1660 ´ Blanka Vochyanov a´ 1 1 ˇ Frantisek Opekar 2 ˚ Petr Tuma 1 Faculty
of Science, Department of Analytical Chemistry, Charles University in Prague, Albertov, Prague, Czech Republic 2 Third Faculty of Medicine, Institute of Biochemistry, Cell and Molecular Biology, Charles ´ University in Prague, Ruska, Prague, Czech Republic
Received October 3, 2013 Revised October 31, 2013 Accepted November 18, 2013
Electrophoresis 2014, 35, 1660–1665
Research Article
Simultaneous and rapid determination of caffeine and taurine in energy drinks by MEKC in a short capillary with dual contactless conductivity/photometry detection A method has been developed for the simultaneous determination of taurine and caffeine using a laboratory-made instrument enabling separation analysis in a short 10.5 cm capillary. The substances are detected using a contactless conductometry/ultraviolet (UV) photometry detector that enables recording both signals at one place in the capillary. The separation of caffeine and taurine was performed using the MEKC technique in a BGE with the composition 40 mM CHES, 15 mM NaOH, and 50 mM SDS, pH 9.36. Under these conditions, the migration time of caffeine is 43 s and of taurine 60 s; LOD for caffeine is 4 mg/L using photometric detection and LOD for taurine is 24 mg/L using contactless conductometric detection. The standard addition method was used for determination in Red Bull energy drink of caffeine 317 mg/L and taurine 3860 mg/L; the contents in Kamikaze drink were 468 mg/L caffeine and 4110 mg/L taurine. The determined values are in good agreement with the declared contents of these substances. RSD does not exceed 3%. Keywords: Caffeine / Dual detection / Energy drinks / Micellar electrokinetic chromatography / Short capillary separation / Taurine DOI 10.1002/elps.201300480
1 Introduction Next to saccharides, caffeine and taurine are the main components in most energy drinks. Taurine has a favorable effect on a number of physiological processes in the organism; among other things, it optimizes the activity of the central nervous system, improves physical comfort, and stimulates thought processes. Its effects substantially augment those of the alkaloid caffeine. Energy drinks containing these substances improve performance and awareness and their consumption frequently increases with demands placed on human activity and generally with an increase in lifestyle intensity. However, an overdose of these substances, especially caffeine, can have undesirable effects [1, 2]. It is thus important to determine the contents of these two substances in energy drinks and a number of analytical methods have been developed for this purpose. General overviews of the methods employed for determining caffeine and taurine in various matrices can be found in a number of reviews, for example [3–5]. The most frequently used methods are LC (HPLC) and CE—for determining caffeine ˚ Correspondence: Dr. Petr Tuma, Institute of Biochemistry, Cell and Molecular Biology, Third Faculty of Medicine, Charles University in Prague, Ruska´ 87, 100 00 Prague 10, Czech Republic E-mail: [email protected] Fax: +420 267 102 460
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
MEKC, or GC. Detection methods include primarily MS and UV/Vis spectrophotometry. The individual components are generally determined separately in beverages and energy drinks. For example, recently, works have been published describing the determination of caffeine based on MEKC with UV detection [6], MEEKC with UV detection [7], UV/Vis derivative spectrophotometry [8], and capillary chromatography in a monolithic column with UV detection [9]. Caffeine has also been determined voltammetrically, for example, on a modified glassy carbon electrode [10] or using a boron-doped diamond electrode [11]. The determination of taurine has recently been described in the publications [12], CE with LIF detection [13], HPLC with precolumn derivatization and UV/Vis detection [14], amino acid analyzer and FTIR spectroscopy [15], MCE after derivatization with fluorescence detection. Simultaneous determination of the two analytes has, as far as we know, been described only in the works HPLC-MS method [16], planar chromatography with multiwavelength absorbance/fluorescence detection [17], and hydrophilic interaction chromatography with UV and evaporative light scattering detection in tandem [18]. Most of the published methods employ financially and technically demanding instrumentation combined with complicated sample pretreatment, usually resulting in low detection limits. However, the concentration of active components is high in energy drinks and thus rapid analysis, www.electrophoresis-journal.com
Electrophoresis 2014, 35, 1660–1665
CE and CEC
1661
undemanding sample pretreatment and low price are far more important for this type of determination than low LOD. The submitted work describes a method for the rapid simultaneous determination of caffeine and taurine in energy drinks using MEKC in a short 10.5 cm capillary with dual detection: optical detection, UV/Vis, is used for detection of caffeine and contactless conductivity detection, C4 D, for determining taurine. The detection takes place at a single site in the separation capillary. The method was tested on the determination of both analytes in the energy drinks Red Bull and Kamikaze. CE in short capillary represents efficient analytical tool especially in combination with multiple detection. The fundamental methodology of rapid analyses by electrophoresis along short separation pathways has been discussed from many aspects in reviews [19, 20] as well as the various combinations of detectors [21].
2 Materials and methods 2.1 Laboratory-made CE apparatus with dual contactless conductivity/photometry detection The separation and detection apparatus is located on the front panel of a Sapphire (ECOM, Czech Republic) UV/Vis detector under the optical and electrostatic shielding cover, Fig. 1. In this compact arrangement, the separation capillary (1), injection (2) and terminal (3) vessels, dual detection cell (4), electronic of the contactless conductivity detector (5), and (6) are placed under the shielding cover. The structural design of the C4 D is taken from [22]. Two semitubular electrodes (4a) are cut from self-sticking copper foil with a thickness of 35 m, electrode width of 1.7 mm, and distance between electrodes of 0.85 mm. Alternating current is fed to one of the electrodes from a sinus voltage generator (5). The signal passing through the cell is recorded by the second electrode and, following amplification and rectification (6), is fed into a computer for registration. In the optical part of the detection cell, radiation is brought by an optical fiber from the monochromator in the spectrophotometric detector to the detection window of the capillary in the space between the conductivity electrodes. The radiation passing through the capillary is registered by a large-area diode located opposite the optical cable in the lid of the detection cell, which is part of the Sapphire UV/Vis detector. The injection of the sample into the short capillary and its washing is based on the procedure described in detail in [23, 24] and will thus be mentioned only briefly here. The sample is injected into the capillary in such a manner that, simultaneously with turning the six-way injection valve (9) to the “inject” position, the separation electrolyte is fed through tube (10) to the injection valve at a constant rate and for a defined time. The stream of electrolyte carries defined volume of the sample from the sampling loop (11) around the injection end of the capillary, which is inserted to a depth C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 1. Scheme of the apparatus for MEKC in a short capillary with dual detection. The components of the apparatus depicted in the box are located under the shielding cover on the front panel of the spectral detector. Separation capillary (1), injection vessel (2), terminal vessel (3), dual detection cell (4), sinus voltage generator (5), alternating current meter and rectifier (6), outlet for further processing of the signal (7), high-voltage electrodes (8), injection valve (depicted in the “inject” position (9)), inlet for separation electrolyte during sample injection (10), sampling loop (11), injection tube (12), waste outlet for excess electrolyte during injection (13), and inlet for connection to the underpressure source for washing or activation of the capillary (14).
of approx. 0.5 mm into the feed PTFE tube (12) in the injection vessel (2). The injection period is controlled by the time during which the zone of sample solution flows around the capillary injection end; excess separation electrolyte flows out into waste (13) during the injection period. The injection takes place without interruption of the high voltage on the electrophoretic electrodes (8). After the end of the separation, the capillary is rinsed with the separation electrolyte by decreasing the pressure in the terminal vessel (14); the NaOH activation solution or the wash water was drawn into the capillary also by underpressure in the terminal vessel. The injection loop is filled using an injection syringe. Experimental parameters used in the measurement: standard fused silica capillary (Polymicro Technologies, USA) with an id of 50 m, overall length of 10.5 cm, and length to the detector of 8 cm; detection cell, C4 D, 450 kHz/17 Vpp (peak to peak), UV 216 nm; sampling loop volume, 16 L. The injection parameters were set so that the time of passage of the sample zone around the injection end www.electrophoresis-journal.com
1662
´ B. Vochyanov a´ et al.
of the capillary was about 0.3 s. Separation was performed at a voltage of +5 kV. The time of washing the capillary by the separation electrolyte after the end of the separation equaled 30 s.
Electrophoresis 2014, 35, 1660–1665
added to the analyzed solutions, so that their concentrations could be evaluated from a single recording.
3 Results and discussion 2.2 Chemicals and samples The following were used to prepare the BGEs: CHES (Roth, Germany), CAPS (Aldrich), NaOH (Sigma), and SDS (Sigma). The optimized BGE, see later, has the composition 40 mM CHES, 15 mM NaOH, and 50 mM SDS, pH 9.36; the electrolyte was prepared in DI water (Milli-Q Plus, Millipore, USA). Standard solutions of caffeine (Sigma-Aldrich) and taurine (Roth) with concentrations of 992 and 1250 mg/L, respectively, were prepared in DI water. Analyzed samples of energy drinks are commonly commercially available (producer or distributor in parentheses) with the caffeine and taurine contents given on the label: Red Bull (Red Bull, Austria) − caffeine, 80 mg in 250 mL (i.e., following recalculation 320 mg/L), taurine, 0.4% (4000 mg/L) Kamikaze (Tecfood, CR) − caffeine, 120 mg in 250 mL (480 mg/L), taurine, 4000 mg/L The only pretreatment of the beverage samples consisted in ten-min sonication to remove dissolved gases. The samples were then stored in a refrigerator and were diluted 10× for all the measurements using the separation electrolyte. Solutions of standards and energy drinks prepared for analysis by dilution with water did not yield satisfactorily reproducible results. We suppose that in the described dynamic injection method, combining moving sample zone with electrokinetic sampling, two basic factors can influence the reproducibility of the injection: (i) irreproducible mixing of the aqueous sample with the separation electrolyte in the inlet tube before reaching the sampling end of the capillary and (ii) undefined sample stacking. These effects lead to various amount of analyte injected into capillary and consequently to unsatisfactorily reproducible result.
2.3 Experimental procedure Prior to the separation, the capillary was activated for 10 min with a solution of 0.1 M NaOH, thoroughly washed with water and then filled with the separation electrolyte. Activation was performed twice daily. The current after connection of the separation voltage of 5 kV equalled 32–33 A. The separation voltage was not turned off for the entire measuring period. The calibration graph method and the method of standard additions were tested for determining caffeine and taurine. In both cases, standard solutions of both analytes were C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3.1 Separation of taurine and caffeine Taurine has two dissociated groups: an acidic sulpho-group with pKA 1.5 and a basic amino-group with pKA 9.2. In the pH range from pH 2.5 (pKA of the sulfo-group +1) to 8.2 (pKA of the amino-group −1), taurine is present as a zwitterion and does not move in the electric field. CE separation of taurine must be performed in BGE with pH greater than 8.7 to ensure sufficient separation of the taurine anion from the zone of electroneutral substances. Two BGEs were tested for the separation: CHES/NaOH in the range pH 8.7–10 and CAPS/NaOH at pH 10–11; pKA (CHES) = 9.4, pKA (CAPS) = 10.5. Experiments have shown that it is preferable to perform the separation of taurine in CHES/NaOH medium, where taurine is sufficiently separated from the zone of electroneutral substances and the sensitivity C4 D is twice as great as that for separation in CAPS/NaOH. In the positive separation mode, the peak of the water gap appears first on the electropherogram, followed by the positive peak of taurine, which, as an anion, moves in the opposite direction to EOF. At pH above 8.7 it is, however, not possible to determine caffeine by the CE technique as it is not dissociated under these conditions and moves together with the other electroneutral substances in the water gap zone. Consequently, SDS with the normally used concentration of 50 mM was added to CHES/NaOH and the MEKC technique was employed. Hydrophobic caffeine enters the anionic migrating micelle and is separated in this process from the zone of electroneutral substances. The presence of SDS in the electrolyte does not have any effect on the migration of hydrophilic taurine. On the electropherogram of energy drink in BGE with optimum composition 40 mM CHES, 15 mM NaOH, and 50 mM SDS, pH 9.36, the peak of EOF is visible first, followed by the peak of caffeine, and finally by the peak of taurine, see Fig. 2. The difference between electropherogram obtained with the sample and that obtained with the standard mixture is negligible. A dual C4 D/UV detector was used for the simultaneous determination of taurine and caffeine. Taurine, which does not absorb in the UV/Vis part of the spectrum, is detected as a positive peak using C4 D (the mobility of the taurine anion is greater than the mobility of the BGE co-ions). On the other hand, caffeine with purine structure is detected photometrically at the optimized wavelength of 216 nm. Caffeine also yields a signal in universal C4 D. However, it was verified experimentally that the response of C4 D to caffeine changes very little with a change in its concentration. It can be assumed that the zone caffeine/micelles affects primarily the permittivity of the solution, to which C4 D reacts [25] and not the electrical conductivity. www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 1660–1665 Table 1. Reproducibility of the migration times and peak areas of caffeine and taurine with concentrations of 99.2 and 125.0 mg/L, respectively, in a separation electrolyte solution of 40 mM CHES, 15 mM NaOH, and 50 mM SDS
Parameter
Caffeine
Taurine
Migration time (s) RSD (%) Peak area (mV s or mAU s) RSD (%)
43.4 ± 0.3 0.96 11.7 ± 0.8 5.7
60.0 ± 0.5 1.26 16.1 ± 0.7 3.4
The results were obtained from five repeated measurements.
1663
Table 2. Parameters of the linear calibration dependences for the determination of caffeine and taurine by the calibration graph method
Caffeine Concentration range (mg/L) Number of calibration points Slope (mAU s L/mg) Intercept (mAU s) Coefficient of determination, R2 LOD (mg/L (mol/L)) Taurine Concentration range (mg/L) Number of calibration points Slope (mV s L/mg) Intercept (mV s) Coefficient of determination, R2 LOD (mg/L (mol/L))
19.8–496.0 5 0.116 (0.001)a) −0.447 (0.310)a) 0.9996 4 (2 × 10−5 ) 25.0 to 625.0 5 0.159 (0.006)a) −3.511 (1.832)a) 0.9958 24 (2 × 10−4 )
a) Standard deviation in parenthesis.
the concentrations of both analytes are an order of magnitude higher in the samples for which the method is being tested.
3.3 Standard addition method Linear concentration dependences permit the use of the standard addition method for determining both analytes. Two standard additions were employed: for determining caffeine 99.2 and 198.4 mg/L and for determining taurine 125.2 and
Figure 2. Electropherogram of the energy drink Kamikaze with C4 D recording (thin line) and UV recording at 216 nm (bold line). Experimental conditions, BGE 40 mM CHES +15 mM NaOH + 50 mM SDS (pH 9.36); +5 kV/+33 A; sample diluted tenfold with BGE. Peak identification: caffeine (1), taurine (2).
The reproducibility of the migration times and peak areas were tested on an artificial sample of a mixture of caffeine and taurine. It is apparent from the results given in Table 1 that the reproducibility of the tested parameters is quite satisfactory for analytical purposes.
3.2 Calibration graph method The dependences of the peak areas for caffeine and taurine on the concentrations were linear in the entire tested concentration interval. A regression straight line was drawn through the calibration points (average of three repeated measurements), where the parameters for both analytes are given in Table 2. The results of the determination are listed in Table 3. The LOD was estimated from three times the amplitude of the noise in the two detection systems (0.3 mV and 0.03 mAU) and the parameters of the relevant calibration straight lines. The LOD values are relatively high; however, C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Table 3. Results of the determination of caffeine and taurine in energy drinks using the calibration graph and standard addition methods
Red Bull
Kamikaze
Calibration graph method Caffeine concentration (mg/L) Percent of declared value (%) Range, w (mg/L) RSD (%) Taurine concentration (mg/L) Percent of declared value (%) Range, w (mg/L) RSD (%)
324.7 ± 1.2 101.5 0.9 0.2 4123.9 ± 139.9 103.1 107.6 1.5
459.2 ± 13.5 95.7 10.4 1.3 4178.6 ± 339.3 104.5 261.0 3.7
Standard addition method Caffeine concentration (mg/L) Percent of declared value (%) Range, w (mg/L) RSD (%) Taurine concentration (mg/L) Percent of declared value (%) Range, w (mg/L) RSD (%)
317.0 ± 19.2 99.1 14.8 2.7 3857.0 ± 235.3 96.4 180.5 2.8
467.7 ± 11.2 97.5 21.9 2.0 4110.4 ± 224.0 102.8 172.3 2.5
The declared values for the contents of caffeine are equal to 320 and 480 mg/L in the Red Bull and Kamikaze beverages, respectively, while that of taurine is 4000 mg/L in both beverages.
www.electrophoresis-journal.com
1664
´ B. Vochyanov a´ et al.
Electrophoresis 2014, 35, 1660–1665
250.4 mg/L. The results of the determinations are also given in Table 3. All the results are the average of three parallel determinations: the range of results, that is, the difference between the largest and smallest determined values, w, is also given. The confidence interval was calculated for the 95% significance level. The relationship between the experimentally determined concentration values and the values declared on the label of the beverage tin is expressed as the percentage fraction of the determined and declared values. 3.4 Evaluation of the results It is apparent from Table 3 that, in most cases, the declared values of the concentrations of the two analytes are within the reliability interval of the experimentally determined concentrations. Nevertheless, a simple statistical TI test [26] was used to compare the experimental values with the declared values, TI = |x¯ − V | /w, where x¯ is the experimentally determined average value, V is the standard value, in this case the declared value, and w is the range. If the determined TI value is smaller than the critical value, the experimental result is not statistically different from the declared value. The results determined by the calibration graph method and the standard addition method were compared using Lord’s test [26], L = |x¯1 − x¯2 | /(w1 + w2 ), where x¯i and wi are the average values of the results obtained by the relevant calibration method and the corresponding range of results. All the data required for performance of the tests are given in Table 3 and the results of the testing are listed in Table 4. Values of the statistical parameters exceeding the critical value are designated in bold in Table 4. It can be seen that, when the standard addition method is employed, none of the Table 4. Results of the statistical testing of similarity of the experimentally determined concentration values to the declared values, TI test, and test of the similarity of the concentrations determined by the calibration graph method and the method of standard additions, L test
Red Bull TI values Calibration graph method versus declared value Caffeine 5.20a) Taurine 1.15 Standard addition method versus declared value Caffeine 0.20 Taurine 0.79 L values Comparison calibration graph vs. standard addition Caffeine 0.49 Taurine 0.93
Kamikaze
2.0 0.68 0.56 0.64
0.26 0.16
Critical values for three repeated measurements and the 95% significance level are 1.304 and 0.64 for the TI test and L test, respectively. Values in bold exceed the parameters’ critical value. a) The high TI value is caused by the extremely low value of the range of the determined concentrations in the test series, see Table 3.
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
determined concentrations are statistically different from the declared values. However, when the calibration graph method is employed, the determined values of the caffeine content are statistically different from the declared values. It follows from the Lord’s test values that the two calibration methods yield comparable results that were statistically different only in the determination of taurine. On the basis of the results of the test, the more reliable calibration method of standard additions can be recommended for the determination of caffeine and taurine in energy drinks. The long-term reproducibility of the described methods was tested by determining caffeine and taurine by the standard addition method in the Red Bull beverage. The results of the analysis mentioned in Table 3 were compared with the results obtained by repeated sample analysis after ten days. The determined contents of caffeine, 315.5 ± 26.0 (RSD 3.7%, recovery 98.6%) and taurine, 3987.9 ± 263.9 (RSD 3.0%, recovery 99.7%), are fully comparable with the results in Table 3. Between analyses, the beverage sample was stored in the refrigerator. All the results presented have been obtained without necessity to exchange the capillary. Standard addition calibration method eliminates the possible variations in sample injection when the capillary has to be changed.
4 Concluding remarks The determination of caffeine and taurine in energy drinks by separation in a short capillary with dual detection is a useful extension of the formerly described determination of saccharides in these beverages [24]. Pretreatment of the sample is simple and the determination is rapid; under the described experimental conditions, the separation requires about 1 min. When the standard addition method is used, the determined contents of the two analytes are statistically identical with the contents declared on the labels of the beverages. The method is a suitable alternative to the previously described methods permitting simultaneous determination of both active components of energy drinks. The described method is based on the use of a specialized apparatus. It can be anticipated that the method could be adapted for a commercial electrophoresis apparatus that would enable sample injection into the short end of a capillary and would be equipped with C4 D and UV detection. In these apparatuses, the two detection systems are usually at different places in the capillary so that the substances are detected at different degrees of separation [27]. However, this in no way affects the determination. The authors wish to thank Ecom s.r.o., Czech Republic, for cooperation in development of the apparatus and the Ministry of Education, Youth and Sports, Czech Republic, Research Project No. MSM0021620857 for financial support. This work was also supported by Charles University in Prague, project SVV and PRVOUK 31. The authors have declared no conflict of interest. www.electrophoresis-journal.com
Electrophoresis 2014, 35, 1660–1665
5 References [1] Huxtable, R. J., Physiol. Rev. 1992, 72, 101–163. [2] Nehlig, A., Daval, J. L., Debry, G., Brain Res. Rev. 1992, 17, 139–169. ˇ ˇ ık, P., Sv´ıtkova, ´ J., Rievaj, M., Bustin, D., [3] Svorc, L., Tomc´ Chem. Listy 2013, 107, 530–536. [4] Talik, P., Krzek, J., Ekiert, R. J., Sep. Purif. Rev. 2012, 41, 1–61. [5] Mou, S., Ding, X., Liu, Y., J. Chromatogr. B 2002, 781, 251–267. [6] Ballus, C. A., Meinhart, A. D., Bizzotto, C. S., Teixeira Filho, J., Godoy, H. T., Ciencia Technol. Alime. 2012, 32, 401–404. [7] Liotta, E., Gottardo, R., Seri, C., Rimondo, C., Miksik, I., Serpelloni, G., Tagliaro, F., Forensic Sci. Int. 2012, 220, 279–283. [8] Pieszko, C., Baranowska, I., Flores, A., J. Anal. Chem. 2010, 65, 1228–1234. [9] Al-Othman, Z. A., Aqel, A., Alharbi, M. K. E., BadjahHadj-Ahmed, A. Y., Al-Warthan, A. A., Food Chem. 2012, 132, 2217–2223. [10] Zhao, F., Wang, F., Zhao, W., Zhou, J., Liu, Y., Zou, L., Ye, B., Microchim. Acta 2011, 174, 383–390. ˇ ˇ ık, P., Sv´ıtkova, ´ J., Rievaj, M., Bustin, D., [11] Svorc, L., Tomc´ Food Chem. 2012, 135, 1198–1204. [12] Zinellu, A., Sotgia, S., Bastianina, S., Chessa, R., Gaspa, L., Franconi, F., Deiana, L., Carru, C., Amino Acids 2009, 36, 35–41.
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
CE and CEC
1665
[13] Sawabe, Y., Tagami, T., Yamasaki, K., J. Health Sci. 2008, 54, 661–664. [14] Triebel, S., Sproll, C., Reusch, H., Godelmann, R., Lachenmeier, D. W., Amino Acids 2007, 33, 451–457. ¨ [15] Gotz, S., Revermann, T., Karst, U., Lab. Chip 2007, 7, 93–97. [16] Marchei, E., Pellegrini, M., Pacifici, R., Palmi, I., Pichini, S., J. Pharm. Biomed. Anal. 2005, 37, 499–507. [17] Aranda, M., Morlock, G., J. Chromatogr. A 2006, 1131, 253–260. [18] Chirita, R., Dascalu, C., Gavrila, L., Elfakir, C., Rev. Chim.: Bucharest 2010, 61, 1173–1176. ˇ ık, K., Chem. Rev. 2009, 109, [19] Opekar, F., Coufal, P., Stul´ 4487–4499. [20] Matysik, F. M., Anal. Bioanal. Chem. 2010, 397, 961–965. ˇ ık, K., Electrophoresis 2011, 32, [21] Opekar, F., Stul´ 795–810. ˚ [22] Tuma, P., Opekar, F., Jel´ınek, I., Electroanalysis 2001, 13, 989–992. [23] Opekar, F., Chem. Listy 2012, 106, 289–292. ˇ ık, K., Anal. ´ ´ B., Opekar, F., Tuma, ˚ [24] Vochyanov a, P., Stul´ Bioanal. Chem. 2012, 404, 1549–1554. ˇ ık, K., Sensors 2013, 13, ˚ [25] Opekar, F., Tuma, P., Stul´ 2786–2801. [26] Miller, J. C., Miller, J. N., Statistics for Analytical Chemistry, Ellis Horwood, Chichester 1986, p. 124. ˇ ık, K., Anal. Chim. Acta 2011, ˚ ´ E., Stul´ [27] Tuma, P., Samcova, 685, 84–90.
www.electrophoresis-journal.com
of Science, Department of Analytical Chemistry, Charles University in Prague, Albertov, Prague, Czech Republic 2 Third Faculty of Medicine, Institute of Biochemistry, Cell and Molecular Biology, Charles ´ University in Prague, Ruska, Prague, Czech Republic
Received October 3, 2013 Revised October 31, 2013 Accepted November 18, 2013
Electrophoresis 2014, 35, 1660–1665
Research Article
Simultaneous and rapid determination of caffeine and taurine in energy drinks by MEKC in a short capillary with dual contactless conductivity/photometry detection A method has been developed for the simultaneous determination of taurine and caffeine using a laboratory-made instrument enabling separation analysis in a short 10.5 cm capillary. The substances are detected using a contactless conductometry/ultraviolet (UV) photometry detector that enables recording both signals at one place in the capillary. The separation of caffeine and taurine was performed using the MEKC technique in a BGE with the composition 40 mM CHES, 15 mM NaOH, and 50 mM SDS, pH 9.36. Under these conditions, the migration time of caffeine is 43 s and of taurine 60 s; LOD for caffeine is 4 mg/L using photometric detection and LOD for taurine is 24 mg/L using contactless conductometric detection. The standard addition method was used for determination in Red Bull energy drink of caffeine 317 mg/L and taurine 3860 mg/L; the contents in Kamikaze drink were 468 mg/L caffeine and 4110 mg/L taurine. The determined values are in good agreement with the declared contents of these substances. RSD does not exceed 3%. Keywords: Caffeine / Dual detection / Energy drinks / Micellar electrokinetic chromatography / Short capillary separation / Taurine DOI 10.1002/elps.201300480
1 Introduction Next to saccharides, caffeine and taurine are the main components in most energy drinks. Taurine has a favorable effect on a number of physiological processes in the organism; among other things, it optimizes the activity of the central nervous system, improves physical comfort, and stimulates thought processes. Its effects substantially augment those of the alkaloid caffeine. Energy drinks containing these substances improve performance and awareness and their consumption frequently increases with demands placed on human activity and generally with an increase in lifestyle intensity. However, an overdose of these substances, especially caffeine, can have undesirable effects [1, 2]. It is thus important to determine the contents of these two substances in energy drinks and a number of analytical methods have been developed for this purpose. General overviews of the methods employed for determining caffeine and taurine in various matrices can be found in a number of reviews, for example [3–5]. The most frequently used methods are LC (HPLC) and CE—for determining caffeine ˚ Correspondence: Dr. Petr Tuma, Institute of Biochemistry, Cell and Molecular Biology, Third Faculty of Medicine, Charles University in Prague, Ruska´ 87, 100 00 Prague 10, Czech Republic E-mail: [email protected] Fax: +420 267 102 460
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
MEKC, or GC. Detection methods include primarily MS and UV/Vis spectrophotometry. The individual components are generally determined separately in beverages and energy drinks. For example, recently, works have been published describing the determination of caffeine based on MEKC with UV detection [6], MEEKC with UV detection [7], UV/Vis derivative spectrophotometry [8], and capillary chromatography in a monolithic column with UV detection [9]. Caffeine has also been determined voltammetrically, for example, on a modified glassy carbon electrode [10] or using a boron-doped diamond electrode [11]. The determination of taurine has recently been described in the publications [12], CE with LIF detection [13], HPLC with precolumn derivatization and UV/Vis detection [14], amino acid analyzer and FTIR spectroscopy [15], MCE after derivatization with fluorescence detection. Simultaneous determination of the two analytes has, as far as we know, been described only in the works HPLC-MS method [16], planar chromatography with multiwavelength absorbance/fluorescence detection [17], and hydrophilic interaction chromatography with UV and evaporative light scattering detection in tandem [18]. Most of the published methods employ financially and technically demanding instrumentation combined with complicated sample pretreatment, usually resulting in low detection limits. However, the concentration of active components is high in energy drinks and thus rapid analysis, www.electrophoresis-journal.com
Electrophoresis 2014, 35, 1660–1665
CE and CEC
1661
undemanding sample pretreatment and low price are far more important for this type of determination than low LOD. The submitted work describes a method for the rapid simultaneous determination of caffeine and taurine in energy drinks using MEKC in a short 10.5 cm capillary with dual detection: optical detection, UV/Vis, is used for detection of caffeine and contactless conductivity detection, C4 D, for determining taurine. The detection takes place at a single site in the separation capillary. The method was tested on the determination of both analytes in the energy drinks Red Bull and Kamikaze. CE in short capillary represents efficient analytical tool especially in combination with multiple detection. The fundamental methodology of rapid analyses by electrophoresis along short separation pathways has been discussed from many aspects in reviews [19, 20] as well as the various combinations of detectors [21].
2 Materials and methods 2.1 Laboratory-made CE apparatus with dual contactless conductivity/photometry detection The separation and detection apparatus is located on the front panel of a Sapphire (ECOM, Czech Republic) UV/Vis detector under the optical and electrostatic shielding cover, Fig. 1. In this compact arrangement, the separation capillary (1), injection (2) and terminal (3) vessels, dual detection cell (4), electronic of the contactless conductivity detector (5), and (6) are placed under the shielding cover. The structural design of the C4 D is taken from [22]. Two semitubular electrodes (4a) are cut from self-sticking copper foil with a thickness of 35 m, electrode width of 1.7 mm, and distance between electrodes of 0.85 mm. Alternating current is fed to one of the electrodes from a sinus voltage generator (5). The signal passing through the cell is recorded by the second electrode and, following amplification and rectification (6), is fed into a computer for registration. In the optical part of the detection cell, radiation is brought by an optical fiber from the monochromator in the spectrophotometric detector to the detection window of the capillary in the space between the conductivity electrodes. The radiation passing through the capillary is registered by a large-area diode located opposite the optical cable in the lid of the detection cell, which is part of the Sapphire UV/Vis detector. The injection of the sample into the short capillary and its washing is based on the procedure described in detail in [23, 24] and will thus be mentioned only briefly here. The sample is injected into the capillary in such a manner that, simultaneously with turning the six-way injection valve (9) to the “inject” position, the separation electrolyte is fed through tube (10) to the injection valve at a constant rate and for a defined time. The stream of electrolyte carries defined volume of the sample from the sampling loop (11) around the injection end of the capillary, which is inserted to a depth C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 1. Scheme of the apparatus for MEKC in a short capillary with dual detection. The components of the apparatus depicted in the box are located under the shielding cover on the front panel of the spectral detector. Separation capillary (1), injection vessel (2), terminal vessel (3), dual detection cell (4), sinus voltage generator (5), alternating current meter and rectifier (6), outlet for further processing of the signal (7), high-voltage electrodes (8), injection valve (depicted in the “inject” position (9)), inlet for separation electrolyte during sample injection (10), sampling loop (11), injection tube (12), waste outlet for excess electrolyte during injection (13), and inlet for connection to the underpressure source for washing or activation of the capillary (14).
of approx. 0.5 mm into the feed PTFE tube (12) in the injection vessel (2). The injection period is controlled by the time during which the zone of sample solution flows around the capillary injection end; excess separation electrolyte flows out into waste (13) during the injection period. The injection takes place without interruption of the high voltage on the electrophoretic electrodes (8). After the end of the separation, the capillary is rinsed with the separation electrolyte by decreasing the pressure in the terminal vessel (14); the NaOH activation solution or the wash water was drawn into the capillary also by underpressure in the terminal vessel. The injection loop is filled using an injection syringe. Experimental parameters used in the measurement: standard fused silica capillary (Polymicro Technologies, USA) with an id of 50 m, overall length of 10.5 cm, and length to the detector of 8 cm; detection cell, C4 D, 450 kHz/17 Vpp (peak to peak), UV 216 nm; sampling loop volume, 16 L. The injection parameters were set so that the time of passage of the sample zone around the injection end www.electrophoresis-journal.com
1662
´ B. Vochyanov a´ et al.
of the capillary was about 0.3 s. Separation was performed at a voltage of +5 kV. The time of washing the capillary by the separation electrolyte after the end of the separation equaled 30 s.
Electrophoresis 2014, 35, 1660–1665
added to the analyzed solutions, so that their concentrations could be evaluated from a single recording.
3 Results and discussion 2.2 Chemicals and samples The following were used to prepare the BGEs: CHES (Roth, Germany), CAPS (Aldrich), NaOH (Sigma), and SDS (Sigma). The optimized BGE, see later, has the composition 40 mM CHES, 15 mM NaOH, and 50 mM SDS, pH 9.36; the electrolyte was prepared in DI water (Milli-Q Plus, Millipore, USA). Standard solutions of caffeine (Sigma-Aldrich) and taurine (Roth) with concentrations of 992 and 1250 mg/L, respectively, were prepared in DI water. Analyzed samples of energy drinks are commonly commercially available (producer or distributor in parentheses) with the caffeine and taurine contents given on the label: Red Bull (Red Bull, Austria) − caffeine, 80 mg in 250 mL (i.e., following recalculation 320 mg/L), taurine, 0.4% (4000 mg/L) Kamikaze (Tecfood, CR) − caffeine, 120 mg in 250 mL (480 mg/L), taurine, 4000 mg/L The only pretreatment of the beverage samples consisted in ten-min sonication to remove dissolved gases. The samples were then stored in a refrigerator and were diluted 10× for all the measurements using the separation electrolyte. Solutions of standards and energy drinks prepared for analysis by dilution with water did not yield satisfactorily reproducible results. We suppose that in the described dynamic injection method, combining moving sample zone with electrokinetic sampling, two basic factors can influence the reproducibility of the injection: (i) irreproducible mixing of the aqueous sample with the separation electrolyte in the inlet tube before reaching the sampling end of the capillary and (ii) undefined sample stacking. These effects lead to various amount of analyte injected into capillary and consequently to unsatisfactorily reproducible result.
2.3 Experimental procedure Prior to the separation, the capillary was activated for 10 min with a solution of 0.1 M NaOH, thoroughly washed with water and then filled with the separation electrolyte. Activation was performed twice daily. The current after connection of the separation voltage of 5 kV equalled 32–33 A. The separation voltage was not turned off for the entire measuring period. The calibration graph method and the method of standard additions were tested for determining caffeine and taurine. In both cases, standard solutions of both analytes were C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3.1 Separation of taurine and caffeine Taurine has two dissociated groups: an acidic sulpho-group with pKA 1.5 and a basic amino-group with pKA 9.2. In the pH range from pH 2.5 (pKA of the sulfo-group +1) to 8.2 (pKA of the amino-group −1), taurine is present as a zwitterion and does not move in the electric field. CE separation of taurine must be performed in BGE with pH greater than 8.7 to ensure sufficient separation of the taurine anion from the zone of electroneutral substances. Two BGEs were tested for the separation: CHES/NaOH in the range pH 8.7–10 and CAPS/NaOH at pH 10–11; pKA (CHES) = 9.4, pKA (CAPS) = 10.5. Experiments have shown that it is preferable to perform the separation of taurine in CHES/NaOH medium, where taurine is sufficiently separated from the zone of electroneutral substances and the sensitivity C4 D is twice as great as that for separation in CAPS/NaOH. In the positive separation mode, the peak of the water gap appears first on the electropherogram, followed by the positive peak of taurine, which, as an anion, moves in the opposite direction to EOF. At pH above 8.7 it is, however, not possible to determine caffeine by the CE technique as it is not dissociated under these conditions and moves together with the other electroneutral substances in the water gap zone. Consequently, SDS with the normally used concentration of 50 mM was added to CHES/NaOH and the MEKC technique was employed. Hydrophobic caffeine enters the anionic migrating micelle and is separated in this process from the zone of electroneutral substances. The presence of SDS in the electrolyte does not have any effect on the migration of hydrophilic taurine. On the electropherogram of energy drink in BGE with optimum composition 40 mM CHES, 15 mM NaOH, and 50 mM SDS, pH 9.36, the peak of EOF is visible first, followed by the peak of caffeine, and finally by the peak of taurine, see Fig. 2. The difference between electropherogram obtained with the sample and that obtained with the standard mixture is negligible. A dual C4 D/UV detector was used for the simultaneous determination of taurine and caffeine. Taurine, which does not absorb in the UV/Vis part of the spectrum, is detected as a positive peak using C4 D (the mobility of the taurine anion is greater than the mobility of the BGE co-ions). On the other hand, caffeine with purine structure is detected photometrically at the optimized wavelength of 216 nm. Caffeine also yields a signal in universal C4 D. However, it was verified experimentally that the response of C4 D to caffeine changes very little with a change in its concentration. It can be assumed that the zone caffeine/micelles affects primarily the permittivity of the solution, to which C4 D reacts [25] and not the electrical conductivity. www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 1660–1665 Table 1. Reproducibility of the migration times and peak areas of caffeine and taurine with concentrations of 99.2 and 125.0 mg/L, respectively, in a separation electrolyte solution of 40 mM CHES, 15 mM NaOH, and 50 mM SDS
Parameter
Caffeine
Taurine
Migration time (s) RSD (%) Peak area (mV s or mAU s) RSD (%)
43.4 ± 0.3 0.96 11.7 ± 0.8 5.7
60.0 ± 0.5 1.26 16.1 ± 0.7 3.4
The results were obtained from five repeated measurements.
1663
Table 2. Parameters of the linear calibration dependences for the determination of caffeine and taurine by the calibration graph method
Caffeine Concentration range (mg/L) Number of calibration points Slope (mAU s L/mg) Intercept (mAU s) Coefficient of determination, R2 LOD (mg/L (mol/L)) Taurine Concentration range (mg/L) Number of calibration points Slope (mV s L/mg) Intercept (mV s) Coefficient of determination, R2 LOD (mg/L (mol/L))
19.8–496.0 5 0.116 (0.001)a) −0.447 (0.310)a) 0.9996 4 (2 × 10−5 ) 25.0 to 625.0 5 0.159 (0.006)a) −3.511 (1.832)a) 0.9958 24 (2 × 10−4 )
a) Standard deviation in parenthesis.
the concentrations of both analytes are an order of magnitude higher in the samples for which the method is being tested.
3.3 Standard addition method Linear concentration dependences permit the use of the standard addition method for determining both analytes. Two standard additions were employed: for determining caffeine 99.2 and 198.4 mg/L and for determining taurine 125.2 and
Figure 2. Electropherogram of the energy drink Kamikaze with C4 D recording (thin line) and UV recording at 216 nm (bold line). Experimental conditions, BGE 40 mM CHES +15 mM NaOH + 50 mM SDS (pH 9.36); +5 kV/+33 A; sample diluted tenfold with BGE. Peak identification: caffeine (1), taurine (2).
The reproducibility of the migration times and peak areas were tested on an artificial sample of a mixture of caffeine and taurine. It is apparent from the results given in Table 1 that the reproducibility of the tested parameters is quite satisfactory for analytical purposes.
3.2 Calibration graph method The dependences of the peak areas for caffeine and taurine on the concentrations were linear in the entire tested concentration interval. A regression straight line was drawn through the calibration points (average of three repeated measurements), where the parameters for both analytes are given in Table 2. The results of the determination are listed in Table 3. The LOD was estimated from three times the amplitude of the noise in the two detection systems (0.3 mV and 0.03 mAU) and the parameters of the relevant calibration straight lines. The LOD values are relatively high; however, C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Table 3. Results of the determination of caffeine and taurine in energy drinks using the calibration graph and standard addition methods
Red Bull
Kamikaze
Calibration graph method Caffeine concentration (mg/L) Percent of declared value (%) Range, w (mg/L) RSD (%) Taurine concentration (mg/L) Percent of declared value (%) Range, w (mg/L) RSD (%)
324.7 ± 1.2 101.5 0.9 0.2 4123.9 ± 139.9 103.1 107.6 1.5
459.2 ± 13.5 95.7 10.4 1.3 4178.6 ± 339.3 104.5 261.0 3.7
Standard addition method Caffeine concentration (mg/L) Percent of declared value (%) Range, w (mg/L) RSD (%) Taurine concentration (mg/L) Percent of declared value (%) Range, w (mg/L) RSD (%)
317.0 ± 19.2 99.1 14.8 2.7 3857.0 ± 235.3 96.4 180.5 2.8
467.7 ± 11.2 97.5 21.9 2.0 4110.4 ± 224.0 102.8 172.3 2.5
The declared values for the contents of caffeine are equal to 320 and 480 mg/L in the Red Bull and Kamikaze beverages, respectively, while that of taurine is 4000 mg/L in both beverages.
www.electrophoresis-journal.com
1664
´ B. Vochyanov a´ et al.
Electrophoresis 2014, 35, 1660–1665
250.4 mg/L. The results of the determinations are also given in Table 3. All the results are the average of three parallel determinations: the range of results, that is, the difference between the largest and smallest determined values, w, is also given. The confidence interval was calculated for the 95% significance level. The relationship between the experimentally determined concentration values and the values declared on the label of the beverage tin is expressed as the percentage fraction of the determined and declared values. 3.4 Evaluation of the results It is apparent from Table 3 that, in most cases, the declared values of the concentrations of the two analytes are within the reliability interval of the experimentally determined concentrations. Nevertheless, a simple statistical TI test [26] was used to compare the experimental values with the declared values, TI = |x¯ − V | /w, where x¯ is the experimentally determined average value, V is the standard value, in this case the declared value, and w is the range. If the determined TI value is smaller than the critical value, the experimental result is not statistically different from the declared value. The results determined by the calibration graph method and the standard addition method were compared using Lord’s test [26], L = |x¯1 − x¯2 | /(w1 + w2 ), where x¯i and wi are the average values of the results obtained by the relevant calibration method and the corresponding range of results. All the data required for performance of the tests are given in Table 3 and the results of the testing are listed in Table 4. Values of the statistical parameters exceeding the critical value are designated in bold in Table 4. It can be seen that, when the standard addition method is employed, none of the Table 4. Results of the statistical testing of similarity of the experimentally determined concentration values to the declared values, TI test, and test of the similarity of the concentrations determined by the calibration graph method and the method of standard additions, L test
Red Bull TI values Calibration graph method versus declared value Caffeine 5.20a) Taurine 1.15 Standard addition method versus declared value Caffeine 0.20 Taurine 0.79 L values Comparison calibration graph vs. standard addition Caffeine 0.49 Taurine 0.93
Kamikaze
2.0 0.68 0.56 0.64
0.26 0.16
Critical values for three repeated measurements and the 95% significance level are 1.304 and 0.64 for the TI test and L test, respectively. Values in bold exceed the parameters’ critical value. a) The high TI value is caused by the extremely low value of the range of the determined concentrations in the test series, see Table 3.
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
determined concentrations are statistically different from the declared values. However, when the calibration graph method is employed, the determined values of the caffeine content are statistically different from the declared values. It follows from the Lord’s test values that the two calibration methods yield comparable results that were statistically different only in the determination of taurine. On the basis of the results of the test, the more reliable calibration method of standard additions can be recommended for the determination of caffeine and taurine in energy drinks. The long-term reproducibility of the described methods was tested by determining caffeine and taurine by the standard addition method in the Red Bull beverage. The results of the analysis mentioned in Table 3 were compared with the results obtained by repeated sample analysis after ten days. The determined contents of caffeine, 315.5 ± 26.0 (RSD 3.7%, recovery 98.6%) and taurine, 3987.9 ± 263.9 (RSD 3.0%, recovery 99.7%), are fully comparable with the results in Table 3. Between analyses, the beverage sample was stored in the refrigerator. All the results presented have been obtained without necessity to exchange the capillary. Standard addition calibration method eliminates the possible variations in sample injection when the capillary has to be changed.
4 Concluding remarks The determination of caffeine and taurine in energy drinks by separation in a short capillary with dual detection is a useful extension of the formerly described determination of saccharides in these beverages [24]. Pretreatment of the sample is simple and the determination is rapid; under the described experimental conditions, the separation requires about 1 min. When the standard addition method is used, the determined contents of the two analytes are statistically identical with the contents declared on the labels of the beverages. The method is a suitable alternative to the previously described methods permitting simultaneous determination of both active components of energy drinks. The described method is based on the use of a specialized apparatus. It can be anticipated that the method could be adapted for a commercial electrophoresis apparatus that would enable sample injection into the short end of a capillary and would be equipped with C4 D and UV detection. In these apparatuses, the two detection systems are usually at different places in the capillary so that the substances are detected at different degrees of separation [27]. However, this in no way affects the determination. The authors wish to thank Ecom s.r.o., Czech Republic, for cooperation in development of the apparatus and the Ministry of Education, Youth and Sports, Czech Republic, Research Project No. MSM0021620857 for financial support. This work was also supported by Charles University in Prague, project SVV and PRVOUK 31. The authors have declared no conflict of interest. www.electrophoresis-journal.com
Electrophoresis 2014, 35, 1660–1665
5 References [1] Huxtable, R. J., Physiol. Rev. 1992, 72, 101–163. [2] Nehlig, A., Daval, J. L., Debry, G., Brain Res. Rev. 1992, 17, 139–169. ˇ ˇ ık, P., Sv´ıtkova, ´ J., Rievaj, M., Bustin, D., [3] Svorc, L., Tomc´ Chem. Listy 2013, 107, 530–536. [4] Talik, P., Krzek, J., Ekiert, R. J., Sep. Purif. Rev. 2012, 41, 1–61. [5] Mou, S., Ding, X., Liu, Y., J. Chromatogr. B 2002, 781, 251–267. [6] Ballus, C. A., Meinhart, A. D., Bizzotto, C. S., Teixeira Filho, J., Godoy, H. T., Ciencia Technol. Alime. 2012, 32, 401–404. [7] Liotta, E., Gottardo, R., Seri, C., Rimondo, C., Miksik, I., Serpelloni, G., Tagliaro, F., Forensic Sci. Int. 2012, 220, 279–283. [8] Pieszko, C., Baranowska, I., Flores, A., J. Anal. Chem. 2010, 65, 1228–1234. [9] Al-Othman, Z. A., Aqel, A., Alharbi, M. K. E., BadjahHadj-Ahmed, A. Y., Al-Warthan, A. A., Food Chem. 2012, 132, 2217–2223. [10] Zhao, F., Wang, F., Zhao, W., Zhou, J., Liu, Y., Zou, L., Ye, B., Microchim. Acta 2011, 174, 383–390. ˇ ˇ ık, P., Sv´ıtkova, ´ J., Rievaj, M., Bustin, D., [11] Svorc, L., Tomc´ Food Chem. 2012, 135, 1198–1204. [12] Zinellu, A., Sotgia, S., Bastianina, S., Chessa, R., Gaspa, L., Franconi, F., Deiana, L., Carru, C., Amino Acids 2009, 36, 35–41.
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
CE and CEC
1665
[13] Sawabe, Y., Tagami, T., Yamasaki, K., J. Health Sci. 2008, 54, 661–664. [14] Triebel, S., Sproll, C., Reusch, H., Godelmann, R., Lachenmeier, D. W., Amino Acids 2007, 33, 451–457. ¨ [15] Gotz, S., Revermann, T., Karst, U., Lab. Chip 2007, 7, 93–97. [16] Marchei, E., Pellegrini, M., Pacifici, R., Palmi, I., Pichini, S., J. Pharm. Biomed. Anal. 2005, 37, 499–507. [17] Aranda, M., Morlock, G., J. Chromatogr. A 2006, 1131, 253–260. [18] Chirita, R., Dascalu, C., Gavrila, L., Elfakir, C., Rev. Chim.: Bucharest 2010, 61, 1173–1176. ˇ ık, K., Chem. Rev. 2009, 109, [19] Opekar, F., Coufal, P., Stul´ 4487–4499. [20] Matysik, F. M., Anal. Bioanal. Chem. 2010, 397, 961–965. ˇ ık, K., Electrophoresis 2011, 32, [21] Opekar, F., Stul´ 795–810. ˚ [22] Tuma, P., Opekar, F., Jel´ınek, I., Electroanalysis 2001, 13, 989–992. [23] Opekar, F., Chem. Listy 2012, 106, 289–292. ˇ ık, K., Anal. ´ ´ B., Opekar, F., Tuma, ˚ [24] Vochyanov a, P., Stul´ Bioanal. Chem. 2012, 404, 1549–1554. ˇ ık, K., Sensors 2013, 13, ˚ [25] Opekar, F., Tuma, P., Stul´ 2786–2801. [26] Miller, J. C., Miller, J. N., Statistics for Analytical Chemistry, Ellis Horwood, Chichester 1986, p. 124. ˇ ık, K., Anal. Chim. Acta 2011, ˚ ´ E., Stul´ [27] Tuma, P., Samcova, 685, 84–90.
www.electrophoresis-journal.com
