Capillary Electrophoresis in Food Analysis

Capillary Electrophoresis: Basic Principles

Zeynep Kalaycıoğlu¹, F. Bedia Erim¹, *

¹ Department of Chemistry, Istanbul Technical University, Maslak, Istanbul, Turkey

Abstract

Capillary Electrophoresis (CE) is a powerful separation and analysis technique that has been rapidly progressing since it was first introduced. The application range of CE is so diverse that it ranges from small analytes to large and complex macromolecules. This chapter aims to provide a deep understanding of the basic principles of CE. The first part of the chapter involves the theoretical basis, instrumentation, and separation mechanism of CE. The second part focuses on capillary electrophoretic separation modes and the third part describes the detection methods in CE. The fourth and final part covers capillary electrophoretic strategies for specific analyte groups.

Keywords: Capillary electrophoresis, CEC, CGE, Chiral separation, Contactless conductivity detector, CZE, Indirect detection, LIF, MEKC, Proteins, Sample stacking, Small anions, Small organic acids.

* Corresponding author F. Bedia Erim: Department of Chemistry, Istanbul Technical University, Maslak, Istanbul, Turkey; E-mail: [email protected]

INTRODUCTION

Capillary Electrophoresis (CE) is a separation method that takes the basis of separation principle from classical electrophoresis and has the device design of modern chromatographic techniques. CE was first described as a new separation technique between 1980-1990, then a relatively new separation technique by comparison with High-Performance Liquid Chromatography (HPLC) and Gas Chromatography (GC). Today, CE has been widely applied to real samples and can be defined as a modern separation technique.

Classical electrophoresis was first introduced by the Swedish chemist Arne Tiselius in 1930 [1]. This invention earned Tiselius a Nobel Prize in 1948. Electrophoresis, in its simplest definition, means the migration of charged particles under the electric field.

Classical electrophoresis experiments began in U-shaped tubes with electrodes attached to both ends, containing an electrolyte inside. Later, the separation media preferred were gels, as the reproducibility of migration velocities in open solution was low. This method, is known today as slab-gel electrophoresis

and it is an essential tool of biochemistry laboratories, especially in protein separations. The technology has developed over the years and made it possible to produce silica columns with a diameter of micrometers. Thus, reproducible migration velocities have been obtained in open solutions in capillary columns. J.

W. Jorgensen and K. D. Lukacs introduced the CE technique in 1981 with an article published in Analytical Chemistry [2]. The method first evolved with hand-made devices in research labs, and later, modern CE devices were introduced to the market.

The present chapter describes the basic principles and mechanisms of capillary electrophoretic techniques. It covers instrumentation, different separation modes, and detection methods in CE, and separation strategies for specific analyte groups, such as small inorganic anions and organic acids, proteins, metal ions, and chiral molecules.

SEPARATION MECHANISM

In this section, general and some unique principles, strengths, and weaknesses of the CE technique will be given. Although the CE separation principle is based on classical electrophoresis, there are essential differences between the two methods. The high electrical resistance of the capillary column makes it possible to apply a very high voltage (HV) between the two electrodes. In commercial CE devices, up to 30 kV voltage can be achieved. The high voltage applied gives the CE technique a tremendous separation speed. This high speed also increases the separation efficiency of the method. An advantage of CE over other chromatographic techniques is that separation can be done in an aqueous medium. If necessary, hydrophilic organic solvents can also be added to the aqueous separation medium. Another advantage of CE is that different types of analytes can be separated in the same silica column using the same instrument design. One designation that has been used since the early years of CE is that the species can be separated in a broad spectrum from Li-ion to DNA by capillary electrophoretic methods. Indeed, many species in inorganic, organic, and macromolecular structures have been shown to be separated by CE. The different applications of CE can be seen in the most recent reviews [3-5]. Another advantage of CE compared to liquid chromatography methods is that silica capillary columns are less costly. Since the injection volume in CE is in nL size, it is an economical method in terms of capillary columns and the small amounts of sample and chemical consumption. The capillary column can be regenerated by washing with separation electrolytes and suitable solvents

between each injection. Due to this advantage, many sample solutions can be injected directly into the capillary column without the need for pre-purification.

The weakness of CE is the detection limit. In CE, the separation column is also the detection cell. Since the light path of the detection cell is in micrometers, if the molar absorptivity of the analyte is not high enough in UV detection, detection limits will be high. There are unique methods to overcome this obstacle in CE. These methods will also be mentioned in this section. The schematic device design of CE is shown in Fig. (1).

Fig. (1))

Basic components of capillary electrophoresis instrumentation.

Fig. (1) shows two small buffer vials and a sample vial. The ends of a fused-silica column are immersed in buffer vials during electrophoresis. For electrophoretic separation, the separation medium must be an electrolyte solution with electrical conductivity. In CE separations, the use of a buffer solution for the background electrolyte medium is preferred due to the formation of H+ and OH- ions in the electrode regions as a result of electrolysis. H+ and OH- concentrations due to electrolysis are at the micromolar levels. However, the migration mobilities of pH- sensitive analytes and the surface charge of silica capillary column are affected by the pH change of the medium. Therefore, it is common practice to use a buffer as the separation medium in CE applications, with some exceptions.

The two electrodes are connected to a high voltage (HV) source. Voltage up to 30 kV can be provided in CE devices. Capillary columns are generally fused-silica columns. The outer layer of a fused-silica capillary is coated with a thin layer of polymer to prevent breakage and provide flexibility while using it. The detection

window is opened by burning the polymer layer about 0.5 cm in length. As seen in Fig. (1), in general practice, the detector is placed close to the cathode side. Injections are performed from the anodic side. All the injected species, i.e., positively charged, negatively charged, and neutral species, migrate inside the capillary from the anodic side to the cathodic side, regardless of their charges.

During this migration, the analytes separate from each other, reach the detector, and the data is evaluated with the help of a software.

Upon the application of high voltage between electrodes, ionic species experience an electric force given by eq. (1):

where q is ion charge and E is applied electric field which is given by eq. (2):

where V is the voltage applied across the capillary, and L is the capillary length.

Ionic species in the capillary accelerate towards electrically opposite electrodes. While the velocity of the charged species increase, the friction force caused by the effect of the environment slows down the speed of the species. According to Stokes' law, the friction force for a spherical molecule can be expressed by eq. (3):

where η is the viscosity of solution inside the capillary, r is the radius of the ion, and v is the velocity of the ion.

When a steady state is reached, the forces are equal but opposite, and the ions move with a constant velocity, which is proportional to the electrical field.

where the proportionality constant µe here is called electrophoretic mobility (cm2V-¹s-1). After steady-state, i.e. FE = FF, by the combination of equations (1-4), the electrophoretic mobility µe of a charged particle is expressed as

The separation of the species from each other is due to the mobility differences during migration in the capillary column. As seen from eq. (5), in a medium where viscosity is constant, the electrophoretic separation of the species is based on differences in their charge/size ratios.

We have mentioned that the injection is generally done from the anodic side, and all species migrate towards the cathode. It is necessary to look inside the capillary to understand the reason for this, which is contrary to Coulomb's law at first sight. The column consists of fused-silica capillaries. When an aqueous solution of electrolytes at around pH ≥ 1.5 is in contact with the capillary wall, the silanol groups of the silica on the surface of the capillary inner wall are deprotonated, leaving negative groups on the wall. Hence the inner wall of fused-silica capillary is now negatively charged. The excess of positive ions is located near the capillary wall due to the electrostatic attraction between the negatively charged wall and positively charged ions inside the capillary.

Capillary wall-SiOH + H2O ↔ Capillary wall-SiO- + H3O+

When a high potential is applied between the ends of the capillary, excess positive charge inside the capillary migrates to the cathodic side and drags their hydrate solvent molecules. Thus, a solution flow occurs in the capillary. This flow is called electroosmotic flow (EOF). In CE separation techniques, the driving force is EOF. The magnitude of the EOF can be expressed by electroosmotic mobility (EOM). Generally, the mobility of the electroosmotic flow is greater than the electrophoretic mobilities of the injected species. While positively charged species injected from the anodic side move towards the cathode with their electrophoretic mobility, they are also dragged by the electroosmotic flow in the same direction. So the apparent mobility for these species is given with the following equation:

where

µa: apparent mobility

µe: electrophoretic mobility µeo: electroosmotic mobility

In contrast, while the negatively charged species try to move towards the injection direction, that is, the anodic side, with their electrophoretic mobility, the strong electrosmotic flow in the capillary drags them to the cathode direction. The apparent mobility of negatively charged species within the capillary is given by the following equation:

This process is illustrated in Fig. (2). The charged particles will separate from each other during their migration in the cathode direction within the capillaries, if there is a difference in their charge/size ratios. However, all of the injected uncharged species will move to the detector together with apparent mobility equal to electroosmotic mobility, i.e. µa = µeo.

Fig. (2))

Migration order of ionic and neutral species with the effect of EOF.

EOF can be easily adjusted with some experimental variables. However, while changing EOF, positive and negative effects of variables on separation should be taken into account. Experimental parameters affecting the speed of EOF are buffer pH, buffer concentration, applied electric field, temperature, buffer additions, and coating the inner wall of the capillary column. The change of buffer pH most easily and effectively controls the speed of the electroosmotic flow in the capillary. As the pH of the separation electrolyte increases, the ionization in the capillary inner wall increases hence the negative charge of the wall increases and EOF accelerates.

Increasing the electric field causes an increase in EOF. However, an increase in the electrical field may increase the current and, consequently, the Joule heating in the capillary. Reducing the electric field reduces separation efficiency and resolution.

EOF decreases with increasing buffer concentration. Increased buffer concentration causes an increased current in the capillary and possible Joule heating. On the other hand, a very low concentration of buffer increases the risk of wall adsorption of some analytes. Increased temperature lowers the viscosity of the separation electrolyte and EOF increases.

Organic solvents, additives, especially surfactants which are added into the buffer, significantly affect the speed of the EOF. Finally, the temporary or permanent coating of the capillary interior wall is an effective method for EOF control.

The most basic CE type, in which charged particles are separated according to mobility differences in free solution, is called Capillary Zone Electrophoresis (CZE). The uncharged species can be separated by other CE types, as will be explained later.

SEPARATION PARAMETERS

Migration Time

The migration time of species in the capillary column is calculated according to the eq. (8):

where

t: migration time (s)

L: capillary length (cm)

L': capillary effective length (the length of capillary from injection point to the detection point) (cm)

V: separation voltage (V).

Separation Efficiency and Peak Broadening

The expression of high separation efficiency is used in defining the CE methods. One of the reasons for the high separation efficiency of CE is the flat profile of EOF. Fig. (3A) schematically shows the flat profile of EOF.

Fig. (3))

Flow profiles and corresponding peak shapes.

The pressure-driven flow is a laminar flow resulting in broader peaks. More than one analyte peak may be hidden in the peak in Fig. (3B). However, the peaks formed in a flat flow are sharp peaks, and the resolution of the two peaks within a few seconds can be achieved. Peak broadening is an undesirable phenomenon that reduces the separation efficiency in all chromatographic methods. One of the reasons for peak broadening is the diffusion of the injected zone during migration. However, since the capillary has a very small diameter, the diffusion of the injection zone in the direction of the capillary walls can be neglected. The diffusion can be in the longitudinal direction. The smaller the standard deviation of the data of a peak in the form of a Gaussian curve results in a narrow peak. The effect of longitudinal diffusion to the peak broadening is given as the variance of the peak in eq. (9):

where

D: diffusion coefficient of the species.

t: migration time.

The very high electric field applied in the CE method causes short migration times. As t decreases, the effect of longitudinal peak broadening will decrease. The peaks of macromolecules such as protein and DNA with low D values will be very narrow, and high separation efficiencies are achieved for these molecules.

Apart from its longitudinal diffusion, another factor that causes peak broadening is the formation of Joule heating in the capillary. Although the electrical current passing through the capillary is very low, there may be a heat generation in the capillary called Joule heating. The Joule heating formed in the capillaries causes the electrolyte viscosity in the capillary center to be lower than the edges, and the flat profile flow within the capillaries turns into a laminar flow, which causes zone broadening. The way to reduce Joule heating is to lower the operating voltage and buffer concentration. However, a low electric field will reduce separation efficiency and resolution.

On the other hand, lowering the buffer concentration or ionic strength will both decrease the buffer capacity and increase the risk of possible sample adsorption to the wall. Another method for controlling Joule heating is to reduce the internal

capillary diameter. However, with the use of narrow capillaries, sensitivity will decrease for analytes with low molar absorptivity, especially in UV detection. Joule heating is dissipated through the capillary wall if the device has active temperature control.

Sample adsorption to the capillary wall may cause peak tailing. This effect usually results from the electrostatic interaction between positively charged analytes and the negatively charged capillary wall. Increasing the buffer concentration prevents adsorption by shielding the wall charge. However, since CE separation is very fast, there is no wall adsorption problem for small particles. Adsorption is a problem for proteins carrying a large number of charges on them. In protein separations, coating the capillary inner wall with an uncharged or positively charged polymer according to the type of protein is a method applied to prevent wall adsorption.

If the conductivity of the sample zone and the separation buffer are very different from each other, this may cause distortion and tail in the peak shapes. Although this effect is smaller than the other effects that cause peak broadening, it can be effective in low-resolution separations. Additionally, it can be turned into an advantage to increase the sensitivity of some analytes in the form of sample stacking, which will be explained soon.

Precision in Migration Times and Peak Areas

Capillary columns provide reproducible results by rinsing the column with a washing solution between each run. It is sufficient to wash only with the working buffer for some analytes, while special washing strategies should be developed for some others. Sodium hydroxide, hydrochloric acid, and some organic solvents such as methanol are usual washing solutions. Since the EOM is very sensitive to pH, buffer capacity is a critical factor for obtaining reproducible migration times. In the case of initial concentrations of conjugate acid-base pair of the buffer are equal to each other, the buffer capacity is maximum. The buffer works effectively when the pH of the buffer is around ± 1 unit of the pKa value of the conjugated acid. A high concentration of buffers is also required for a good buffer capacity. However, buffer concentration in CE is limited due to possible Joule heating. The CE device automatically cuts off or lowers the voltage when the current limit value is exceeded.

Zwitterionic buffer solutions can be used in relatively high concentrations without increasing the CE current. Tris, MES, CAPS, CAPSO are the most used zwitterionic buffers in CE. A buffer with a low absorbance at the wavelength

studied is a desired feature. Phosphate and borate buffers are the most used buffers in capillary electrophoretic separations with their low absorbances and high buffer capacities. Even if the buffer has a high capacity, the buffer containers need to be refreshed occasionally. When the buffer contains an organic solvent, the polymer layer at the capillary end can creep and consequently partially close the capillary entrance, which affects injection volume reproducibility. This problem can be solved by burning the polymer layer at the capillary end approximately half a cm.

pH difference between the washing solution and the buffer solution can cause hysteresis in the capillary column. In such cases, it is preferred to apply an electric field for a while without injecting the new solution or to ignore the first injection results.

The internal standard increases peak area reproducibility. In this case, the calibration curve is drawn as peak areas/internal standard area ratios vs. analyte concentrations. However, it is not always possible to find a stable and suitable internal standard that migrates near analyte peaks. In this case, the calibration curve can be created between the corrected peak areas and concentrations. Corrected peak areas are obtained by dividing the area of each peak by migration time.

The most common solutions for reproducibility of migration time and peak area are summarized in Tables 1 and 2, respectively.

Table 1 Common solutions to increase migration time reproducibility.

Table 2 Common solutions to increase peak area reproducibility.

MODES OF CE

Capillary Zone Electrophoresis (CZE)

The separation of analytes in free solution according to mobility differences, as described above, is CZE. CZE is the basic separation mode of CE. Other capillary modes use the device design of CZE.

Micellar Electrokinetic Chromatography (MEKC)

If the separation is mainly based on the interaction between analytes and some separation medium additives, such separations in CE are commonly referred to as electrokinetic chromatography or electrokinetic capillary chromatography. Micellar Electrokinetic Chromatography (MEKC) is the dominantly used form of electrokinetic chromatography.

After Terabe et al. introduced MEKC [6], it has become a CE method primarily used in the separation of uncharged species. In MEKC, a surfactant is added to the separation buffer above its critical micelle concentration (CMC). Aggregates called micelles are formed inside the capillaries, and these aggregates act as a fixed phase in the liquid chromatography. However, since the micelles also have an electrophoretic mobility, they move within the capillary. Micellar aggregates can be called a pseudo-stationary phase. Sodium dodecyl sulfate (SDS) is the most widely used surfactant in MEKC applications. The hydrophilic ends of the micelles formed in the aqueous buffer medium are directed towards the aqueous phase and the hydrophobic tails towards the middle of the micelle.

A schematic separation scheme of MEKC is given in Fig. (4). A small neutral hydrophilic molecule injected from the anodic side will be dragged to the detector with the electroosmotic flow (t0). The peak will appear in the electropherogram at t0 time. Neutral markers such as acetone, formamide, mesityl oxide can be used in order to determine the mobility of the EOF in both MEKC and CZE. The negatively charged micelles, in that example, tend towards the anodic side with their electrophoretic mobility. However, they will be driven towards the cathodic side, that is, the detector direction with the electroosmotic flow in the capillary.

When a very hydrophobic dyestuff is injected into the medium, it will remain in the hydrophobic core of the micelle. It will appear as a peak in the electropherogram when the micelle reaches the detector (tm). Sudan dye is commonly used as a marker for this purpose. In MEKC separations, the partition of species between the hydrophobic core of micelles and the solution phase occurs. All uncharged analytes will migrate between t0 and tm window. Those with a more hydrophilic structure will be seen in areas closer to t0, and those that are more hydrophobic will appear in areas close to tm. If all hydrophobic analytes tend to spend most of their time in the micelle phase, a certain proportion of water-compatible organic solvent can be added to the buffer to increase the tendency of these analytes to the aqueous phase. Alcohols and acetonitrile are the most commonly used organic solvents in the studies. In MEKC, the organic solvent added should not exceed 20-30% concentration because micelles can be dispersed in the organic solvent medium.

Fig. (4))

Separation mechanism of Micellar Electrokinetic Chromatography (MEKC).

Alternatively, by adding cyclodextrins to the buffer together with a surfactant, separation of hydrophobic molecules can be achieved by the partition of molecules between the micellar phase and cyclodextrin cavity. MEKC has been used in the analysis of a large number of compounds in food, plant, biological, and environmental samples.

Nonaqueous Capillary Electrophoresis (NACE)

In NACE, the separation medium is an organic solvent instead of water. Methanol, acetonitrile, formamide, dimethylformamide are the most common organic solvents in NACE studies. Organic solvent affects the solubility of analytes. However, an important factor in NACE separations is the significant change of acid-base properties of analytes relative to the water environment. The organic solvent also affects the ionization of the capillary wall. Since the conductivity of the organic solvents used is lower compared to water, the high electric field in the capillary increases the separation speed. of HClO4. Thus, with the selected NACE electrolyte, acrylamide migrates electrophoretically in the capillary as a positively charged molecule. Fig. (1) NACE analysis of acrylamide (A) a French fry extract and (B) a spiked extract with 2 mg/L acrylamide standard. * shows the acrylamide peak. Buffer: 30 mmol/L HClO4 , 218 mmol/L CH3COOH, and 0.129% w/v water. Voltage: 25 kV (injection is from the positive electrode side). Capillary, 75 mm, (52.5 x 44) cm; wavelength: 200 nm; injection: hydrodynamic, 40 mbar, 6 s. Reprinted from [7] with permission.

Fig. (5))

NACE analysis of acrylamide (A) a French fry extract and (B) a spiked extract with 2 mg/L acrylamide standard. * shows the acrylamide peak. Buffer: 30 mmol/L HClO4, 218 mmol/L CH3COOH, and 0.129% w/v water. Voltage: 25 kV (injection is from the positive electrode side). Capillary, 75 mm, (52.5 x 44) cm; wavelength: 200 nm; injection: hydrodynamic, 40 mbar, 6 s. Reprinted from [7] with permission.

Capillary Gel Electrophoresis (CGE)

Capillary Gel Electrophoresis (CGE) is a type of CE mostly used to separate proteins and DNA according to their size differences. DNA is negatively charged due to its phosphate groups, whereas proteins gain negative charge by being treated with SDS. Since the sizes of macromolecules such as DNA are huge, their electrophoretic separation based on charge/size differences is not possible. These molecules are separated according to their size differences only in a gel environment [8].

In CGE, a gel medium exists in the capillary column. The most commonly used gel medium in CGE applications is crosslinked polyacrylamide (PAA). The separation in the CGE method is based on the ability of the molecules to pass through the pores of the gel. Macromolecules having small sizes can easily migrate through the pores of the gel, whereas the bigger molecules look for pores that can pass through, and their migration paths increase. In CGE, the movement of species is either by the combination of electrophoretic mobility and electroosmotic mobility, or by coating the capillary wall with an uncharged polymer, EOF is suppressed, and migration occurs only with electrophoretic mobilities. In capillary gel electrophoresis, the separation medium is not always solid gel. Entangled linear polymer solutions can be used as separation media. The separation mechanism in the non-crosslinked polymer network is identical to that of solid gel.

Capillary Electrochromatography (CEC)

CEC is a relatively newly developed hybrid separation method of CE and liquid chromatography. In this mode, the partition of species is between the fixed stationary phase and solution phase. Here again, the driving force is EOF. CEC can be performed in packed, monolithic, and open-tubular columns. As can be seen in the recent review, the development of new fixed phases in CEC is an evolving field [9].

Capillary Isotachophoresis (CITP)

Two buffer systems are used in CITP. The leader electrolyte (LE) has higher mobility than the terminating electrolyte (TE). With the application of voltage, a non-uniform electric field is formed inside the capillary and the analyte zone stacks between LE and TE. CITP provides concentration and separation of the analyte bands at the same time. The overall progress in CITP has been collected in a recent review article [10].

Capillary Isoelectric Focusing (CIEF)

CIEF is mostly used to separate proteins and peptides based on their isoelectric points (pI) [11]. Ampholyte mixtures establish a pH gradient inside the capillary. The protein migrates until it reaches the pH area corresponding to its pI, where it becomes uncharged and immobilized.

ON-LINE SENSITIVITY ENHANCEMENT: SAMPLE STACKING

In CZE, the sample is injected into the capillary column as a short zone. When voltage is applied, analyte zones are separated from each other along the capillary column. Rapid migration time due to the high voltage applied causes the analyte zone to spend less time in the capillaries and less exposure to peak broadening due to the longitudinal diffusion along the capillary. Analyte zones reach the detector as narrow peaks under the influence of the flat character of the EOF. However, if the molar absorptivities of analytes are low in UV detection, their concentrations may be below the LOD of the analysis method. Injecting larger sample zones to increase detection sensitivity causes broad and overload peaks.

A practical method to increase detection sensitivity in CE includes sample stacking methods where a long sample zone with low conductivity is injected into the capillary column and a high conductivity buffer is selected as the separation buffer. Since the electric field is high in the low conductivity zone, the charged particles move with high mobilities in this zone under the applied voltage. The electric field drops in the high conductivity buffer zone and the analytes slow down at the border of two zones. The analyte molecules are stacked into a narrow zone. This concentrated zone appears in the detector with increased sensitivity.

Sample stacking is an easily applicable technique and different stacking applications can be developed according to the types of samples. For example, the low conductivity sample region can be discarded through the capillary by changing the polarity for a short period of time, or in electrokinetic injection, a short frontal buffer or acetonitrile zone can be hydrodynamically injected. Different names have been used for stacking procedures in the literature, such as Large-Volume Sample Stacking (LVSS), Field-Amplified Sample Stacking (FASS), Field-Amplified Sample Injection (FASI). However, the method used is generally referred to as sample stacking in recent years. The details of the stacking methods are explained in the articles.

In Fig. (6A and B), a sample stacking effect is seen for nitrite and nitrate ions for a method that was developed for the simultaneous determination of these ions in canned fish samples [12]. In the method, keeping the pH of the separation electrolyte low, the anions injected from the cathodic side are enabled to move against EOF. The addition of sodium sulfate increases buffer conductivity. The anions moving rapidly in the direction of the anode in the low conductivity large sample zone are slowed down when they reach the high conductivity buffer border. The schematic view of the stacking is given in Fig. (7). In the study, sensitivities are increased by 30 times for both ions due to the sample stacking. Since the EOF is to the cathode direction in the method, the sample zone is discarded from the cathode side during electrophoresis.

Fig. (6))

Comparison of the sample stacking experiment and small-volume injection of a standard sample containing 100 μmol/L nitrate and nitrite. A) Injection: 50 mbar, 160 s, Buffer: 30 mmol/L formic acid, 30 mmol/L Na2SO4, pH:4. B) Injection: 50 mbar, 6 s, Buffer: 30 mmol/L formic acid pH:4. Voltage: −25 kV. Peaks: 1:nitrate, 2:nitrite. Reprinted from [12] with permission.

Fig. (7))

Schematic illustration of sample stacking method for nitrate and nitrite ions.

Although the easiest and most applied method of stacking is based on the concentration difference of the sample and buffer zone, pH-mediated stacking techniques can be applied for the pH-dependent analytes, where the pH difference of the sample zone and buffer zone is created. Stacking can be carried out by the sweeping method in EKC, especially in MEKC. In this method, the sample zone does not include the micelle phase. When voltage is applied, micelles in the separation electrolyte enter the sample zone and collect analytes with micelle- analyte interaction. The analytes are concentrated at the front of micelles as a short zone. The sample stacking studies in recent years have been collected in review articles [13, 14].

DETECTION METHODS

UV and Indirect UV Detection

The most widely used CE detectors in CE applications are the variable-wavelength ultraviolet (UV) detectors and diode-array (DAD) detectors. The analytes with UV active functional groups can be detected with these detectors during electrophoretic separation. For species that do not contain chromophore groups, a useful and widely used application in CE is the indirect detection method where a strong chromophore group-containing compound is added into the separation buffer. The detector's wavelength is adjusted to the maximum wavelength of the chromophore. The detector sees the baseline signal of the chromophore. Analyte ions displace chromophore ions, and a decrease in absorbance occurs as the analyte zone passes across the detector. The area of this negative peak depends on the analyte concentration [15]. The formation of indirect detection peaks and their symmetry depend on chromophore concentration and mobility. In indirect detection, analyte mobility and chromophore mobility should be compatible with each other. While negative peaks were given on the electropherogram in the first studies with indirect detection, the display of negative peaks in the form of positive peaks with the adjustment of software has been preferred in the last years. The indirect detection method is widely used for organic acid determinations in food samples. Organic acids are present in many food products in small or large amounts. While it gives a sour taste to the juices, the amounts in the commercial juices are also adulteration markers. Fig. (8) shows the separation of malic acid and citric acid in a commercial pomegranate juice by indirect detection. 2,6-pyridine dicarboxylic acid (PDC) was used as the chromophore. In this study, 7 commercial pomegranate juices were analyzed, and the significantly high value of the malic acid amount of one pomegranate juice was interpreted as a sign of its adulteration with apple juice [16]. With an additional analysis, the significant difference in Fructose/Glucose ratio, total phenolic amount, and ferric reducing antioxidant power (FRAP) value of this specific juice from others supported adulteration.

Fig. (8))

Separation and indirect detection of organic acids in a commercial pomegranate juice. Separation buffer: 10 mmol/L of PDC + 0.1 mmol/L CTAB. pH: 5.6. Detection wavelength: 350 nm with a reference at 200 nm. Injection: 50 mbar, 6 s. Voltage: -25 kV. Reprinted from [16] with permission.

Fig. (9))

Separation and indirect detection of fatty acids. (A): the standard solutions of C8-C20 fatty acids, (B): Unsaturated fatty acids in an extract of dairy-fresh butter. Background electrolyte: 10 mM SDBS + 50% acetonitrile + 30 mM Brij. Reprinted from [17] with permission.

Fluorescent and Laser-Induced Fluorescence Detection

Fluorescent and Laser-Induced Fluorescence (LIF) detectors are used for analytes that naturally show fluorescent properties or can be derivatized with a fluorescent dye. A fluorescent detector provides a lower detection limit (LOD) than a UV detector. On the other hand, very low LOD values can be reached with LIF detectors. For example, riboflavin (vitamin B2) which is a fluorescent substance and an essential vitamin, cannot be synthesized in the body so it must be taken from food products. However, the content of riboflavin in many foods is too low to allow detection by a UV detector.

In Fig. (10), an electropherogram showing vitamin B2 separation and LIF detection in a saffron extract is given [18].

Fig. (10))

Separation and LIF detection of vitamin B2 in a saffron sample. Buffer: 20 mM borate at pH 9.5. Injection 50 mbar, 6 s; voltage: 25 kV; Detection: excitation at 488 nm and emission at 520 nm. Riboflavin peak is shown with an asterisk. Reprinted from [18] with permission.

On the other hand, fluorescent detectors are not as easy to handle as UV because some separation medium can result in fluorescent quenching. However, there are cases where fluorescence intensity increases depending on the content of the separation medium. Curcumin is a bioactive substance of turmeric spice that has been very popular in recent years. However, there are two more curcuminoids beside curcumin in turmeric and each has different bioactivities [19].

Fig. (11))

Separation of curcuminoids and the effect of 2-HP-β-CD in buffer on fluorescence intensities of curcuminoids. (A) Buffer: 25 mM borate, 35 mM 2-HP-β-CD; (B) Buffer: 25 mM borate. Injection: 50 mbar 6s, Voltage: 25 kV, detection: excitation at 488 nm and emission at 520 nm; 1: BDMC (1.5 µg/mL); 2: DMC (20 µg/mL); 3: Curcumin (20 µg/mL). Reprinted from [20] with permission.

Similarly, the fluorescence intensities of biogenic amines, which are derivatized by fluorescein isothiocyanate (FITC) in MEKC-LIF analysis, increase to between 4 and 21 folds when Brij 35 is substituted for SDS in the same separation medium as seen in Fig. (12).

Fig. (12))

Effect of Brij 35 in buffer on fluorescence intensities of biogenic amines derivatized with FITC.

(A) Buffer: 75 mM borate, 20 mM SDS, pH 9.7 (B) 75 mM borate, 10 mM Brij, pH 9.7. Voltage: 25 kV, capillary: 50 µm, (63x47) cm; injection: hydrodynamic, 50 mbar, 6s. 1:Trypthophan, 2:Tyrosine, 3:Cadaverine, 4:Spermidine, 5:Histidine, 6:Putrescine. Concentrations: 0.3 µM for each. *:FITC. Reprinted from [21] with permission.

MS Detection

Combining CE with MS allowed many compounds to separate with high separation efficiency and obtain molecular mass information of the separated analytes. Although the most used interface in CE-MS applications until today is Electrospray Ionization Interface (ESI), there are also studies where CE is combined with MALDI-MS and some other ionization techniques. CE-MS combination has been mostly used in proteomics, foodomics, pharmaceutical, and clinical researches. Recent developments and applications in CE-MS have been collected in review articles [22, 23]. Although less common, inductively coupled plasma mass spectrometry coupled with CE (CE-ICP-MS) has also been used for metal- biomolecule relationships [24].

Contactless Conductivity Detection

A Contactless Conductivity Detector (C⁴D) for capillary electrophoresis was first introduced in 1998 [25]. Since then, this method has been applied for the detection of inorganic and organic ions in electrophoresis [26]. C⁴D allows simple combination with commercial CE instruments and can be performed in an on- capillary mode. The detector contains two electrodes that are placed around the polyimide coating of the silica capillary column. The electrical conductivity changes of the solution between two electrodes are measured.

Fig. (13))

Separation and CE-C⁴D detection of biogenic amines. 1:spermidine, 2: putrescine, 3: histamine, 4: cadaverine, 5: tyramine, and ‘*’ represents system peaks. The concentration of each amine is 2 mg/L, except tyramine. Its concentration is 4 mg/L. Separation electrolyte: 500 mM hydroxyisobutyric acid (HIBA), pH 2.05. Separation conditions: 28 kV, 20 °C. injection: hydrodynamic, 60 mbar for 60 s from the anodic side. Reprinted from [27] with permission.

Other Detection Methods

Although they are not routinely used as the above detection methods, different detection systems have been combined with CE such as amperometric detection, especially in research studies. Amperometric detection is particularly suitable for microchip CE.

2.05. Separation conditions: 28 kV, 20 °C. injection: hydrodynamic, 60 mbar for 60 s from the anodic side. Reprinted from [27] with permission.

SEPARATION STRATEGIES FOR SOME ANALYTE GROUPS

CE is a separation method that can be applied to samples in a broad spectrum with different separation modes. It has been used in sample analysis in many different fields due to its high speed, easy method development, and high separation efficiency. In this section, the separation strategies applicable for some special groups will be briefly mentioned.

Small Inorganic Anions and Organic Acids

As mentioned before, injections are generally performed from the anodic side in CE separations. All the species move towards the cathode direction and reach a detector located near the cathode end of the capillary. The migration behaviors of small inorganic anions and organic acids are exceptions. Due to the high electrophoretic mobility of both groups, EOF can not easily drag them into the cathodic side. Thereby, they cannot reach the detector, or their migration times become very long. Separation options for small negatively charged ions are given schematically in (Fig. 14A-C). In Fig. (14A), a cationic surfactant below its critical micelle concentration (CMC) is added to the separation buffer. The monomers of cationic surfactants are attracted to the negatively charged capillary wall. The hydrophobic tails of surfactant monomers adhering to the wall interact with the tails of the free surfactant molecules in the solution, forming a double layer on the wall. Thus, the capillary inner wall is positively charged. In this case, the direction of the EOF in the capillary is reversed from the cathode to the anode. It is possible to change the polarity of the electrodes in CE devices. Changing the polarity, the injection is made from the cathodic side, and the detector is on the anodic side. When negatively charged small ions are injected from the cathodic side, they migrate to the anodic side due to their electrophoretic mobility. Since the EOF in the capillary is also directed to the anode, a high-speed separation of these ions takes place.

Cetyltrimethylammonium bromide (CTAB) is the most commonly used surfactant for dynamically coating the capillary wall. Fig. (15) shows the separation of small organic acids in a pollen sample using CTAB in the separation medium [28].

Fig. (14))

Electromigration of small negatively charged ion. A) Reversed EOF with a cationic surfactant in buffer, B) Reversed EOF in a coated capillary with a positively charged polymer. C) Reduced EOF.

Fig. (15))

Separation of organic acids in a pollen extract (Buckwheat pollen). Conditions: capillary 56.5 cm effective length 50 μm I.D; Separation electrolyte: 5 mmol/L of PDC + 0.1 mmol/L CTAB, pH: 5.6. Detection wavelength: 350 nm with a reference at 200 nm. Voltage: -25 kV. Peaks: 1: Oxalic acid, 2: Tartaric acid, 3: Malic acid, 4: Citric acid, 5: Succinic acid, 6: Acetic acid, 7: Lactic acid, 8: Gluconic acid. The on-set electropherogram shows the analysis of the diluted pollen extract for the quantification of gluconic acid. Reprinted from [28] with permission.

As shown schematically in Fig. (14B), alternatively, by physically adhering or covalently bonding a positively charged polymer to the wall, the capillary inner wall is positively charged for a long time. (Fig. 16) shows the separation of nitrate and nitrite ions in a physically coated capillary column with high molecular weight PEI. Thiocyanate ion has been used as the internal standard. As can be seen, separation is extremely fast. This method has been used for the quantification of both ions in processed meat products and some vegetables [29].

Fig. (16))

Simultaneous separation of nitrate and nitrite ions in PEI coated capillary. Buffer: 20 mmol/L Tris, pH: 7.5. Voltage: -28 kV, Injection: 4.10-3 MPa. Total length 75 cm, effective length 60 cm. i.d. 75 μm. Direct UV detection at 210 nm. Peaks: 1:nitrite, 2:nitrate, 3:thiocyanate (internal standard). Reprinted from [29] with permission.

Another option is to suppress EOF in the capillaries and allow ions to move against EOF. As the pH of the working buffer decreases, the negative charge density in the capillary wall decreases, as a result of which the speed of the EOF decreases. The pH of the separation buffer is chosen at low values. The sample containing negatively charged small ions is injected from cathodic side. Negatively charged ions move towards the anode side. Meanwhile, although there is an EOF in the direction of the cathode in the capillary, its speed has decreased very much. Small ions with high electrophoretic mobility move rapidly towards the anodic side against EOF. This form of separation is shown schematically in Fig. (14C). Fig. (6) shows the separation of nitrate and nitrite ions in suppressing EOF mode.

Metal ions

The general approach to separating metal ions, whose charge and mass ratios are close together, is to complex them with a suitable ligand. The separation of metal ions is based on the differences in the charge and size ratios of the complexes as a result of the differences in the complex formation constants. If the ligand used is UV active, direct UV detection is possible. Metal complexes can be determined indirectly using a suitable chromophore together with a UV inactive ligand. Two different methods are applied to the separation of metal ions. If the stability constants of the complexes are high, precolumn complexation is preferred, and the metal ion-ligand mixture is injected into the capillary column. In the case of weak complexes, an on-column complexation is preferred. A ligand is added into the buffer solution, and metal ions are injected into this separation medium. More than one ligand can be used if the charges, ionic diameters of metal ions and the formation constants of complexes formed by a ligand are very close to each other. Separation is accomplished by competing ligands for metal ions. Lanthanide group metal ions are one of the examples for this situation, as seen in Fig. (17).

Fig. (17))

Electrophoregram of 14 lanthanide group metal ions. Buffer: 15 mmol/L HIBA, 13 mmol/L Tris, 0.1 mmol/L cupferron, pH 4.9. Injection 40 mbar, 0.06 min. Run voltage 28 kV. 1: La, 2: Ce, 3: Pr, 4: Nd, 5: Sm, 6: Eu, 7: Gd, 8: Tb, 9: Dy, 10: Ho, 11: Er, 12: Tm, 13: Yb, 14: Lu. Reprinted from [30] with permission.

The separation of lanthanide group metal ions is a challenge of chromatographic methods considering their similarity. In this study, 14 lanthanide group metal ions were separated by the addition of two ligands, namely HIBA and cupferron, into the separation buffer [30]. Direct UV detection of lanthanides complexed with cupferron was possible since cupferron is a UV-active ligand. Since the weak complexes are formed, both ligands were added into the separation medium, and lanthanide ions were then injected into the capillary.

Metal complexes can be determined with conductivity detectors without requiring a UV active ligand or chromophore. Fourteen lanthanide ions, complexed with HIBA and acetic acid, were separated and determined by the CE-C⁴D system [31].

Proteins

There are many applications in which proteins are separated by different electromigration modes such as CZE, CITP, CIEF, CEKC, CEC collected in a recent review article [32]. As stated in the CGE section, SDS-treated proteins can also be separated by size differences. In protein separation with CE, adsorption of proteins to the negatively charged capillary wall should be prevented. One method is to suppress the positive charges of the proteins using a separation medium having a pH well above the isoelectric points of the proteins. The more effective and widely used method is to coat the capillary inner wall with a suitable polymeric layer according to the type of protein. Dynamic and permanent coating methods for protein separation in CE and new coating materials with nanoparticles were collected in the last review article [33].

(Fig. 18) shows effective separation of basic proteins in a capillary coated with a positively charged polymer (PEI) [34].

Fig. (18))

Separation of five basic proteins in PEl-coated capillary. Total length 63.5 cm, effective length 46.5 cm, I.D. 75 μm. Buffer: 50 mM acetate, pH 5.5. Run voltage -28.8 kV. Injection 10 mbar for 6 s. Sample: 1: trypsinogen 0.32 mg/mL, 2: α-chymotrypsinogen 0.32 mg/mL, 3: ribonuclease A 0.72 mg/mL, 4: cytochrome C 0.32 mg/mL and 5: lysozyme 0.32 mg/mL. UV detection at 214 nm. Reprinted from [34] with permission.

Chiral Molecules

In CE, chiral separation is performed in electrokinetic chromatography mode. A soluble chiral selector in the separation buffer forms a pseudo-stationary phase. During migration, enantiomers are shared between the chiral selector and the solution phase. Interaction differences between the enantiomers and the chiral selector provide the enantioselectivity [35]. Chiral separation with CE is very easy to perform, and high separation efficiencies are achieved. Chiral selective cyclodextrins are the most commonly used in CE separations. Besides numerous commercially available cyclodextrins, research is on the development of new cyclodextrins. However, there are also studies in CE separations where different potential selectors such as ionic liquids are tested [36]. Analytical chiral separation is especially important in the pharmaceutical industry. Therefore, most CE applications in the literature are on chiral drug analysis. In recent years, chiral separations have gained importance also in the field of food [37].

Fig. (19). shows the separation of four pairs of chiral aminoacid with the addition of 10 mM betacyclodextrin (β-CD) to the separation medium as the chiral selector.

Fig. (19))

Electropherogram of amino acids in 50 mM borate buffer at pH 9.6 containing 5 mM SDBS, 10 mM β-CD; Voltage: 25 kV. Temperature: 25 °C. Capillary 50 μm, (60 × 40) cm. Injection: hydrodynamic, 50 mbar, 6 s. 1: D-Arg, 2: L-Arg, 3: L-Pro, 4: D-Pro, 5: D-Ala, 6: L-Ala, 7: D-Asp, 8: L-Asp, ∗: peaks from FITC. Concentrations: 0.6 μM for each, except Asp as 1.2 μM. Reprinted from [38] with permission.

CONCLUDING REMARKS

CE is a versatile analytical technique because of its different operation modes and plays a great role in the field of separation science based on its multiple applications. Moreover, some of its advantages such as easy operation modes, different detection techniques, enhanced separation efficiency, and separation speed make CE a sophisticated technique. Our approach in this chapter was to give the basic principles of CE and some separation applications for food samples mostly. It appears that CE continues to be developed and be utilized in all branches of chemistry, to expand its application areas.

CONSENT FOR PUBLICATION

Not applicable.

CONFLICT OF INTEREST

The author declares no conflict of interest, financial or otherwise.

ACKNOWLEDGEMENTS

F.B. Erim would like to commemorate the late Prof. J.C. Kraak, who introduced her to CE for the first time in 1993, and to thank Istanbul Technical University, which gave her the opportunity to establish a CE research laboratory.

REFERENCES

Sample Preparation in Capillary Electrophoresis for Food Analysis

Ling Xia¹, Simin Huang¹, Gongke Li¹, *

¹ School of Chemistry, Sun Yat-sen University, Guangzhou, 510275, China

Abstract

This chapter introduces sample preparation techniques in Capillary Electrophoresis (CE) for food analysis. Food sample preparation prior to CE analysis aims to transfer target analytes from random statuses in the original food matrix to highly ordered pre-detection statuses, which is an entropy reduction procedure and cannot happen spontaneously. Generally, this is a time-consuming, labor-intensive, and error-prone step in complex sample analysis, especially in food analysis. Nevertheless, to match the fast analysis nature of CE, food samples have to be prepared efficiently in a relatively short time. Therefore, many highly efficient and fast sample preparation techniques were applied in CE for food analysis, including phase separation, field-assisted extraction, membrane separation, chemical conversion, and online coupling of sample preparation/analysis techniques. The principles and operation of each of the above-listed sample preparation techniques and some application examples are shown in different sections.

Keywords: Chemical conversion, Field- assisted extraction, Food analysis, Membrane separation, Sample preparation, capillary electrophoresis, phase partition.

* Corresponding author Gongke Li: School of Chemistry, Sun Yat-sen University, Guangzhou, 510275, China; E-mail: [email protected]

INTRODUCTION

Sample preparation is a critical step in capillary electrophoresis for food analysis, which affects sensitivity, selectivity, accuracy, and speed of analytical results. From an analytical science viewpoint, foodstuff is a type of sample with wide varieties and high complexity, which contain a large number of target analytes and in a very complex matrix [1]. Moreover, these target analytes may be extremely similar to each other but present a trace amount in the food sample matrix [2-4]. Before injecting foodstuff into capillary, the preparation process can isolate and/or preconcentrate target analytes from food matrix based on their different physical, chemical, and biological properties, making them suitable for

electrophoretic analysis [5]. Sample preparation is a time-consuming process, since transferring analytes from random status in the original sample matrix to highly ordered pre- detection statuses is an entropy reduction procedure which cannot occur spontaneously [6, 7]. Nevertheless, to match the fast analysis nature of capillary electrophoresis (CE), food samples have to be efficiently prepared in a relatively short time [8]. Therefore, many high efficiency and fast sample preparation methods and techniques were applied in CE for food analysis as shown in Fig. (1), including phase separation, field-assisted extraction, membrane separation, chemical conversion,