interface Science 300 (2006) 655–662 659

Table 2
Values of electrochemical parameters obtained by fitting the experimental im-
pedance data using the suggested equivalent circuit with concentration of the
KCl solution and existence of stirring
Solution Stir- RSM RHT QHT a RDBL QDBL a
(M KCl) ring ( cm ) ( cm ) Y ∗
2 2
n ( cm ) Y ∗
2
0,HT 0,DBL n
0.01 No 134.6 42.59 1.883 1.0 17.60 0.231 0.59
Yes 135.9 26.16 0.358 0.79 – – –

0.025 No 50.29 15.01 5.981 1.0 9.467 0.578 0.53

0.5 No 6.476 0.793 16.29 0.73 – – –
Yes 6.585 0.375 13.55 0.70 – – –
Note. Subscript SM, effects of the membrane immersed in solution; HT, effects
of heterogeneous transport; DBL, effects of diffusion boundary layer.
a The dimensions are S sn /cm2 ; if n = 1, they are F/cm2 .

(a)
be calculated by the equation

σi = σi∗ Ci A/d, (6)

(b)
Fig. 7. Impedance spectra of the AMX membrane in (a) 0.01 M KCl and
(b) 0.025 M KCl using the two-compartment electrolytic cell under no stirring.
The impedance data of a high-frequency region (above 139 Hz) were omitted
(the first capacitive and following inductive loops as shown in Fig. 1). The val-
ues in each figure represent the value of real impedance when the phase angle
is zero, that is, the resistive component.
below 16.8 Hz, the fractional power of the reactance was to
be approximately 0.5, implying the Warburg impedance, which
was proportional to ω−0.5 (see Fig. 6b). Thus, the capacitor and
diffusion characteristics might result from effects of the het-
erogeneous ionic transport and the DBL effect, respectively.
Hence, it can be suggested that the three components (solu-
tion/membrane, heterogeneous transport, and DBL) are electro-
chemically connected in series in the IEM system, as illustrated
in Fig. 3. Figs. 7a and 7b show good agreement between the
experimental data and the fitting results using the equivalent
circuit for the IEM systems in the 0.01 and 0.025 M KCl solu-
tions.
Table 2 summarizes the fitting results obtained using the
equivalent circuit for 0.01, 0.025, and 0.5 M KCl. The im-
pedance values (i.e., RSM ) at zero phase angle represent the
pure resistance of the solution and membrane, as discussed ear-
lier. In fact, the theoretical conductance of a KCl solution can
where σi∗ is the equivalent conductivity of the i solution (i.e.,
0.01035 S eq/m2 for KCl), Ci the concentration of the solu-
tion i, A the area of the membrane, and d the distance between
the two reference electrodes (1.4 mm in Fig. 7). The theoret-
ical resistances of the 0.01 and 0.025 M KCl including the
membrane (i.e., 0.398 S/cm2 ) were 135.3 and 54.10  cm2 , re-
spectively. The results agree well with the resistance (i.e., RSM )
as shown in Table 2.
We examined the effects of electrolyte concentration and
stirring and confirmed several basic phenomena of IEM sys-
tems. Stirring does not change the resistance of the solution
and the membrane itself. It diminishes the effects of DBLs in
the IEM systems in that the resistance and constant-phase ele-
ment of DBLs disappear. At a high concentration (0.5 M KCl),
construction of DBLs in the IEM system is insignificant. It is
inferred that a high-concentration feed into real applications
(e.g., electrodialysis) has advantages such as low resistance of
the feed, as well as insignificant concentration polarization. The
n value of ∼0.7 in the constant phase element provides an
independent verification of the absence of DBLs in the IEM
systems. Moreover, stirring decreases the resistance and the
capacitance of heterogeneous transport, which shows nonlin-
earity with the concentration. Based on this nonlinearity, it is
found that they are consistent with the concentration depen-
dence of the Debye equation, i.e., RHT ∝ concentration−1/2
and CHT (QHT when n = 1) ∝ concentration1/2 . This implies
that an electrical double layer formed at the surface of the ion-
exchange membrane dominantly influences dispersions of the
conductance and capacitance of the heterogeneous transport,
which is consistent with the previous study [25].
The capacitances in Table 2 are high compared to the di-
electric capacitance (e.g., 10−11 –10−12 F) of a solution or a
membrane. Smith [30] has calculated the dependence of the
α-dispersion (see Ref. [31]) occurring in unstirred layers on
membrane conductance. He found that the diffusion polariza-
tion due to difference between the ionic transport number in
unstirred layers and the electroosmosis one generated very
high reactance, and the dispersion by the polarization moved
660 J.-S. Park et al. / Journal of Colloid and Interface Science 300 (2006) 655–662

to lower frequencies as the membrane conductance decreased

in biological tissues. It should be noted that the α-dispersion
might be dominant in a wide frequency range for the IEM sys-
tem, which is highly conductive. In fact, the high capacitance
measured at low frequencies in this study was also affected
by the α-dispersion arising from the diffusion polarization due
to the difference in ionic transport number and electroosmosis,
which can be related not to dielectric properties but to phenom-
enological effects to make the impedance analyses complicate.
In this study, high phenomenological capacitance was princi-
pally manifested by the DBL and heterogeneous transport (see
Table 2).

3.4. Analogy between the impedance spectra and

current–voltage curves for the IEM systems
(a)

In general, the current–voltage curve of an ion-exchange

membrane shows an S shape with three distinctive regions:
the first region of approximately ohmic behavior, the second
region corresponding to the “plateau,” followed by the third re-
gion of rapid current increase [12,13,32]. The classical theory
of concentration polarization, based on the concepts of an un-
stirred layer and local electroneutrality, predicts true saturation
of the current–voltage curve. Barrangán and Ruíz-Bauzá [12]
proposed an equation applicable up to the limiting current den-
sity (LCD). From the data obtained in the ohmic region, the
limiting current density could be obtained for each experimen-
tal condition examined in this study by fitting the data to the
equation
 
2RT RT  
ΦT = Ro − I+ (2 ti ) + (2 ti )−1
F IL F
  (b)
1 + (I /IL )
× ln , (7) Fig. 8. I –V curves of the AMX membrane in (a) 0.01 M and (b) 0.025 M
1 − (I /IL ) KCl using the two-compartment electrolytic cell under no stirring and their
where ΦT is the electric potential difference between two fixed fittings by Eq. (9) when dc power is applied by a galvanostatically linear sweep
voltammetric method (current scan rate: 1 µA/s).
points placed on the two sides of the membrane in the bulk so-
lution, Ro the ohmic resistance before the polarization layers
are formed, R the universal gas constant, F the Faraday con- small due to the characteristics of dc power, in which electro-
stant, IL the limiting current density, I the current density, and migration by the counterions dominates in the IEM system.
ti (= tim − ti ) the difference of transport number between the Based on the classical concentration polarization theory, no
membrane and solution. current higher than the limiting current can be expected, since
Figs. 8a and 8b show the typical S shaped current–voltage the concentration near the membrane has reached zero. How-
curves and their fitting results for the 0.01 and 0.025 M KCl. ever, it has been recognized that in practice an overlimiting cur-
The slope in the first region was determined by the ohmic rent was obtained. Many researchers have reported that such an
behavior of the bulk solution, membrane, DBL, and heteroge- overlimiting current arises from co-ion leakage, water-splitting
neous transport involved between the two reference electrodes. phenomena, or electroosmotic convection. Recently, the most
The resistances in the first region of the I –V curves of the 0.01 promising theory for explaining the overlimiting current phe-
and 0.025 M KCl are 138.9 and 51.04  cm2 , respectively, nomena is so-called electroconvection (i.e., microfluid flows),
as shown in Fig. 8. It is noted that the first region resistances which is caused by heterogeneity of the surface of the ion-
are slightly higher than the pure resistances obtained from the exchange membranes [33]. According to the theory proposed
impedance measurements (see Figs. 7a and 7b). As discussed by Rubinstein [33], electric field lines in the solution layer adja-
earlier, the pure resistances obtained from the impedance spec- cent to the surface of the ion-exchange membranes comprising
tra include only the impedance arising from the solution and patches of varying ion conductance are distorted. The inter-
the membrane in the IEM systems. Thus, the difference is con- action of space charges with the electric field gives rise to a
tributed by the resistances arising from heterogeneous transport spatially inhomogeneous bulk force, which is bound to set in
and DBL in the IEM systems. However, the difference is very motion the fluid in the depletion diffusion layer. As a result, an
 J.-S. Park et al. / Journal of Colloid and Interface Science 300 (2006) 655–662
Table 3
Values of the fractional power (n) of each constant phase element obtained by fitting the experimental impedance data for the AMX membrane systems in the
absence of stirring using the suggested equivalent circuit as a function of current density
Region
Under LCD
Superimposed direct 0 6.37 19.1
current density (A/m2 )
SM n 1 1 1
HT n 1 1 1
DBL n 0.55 0.48 0.55
Note. SM, effects of the membrane immersed in solution; HT, effects of heterogeneous transport; DBL, effects of diffusion boundary layer.
Fig. 9. Dispersions of the conductance with frequency for the AMX an-
ion-exchange membrane system in 0.025 M KCl (limiting current density,
30.9 A/m2 ) as a function of current density (offset current indicated).
additional driving force for ion transport toward the membrane
surface is created, leading to the possibility of an overlimiting
current. The electroconvective effects over the LCD may dis-
turb the concentration gradient in the DBL, resulting in attenu-
ating concentration polarization. This was confirmed by direct
measurements of concentration gradients in the DBL using a
microelectrode [27]. Hence, it is expected that the electrocon-
vective effects may decrease the resistances arising from the
heterogeneous transport and DBL due to the disturbance of the
formation of a space charge region and the relieved concentra-
tion gradient, respectively.
Fig. 9 shows dispersions of the conductance with frequency
for the IEM systems in 0.025 M KCl. Each spectrum was mea-
sured using ac superimposed on the indicated offset dc. The
indicated dc can be categorized into three groups (i.e., un-
der LCD, LCD, and over LCD; in this study, the LCD was
30.9 A/m2 ). The spectrum measured at the LCD shows the dis-
tinguishable conductance change with frequency (see Fig. 9). It
is found that LCDs for IEM systems can also be identified using
impedance spectroscopy.
For the IEM systems over the LCD, dispersions of the con-
ductance in the frequency range from 100 to 0.001 Hz notably
support the theory on electroconvective effects. The constant
conductance increases with increasing current density. It is ex-
pected that increasing current density gives rise to pronounced
electroconvection adjacent to the membrane surface. In fact,
661
LCD Over LCD
30.9 33.1 34.4 35.6
1 1 1 1
1 0.94 0.98 1
0.55 0.49 0.75 0.79
the conductance variation measured at current density 38.2 and
39.5 A/m2 increased with decreasing frequency, nonsensically
based on the typical impedance behavior that conductance de-
creases with decreasing frequency due to the contribution of
increasing the resistance of interfacial layers to total impedance.
Note that the n value of ∼0.7 in the constant phase element
provides an independent verification of the presence of stirring
in the IEM systems as discussed earlier (see Table 2). Simi-
larly, it is observed in Table 3 that the n values obtained in
the over-LCD range change from ∼0.5 to ∼0.7. Such an in-
crease is associated with the presence of stirring. Hence, it is
believed that the increase may arise from mixing effects of the
fluid adjacent to the membrane by the microfluid flows (i.e.,
electroconvection).
4. Conclusions
It was confirmed that the IEM systems dominantly com-
prised three sublayers, the ion-exchange membrane immersed
in the solution, the electrical double layers (i.e., heterogeneous
transport effect), and the diffusion boundary layers, based on
several impedance spectra of the IEM systems and their fitting
results using the suggested equivalent circuit. Some experimen-
tal comparisons proved the analogy between impedance spectra
and current–voltage curves. It is found that impedance spec-
troscopy can be an alternative way to characterize IEM systems.
Especially, impedance spectroscopy makes possible the quan-
titative analyses of the IEM systems, which are separated into
main components using a well-proposed equivalent circuit. Fur-
thermore, the separate analyses make possible investigation of
microscopic phenomena (e.g., electroconvection). In this re-
gard, it is considered that the impedance study can provide
useful information on the interfacial phenomena of IEM sys-
tems.
Acknowledgment
This work was supported by the Core Technology Devel-
opment Program for Fuel Cells of the Ministry of Commerce,
Industry and Energy, Republic of Korea.
References
[1] E.G. Lee, S.-H. Moon, Y.K. Chang, I.-K. Yoo, H.N. Chang, J. Membr.
Sci. 145 (1998) 53.
[2] J.-H. Choi, S.-H. Kim, S.-H. Moon, Sep. Purif. Technol. 28 (2002) 69.
662 J.-S. Park et al. / Journal of Colloid and Interface Science 300 (2006) 655–662
[3] M.-S. Kang, S.-H. Moon, Y.-I. Park, K.-H. Lee, Sep. Sci. Technol. 37
(2002) 1789.
[4] M.-S. Kang, Y.-J. Choi, S.-H. Moon, J. Membr. Sci. 5317 (2002) 1.
[5] J.-S. Park, S.-H. Moon, Korean J. Chem. Eng. 19 (2002) 797.
[6] J.-H. Choi, S.-J. Oh, S.-H. Moon, J. Chem. Technol. Biotechnol. 77 (2002)
785.
[7] Y.-H. Kim, S.-H. Moon, J. Chem. Technol. Biotechnol. 76 (2001) 169.
[8] M.-S. Kang, K.-S. Yoo, S.-J. Oh, S.-H. Moon, J. Membr. Sci. 188 (2001)
61.
[9] S.-J. Oh, S.-H. Moon, T. Davis, J. Membr. Sci. 169 (2000) 95.
[10] S. Logette, C. Eysseric, G. Pourcelly, A. Lindheimer, C. Gavach,
J. Membr. Sci. 144 (1998) 259.
[11] H. Strathmann, in: W.S. Winston Ho, K.K. Sirkar (Eds.), Membrane Hand-
book, Van Nostrand Reinhold, New York, NY, 1992, pp. 230–245.
[12] V.M. Barrangán, C. Ruíz-Bauzá, J. Colloid Interface Sci. 205 (1998)
365.
[13] J.-H. Choi, Transport phenomena in ion-exchange membrane at under-
and over-limiting current regions, Ph.D. dissertation, Kwangju Institute
of Science and Technology (K-JIST), Republic of Korea, 2002.
[14] I. Rubinstein, B. Zaltzman, O. Kedem, J. Membr. Sci. 125 (1997) 17.
[15] A. Alcaraz, H. Holdik, T. Ruffing, P. Ramírez, S. Mafé, J. Membr. Sci. 150
(1998) 43.
[16] J.R. Smith, R. Simons, J. Weidenhaun, J. Membr. Sci. 140 (1998) 155.
[17] T.C. Chilcott, H.G.L. Coster, E.P. George, J. Membr. Sci. 108 (1995)
185.
[18] T. Osaki, A. Tanioka, J. Colloid Interface Sci. 253 (2002) 88.
[19] T. Osaki, A. Tanioka, J. Colloid Interface Sci. 253 (2002) 94.
[20] D.-J. Lee, Y.-K. Choi, S.-B. Lee, K.-H. Ahn, B.-R. Min, J. Membr.
Sci. 150 (1998) 9.
[21] J. Benavente, A. Cañas, M.J. Ariza, A.E. Lozano, J. de Abajo, Solid State
Ionics 145 (2001) 53.
[22] T.C. Chilcott, M. Chan, L. Gaedt, T. Nantawisarakul, A.G. Fane, H.G.L.
Coster, J. Membr. Sci. 195 (2002) 153.
[23] L. Gaedt, T.C. Chilcott, M. Chan, T. Nantawisarakul, A.G. Fane, H.G.L.
Coster, J. Membr. Sci. 195 (2002) 169.
[24] J.-S. Park, J.-H. Choi, K.-H. Yeon, S.-H. Moon, J. Colloid Interface Sci.
294 (2006) 129.
[25] J.-S. Park, T.C. Chilcott, H.G.L. Coster, S.-H. Moon, J. Membr. Sci. 246
(2005) 137.
[26] Tokuyama Soda Co. Ltd., Products brochure, 1997.
[27] J.-H. Choi, J.-S. Park, S.-H. Moon, J. Colloid Interface Sci. 251 (2002)
311.
[28] H.G.L. Coster, T.C. Chilcott, A.C.F. Coster, Bioelectrochem. Bioenerget.
40 (1996) 79.
[29] A.K. Jonscher, Dielectric Relaxation in Solid, Chelsea Dielectric Press,
London, 1983.
[30] J.R. Smith, Electrical characteristics of biological membranes in differ-
ent environments, Ph.D. dissertation, University of New South Wales,
Australia, 1977.
[31] H.P. Schwan, Advances in Biological and Medical Physics, vol. 5,
Academic Press, New York, 1957, pp. 147–209.
[32] M. Mulder, Basic Principles of Membrane Technology, Kluwer Acad-
emic, Dordrecht, Netherlands, 1996.
[33] I. Rubinstein, Phys. Fluids A 3 (1991) 2031.