Uploaded by Игорь Новожилов

The preparation of background electrolytes in capillary zone electrophoresis: Golden rules and pitfalls

advertisement
518
Electrophoresis 2003, 24, 518–535
Review
Jozef L. Beckers1
Petr Boček2
1
Eindhoven University
of Technology,
Dempartment
of Chemistry (SPO),
Eindhoven, The Netherlands
2
Institute of Analytical Chemistry,
Academy of Sciences
of the Czech Republic,
Brno, Czech Republic
The preparation of background electrolytes in
capillary zone electrophoresis: Golden rules and
pitfalls
In this article the methodology of the design of suitable background electrolytes (BGEs)
in capillary zone electrophoresis (CZE) is described. The principal aspects of the role
of a BGE in CZE are discussed with respect to an appropiate migration behavior of
analytes, including the transport of the electric current, the buffering of pH, the Joule
heat, the electro-endosmotic flow (EOF) and the principal migration and detection
modes. The impact of the composition of the BGE upon migration and detection is
discussed. It is shown that the total concentration of the BGE is a principal factor and
the adjustment of migrating analyte zones according to the Kohlrausch regulating
function (KRF) is the principal effect in most of the sample stacking techniques. The
number of co-ions and their properties are of key importance for peak shapes of the
analyte peaks and for the existence of system zones. The detection of UV-transparent
analytes may advanteously be done in the indirect UV mode, by using UV-absorbing
co-ions, however, both peaks and dips may be expected in the UV trace in case of
multiple co-ionic BGEs. Properties of BGEs can be predicted applying mathematical
models and it is shown that with SystCharts, predictions can be given concerning
the existence of system zones, detection modes and the peak shapes of analytes for
a given BGE. Practical examples of methodological considerations are given in the
design of suitable BGEs for four principal combinations of migration and detection
modes. The properties of the BGEs selected are exemplified with experimental results.
Golden rules are summarized for the preparation of suitable BGEs in CZE.
Keywords: Background electrolyte / Capillary zone electrophoresis / Review
Contents
1
2
2.1
2.2
3
3.1
3.2
3.3
3.4
3.5
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . .
Material and methods . . . . . . . . . . . . . . . . . . .
Instrumentation . . . . . . . . . . . . . . . . . . . . . . . .
Chemicals . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The role of a BGE . . . . . . . . . . . . . . . . . . . . . . .
Transport of the electric current . . . . . . . . . . .
pH of the BGE . . . . . . . . . . . . . . . . . . . . . . . . .
Joule heating . . . . . . . . . . . . . . . . . . . . . . . . . .
The electroosmotic flow . . . . . . . . . . . . . . . . .
Principal migration and detection modes . . . .
4
519
519
519
519
519
519
520
521
521
522
Correspondence: Prof. Petr Boček, Institute of Analytical Chemistry, Academy of Sciences of the Czech Republic, Veveři 97,
CZ-61142 Brno, Czech Republic
E-mail: bocek@iach.cz
Fax: +420-5-41212113
Abbreviations: KRF, Kohlrausch regulating function; SP, system
peak; SZ, system zone; TBA, tetrabutylammonium; ZE, zone
electrophoresis
 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
4.1
4.2
4.3
4.4
4.5
4.6
5
5.1
5.2
6
6.1
6.2
6.3
EL 5256
Impact of the composition of a BGE upon
migration and detection . . . . . . . . . . . . . . . . . .
The concentration of a BGE and stacking
phenomena . . . . . . . . . . . . . . . . . . . . . . . . . . .
Electrodispersion and peak shape . . . . . . . . .
System zones in CZE . . . . . . . . . . . . . . . . . . . .
The use of multiple co-ionic BGEs . . . . . . . . .
The use of weak multivalent ionic species
in BGEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Peaks and dips . . . . . . . . . . . . . . . . . . . . . . . . .
Prediction of properties of BGEs . . . . . . . . . .
General considerations . . . . . . . . . . . . . . . . . .
SystCharts . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Examples of BGEs . . . . . . . . . . . . . . . . . . . . . .
BGE for analysis of cations with direct UV
detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
BGE for analysis of cations with indirect UV
detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
BGE for analysis of anions with direct UV
detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
523
523
524
525
526
526
526
527
527
528
530
530
531
532
0173-0835/03/0302–518 $17.501.50/0
Electrophoresis 2003, 24, 518–535
7
8
BGE for analysis of anions with indirect UV
detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 533
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 533
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 534
1 Introduction
If electrophoresis is applied in order to solve an analytical
problem, important questions are “which electrophoretic
technique should be used” and “which electrolyte system
is needed”. Especially the latter question is still one of
the key problems in the application of electrophoresis in
analytical chemistry since its role is very complex. In this
article we will discuss basic principles of the choice of
background electrolytes (BGEs) in CZE. The BGE should
primarily provide an appropiate migration of the analytes,
i.e., a quantitative transport of their concrete zones along
the detector in a reasonable time. Moreover, extraneous
peak broadening and other migration interferences should
not be present. When the primary requirements are fulfilled, and, a suitable BGE is designed which provides a
suitable migration and detection behavior, the second
question is how to reach the required selectivity in order
to separate the analytes. The selectivity is not dealt with
in this contribution since it is beyond the scope and size
of this article and it may be found elsewhere [1]. This
article deals with the primary task and describes how to
design a BGE for a good migration in CZE. At first, we
describe the main purposes of the BGE, the requirements
which should be met and the principal CZE modes which
can be applied. After that, the main consequences of the
composition of the BGE will be discussed from peak and
dips in the indirect UV mode to the problems of system
zones (SZs). Then, we give a short overview of mathematical models, which can be used to get insight in the properties of BGEs and the SystChart has been selected as
an actual example. Finally, we describe the principles
how to choose a BGE for an application and also some
pitfalls are shown. All the above items are supplemented
with theoretical and experimental results. For the experiments we used simple standard solutions to show clearly
the actual individual effects and to give a telling description and explanation of them.
and detection Ld of 50.0 cm. The wavelength of the UV
detector was set at 214 nm. All experiments were carried
out at an operating temperature of 257C and applying
10 kV, unless otherwise stated. Sample introduction was
performed applying 5 s pressure injections at 3.46103 Pa
(0.5 psi). Data analysis was performed using laboratorywritten data analysis program CAESAR.
2.2 Chemicals
All chemicals were of analytical-reagent grade. Deionized
water was used for the preparation of all buffer and
sample solutions.
3 The role of a BGE
The main purposes of a BGE are to provide the transport
of electric current and the separation of the analytes. But
if an electric current passes through a BGE, some additional phenomena occur, such as the electroosmotic flow
(EOF) and the Joule heat, which play an important role in
electrophoretic processes. In this section we will discuss
the most important aspects of using a BGE.
3.1 Transport of the electric current
In zone electrophoresis (ZE) the whole system is filled with
a BGE, the purpose of which is to transport the electric
current if a voltage is applied over the system and to provide an electric field strength E, according to modified
Ohm’s law:
Es = j
2.1 Instrumentation
For all CZE experiments the P/ACE System 5000 HPCE
(Beckman-Coulter, Fullerton, CA, USA) was used, applying a Beckman eCAP capillary tubing (75 mm ID) with a
total length Lt of 57.0 cm and a distance between injection
(1)
where s refers to the specific conductivity (1/Om) and j
to the current density (A/m2).
In order that an electric current can pass through a solution if a voltage is applied over the system, the solution
must contain charged particles and, i.e., that the BGE
must consist of at least a cation and an anion. The specific conductivity s of a BGE, consisting of strong electrolytes, is given as:
sˆF
2 Materials and methods
519
n
X
c i jz i mi j
(2)
iˆ1
In this equation, F represents Faraday’s constant (96 490
C/mol) and ci (mol/m3), zi and mi (m2/Vs) refer to the concentrations, valences and mobilities of all ionic forms
present in the BGE. Mobilities are defined as the migration velocity per unit of electric field strength and can be
calculated from the ionic conductance or from electrophoretic data (see Section 3.4). In ZE the whole system
CE and CEC
6.4
BGEs in CZE
520
J. L. Beckers and P. Boček
Electrophoresis 2003, 24, 518–535
is filled with the same BGE, in order to keep parameters,
such as the conductivity, the E, temperature and pH, as
constant as possible. In that case any sample component
S migrates through the BGE system with a constant
electrophoretic velocity, vS, defined by:
vs = meff,sE
(3)
where meff,S represents the effective mobility of a sample
component related to a nonmoving solvent (see Section 3.4). The ionic species of the BGE with like charge
as the sample components S are called co-ions, and the
other ionic species are counterions.
3.2 pH of the BGE
The regulation of the pH is an important purpose of the
BGE in order to keep the migration velocity of weak electrolyte components and the velocity of the EOF constant.
In that way a stable and reproducible migration behavior
of the sample components can be obtained. The pH is
of key importance for all electromigration phenomena in
systems with weak electrolytes, i.e., weak bases, weak
acids and ampholytes. The concepts of ionic mobility
and effective mobility are used to describe the migration
behavior, where the ionic mobility relates to the electromigration of a fully ionized substance and the effective
mobility relates to the electromigration of a partially ionized
substance. The effective mobility of a weak monovalent
acid HA is given by [2]:
meff,A = aAmA
(4)
where aA represents the degree of dissociation of HA
and mA is the ionic mobility of A2. The effective mobility
of a weak acid base B is given by
meff,B = (1 2 aA)mB
(5)
where aB represents the degree of dissociation of BH1
and mB is the ionic mobility of BH1. Obviously, for weak
anionic and cationic ionic species the effective mobilities
strongly depend on their pK values related to the pH of
the BGE. As an example, in Fig. 1 the calculated relationships between effective mobilities and pH are given
for (solid lines) a component BH1 (pK = 4.0, mB =
3061029 m2/Vs), a component HA (pK = 5.0, mA =
23061029 m2/Vs), and for an amphiprotic component
(dashed line) containing both an acid group and a base
group with ionic mobilities of 23061029 m2/Vs and
3061029 m2/Vs, respectively, and pK values of 3 and 10.
It can be seen that the weak base (B), is practically fully
protonated (BH1) at pH , pKB 2 2 and its effective mobility equals then the ionic mobility. For pH . pHB 1 2, the
base (B) is practically neutral (nonprotonated) and its
effective mobility is zero. A weak acid (HA) has an effec-
Figure 1. Calculated relationships between effective
mobility and pH for (solid lines) a component BH1 (pK =
4.0, mB = 3061029 m2/Vs), a component HA (pK = 5.0,
mA = 23061029 m2/Vs) and for an amphiprotic component (dashed line) containing an acid group and a base
group with ionic mobilities of 23061029 m2/Vs and
3061029 m2/Vs, respectively, and pK values of 3 and 10.
tive mobility of zero for pH , pKA 2 2 and is practically
fully ionized at a pH . pHA 1 2. Ampholytes behave like
a weak base at low pH and like a weak acid at high pH.
From the practical point of view, the larger the effective
mobilities of substances, the faster their electrophoretic
migration, and, in a suitable experimental arrangement,
the shorter the time of analysis. However, it should be
emphasized, that even substances with zero effective
mobilities may move in the capillary due to the EOF,
(see Section 3.4), and this EOF is also strongly dependent
on the pH of the BGE used. Further in analogy with the
curves in Fig. 1, it is obvious that bases with high pK
values and acids with low pK values are practically fully
ionized at moderate pH values and for these components
the pH of the BGE is not so important for their migration
behavior. Concerning the separation of substances, the
pH of the BGE is of key importance, too, since the prerequisite for the separation of a couple of substances is that
these substances differ sufficiently in their effective mobilities. This topic is not, however, dealt with here since its
complexity is beyond the scope of this article, and, it can
be found elsewhere [1]. Concerning the pH of a BGE, one
aspect should be mentioned here. The BGE must have
some buffering capacity at the selected pH. It means
that one of the ionic species of the BGE should have a
buffering capacity, thus it must have a pK value near the
desired pH. It is of no importance whether this is a co-ion
or a counterion. It should be mentioned here that there are
also procedures to reach fast migration, good separation
and/or concentration by stacking where the BGE is not
kept constant but is subjected to dynamic changes [3].
Electrophoresis 2003, 24, 518–535
BGEs in CZE
Here, usually a pH change, having the character of a step,
gradient, or pulse is induced at one side of the capillary
and this change migrates along the capillary due to the
applied electric field strength [4–10]. Thus, it brings dynamic changes in the effective mobilities of sample components as well as of the BGE constituents as it meets
them during its migration. In this way very selective
separations and other migration effects may be reached.
The formation of pH changes at one side of a capillary is
easily done either by hydrodynamic modification of the
composition of an electrolyte in an appropriate electrode
chamber [4], or by a sophisticated automated electromigration three-pole-column system [5, 6]. In this system,
the capillary is equipped with two electrode chambers at
one end, with electrodes of the same polarity, and, by
controlling the individual electric currents passing through
these individual electrode chambers filled with mutually
different electrolytes, one can control the fluxes of electromigrating ionic species (including H1 or OH2) passing
through the capillary. It seems that these principles
attracted recently new attention [11].
3.3 Joule heating
If a voltage is applied across the capillary and an electric
current passes through the capillary, the temperature of
the solution in the capillary increases due to the dissipation of electric energy, i.e., production of the Joule heat.
The Joule heat is conducted through the capillary wall
into the thermostating medium (circulating air or liquid
coolant) surrounding the capillary. It has already been
shown that radial temperature profiles in the solution are
negligible in comparison to the temperature drop across
the capillary wall [12]. Further it has been shown that the
elevation of the mean temperature nT = T2To, where T is
the mean temperature of the solution and To is the temperature of the coolant, is proportional to the Joule heat
produced per unit length of the capillary used [13–15],
nT = QEI
(6)
where I represents the electric current (A) and the quotient of proportionality Q depends first of all upon the
coolant used (air or liquid) and for present commercial
instruments with liquid cooling and the usual capillaries
of 50 up to 100 mm ID and 360 mm OD, it amounts
1–2 Km/W.
Obviously, the power used in the capillary should not
exceed 0.5 W/m if the elevation of the temperature should
be under 1 K [15]. It is a general rule that the speed of
the analysis is proportional to E (V/m), and, the Joule
heat is proportional to EI (W/m). Hence, low-conductive
BGEs, which bring low currents, are recommended. If
521
the specific conductance of the BGE is too high, the heat
production will increase, and increasing temperature of
the system brings extra peak broadening and a decrease
in resolution. Recently, on-line monitoring of the temperature in CZE capillaries [16] reported on values for
nT up to 20 K.
3.4 The electroosmotic flow
If a BGE is in contact with the inner capillary wall, an electrical double layer is formed. If a voltage is applied across
the separation capillary the BGE solution migrates with a
velocity directly proportional to the electric field strength
E, the so-called EOF, and analogous to the ionic mobility,
the mobility of the EOF, mEOF, can be defined as:
mEOF ˆ
vEOF
E
(7)
The mEOF mainly depends on pH, composition and concentration of the BGE and for silica capillaries the mEOF
has a positive value, i.e., the EOF shows a cathodic
migration. The consequence is that the apparent migration velocity of a species S, vapp, S, which is related to the
capillary as a nonmoving frame, can be expressed as:
vapp; S ˆ
V
Ld
ˆ mapp; S E ˆ meff; S ‡ mEOF
tS
Lt
(8)
where Ld is the distance between injection and detection,
Lt is the total length of the capillary, V is the voltage
applied across the capillary and tS is the migration time
of an ionic species S.
From Eqs. (7) and (8) it can be concluded that the mEOF
can be calculated from the migration time of the EOF by:
mEOF ˆ
Ld Lt
VtEOF
(9)
and the effective mobility of an ionic species can be calculated from its migration time tS by:
meff; S ˆ mapp; S
mEOF ˆ
Ld Lt
VtS
Ld Lt
VtEOF
(10)
A benefit of the EOF in bare fused-silica capillaries is,
that it causes movement of cations, neutral components,
anions (if the absolute values of their anionic mobilities
are smaller than that of the mEOF) and even a micellar
phase into the same cathodic direction, and, thus, brings
them to pass the detector. Obviously, for cathodic EOF,
mEOF . 0, and the EOF accelerates cations. Anions with
umeffu , mEOF are driven cathodic and may be detected
in cathodic runs. Anions with umeffu . mEOF do not reach
the detector at all. Because the mEOF affects the migration behavior of analytes, it is often important to control
the EOF, i.e., we want to change or even to reverse the
522
J. L. Beckers and P. Boček
mEOF. For that purpose surfactants can be added to the
BGE. To show their effect on the EOF, the measured relationships between mEOF and the pH of BGEs are given in
Fig. 2A, for (a) BGEs prepared from a solution of 0.01 M
Tris adjusted to the desired pH by adding acetic acid
(dashed line for the addition of formic acid) and (b) the
same BGEs under the addition of 561024 M CTAB. For
the determination of the mEOF, a 5 s pressure injection of
a solution of mesithyloxide (30 mL/100 mL) was applied.
It can be seen that the values of the mEOF vary between
ca. 20–10061029 m2/Vs. The mEOF is strongly increasing
at high pH owing to the twofold effect of increasing pH
Electrophoresis 2003, 24, 518–535
and decreasing ionic strength. For BGEs at very high pH of
11–12, mEOF values up to 110–12061029 m2/Vs can be
reached. For BGEs with the addition of CTAB, the mEOF
varies between ca. 250 to 210061029 m2/Vs. In order to
show the influence of the concentration of surfactants
on the mEOF, we give in Fig. 2B the relationship between
mEOF and the concentration of the surfactant CTAB in
several BGEs at various pHs. Besides the concentration
of the surfactant, also the composition of the BGE is
important [17]. It should be mentioned here that the
reversal of the EOF by adding CTAB may often be
accompanied by the existence of a micellar phase due
to its low critical micellar concentration (cmc).
3.5 Principal migration and detection modes
Figure 2. (A) Relationship between measured mobilities
of the EOF and the pH of (a) diverse BGEs, (b) the same
BGEs with the addition of 561024 M CTAB. (B) Measured
relationships between mobilities of the EOF vs. the concentrations CTAB in BGEs consisting of (d) 0.01 M Tris
adjusted to a pH of 4.5 by adding acetic acid; (s) 0.02 M
MES adjusted to pH 6 by adding 4 M NaOH; (n) 0.01 M Tris
adjusted to pH 7.5 by adding acetic acid; and (m) 0.005 M
(NH4)2HPO4 adjusted to pH 9.0 by adding 1 M NaOH.
Applied Voltage, 15 kV.
There are two principal migration and two principal detection modes in ZE. Often this is not clearly indicated in
published papers and, instead, practical names are used
which may be misleading. In order to avoid misunderstandings, we will characterize the above modes. With
respect to the polarity of a voltage across the capillary,
two modes can be recognized. The cathodic mode: In
this mode the anode is placed at the inlet and the cathode
at the outlet. Because generally the mEOF is into the direction of the cathode, cations can be determined and even
anions if the absolute values of their mobilities are smaller
than that of the mEOF. This mode is often denoted by the
terms: normal polarity, direct polarity or positive (1) voltage applied. The anodic mode: In this mode the cathode
is placed at the inlet and the anode at the outlet. Because
in fused-silica capillaries, the EOF mostly migrates into the
direction of the cathode, only such anionic species can
reach the detector, which have the absolute values of
their mobilities higher than mEOF. In order to detect anions
with a lower mobility, the EOF has to be suppressed or
even reversed. This can be done by coating the inner surface of the capillary with an electroosmotically nonactive
layer or by a dynamic coating, i.e., by adding surfactants
to the BGE, such as FC 127 or CTAB. The anodic mode is,
in laboratory practice, often called: reversed mode or
negative-voltage mode. Sometimes, independently of the
polarity used and types of analytes, the terms upstream
and downstream are used and they denote the migration
of analytes against or along the direction of the EOF.
With respect to the detection we can distinguish between
the direct and indirect mode. In the direct mode, the
detector monitors directly the sample zones and the
BGE serves as a blank. In the indirect mode, the detector
monitors a suitable ionic component of the BGE and
substitution of this component by ionic analytes. It should
be mentioned that according to KRF, neutral components
of the BGE are not replaced by ionic analytes. In case of
Electrophoresis 2003, 24, 518–535
a UV detector, the indirect UV mode is applied for the
detection of UV-transparent ions and a BGE is used with
UV-absorbing properties. Generally, a co-ion is chosen,
which is UV-absorbing [18].
4 Impact of the composition of a BGE upon
migration and detection
In Section 3 the main purposes of a BGE are discussed
and the effect of the composition of the BGE on migration
behavior is described. Sometimes, some other phenomena can occur and they can facilitate or disturb the separation process. In this section some phenomena are described, affecting the electrophoretic separation process.
4.1 The concentration of a BGE and stacking
phenomena
The concentration of the BGE can play an important
part in the resolution of a separation in electrophoresis,
dependent on several aspects. If a constant voltage is
applied over the capillary, the electric current will increase
for higher concentrations of the BGE and this can cause
extra peak broadening due to a strong increase in the
Joule heat. Further a higher concentration of the BGE
gives a lower mobility of the EOF and by this, fast and
medium fast anions cannot be determined in the cathodic
mode. The mobilities of analytes are affected by the concentration of the BGE according to the Debye-HückelOnsager effect [19] and sometimes this can improve the
separation, especially for multivalent analytes. Often there
is a strong attraction force between positively charged
analytes and the negatively charged wall of the silica
capillary resulting in broad tailing sample zones. A high
concentration of the BGE can decrease this adhesive
behavior of the sample ions by a competitive adsorption
of the co-ions of the BGE, thus increasing the resolution.
The counterions of the BGE can show attraction forces
with analytes and this complexation can be stronger at a
high concentration of the BGE. There are, however, phenomena which are of paramount importance for migration
behavior of analytes and are directly dependent on the
concentration of the BGE. These phenomena are stemming from the existence of KRF. The KRF [20] (for recent
paper on KRF see [21]) prescribes that the numerical
value o (molVs/m5), for fully ionized monovalent ionic constituents, defined as:
X ci
(11)
oˆ
jmi j
i
is locally invariant in time, i.e., once given for any position
along the capillary, prior to applying the voltage, it does
not change with time during the electrophoretic run. In
BGEs in CZE
523
our case, the initial value of o is given by the BGE used
to fill the capillary prior to the electrophoretic run. During
the electrophoretic process, sample ions migrate into the
separation capillary, substitute the co-ions of the BGE
and their concentrations adjust to the o value of the BGE
electrolyte. This process is called adjustment of the concentrations and belongs to the well-known characteristics of ITP [22–25]. In case that an analyte is introduced
at low concentration as a broad sample pulse, the adjustment of its concentration to that of the original BGE
manifests itself as the accumulation of this analyte into a
narrow band. This process is called stacking [26]. In many
cases, the term stacking is supplemented with specific
adjectives, e.g., field-enhanced sample stacking (FESS),
field-amplified sample stacking (FASS) or large-volume
sample stacking (LVSS) [26]. However, the principle is
always the same, the adjustment of concentrations and
it is an intrinsic property of CE. Sometimes, one of the
sample ionic components is present at a very high
concentration, through which a complete separation is
impossible. Higher concentrations of the BGE can give a
better stacking process, resulting in a higher resolution.
In order to elucidate this stacking mechanism, we give
in Fig. 3 the simulated cationic concentration profiles at
four different times, for electrophoresis of a sample of
0.001 M ephedrine. The model BGE consists of 0.01 M of
Figure 3. Simulated cationic concentration profiles for an
electrophoretic process of a sample of 0.001 M ephedrine
(Eph, solid line) and 0.01 M of a model co-ion (dashed
line). (a) Original concentration profile prior to the electrophoretic process. (b) After a short time the ephedrine zone
is partially replaced. (c) Ephedrine is just totally stacked.
(d) Ephedrine migrates in a zone electrophoretic way.
After the ephedrine zone has passed, the BGE zone is
restored.
524
J. L. Beckers and P. Boček
Electrophoresis 2003, 24, 518–535
a co-ion and counterion with equal assumed mobilites of
2661029 m2/Vs. We applied all ionic mobilities at infinite
dilution in the simulation. In Fig. 3a the initial concentration profiles are given. The original ephedrine sample
solution is 0.001 M (solid line) and the concentration of
the co-ion is 0.01 M (dashed line). After a short time (see
Fig. 3b), the front of the sample cation passes through the
original stationary boundary (between original sample
zone and BGE), and, the sample is concentrated owing
to the adaptation to the o value of the BGE. Simultaneously, at the rear boundary, the sample is partially
replaced by the co-ion. Its concentration is adjusted to
the value of o, valid in the original sample zone, according
to KRF and is a little bit lower than that of ephedrine, because its mobility is lower than that of ephedrine (see
Table 1). In Fig. 3c all amount of ephedrine is just stacked,
and, although the destacking has already started, the
concentrated ephedrine is practically 0.01 M in its zone
according to the KRF. It should be stressed here that the
sample is stacked only after it has passed its own original
stationary front boundary, i.e., the sample stack is formed
only at a site which was previously occupied by the BGE.
In Fig. 3d we see that behind the migrating ephedrine
zone, the concentration of the co-ion of the BGE has
been restored at its original value of 0.01 M. There are
four main consequences of KRF viz., (i) if a sample solution is introduced with an o value different from that of
Table 1. List of cations and anions often used for the preparation of BGEs
m6109 (m2/Vs)
pK
UV-absorbing
Creatinine
Ephedrine
Histamine
37.2
28.3
65
32
4.828
10.1
9.95
6.0
1
1
1
1
Histidine
Imidazole
29.6
50.4
6.04
7.0
1
1
Lithium
Potassium
Sodium
Tris
40.1
76.2
51.9
29.5
14
14
14
8.08
2
2
2
2
Benzoic acid
Phthalic acid
233.6
228.1
252.9
4.203
2.81
5.1
1
1
1
Phenyl acetic acid
Sulfanilic acid
Sulphamic acid
231.7
233.7
250.3
4.405
3.227
22
1
1
1
Acetic acid
Formic acid
Hydrochloric acid
Phosphoric acid
242.4
256.6
279.1
234.6
261.4
271.5
4.756
3.752
22
2.16
7.21
12.67
2
2
2
2
2
the BGE this initial o value remains valid at the site of
injection during the whole experiment; (ii) the site where
the original sample was injected, shows always stationary boundaries; (iii) diluted sample zones will be concentrated and stacked at the beginning of this electromigration, i.e., when they cross their original front boundary;
and (iv) when the co-ion of the BGE starts to penetrate
into the sample zone, the destacking process starts, i.e.,
the sample zone is going to migrate in a zone electrophoretic way. Of course, when an EOF is present, the
injection site migrates with the velocity of the EOF through
the capillary and this is the reason why the EOF can be
observed as a peak or as a dip in the baseline. If the initial
o value of the sample is lower than that of the BGE a
“waterdip” is visible, if this o value is higher a peak can
be observed. Further it can be seen that the boundaries
marginating the sample zones are moving and can be
electrophoretically stabilised or not stabilized [26]. The
migrating ephedrine zone is a moving zone with a stabilized rear boundary whereas the front boundary has a
nonstabilized fronting character, although the peak shape
is practically symmetrical because of the small difference
between the mobilities of ephedrine and co-ion.
4.2 Electrodispersion and peak shape
Zone boundaries can be stabilized or nonstabilized [26]
and generally this behavior is regulated by Ohm’s law.
If an ionic species leaves its zone in the direction of its
migration, and it reaches a site with higher electrical conductivity, i.e., a locally lower E, its velocity decreases and
it will be overtaken by its own zone. In that case, we have
a stabilized zone boundary. If it reaches a site with higher
E, its velocity increases and a nonstabilized diffuse
boundary is the result. In that case, the electrophoretic
migration contributes to the broadening of the migrating
zone and this is called electrodispersion. The electrodispersive effect is generally larger than peak broadening by diffusion. The degree of electrodispersion is related
to the differences between the mobilities of sample ion
and co-ion. For a large difference between these mobilities broad, asymmetric tri-angled peaks are obtained. As
a rule of thumb, it can be stated that a peak is fronting, if
the mobility of an analyte is higher than that of the co-ion,
otherwise it is tailing. If the sample solution contains only
a few closely spaced analytes of interest, it is possible to
optimize the BGE by mobility matching. Here, a co-ion is
chosen with a mobility close to that of the analytes of
interest. In that case, the electrodispersion is minimum
and the peaks are sharp and symmetric [18]. The above
rules are fully correct for strong electrolyte systems. For
weak electrolyte systems the rules are for orientation
only, since a more sophisticated approach is then needed,
see Section 5.
Electrophoresis 2003, 24, 518–535
4.3 System zones in CZE
In CZE there are often zones, which migrate through an
electrophoretic system and which do not contain any
of the analytes and contain only ionic species of the
BGE. These zones migrate with a specific mobility
through the electrophoretic system and are often called
“eigen zones”, “system zones” or “system peaks” [27–
32]. SZs play a very important role in the electrophoretic
process, and, their behavior is frequently decisive for the
failure or success of the analysis. The composition of
these zones differ from the composition of the BGE with
regard to the concentrations of the ionic species and the
pH. One of the properties of SZs is that they interact with
sample zones with a mobility close to that of the SZ and
they deform the peak shape of the sample components
and disturb the separation process by a strong peak
broadening. Important questions are (i) when do SZs
exist, (ii) what are their mobilities and (iii) what is their
effect on the electrophoretic process. At this very moment
we know that SZs are formed (1) in BGEs containing two
co-ions, (2) in BGEs at very high or very low pH, whereby
OH2 or H1 ions act as a second co-ion and (3) BGEs
containing multivalent weak acids and bases at a pH
round a pK value, in which two ionic forms of that component act as two co-ions. For the first case, when BGEs
consisting of two co-ionic species are used, the mobilities
of the SZs are lying in-between the mobilities of these
two co-ions. The actual mobility of a SZ, mSZ, is closer
to the mobility of the co-ion which is present in minority.
For all other cases, no rules of thumb can be given, and
the mSZ must be calculated with mathematical models
or simulation programs (see Section 5). Generally, the
BGEs in CZE
525
mSZ increases for lower pH and higher pH of the BGEs,
respectively. BGEs containing weak multivalent cations
and anions show a maximum mSZ around the pK values
of the multivalent ionic species, where two ionic forms of
that ionic species are present. In that case, both ionic
forms of that multivalent ionic species seem to act like
two co-ions.
In order to give an impression of the mSZ, we give in
Fig. 4A the calculated relationship (solid line) and
measured values of the mSZ for several BGEs containing the two co-ionic species (m) potassium and histidine and (n) potassium and imidazole. The total concentration of the two co-ions was always 0.01 M and
the composition was expressed as molar ratio (%) of
potassium. The pH was adjusted to pH 5 by adding
acetic acid in all cases. For a concentration of the coion potassium approaching zero, the mSZ is equal to
the mobility of potassium, whereas for a potassium
concentration approaching 100%, the mSZ is equal to
the mobility of the minor component (imidazole or
histidine). In Fig. 4B calculated values (solid lines) and
measured values (m, n) are given for the mSZ of SZs
present in electropherograms applying BGEs consisting of (a) 0.01 M histidine and (b) 0.006 M histidine adjusted
to different pH by adding formic acid. They always show
an increase of mSZ for low pH of the BGE. It is clear that
SZs are only visible in electropherograms as system
peaks (SPs)/dips if at least one of the ions of the BGE
is UV-absorbing. But even if SZs are invisible, their
effects on peak shapes of sample components are still
visible in the electropherograms. Some examples are
given in Section 6.
Figure 4. (A) Calculated values
(solid line) and measured values
of the mSZ of system zones for
different compositions of BGEs
containing the two co-ionic species (m) potassium and histidine
and (n) potassium and imidazole. The total concentration of
the two co-ions was always
0.01 M and the composition is
expressed as molar ratio (%) of
potassium. The pH was adjusted
to pH 5 by adding acetic acid.
(B) Calculated values (solid lines)
and measured values (m, n) for
the mSZ of SZs in electropherograms applying BGEs consisting
of (a) 0.01 M histidine and (b)
0.006 M histidine adjusted to different pH by adding formic acid.
526
J. L. Beckers and P. Boček
4.4 The use of multiple co-ionic BGEs
It has already been mentioned that a strong peak broadening owing to electrodispersion is obtained if there is a
big difference between the mobilities of sample components and co-ions of the BGE. Therefore, it might be concluded that the use of a multiple co-ionic BGE would be
favorable for the determination of sample solutions containing several ionic species with a diversity of mobilities,
because then there are multiple centers of symmetry [29,
33], in which the electrodispersion is minimum. Sorry to
say, it is just only half of the truth. If a BGE contains n
co-ions, n-1 SZs are present, which interact with sample
zones having equal mobilities and deform their peak
shapes. The mobility of SZs is already shown in Figs. 4A
and B for BGEs containing two co-ions and for BGEs
at extreme pHs. To demonstrate this phenomenon, we
separated an equimolar sample mixture of 0.0005 M of
K1, Na1, Li1 and Tris1 ions, applying a BGE consisting of
(a) 0.01 M histidine and (b) a mixture of 0.005 M histidine
and 0.005 M imidazole, both adjusted to pH 5 by adding
acetic acid. The electropherograms for 5 s pressure injections are shown in Fig. 5. For the BGE containing a single
co-ion histidine, the Tris1 zone is very sharp because its
mobility equals that of histidine. All other zones are fronting and show an increased peak broadening owing to the
larger differences between the mobilities of sample ions
and histidine. Applying a BGE containing the two co-ions
histidine and imidazole, both the Na1 zone and the Tris1
Electrophoresis 2003, 24, 518–535
zone are sharp, because their mobilities correspond with
that of the co-ion imidazole and histidine, respectively.
A small SZ is visible in the electropherogram 5b just
behind the lithium dip. The calculated effective mobility
of the SZ according to the SystChart of the BGE is ca.
3361029 m2/Vs (see Section 5 for the concept SystChart)
and this value corresponds quite well with the value of
ca. 3061029 m2/Vs obtained from Peakmaster [34]. The
SZ will be more pronounced if its mobility is closer to
that of a sample peak and if only one of the co-ions is
UV-absorbing. The presence of more counterions has
no implications for the presence of SZs.
4.5 The use of weak multivalent ionic species
in BGEs
Although weak multivalent ionic species, present in a
solution in different ionic forms in fast dynamic equilibrium
with each other, generally migrate as a uniform single
zone [15] with an average mobility defined by:
meff ˆ
n
X
iˆ0
a i mi ˆ
n
X
ci
iˆ0
ct
mi
(12)
in which the subscript i refers to all ionic forms and ct to
the total concentration of the ionic species, they sometimes seem to behave as a mixture of independent ionic
species. In a BGE, prepared at a pH around a pK of a
weak multivalent ionic species, two ionic forms of that
ionic species are present simultaneously and SZs often
exist [35–37]. The two ionic forms seem to act as two
co-ions. By this effect, BGEs with multivalent ions are
not useful for the complete mobility window of the analytes. In Section 6, we give examples of BGEs containing
weak multivalent cationic and anionic species and present calculated mSZ values for these BGEs. In order to
avoid the influence of SZs, it is advisable to use BGEs
in its safe region. Unsafe regions can be defined as the
mobility window of analytes with values of mSZ 6 10% [38].
4.6 Peaks and dips
Figure 5. Separation of an equimolar sample mixture of
0.0005 M of K1, Na1, Li1 and Tris1 ions applying a BGE
consisting of (a) 0.01 M histidine and (b) a mixture of
0.005 M histidine and 0.005 M imidazole, both adjusted to
pH 5 by adding acetic acid. For further information see
Section 4.4.
Applying a UV detector, the detector signal depends on
the change in concentration of a UV-absorbing component (chromophore). For UV-absorbing sample components using a UV-transparant BGE, all components are
visible in electropherograms as peaks. For the detection
of UV-transparant sample components, the indirect UV
mode has to be applied. With indirect detection, a co-ion
is chosen as chromophore. The sample components displace this chromophore present in the BGE and their
zones are always visible as dips in the electropherogram.
It should be mentioned, however, that a one-to-one
displacement occurs only if the mobilities of co-ion and
Electrophoresis 2003, 24, 518–535
BGEs in CZE
527
sample component are equal. In that case, the electrodispersion is minimum resulting in a minimum peak
broadening and large plate numbers. Generally, the transfer ratio, indicating the number of moles of a component i
of the BGE displaced by a mole of the sample component
S, can be defined as [39]:
TRi ˆ
cSi
cBGE
i
cSS
ˆ
Dci
cSS
(13)
where c refers to the total concentration of an ionic species, the subscripts refer to the sample component S and
an ionic species i of the BGE and the superscripts refer
to the sample zone S and the BGE, respectively. Large
transfer ratios are to be preferred in order to get a large
UV signal. These transfer ratios can be calculated with
mathematical models or determined with simulation programs. The component i of the BGE can be either a co-ion
or a counterion and the definition (13) can even be used for
complicated BGEs containing several co-ions or counter
ions.
Calculations of transfer ratios show that the values can be
either positive (increasing concentration of component i
in the sample peak) or negative, for both co-ions and
counterions and this can give a very complicated case
for the UV signal, dependent on which ions are chromophores [39, 40]. Especially in case of two co-ions, remarkable TR values for the co-ions can be obtained, in which
the concentration of one of the co-ions increases in the
sample zone (positive TR) and the concentration of the
other co-ions decreases (negative TR). This means that if
the first co-ion is a chromophore a peak is visible in the
electropherogram and if the second is a chromophore
a dip will be obtained in the electropherogram. If more
than one of the constituents of the BGE are chromophores the result is difficult to predict without knowledge
of all physicochemical data and calculations. As an example of such a complicated situation, we give in Fig. 6
the electropherogram for the separation of a mixture of
0.0005 M Li1 and Na1 ions, applying a BGE consisting of
a mixture of 0.005 M of the co-ions potassium and histidine adjusted to pH 3.43 by adding formic acid. In the
electropherogram Na1 is a peak and Li1 is a dip and three
SPs are visible, viz. the EOF dip SP1, a system peak SP2
due to the low pH of the BGE and an SP3 owing to the use
of two co-ions. Because the mobilities of Na1 and SP3 are
close together, the zones show an extra peak broadening.
5 Prediction of properties of BGEs
5.1 General considerations
A particular BGE is characterized by its composition,
i.e., by the selected co-ions and counterions and their
concentrations. By this choice, all other parameters such
Figure 6. Separation of a mixture of 0.0005 M Li1 and Na1
ions applying a BGE consisting of a mixture of 0.005 M of
the co-ions potassium and histidine adjusted to pH 3.43
by adding formic acid. Measuring conditions: 3 s pressure
injection; voltage, 5 kV; capillary length, 36.7 cm; distance
between injection and detection, 30.0 cm.
as the pH of the BGE, the ionic strength and the specific conductivity are fixed and can be calculated from
the mobilities and pK values of all constituents. The
mEOF can experimentally be determined applying such
a BGE for a given capillary. If all foregoing characteristics of a BGE, as well as the mobilities and pK values
of the analytes, are known and providing that we consider analytes having no specific mutual interactions
and no interactions with the ionic species of the BGE,
the migration behavior of analytes (apparent mobilities)
and peak shapes can be predicted. Even the existence
and mobilities of SZs can be predicted. Predictions can
be made using (i) rules of thumb, (ii) mathematical models [40–45] and (iii) simulation programs [34, 46–48]
based on mathematical models. A problem is often
how to present the information obtained with mathematical models in such a way that the reader can use
this information in a general way. In this section, we
will present an example of a SystChart, in which all
parameters calculated with a mathematical model [40,
41] are visualized in eight panels, and the calculated
data are confirmed with a measured electropherogram.
In all panels the relationship between a specific parameter versus the mobilities of the sample analytes are
given for sample zones at a sample concentration of
561024 M.
528
J. L. Beckers and P. Boček
5.2 SystCharts
As an example, we will discuss the SystChart of a BGE
consisting of a mixture of 0.005 M potassium and 0.005 M
histidine and 0.02 M acetic acid at a pH of ca. 4.7. This
BGE is useful for both UV-transparent cations and UVabsorbing anions and in this system acetic acid acts as
buffering ion. The SystChart is calculated for fully ionized
analytes. In Table 1, all mobilities at infinite dilution and pK
values are given for the ionic species used in the calculations. The setup of a SystChart [49] is always as follows.
The panels (A)–(D) always describe the electrophoretic
behavior of cationic analytes and the panels (E)–(H) that
of anionic analytes for the specified BGE. In all relationships in the panels (A)–(D), discontinuities are present
indicating the existence of an SZ, with a mobility (at
infinite dilution) of ca. 4661029 m2/Vs. In all panels the
dashed vertical lines refer to the mobilities of the coions, the dashed arrows indicate the mSZ and the dashed
horizontal lines indicate the values of the given parameter in the BGE. The set of panels for this BGE is given
in Fig. 7.
In the panels (A) and (E) the pH in the sample zones
are given, calculated for a sample zone with a cS of
561024 M. The calculated pH in the sample zones in this
BGE do not differ so much from the pH of the BGE,
except for sample cations with mobilities close to that
of the SZ. In the panels (C) and (G) the ratios E1m1/E2m2
are given where the subscripts 1 and 2 refer to the parameters in the pure BGE and in a sample zone with a cS
of 561024 M, respectively. The m refers to the values of
the mobilities of a particular sample ionic species in the
pure BGE zone and in the sample zone of 561024 M,
respectively. This ratio can better be applied than often
used ratios E1/E2 because for “weak” acids and bases,
small shifts in pH can result in a large change in effective
mobility and this effect often overrules the changes in
E ratio. If the ratio E1m1/E2m2 is larger than unity the
sample component in the BGE moves faster than that
in the sample zone at a concentration of 561024 M and
because the lowest concentration segment of the peak
moves fastest, the peaks will be fronting. If the ratio is
smaller than unity tailing peaks are the result. For fully
ionized sample cations and anions, having mobilities
mS = mco-ion, the ratio is unity.
In the panels (B) and (F) the total concentrations are given
of all cationic species of the BGE present in the sample
peaks and in the panels (D) and (H) the total concentrations are given of all anionic species of the BGE present
in the sample peaks. The panels (B) and (H) describe the
concentrations of the co-ions and the panels (D) and (F)
those of the counter ionic species. In the panels (B) and
(H) a second horizontal dashed line indicates the total
Electrophoresis 2003, 24, 518–535
concentration of the co-ions for a TR of unity, i.e., with a
concentration value of 561024 M lower than that of the
concentration in the BGE. The turn-over point for TR,
i.e., the mobility of the sample ions at which the value of
TR is unity, is generally not affected by pK values of the
sample ions and equals the mobility of the co-ions. If ionic
mobilities of sample components and co-ions are equal,
there is a one-to-one displacement of sample and co-ions
and in that case the concentration of the counterions
equals that of the BGE. For this BGE, we see a selective
displacement for cationic sample ions if the mobility of
sample ions equals that of one of the co-ions. For sample
cations with a mobility near the mobility of the SZ a
strange effect is visible. One of the co-ions is displaced
at a very large degree whereas the concentration of the
other co-ion increases! If the latter co-ion is UV-absorbing, then this sample zone is visible as a peak in the
electropherogram. We see in panel (B) that this is true
for sample ions with a mobility higher than that of the SZ
and lower than that of potassium. In panel (F) only a single
relationship is given because the calculated concentration for the counterions histidine and potassium are nearly
identical.
To illustrate these theoretical predictions, the experimental electropherogram, for the separation of a mixture of
561024 M of Na1, Li1 and Tris1 ions applying this BGE,
is given in Fig. 8A. Just according to Fig. 7B, the sodium
zone is visible as a peak in the electropherogram and the
components lithium and Tris are dips. The discontinuities
in Fig. 7, indicate the existence of a SZ with a mobility of
ca. 4661029 m2V21s21, and, really a system peak SP2 is
visible just before the lithium dip in Fig. 8A. According to
Fig. 7B clearly can be seen that the area of the lithium
dip is enlarged compared with the area of the sodium
peak and Tris dip. In accord with the theoretical prediction in Fig. 7C, the experiment in Fig. 8A shows that the
sodium zone is visible as a tailing peak, lithium is a fronting dip and Tris is nearly symmetric and sharp. Further a
system peak SP1 is visible corresponding to the low pH
of the BGE with a very low mobility. Besides this presentation of calculated data a lot of other mathematical
models and simulation programs can be used for the
prediction of the suitability of a BGE. A sophisticated
simulation program Peakmaster is recently published by
Gaš et al. [34] and this simulation program can even be
downloaded. In Fig. 8B the electropherogram is given,
simulated with Peakmaster, for the same electrophoretic
run as given in Fig. 8A. The simulated electropherogram
resembles very well the measured electropherogram.
Concerning the anions, the panels (E)–(H) predict that
no SZs are present and this BGE is only suitable for
UV-absorbing anions, because the changes in the UV
signal owing to changes in the concentration of the
Electrophoresis 2003, 24, 518–535
BGEs in CZE
529
Figure 7. SystChart for a BGE consisting of a mixture 0.005 M potassium, 0.005 M histidine and 0.02 M acetic acid at a pH of
4.7. For further information see Section 5.2.
530
J. L. Beckers and P. Boček
Electrophoresis 2003, 24, 518–535
Figure 8. (A) Measured electropherogram for the
separation of an equimolar mixture of 561024 M of
Na1, Li1 and Tris1 ions applying a BGE consisting
of 0.005 M K1 and histidine1 adjusted to a pH of 4.7
by adding acetic acid. (B) Simulation of the same
separation. Description from left to right: sodium
(peak), SZ 2 (vertical line representing discontinuity),
lithium (dip), Tris (dip), SZ 1 (vertical line), and EOF
(vertical line).
UV-absorbing counterion histidine are too small. Examples of these separations are trivial and we do not include
them here.
6 Examples of BGEs
In this part we like to describe the general way of constructing typical BGEs for the principal modes of migration (cathodic and anodic) and of detection (direct and
indirect), and, exemplify these considerations by using
simple model samples. For the preparation of a BGE,
we first have to know whether we want to work in the
direct or indirect UV mode. If the analytes are UV-absorbing, we will work in the direct UV mode and in that case
we can choose a UV-transparent cation and anion for the
BGE. If the analytes are UV-transparent we will work in
the indirect UV mode and in that case it is advisable to
take a UV-absorbing co-ion. In order to obtain a separation, by controlling effective mobilities of analytes, the
analytes must be (partially) charged, i.e., that for cations
a pH will be chosen near or lower than its pK value and for
anions near or higher than its pK value. One of the ionic
species of the BGE should have a pK value near the
desired pH in order to act as buffer.
6.1 BGE for analysis of cations with direct UV
detection
A UV-transparent cation, such as Tris or sodium should
be chosen. Tris can be used as pH buffering co-ion between pH 7–9 because its pK value is 8, in combination
Electrophoresis 2003, 24, 518–535
with all acids with a pK value lower than 8. For pH lower
than ca. 7, the counterion must be buffering and its pK
value defines the useful pH region. If sodium is used as
co-ion, the counterions should always be buffering. All
these BGEs show no SZ except at a very low pH. In order
to obtain an optimum resolution, the effect of electrodispersion should be considered. A minimum peak broadening is obtained if the mobility of the co-ion matches
those of the analytes. If two co-ions are used, there are
invisible SZs present and they can disturb the separation
and deform the peak shapes. To demonstrate this, we
consider the separation of imidazole, histidine and creatinine. This mixture should easily be separated at a pH
where the analytes are completely charged. In Fig. 9, the
measured electropherograms are given for the separation
of an equimolar mixture of 561024 M of imidazole, histidine and creatinine applying BGEs consisting of (a) 0.01 M
potassium, (b) 0.01 M Tris and (c) 0.009 M potassium and
0.001 M Tris, all adjusted to pH 3.5 by adding formic acid.
Applying a BGE containing the co-ion K1 (a), the analyte
peaks are broad, especially histidine, because of the
big difference between the mobilities of K1 and those of
the analytes and all peaks are tailing because the co-ion
has the highest mobility. Applying a BGE containing the
co-ion Tris (b), the analyte peaks are narrower because
the mobilities of analytes do not differ so much from that
of Tris, especially that of histidine. Imidazole and creatinine are fronting, whereas Tris is practically symmetric.
Figure 9. Measured electropherograms for the separation of an equimolar mixture of 0.0005 M of imidazole
(Im), histidine (Hi) and creatinine (Cr) applying BGEs
consisting of (a) 0.01 M potassium, (b) 0.01 M Tris and
(c) 0.009 M potassium and 0.001 M Tris, all adjusted to
pH 3.5 by adding formic acid.
BGEs in CZE
531
Applying a BGE containing two co-ions potassium and
Tris (c) an invisible SZ is present, with a mobility close to
that of creatinine through which the peak shape of creatinine is totally disturbed.
6.2 BGE for analysis of cations with indirect UV
detection
For the separation of cations in the indirect UV mode, it
is advisable to choose a UV-absorbing co-ion, such as
histamin, histidine or imidazole. All these ions are weak
and the buffering capacity of them can be used advantegeously. Especially histamin, with a pK of 9.95, is a UVabsorbing base that can be used at high pH. On the other
hand, these co-ions may be combined with buffering
anions, e.g., acetic acid and then the buffering capacity
of the applied anionic species can be used at pH values
near their pK values. BGEs consisting of histidine acetate
and imidazole acetate give no SZs whereas the divalent
histamine can give SZs as explained in Section 4.5. In
order to demonstrate this behavior, we give in Fig. 10 the
electropherograms for the separation of 5 s pressure
injections of an equimolar mixture of 561024 M of (a, b)
K1, Na1, Li1, Tris1, and tetrabutylammonium (TBA1) ions
and (c, d) K1, Ba21, Na1, Li1, Tris1, and TBA1 ions applying BGEs consisting of (a, c) 0.01 M histidine and (b, d)
0.01 M imidazole, all adjusted to pH 5 by adding acetic
acid. A good separation is easily obtained in Fig. 10 both
applying (a) histidine and (d) imidazole acetate. If we add
Figure 10. Separation of an equimolar mixture of
561024 M of (a, b) K1, Na1, Li1, Tris1, and TBA1 ions, and
(c, d) K1, Ba21, Na1, Li1, Tris1, and TBA1 ions, applying
BGEs consisting of (a, c) 0.01 M histidine and (b, d) 0.01 M
imidazole, all adjusted to pH 5 by adding acetic acid.
532
J. L. Beckers and P. Boček
Electrophoresis 2003, 24, 518–535
adjusted to pH 6.5 by adding acetic acid. At pH 6.5, histamine is for the larger part a monovalent cation and its
competitive adsorption to the wall is reduced, through
which the Ba21 dip is strongly tailing again. Further, the
mobility of the SZ is much higher than it was at pH 5, and
the SP is visible just before the lithium dip, see Fig. 11B (a).
These electropherograms demonstrate that weak multivalent co-ions should be used for BGE preparation with
reservation owing to the existence of SZs. For BGEs
consisting of 0.01 M histamine adjusted to different pH
by adding acetic acid we calculated the relationship between the mobility of the SZ and pH of the BGEs and the
relationships are given in Fig. 12 using (o) the simulation
program Peakmaster [34] and (d) SystCharts [28]. Both
calculations agree perfectly well, although the values
obtained from SystCharts are higher because no corrections are made for relaxation and retardation effects
according to Debye-Hückel-Onsager.
Figure 11. (A) Separation of an equimolar mixture of
561024 M of (a) K1, Na1, Li1, Tris1, and TBA1 ions, and
(b) K1, Ba21, Na1, Li1, Tris1 and TBA1 ions, applying a
BGE consisting of 0.01 M histamine adjusted to pH 5.0
by adding acetic acid. (B) Separation of the same sample
mixtures as used for (A), applying a BGE consisting of
0.01 M histamine adjusted to pH 6.5 by adding acetic
acid.
Ba21 to the sample we see that this divalent ion strongly
adsorbs to the wall causing a fronting Ba21 dip with a very
long tailing backside, and this indicates that a BGE cannot always be used for all samples. The same samples are
analyzed with BGEs consisting of histamine acetate.
In Fig. 11A the electropherograms for the separation of
the same samples as used in Fig. 10 are given for the
BGE consisting of 0.01 M histamine adjusted to pH 5
by adding acetic acid. The Ba21 dip is extremely sharp
because its mobility equals that of the histamine21 ions
and its adsorption to the negative wall charge is strongly
suppressed probably by the competitive adsorption of
the divalent co-ion. Remarkable is the presence of an SP
in the electropherogram, although this SP does not disturb the separation because of its low mobility (see also
Fig. 12). In Fig. 11B the electropherograms for the same
separations are given applying a BGE of 0.01 M histamine
Figure 12. Calculated mobilities of SZs in BGEs consisting of 0.01 M histamine adjusted to different pH by
adding acetic acid, according to (s) Peakmaster and
(d) SystCharts. For further information see Section 6.2.
6.3 BGE for analysis of anions with direct UV
detection
Actually, identical BGEs as used for the determination
of cations in the direct UV mode can be applied. For
example, a cation such as Tris or sodium should be
selected in combination with a UV-transparent acid, in
which one of the ionic species of the BGE should be buffering. Important is whether the separation can be carried
out in the cathodic or anodic mode and this is determined
by the mobility of the EOF. If the absolute value of the mo-
Electrophoresis 2003, 24, 518–535
BGEs in CZE
533
bility of an anionic species is smaller than the mEOF, it can
be analyzed in the cathodic mode otherwise it should be
analyzed in the anodic mode. If the EOF is suppressed by
the addition of an EOF-supressing agent, anions can
always be analyzed in the anodic mode. It should be mentioned that the EOF modifier also can affect the migration
behavior of the analytes [17] and that a second co-ion can
be introduced causing SZs. Popular anionic species for
BGEs are phosphate and borate. In case of multivalent
phosphate buffers, invisible SZs are possible, disturbing
the separation process. Peakmaster can be used for the
calculation of the mobilities of SZs.
6.4 BGE for analysis of anions with indirect UV
detection
Here, a cation such as Tris or sodium in combination with
a UV-absorbing acid as co-ion should be selected. In the
case of Tris a pH between 7–9 is suitable otherwise the
pK value of the acid determines the useful pH region. For
a single monovalent acid no SZs can be expected. For
multiprotic acids SZs are possible near the pK values
where two charged particles of the acid exist. For BGEs
consisting of 0.005 M phthalic acid adjusted to different
pH by adding Tris, the relationship between the mobilities
of the SZs versus the pH is given in Fig. 13, again using (o)
the simulation program Peakmaster and (d) SystCharts.
Although phthalate buffers are very popular for the determination of anionic species [50], the use of these buffers
is limited owing to the presence of SZs, as can be seen in
Figure 14. Electropherogram for the separation of an
equimolar mixture of 0.002 M of tartrate (Ta), citrate (Ci),
succinate (Su), and acetate (Ac) applying a BGE of
0.005 M phthalic acid and 0.0075 M Tris at pH 5.1.
Fig. 13. In Fig. 14 the electropherogram is given for the
separation of an equimolar mixture of 0.002 M of tartrate,
citrate, succinate, and acetate applying a BGE of 0.005 M
phthalic acid and 0.0075 M Tris at pH 5.1. The separation
is carried out in the anodic mode and 561024 M CTAB is
added to the BGE in order to reverse the EOF. This BGE is
often presented as standard BGE for the determination of
anions, but should not be used for anions with mobilities
lower than that of acetate, owing to the presence of a SZ,
visible in the electropherogram just behind acetate (at ca.
3.2 min.). The EOF is visible in Fig. 14 as a peak, because
of the adaption of the concentration of the BGE to the
high o value of the original sample solution according to
KRF and a small SP is visible at ca. 2 min due to the addition of a second co-ion bromide of the CTAB. The latter
SP can be avoided by adding cetyltrimethylammonium
hydroxide (CTAOH) instead of CTAB.
7 Conclusions
Figure 13. Calculated mobilities of SZs in BGEs consisting of 0.005 M phthalic acid adjusted to different pH
by adding Tris, using (s) Peakmaster and (d) SystCharts.
For further information see Section 6.4.
The composition of a BGE, with respect to the number
and sort of co-ions and counterions and their concentrations, is of key importance in CZE. The most important
basic principles to design a suitable BGE, as well as, to
avoid some related pitfalls, may be summarized as
follows:
(i) Suitable BGEs should contain enough ions to conduct
the electric current. The concentration of ions should not
be too high, however, in order to avoid excessive Joule
heating.
(ii) The voltage needed for a specific time of analysis can
be calculated from the mobilities of the analytes. The
corresponding current can be calculated from the electric
conductivity of the BGE.
534
J. L. Beckers and P. Boček
(iii) For a known applied voltage, the electric power
can be calculated and, if needed, the concentration
of the BGE or the diameter of the capillary can be
decreased.
(iv) At least one component of the BGE must have significant buffering capacity at the selected pH, i.e., the pH
of the solution must be within the range pK 6 1 of the buffering component.
(v) A co-ion of the BGE should be selected, in such a
way that its mobility is close to those of sample ions; then
the electrodispersion of sample peaks is minimum and
peaks are practically symmetric and sharp.
(vi) Samples with UV-transparent components may be
detected by using the indirect detection mode with a
BGE where a suitable UV-absorbing co-ion (preferably)
or counterion is used.
(vii) When indirect detection is applied and the BGE
contains two co-ions, the analytes show always selective
displacements of these co-ions, and, the detection
record may show peaks or dips (positive or negative
peaks) for the analyzed components.
(viii) Generally, an EOF is present in fused-silica capillaries migrating to the cathode (cathodic EOF). Thus,
cations and anions can be detected by using cathodic
mode. Anions with mobilities (absolute value) lower than
that of the EOF can be detected. Anions with higher
mobilities should be analyzed in the anodic mode.
(ix) The EOF can be suppressed (zero EOF) or even
reversed (anodic EOF) by adding suitable surfactants,
such as CTAB; however, the risk that it brings some
more co-ions into the BGE should always be considered.
(x) The use of multiple co-ions should be avoided (unless unavoidable for indirect detection) since it brings
the formation of SZs. When two co-ions are used, an
SZ with a mobility in-between the mobilities of these
co-ions is present.
(xi) The use of buffering multivalent weak co-ions should
be avoided for the same reasons.
(xii) BGEs at low and high pHs show the pronounced risk
of disturbances due to the formation of SZs where H1 or
OH2 act as the second co-ion. For BGEs consisting of,
e.g., 1 M acetic acid, H1 ions are the sole cations.
(xiii) Theoretical predictions, based on mathematical
models and computer simulations, are very useful and
enable one to examine the migration behavior of the
BGE with a proposed composition as far as all the
above aspects are concerned.
(xiv) The composition of a BGE should be as simple as
possible, in order to avoid the formation of SZs. That is
why the use of one monovalent cationic and one monovalent anionic component (besides H1 and OH2) is
recommended.
Electrophoresis 2003, 24, 518–535
This work was supported by the Royal Netherlands
Academy of Arts and Sciences (KNAW) and the University
of Technology of Eindhoven (TU/e), Netherlands, within
the framework of the agreement with the Academy of
Sciences of the Czech Republic. Moreover, P. B. acknowledges the support by the Grant Agency of the Czech
Republic, Grant No. 203/01/0401 and 203/02/0023, and
by the Grant Agency of the Academy of Sciences of the
Czech Republic, Grants No. A 4031703.
Received August 20, 2002
8 References
[1] Boček, P., Vespalec, R., Giese, R., Anal. Chem. 2000, 72,
586A–595A.
[2] Vespalec, R., Boček, P., Electrophoresis 1994, 15, 755–762.
[3] Boček, P., Gebauer, P., in: Guzman, N. A. (Ed.), Capillary
Electrophoresis Technology, Marcel Dekker, New York
1993, pp. 261–283.
[4] Šustáček, V., Foret, F., Boček, P., J. Chromatogr. 1989, 480,
271–276.
[5] Pospíchal, J., Deml, M., Gebauer, P., Boček, P., J. Chromatogr. 1989, 470, 43–55.
[6] Boček, P., Deml, M., Pospíchal, J., Sudor, J., J. Chromatogr.
1989, 470, 309–312.
[7] Sudor, J., Stránský, Z., Pospíchal, J., Deml, M., Boček, P.,
Electrophoresis 1989, 10, 802–805.
[8] Boček, P., Deml, M., Pospíchal, J., J. Chromatogr. 1990,
500, 673–680.
[9] Gebauer, P., Deml, M., Pospíchal, J., Boček, P., Electrophoresis 1990, 11, 724–731.
[10] Foret, F., Fanali, S., Boček, P., J. Chromatogr. 1990, 516,
219–222.
[11] Wei, W., Xue., G., Yeung, E. S., Anal. Chem. 2002, 74, 934–
940.
[12] Ryšlavý, Z., Boček, P., Deml, M., Janák, J., J. Chromatogr.
1977, 144, 17–25.
[13] Boček, P., Ryšlavý, Z., Deml, M., Janák, J., Collect. Czech.
Chem. Commun. 1977, 42, 3382 – 3387.
[14] Boček, P., Topics Curr. Chem. 1981, 95, 131–177.
[15] Foret, F., Křivánková, L., Boček, P., Capillary Zone Electrophoresis, VCH Verlagsgesellschaft mbH, Weinheim 1993.
[16] Lacey, M. E., Webb, A. G., Sweedler , J. V., Anal. Chem.
2000, 72, 4991–4998.
[17] Beckers, J., Boček, P., Electrophoresis 2002, 23, 1947–1952.
[18] Foret, F., Fanali, S., Ossicini, L., Boček, P., J. Chromatogr.
1989, 470, 299–308.
[19] Robinson, R. A., Stokes, R. H., Electrolyte Solutions, Butterworths, London 1959, p. 468.
[20] Kohlrausch, F., Ann. Phys. (Leipzig) 1897, 62, 209–239.
[21] Hjelmeland, L.-M., Chrambach, A., Electrophoresis 1982, 3,
9–17.
[22] Everaerts, F. M., Beckers, J. L., Verheggen, T. P. E. M., Isotachophoresis – Theory, Instrumentation and Applications,
Elsevier, Amsterdam, Oxford, New York 1976.
[23] Boček, P., Deml, M., Gebauer, P., Dolník, V., Analytical Isotachophoresis, VCH Verlagsgesellschaft mbH, Weinheim
1988.
Electrophoresis 2003, 24, 518–535
[24] Gebauer, P., Boček, P., Electrophoresis 2000, 21, 3898–3904.
[25] Gebauer, P., Boček, P., Electrophoresis 1997, 18, 2154–2161.
[26] Beckers, J. L., Boček, P., Electrophoresis 2000, 21, 2747–
2767.
[27] Poppe, H., Anal. Chem. 1992, 64, 1908–1919.
[28] Beckers, J. L., Everaerts, F. M., J. Chromatogr. A 1997, 787,
235–242.
[29] Gebauer, P., Borecká, P., Boček, P., Anal. Chem. 1998, 70,
3397–3406.
[30] Beckers, J. L., Gebauer, P., Boček, P., Electrophoresis 2001,
22, 3648–3658.
[31] Mikkers, F. E. P., Anal. Chem. 1999, 71, 522–533.
[32] Štědrý, M., Jaroš, M., Gaš, B., J. Chromatogr. A 2002, 960,
187–198.
[33] Beckers, J. L., Boček, P., Electrophoresis 2002, 23, 1942–
1946.
[34] http://www.natur.cuni.cz/,gas
[35] Gebauer, P., Pantůčková, P., Boček, P., J. Chromatogr. A
2000, 894, 89–93.
[36] Gebauer, P., Boček, P., Electrophoresis 2000, 21, 2809–2813.
[37] Boček, P., Gebauer, P., Beckers, J. L., Electrophoresis 2001,
22, 1106–1111.
BGEs in CZE
535
[38] Beckers, J. L., Electrophoresis 2001, 22, 2684–2690.
[39] Beckers, J. L., J. Chromatogr. A 1994, 679, 153–165.
[40] Beckers, J. L., J. Chromatogr. A 1999, 844, 321–331.
[41] Beckers, J. L., J. Chromatogr. A 1997, 764, 111–126.
[42] Gebauer, P., Boček, P., Anal. Chem. 1997, 69, 1557–1563.
[43] Mosher, R. A., Saville, D. A., Thormann, W., The Dynamics of
Electrophoresis, VCH Verlagsgesellschaft mbH, Weinheim
1992.
[44] Gaš, B., Coufal, P., Jaroš, M., Muzikář, J., Jelínek, I., J. Chromatogr. A 2001, 905, 269–279.
[45] Jaroš, M., Včeláková, K., Zusková, I., Gaš, B., Electrophoresis 2002, 23, 2667–2677.
[46] Reijenga, J. C., Kenddler, J., J. Chromatogr. A 1994, 659,
403–416.
[47] Reijenga, J. C., Kenddler, J., J. Chromatogr. A 1994, 659,
417–426.
[48] Mikkers, F. E. P., Anal. Chem. 1999, 71, 522–533.
[49] Beckers, J. L., Boček, P., Electrophoresis 1999, 20, 518–524.
[50] Altria, K. D., Assi, K. H., Bryant, S. M., Clark, B. J., Chromatographia 1997, 44, 367–371.
Download