3520 Electrophoresis 2002, 23, 3520–3527

Bohuslav Gaš1 Optimization of the high-frequency contactless

Jiří Zuska1
Pavel Coufal1 conductivity detector for capillary electrophoresis
Tom van de Goor2
Two constructions of the high-frequency contactless conductivity detector that are
1
Faculty of Science, fitted to the specific demands of capillary zone electrophoresis are described. The
Charles University, Prague, axial arrangement of the electrodes of the conductivity cell with two cylindrical elec-
Czech Republic trodes placed around the outer wall of the capillary column is used. We propose an
2
Agilent Laboratories,
equivalent electrical model of the axial contactless conductivity cell, which explains
Chemical & Biological Systems
Department, the features of its behavior including overshooting phenomena. We give the computer
Palo Alto, CA, USA numerical solution of the model enabling simulation of real experimental runs. The role
of many parameters can be evaluated in this way, such as the dimension of the sepa-
ration channel, dimension of the electrodes, length of the gap between electrodes,
influence of the shielding, etc. The conception of model allows its use for the optimiza-
tion of the construction of the conductivity cell, either in the cylindrical format or in the
microchip format. The ability of the high-frequency contactless conductivity detector is
demonstrated on separation of inorganic ions.

Keywords: Contactless conductivity detector / Inorganic ions EL 5158

1 Introduction detector (CCD) for isotachophoresis. A specific feature

The contact conductivity detection is commonly used in
ion chromatography, (see, e.g., a review by Rocklin [1]).
The contact conductivity cells have mostly metal elec-
trodes that are immersed directly in the electrolyte solu-
tion. The situation is much more complicated in capillary
electromigration separation methods like capillary iso-
tachophoresis. There the contact conductivity detection
is mostly performed as an on-column detection [2–6]
which means that the detection site is located inside of
the column. Due to the very small dimension of the sens-
ing electrodes and a high superimposed driving electrical
potential there is a big danger of their polarization or dete-
rioration by electrolysis products.
An elegant technical solution of the measurement of elec-
trical conductivity of electrolyte solutions is oscillometry
developed by Pungor [7]. In oscillometry the electrodes
are coupled to the electrolyte without a galvanic contact
by a capacitive coupling through the vessel wall. To get
better coupling, the frequency of the measuring voltage
is considerably higher than is usual in contact conductiv-
ity cells. Gaš et al. [8, 9] used this principle for the first time
in capillary electromigration separation methods and
described the high-frequency contactless conductivity
Correspondence: Dr. Bohuslav Gaš, Faculty of Science, Charles
University, Albertov 2030, CZ-128 40 Prague 2, Czech Republic
E-mail: gas@natur.cuni.cz
Fax: 1420-2-2491-9752
Abbreviations: CCD, contactless conductivitiy detector; HF,
high-frequency
of isotachophoresis is its ability of concentration and
stacking the individual analytes. This way the conductivity
of zones of individual analytes differ considerably. Such a
fact is convenient from a technical point of view as it puts
less demands on signal-to-noise ratio of the evaluating
electronics of the conductivity detector. On the other
hand, the boundaries between zones are mostly very
sharp and their width is often less then one tenth of a milli-
meter so the cell has to have a good axial resolution. The
conductivity cell in [8] is therefore composed of four elec-
trodes arranged radially around the capillary column. The
radial arrangement of electrodes results in a very good
axial resolution, about 0.25 mm. The construction of the
detector was later improved [10] and used for a precise
measurement of physicochemical constants by means of
electromigration methods [11], in isotachophoresis of
lanthanides [12] and in capillary zone electrophoresis of
anions [13–14].
Utilizing contactless conductivity detection in the mode of
capillary zone electrophoresis brings a different technical
problem to solve. The concentration of analytes in their
particular zones are here mostly very small so their pres-
ence in the background electrolyte causes only a very lit-
tle change in the conductivity. On the other hand, the
demands on axial resolution are not that critical as it
was in capillary isotachophoresis as widths of peaks in
standard capillary zone electrophoresis are mostly in the
range of millimeters. To gain more sensitivity of detection
in capillary zone electrophoresis, the axial arrangement of
the electrodes of the contatless cell was proposed by
Zemann et al. [15–17]. Two cylindrical electrodes placed

 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 0173-0835/02/2010–3520 $17.501.50/0
Electrophoresis 2002, 23, 3520–3527 High-frequency contactless conductivity detector 3521

2 Materials and methods

2.1 Construction of conductivity cells

We inspect two constructions of the axial conductivity

cell. Analogously as in some other constructions
described we used shielding between the electrodes to
prevent or at least to diminish too big direct capacitive
Figure 1. Mechanical construction of the contactless coupling by air between both electrodes. A mechanical
conductivity cell 1. 1, Capillary column with outer diam- construction of the conductivity cell 1 is depicted in
eter of 360 mm; 2, brass cylinder electrodes with outer di-
Fig. 1. Two brass cylinders form two electrodes that are
ameter of 4 mm and length of 5 mm; 3, wires for connect-
ing output of the generator and input of the I/U converter; coupled by the capacity of the capillary wall to the elec-
4, cylindrical Teflon insulation; 5, disk Teflon insulation; trolyte solution inside the capillary. The gap between the
6, disk copper shielding connected to the outer brass electrodes is 2.5 mm. A disc copper shielding between
shielding coat; 7, outer brass shielding coat. The gap be- electrodes has a hole with diameter of about the same as
tween electrodes 2 is determined by disk Teflon insulation the outer diameter of the capillary (, 0.4 mm) and almost
5 and is 2.5 mm. touches the outer capillary wall. After computer simula-
tion of the behavior of the cell we found it needs some
around the outer wall of the capillary column form such construction changes to improve its properties so we pro-
type of detector cell. The gap between the electrodes is posed the second construction, the conductivity cell 2,
the measuring active site of the detector and determines which is depicted in Fig. 2. Its fabricating is easier as it
its axial resolution. The electrodes are capacitively smartly uses the double Cu plated glass-epoxy board as
coupled to the solution through the capillary wall. Princi- a mechanical basis of the cell. The gap between elec-
pally the same conductivity cell with two axial cylindrical
electrodes was described by da Silva and do Lago [18–
20]. Chvojka et al. [21] proposed a dual photometric CCD
for capillary zone electrophoresis. The conductivity cell is
formed by two cylindrical electrodes. The gap between the
electrodes as the active site for conductometric detection
is utilized also for photometric detection: an optical fiber is
used to carry light perpendicularly to the capillary in the
gap. A simplification of the measuring electrodes was pro-
posed by Tůma et al. [22] to permit a simple exchange of
the separation capillary. Recently, Pumera et al. [23]

Miniaturization
described the axial CCD implemented in microchips.

What is interesting in almost all records of separation runs

using axial detection cells published so far [15–22] is the
shape of the analyte peaks, which often show over-
shoots. This phenomenon appears on the steeper edge
of such peaks, which have a tendency to form a triangular
shape. It is especially clearly visible, e.g., in Fig. 4 of [16],
Fig. 6 of [20], and Fig. 2 of [22]. It implies an impression
that the overshoots are consequences of the axial config-
uration of the conductivity cell as such phenomena were
never observed with the radial CCD [8, 9].
Figure 2. Mechanical construction of the contactless
In the present paper, we propose an equivalent electrical conductivity cell 2. 1, Glass-epoxy laminate with thick-
model of the axial CCD which completely explains the ness of 1 mm double plated with copper foil; 2, capillary
features of its behavior including overshooting phenom- column with outer diameter of 360 mm; 3, copper foil
shielding connected to ground; 4, brass electrodes with
ena. We give the solution of the model which enables us
bigger outer diameter of 2.5 mm and length of 2.5 mm;
to simulate real experimental runs. The conception of 5, base printed board with other electronics; 6, soldering
model allows its use for the optimization of the construc- of wires; 7, wires for connecting output of the generator
tion of the conductivity cell, either in the cylindrical format and input of the I/U converter. The gap between elec-
or in the microchip format. trodes 4 is 1.5 mm.
3522 B. Gaš et al.
Figure 3. Scheme of the CCD. Resistance of the electro-
lyte in the capillary is depicted as a long rectangle. Both
electrodes and shielding are coupled to the resistance of
the electrolyte via the capacity of the wall.
trodes is 1.5 mm. Unlike the cell 1, here the copper shield-
ing between electrodes has a bigger hole with diameter
of 1 mm for the capillary, so the distance between the
outer capillary wall and the shielding copper foil is greater
than in the cell 1. One of the electrodes in both construc-
tions is the transmitting electrode and is supplied with
a high-frequency voltage from an HF generator. The out-
put voltage of the generator is 50 Veff with frequency of
625 kHz. The second electrode is the receiving electrode
and is connected to the input of the current-to-voltage
(I/U) converter. It has the potential of virtual ground.
The signal from the transmitting electrode is “pumped”
through the capillary wall capacity into the electrolyte
solution, then passes through the solution in the gap of
the cell and is “sensed” by the receiving electrode. The
gap is the active site of the cell. The high-frequency
current flowing to the receiving electrode is dependent
on the conductivity of the electrolyte solution in the gap
between the electrodes. This is the measuring principle.
The electric current is converted by the I/U converter
to the voltage signal, rectified, and its absolute value is
filtered by a low-pass filter with a cut-off frequency of
5 Hz. The signal of such an arrangement is therefore
proportional to the rectified absolute value of the HF
electric current.
2.2 CCD
The conductivity cell together with the I/U converter and
further electronics processing the signal forms the CCD,
which is built into the cassette of the 3DCE instrument
(Agilent Technologies, Waldbronn, Germany). The printed
circuit board with the electronics also fitted into the cas-
sette so all the electronics and the conductivity cell is pre-
cisely thermostated together with the capillary column by
the air stream, which improves the stability of the signal
significantly. In addition, the electronics processing the
signal use an autocalibration principle [11], where a part
of the generator voltage is used as a reference signal,
Electrophoresis 2002, 23, 3520–3527
which controls the gain of the whole signal processing
and rectification. This way the fluctuation of many param-
eters like generator voltage, gain of the signal processing,
etc. are compensated for. When mounting the capillary in
the usual way into the cassette, it can be easily fed
through the hole of the conductivity cell and then fed
through the original alignment interface for the diode-
array detector (DAD). The signal (0–1 V) from the CCD
detector electronics is processed by an Agilent 35900E
dual channel A/D converter (Agilent Technologies). Both
the 3DCE instrument and the A/D converter are controlled
by the standard CE Chemstation software (Agilent Tech-
nologies). In this way the cassette serves as a dual-detec-
tor system, where the detectors, CCD and DAD, are at
the same capillary column. The distance between them
is 5.5 cm.
2.3 Chemicals
Lithium hydroxide monohydrate (for amino acid analysis)
was supplied by Merck (Darmstadt, Germany). Acetic
acid was provided by Lachema (Brno, Czech Republic).
All chemicals used as samples were taken from a current
laboratory stock and were of analytical reagent grade.
Water used for preparation of all solutions was purified
with Milli-Q water purification system (Millipore, Bedford,
MA, USA).
3 Results and discussion
3.1 Discrete model of the axial CCD
The function of the CCD can be schematically depicted
as shown in Fig. 3. It is obvious that the capacitive cou-
pling of the electrodes to the real electrolyte resistance in
the capillary forms a spatially distributed RC network.
Similarly, also capacitive coupling of the shielding with
the electrolyte is spatially distributed. To model such dis-
tributed structures we assume that resistivity of the elec-
trolyte in the capillary can be axially equidistantly divided
into a number of segments forming individual resistors.
All the resistors are connected in series and the i-th of
them has a resistance of Ri. A simplified equivalent circuit
of such a model using division of the cell into nine seg-
ments is shown in Fig. 4. The coupling of both electrodes
with the solution in the capillary is modeled by means of a
series of capacitors Ci. Similarly, the shielding is spatially
coupled with the electrolyte by capacitors Di. Further, U is
the voltage of the HF generator and I10 is the electric
current sensed by the I/U converter. A possible capacitive
crosstalk between both electrodes is modeled by the
capacity Cv.
Electrophoresis 2002, 23, 3520–3527
Such a circuit diagram allows a complete analysis of all
complex currents in all loops, when Cv, all Ri, Di, Ci, the
generator voltage U, and frequency f are known. The
absolute value of the current I10 is in fact the output cur-
rent of the cell sensed by the receiving electrode and pro-
cessed by the I/U converter. To perform the analysis of the
diagram we wrote the computer program HF_Cell. The
program enables to set individually a value of any of the
Ri, Ci, Di elements. Though in Fig. 4 only nine Ri elements
and ten Ci, Di elements are depicted for simplicity, the
number of segments can be arbitrary in the program, the
only limit is computation time and the capacity of the
computer memory available. For all calculations pre-
sented in this paper we divided the cell into 60 segments,
which means that the number of resistors Ri is 59 and the
number of Ci and Di capacitors is 60. The cell current pro-
cessed by the I/U converter is then I60. Any value of the Ci,
Di capacitors can even be set to zero at any position,
which means that there is no capacitive coupling at such
a position. In other words, this means that there is no
electrode at such position (when Ci = 0) or no shielding
(when Di = 0). This way any spatial configuration of the
cell can be analyzed.
Here, we will show the application of the model for the
cylindrical format corresponding with construction of the
cell 1. It has both electrodes of the same length Lc = 5 mm
and the length of the gap between electrodes is Lg =
2.5 mm. The length of the segment is then ls = (2Lc 1 Lg)/
60. Diameters of the capillary are 75/360 mm ID/OD, so
the inner radius is rin = 37.5 mm, the outer radius is rout =
180 mm. The i-th electrode segment is considered as a
cylindrical concentric condenser with capacity Ci with
the solution inside the capillary:
2pe0 er ls
Ci ˆ for i = 1, . . ., 24 or i = 37, . . ., 60
ln r out ln r in
Ci = 0 for i = 25, . . ., 36
where e0 is the permitivity of vacuum, er is the relative per-
mitivity of the silica capillary wall, e0 = 8.854610212 Fm21,
er = 4.4.
High-frequency contactless conductivity detector 3523
Figure 4. Equivalent circuit of
the cell.
The resistances Ri of an individual resistor representing
the solution in the i-th segment are:
1 ls
Ri ˆ for i = 1, . . ., 59 (3)
ki pr 2in
Here, ki is the conductivity of the electrolyte solution in the
i-th segment. The voltage of the generator is U = 50 V, the
frequency of the generator voltage is f = 625 kHz.
The HF_Cell program simulates electrical behavior of the
conductivity cell where it can perform, e.g., the following
tasks: (i) Calculation of the calibration curve showing the
dependence of the cell current I60 on conductivity of a
homogeneous solution in the cell. (ii) Calculation of the
cell current I60 if a certain conductivity profile (say a peak
or a boundary) passes through the CCD. The response on
the conductivity profile will represent the spatial behavior
of the cell.
3.2 Computer simulation of the CCD
The simulation reveals the role of many parameters and
dimensions of the cell, such as the dimensions of the cap-
illary, dimension of the electrodes, length of the gap be-
tween electrodes, influence of the shielding in the gap,
frequency of the generator voltage, etc. It enables to learn
what parameters significantly influence the cell features
and, consequently, it helps to design optimally the config-
uration of the cell to reach the desired function. This is a
complex task so to demonstrate the ability of the model
we will inspect a particular configuration of the cell 1 and
will show and discuss influence of only several param-
eters.
(1)
3.2.1 Influence of crosstalk between electrodes
(2)
The HF electric current sensed from the solution by the
receiving electrode and processed by the I/U converter
is dependent on the resistivity of the electrolyte solution
and has a certain phase shift to the voltage of the genera-
3524 B. Gaš et al. Electrophoresis 2002, 23, 3520–3527

3.2.2 Influence of shielding in the gap

The third configuration is closest to the construction

Figure 5. Calculated calibration curves of the CCD.
Dashed curve: theoretical response of the current I60 for
the cell 1 without crosstalk and without shielding. Dotted
curve: theoretical response of the current I60 for the cell 1
with crosstalk and without shielding. Solid curve: theoret-
ical response of the current I60 for the cell 1 with crosstalk
and with shielding.
tor. A crosstalk is a parasitic signal, which can pass
directly by air or other ways from the transmitting elec-
trode to the receiving electrode via the crosstalk capacity
Cv (see Fig. 3). Its phase shift is therefore 907. The parasi-
tic signal is vectorially added to the “proper” signal and
the cell then exhibits a different behavior.
of the cell 1. In addition to a crosstalk (Cv = 0.02 pF)
there is also a certain though very small capacitive
coupling between the shielding and the solution in the
gap. In the calculation it is simulated with non-zero
values of the Di capacitors: Di = 0.001 pF for i = 25,
. . ., 36, while Di = 0 for i = 1, . . ., 24 or for i = 37, . . .,
60. The simulation of such configuration is depicted by
the solid line in Fig. 5. It is worth noting the course
of the curve, which is even nonmonotonous for small
values of the conductivity in the range of about 0.01–
0.02 S/m, so the curve exhibits a local minimum. It indi-
cates that the shielding of the cell 1 being so close to
the electrolyte in the active site may have an unfavorable
influence on the calibration curve in the range of small
conductivities.
3.2.3 Influence of generator frequency
We have calculated how the signal of the conductivity cell
1 in the third configuration (with a crosstalk and shielding)
is influenced by the frequency of the generator. This
dependence in logarithmic scale is depicted in Fig. 6
(supposing that the cell is filled with solution with conduc-
tivity of 0.1 S/m). The course of the dependence shows
that the signal of the cell monotonously increases without
local minima or maxima in the range of 10 kHz – 10 MHz.
A steeper increase for frequencies higher than about
1 MHz is due to the crosstalk capacity Cv.

The first configuration we consider in simulation is a

hypothetical ideal situation without any crosstalk be-
tween electrodes (Cv = 0) and without any capacitive
coupling of the shielding to the electrolyte in the gap (all
Di = 0). The calibration curve resulting from modeling
such a configuration is depicted in Fig. 5. It is the dashed
curve, which shows the absolute value of the HF cell cur-
rent I60 flowing from the receiving electrode to the virtual
ground in its dependence on the conductivity k of a
homogeneous electrolyte solution filled in the whole cell
(ki = k for any i = 1, . . ., 59). It can be noticed that the cell
current reaches zero (the cell gives the zero signal) for
zero conductivity k = 0, as should be expected. The
second configuration is the case, where there is a cer-
tain small direct crosstalk between electrodes, Cv =
0.02 pF. Such direct capacitive coupling changes the
calibration curve considerably, see the dotted curve in
Fig. 5. Here, the current I60 is 3.93 mA for zero conductiv-
ity and, moreover, the curve has a different shape for low Figure 6. Calculated frequency dependence of the cur-
conductivities. rent I60 for the cell 1 with crosstalk and with shielding.
Electrophoresis 2002, 23, 3520–3527 High-frequency contactless conductivity detector 3525

3.2.4 Spatial behavior of the cell

The spatial resolution of the CCD is an important aspect.

Obviously it should not be worse than the narrowest
attainable peak. In reality, however, the situation is more
complicated. Due to the nonlinear nature of capillary zone
electrophoresis, peaks can be very often distorted by
electromigration dispersion and can attain a characteris-
tic triangular shape: one edge of the peak is dispersed
and broad while the other is steep. The question is what
will be a pattern detected by the cell when such a steep
feature, the width of which is much less than the gap,
passes with some velocity through the conductivity cell.
It is obvious that the detector signal cannot have the
same pattern as the conductivity profile of the peak and
will be distorted in some way due to the finite length Lg.
The model of the cell enables such evaluation of the
detector signal as all conductivities ki of the electrolyte
solution in all segments are input parameters and can be Figure 8. Experimental calibration curves of the CCD.
(u) cell 1; (j) cell 2.
set in an arbitrary way.

To simulate the triangular peak shapes we suppose that

the axial distribution of conductivity is as shown in Fig. 7
by the upper curve. Here, the background conductivity,
which is basically 0.1 S/m, rises up to 0.2 S/m at the top
of the triangle. (Such conductivity profile is far from the
nonmonotonous region at 0.01–0.02 S/m.) We further
suppose that this conductivity profile moves with the
velocity of 1 mm/s in the capillary through the cell 1. The
bottom curve in Fig. 7 shows the time record of the cell
current I60 calculated by the HF_Cell program during the
movement. The pattern of the detector signal is interest-
ing: the right-hand part of the triangle is less steep than
the conductivity profile due to the finite length of the
active site of the cell and, moreover, it is distorted by over-
shoots. It is obvious that the overshooting phenomenon is
an inherent feature of the axial arrangement of the con-
ductivity cell. It rises due to complicated vectorial sum of
complex currents flowing in the spatially distributed RC
network of the cell.

3.3 Real behavior of the CCD

Flushing with water and solutions of KCl with various con-

Figure 7. Calculated time response of the cell current
I60 on the triangular conductivity profile moving through
the cell 1 with velocity of 1 mm/s. Upper and right
axes should be applied for the axial conductivity profile,
bottom and left axes should be applied for the time
response.
ductivities were used for the measurement of a response
of the detector. The open points connected by a straight
line in Fig. 8 present the response or the calibration curve
of the cell 1. As can be noticed, the real CCD behavior is
well forecasted by simulation (cp. the experimental points
in Fig. 8 and the solid line in Fig. 5), as we have even found
here a certain nonmonotonicity at low concentrations with
a local minimum at about 0.01 S/m. According to the
simulation, the nonmonotonicity should be ascribed to
the combined influence of the crosstalk between elec-
trodes and the capacitive coupling of the shielding with
the electrolyte in the gap. This is why we proposed the
new cell 2 (see Fig. 2), where the shielding foil has less
capacitive coupling with the electrolyte in the gap, which
should improve behavior at low conductivities. This was
really found in the reality, as solid experimental points in
Fig. 8 belonging to the cell 2 show.
3526 B. Gaš et al.
Figure 9. Real electrophoretic run of the overloaded
Cs1 ion with the CCD with the cell 1. Sample: 0.5 mM
Cs1 accompanied with K1. Sampling: hydrodynamic,
125 mbar. Background electrolyte: 5 mM acetic acid,
2.5 mM LiOH, pH 4.7. Capillary: 75 mm ID, 360 mm OD,
total length of 80 cm, length to the detector 66 cm. Driving
voltage, 20 kV. Temperature, 307C.
We have never observed local maxima or minima (such as
were demonstrated by, e.g., Chvojka et al. [21]) in fre-
quency dependence of the detector signal in the range
of 100 kHz – 2 MHz. On the contrary, the results of simula-
tion shown in Fig. 6 encouraged us to use even a higher
frequency. In our nowadays detectors equipped with the
cell 2 we use frequency of 1.25 MHz and a gap of 1.5 mm
between electrodes, which gives us substantially better
signal/noise ratio and better axial resolution in compari-
son with the cell 1.
As pointed out in Section 1, many constructions of the
axial conductivity cell described so far (e.g., in [16, 20,
22]) exhibit overshooting phenomena when a steep con-
ductivity profile, the width of which is less than the gap
between electrodes, passes the cell. It was shown by
simulation (see Fig. 7) that this is an inherent feature of
the axial arrangement of the conductivity cell. Both con-
structions of the axial conductivity cell presented here
(Figs. 1 and 2) have the same behavior. Figure 9 is a con-
ductivity record (with the cell 1) of separation of metal
cations with a peak of Cs1 ions that is intentionally over-
loaded to reach a triangular peak shape with one steep
edge. The experimental signal has the same pattern and
exhibits exactly the same overshoots as previously calcu-
lated. Quite analogous overshooting phenomena are
encountered in records (not shown here) of isotacho-
Electrophoresis 2002, 23, 3520–3527
Figure 10. Real electrophoretic separation of cations with
the CCD with the cell 1. Sample: 10 mM Cs1 accompanied
with K1, Ca1, Na1. Sampling: hydrodynamic, 125 mbar.
Background electrolyte: 10 mM acetic acid, 5 mM LiOH,
pH 4.7, experimental conductivity 0 039 S/m. Other con-
ditions as in Fig. 9.
phoretic zones, the boundaries of which are generally
very sharp and are much thinner than the gap between
electrodes. All these observations support strongly a
validity of the proposed model.
The utilization of the cell 2 in practical applications is
demonstrated in another papers by us [24; Coufal et al.,
submitted)]. Here we will demonstrate only a sensitivity
attained with the cell 1. Figure 10 shows a separation of
metal cations as detected by the CCD. One of the peaks
is a peak of cesium ions, which is introduced from a
sample containing 10 mM Cs1 ions. Sampled amount of
Cs1 ions is 130 fmol. Peak height to baseline noise ratio
is 6:1.
In conclusion, the proposed model of the contactless cell
enabled us to understand influence of decisive param-
eters of the detector cell and make construction changes
in the design to improve its properties. This is a promising
feature of the model, which becomes especially useful for
the inspection of various configurations of the detector
cell in the microchip format. There, naturally, a great vari-
ety of shapes and configurations of the channels and
electrodes comes into play, so the theoretical prediction
of the properties of the CCD can spare a substantial
amount of experimental work.
Received June 10, 2002
Electrophoresis 2002, 23, 3520–3527
4 References
[1] Rocklin, R. D., J. Chromatogr. 1991, 546, 175–187.
[2] Everaerts, F. M., Verheggen, T. P. E. M., J. Chromatogr. 1972,
73, 193–210.
[3] Vacík, J., Zuska, J., J. Chromatogr. 1974, 91, 795–808.
[4] Kaniansky, D., Koval, M., Stankoviansky, S., J. Chromatogr.
1983, 267, 67–73.
[5] Haruki, T., Akiyama, J., Anal. Lett. 1973, 6, 985–992.
[6] Boček, P., Deml, M., Janák, J., J. Chromatogr. 1975, 106,
283–290.
[7] Pungor, E., Oscillometry and Conductometry, Pergamon
Press, Oxford 1965.
[8] Gaš, B., Demjaněnko, M., Vacík, J., J. Chromatogr. 1980,
192, 253–257.
[9] Gaš, B., Vacík, J., Chem. listy 1980, 74, 652–658.
[10] Vacík, J., Zuska, J., Muselasová, I., J. Chromatogr. 1985,
320, 233–240.
[11] Gaš, B., Zuska, J., Vacík, J., J. Chromatogr. 1989, 470, 69–78.
[12] Hirokawa, T., Mao, Q., Ikuta, N., Gaš, B., Chromatography
1998, 19, 9–14.
[13] Masár, M., Bodor, R., Kaniansky, D., J. Chromatogr. A 1999,
834, 179–188.
High-frequency contactless conductivity detector 3527
[14] Kaniansky, D., Zelenská, V., Masár, M., Iványi, F., Gazdíková,
Š., J. Chromatogr. A 1999, 844, 349–359.
[15] Zemann, A. J., Schnell, E., Volgger, D., Bonn, G. K., Anal.
Chem. 1998, 70, 563–567.
[16] Mayrhofer, K., Zemann, A. J., Schnell, E., Bonn, G. K., Anal.
Chem. 1999, 71, 3828–3833.
[17] Hilder, E. F., Zemann, A. J., Macka, M., Haddad, P. R.,
Electrophoresis 2001, 22, 1273–1281.
[18] da Silva, J. A. F., do Lago, C. L., Anal. Chem. 1999, 70,
4339–4343.
[19] da Silva, J. A. F., do Lago, C. L., Electrophoresis 2000, 21,
1405–1408.
[20] da Silva, J. A. F., Guzman N., do Lago, C. L., J. Chromatogr.
A 2002, 942, 249–258.
[21] Chvojka, T., Jelínek, I., Opekar, F., Štulík, K., Anal. Chim.
Acta 2001, 433, 13–21.
[22] Tůma, P., Opekar, F., Jelínek, I., Electroanalysis 2001, 13,
989–992.
[23] Pumera, M., Wang, J., Opekar, F., Jelínek, I., Feldman, J.,
Löwe, H., Hardt, S., Anal. Chem. 2002, 74, 1968–1971.
[24] Muzikář, J., van de Goor, T., Gaš, B., Kenndler, E., J. Chro-
matogr. A 2001, 924, 147–154.