864 ´ Mauro Sergio Ferreira Santos Fernando Silva Lopes Ivano Gebhardt Rolf Gutz Instituto de Qu´ımica, ˜ Paulo, Sao ˜ Universidade de Sao Paulo, SP, Brazil
Received September 19, 2013 Revised November 4, 2013 Accepted November 5, 2013
Electrophoresis 2014, 35, 864–869
Research Article
Electrochemical derivatization-capillary electrophoresis-contactless conductivity detection: A versatile strategy for simultaneous determination of cationic, anionic, and neutral analytes The simultaneous determination of cationic, anionic, and neutral analytes in a real sample was demonstrated by coupling electrochemical (EC) derivatization with counter-EOF CE-C4 D. An EC flow cell was used to oxidize alcohols from an antiseptic mouthwash sample into carboxylic acids at a platinum electrode in acid medium. The carboxylates formed in the derivatization process and other sample ingredients, such as benzoate, saccharinate, and sodium ions, were separated in counter-flow mode and detected in one run in Tris-HCl buffer, pH 8.6. Fewer than 5 min were needed to complete each analysis with the automated flow system comprising solenoid pumps for the management of solutions. Insights into the electrochemistry of benzoic acid, present in the sample matrix, were also gained by EC-CE-C4 D; more specifically, by applying potentials higher than 1.47 V to the platinum electrode, some formiate and minute amounts of salicylate were detected. Keywords: Benzoic acid electrooxidation / EC-CE flow system / EC-CE of neutral species / Electrochemical derivatization DOI 10.1002/elps.201300463
1 Introduction Despite the ability of GC to separate and detect volatile species, thermally unstable compounds and metals are out of the scope of this technique if not chemically derivatized prior to separation [1]. HPLC is complementary to GC and has wide applicability, but presents some limitations; for example, for the simultaneous determination of cationic and anionic species by ion chromatography, two different columns are required [2]. In this context, CE is a versatile separation technique allowing the simultaneous determination of cationic and anionic species (organic or inorganic ones). It can be implemented by separation in counter-EOF mode [3, 4], single injection with dual detectors [5], or by dual opposite end injection with separation and detection performed in one run [6]. The determination of neutral analytes is also performed with CE equipment by operating in micellar electrokinetic chromatography mode [7] or by CE after a chemical derivatization step, prior to or concurrently with separation [8]. Precolumn derivatization has wider applicability than on-line Correspondence: Professor Ivano Gebhardt Rolf Gutz, Instituto de ˜ Paulo, Av. Prof. Lineu Prestes, 748, Qu´ımica, Universidade de Sao ˜ Paulo, SP, Brazil 05508–000, Sao E-mail: [email protected]
Abbreviation: EC, electrochemical C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
derivatization, but it increases the ionic strength and complexity of the matrix and augments the risk of interference. Electrochemical (EC) derivatization is a new alternative to deal with neutral species in CE. It consists of the oxidation or reduction of neutral analytes into charged or ionizable ones at an electrode. The first practical example is the formation of carboxylates from alcohols by oxidation in an EC flow cell directly coupled to the CE separation column, allowing separation and C4 D detection based only on the conventional electromigration mechanism [9]. While the quantification of a single alcohol can be done directly by voltammetry, on the basis of the oxidation wave current, many alcohols, sugars, and aldehydes present similar oxidation potentials, thus compromising selectivity. Therefore, voltammetry and particularly amperometry are best applied as detectors to samples containing such compounds after a separation step, for example, by HPLC or CE. In addition to their role as detectors, EC cells are also frequently used preceding separation [10, 11] and spectrometric [12] techniques, mainly for preparation or accumulation purposes. The coupling of EC cells with the inlet of CE equipment, EC-CE, is quite recent, but its versatility has already allowed the development of strategies for on-line sample preconcentration [13,14], the study of an electro-oxidation process by measuring its products by CE-C4 D [15–17], the
Colour Online: See the article online to view Figs. 2–5 in colour.
www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 864–869
formation and detection of anion radicals [18], and the formation of ionizable species amenable to CE separation [9]. This technique may also help to increase ESI efficiency in mass spectrometric detection [19, 20]. Broadening the scope of the previous conceptual short communication on the EC derivatization of alcohols [9], in this work it was first demonstrated that the EC-CE-C4 D system is suitable for the simultaneous determination of cationic, anionic (detected in counter-flow mode), and neutral species (detected after on-line EC derivatization). An antiseptic mouthwash was selected as a complex real sample for this purpose and an automated EC-CE-C4 D flow system facilitated the selection of conditions for the separation of analytes. It was observed that EC derivatization can also occur with other species in the sample formulation besides the intended neutral ones, for example, formic and salicylic acids were detected and attributed to the derivatization of benzoic acid.
2 Materials and methods 2.1 Chemicals and solutions A BGE, composed of 30 mmol/L Tris and 10 mmol/L HCl solution presenting a pH of 8.6, was used in the CE sepaR Tartar Control, ration. The antiseptic mouthwash Listerine used as the sample, was diluted 50-fold in an electrolyte composed of 3.0 mmol/L HNO3 and 1.0 mmol/L HCl (pH = 2.4). All reagents used in electrolyte preparation or as standard solutions were of analytical grade or better.
2.2 Apparatus The EC derivatization was carried out in a three-electrode EC flow system. The propulsion of solutions was performed by three solenoid pumps (two with an output of 10 L per stroke were used for water and BGE delivery and one with 5 L per stroke was used for sample propulsion). The EC flow cell was composed of a block with dual platinum electrodes (MF1000 block from the ESA Company) and a counter block with a spacing gasket in between; and the blocks were joined with screws. The counter rectangular cuboid, self-made from acrylic, comprised the solution inlet and outlet orifices, a pseudo reference electrode (Ag/AgCl) and an orifice normal to the surface of one of the platinum electrodes for the introduction of one-end tip of the capillary of the electrophoresis equipment (Fig. 1). The capillary tip was slant polished as previously described [13] in order to define an average distance of 50 m from the working electrode to the capillary inlet once its edge touches the metal. The second Pt electrode served as the auxiliary electrode during EC derivatization, while during the separation step, it acted as the grounded anode of the electrophoretic cell, powered by a negative high-voltage supply. The cathodic semicell at the opposite end of the capillary consisted of a Pt wire immersed in a vial half-filled with buffered electrolyte. The potentiostat was disconnected from C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
865
Figure 1. Schematic drawing of the EC flow cell coupled to the electrophoresis system, fully interfaced with a microcomputer (connections not shown for clarity). Flow in: Inlet of solutions displaced by solenoid driven pumps; Out: solution outlet; Pseudoreference electrode: Ag/AgCl wire; S1, S2, and S3: software controlled relays to switch all electrodes of the cell to the potentiostat or only the auxiliary electrode to the electrophoresis equipment.
the cell by relays during the CE run to protect it from accidental high voltage leakages (e.g. caused by air bubbles in the system). The EC-CE-C4 D system, mostly built and assembled in the laboratory from mechanical and electro-electronic components, operated under computer software control and data acquisition in the LabView environment. Data and details have been already published for the CE equipment [21], the detector [22], and the automation of the entire EC-CE-C4 D system [14].
2.3 EC-CE-C4 D procedure Before the derivatization step, the EC cell and electrodes were cleaned by pumping 300 L of BGE through the cell and applying a positive potential to the working electrode. After a volume of 200 L of the sample was pumped through the cell, the flow was stopped and a selected positive potential was applied for up to 60 s for derivatization. Afterwards, the software disconnected the potentiostat and hydrodynamically introduced a minute amount of the electrolyzed sample into the capillary. BGE was pumped again through the cell before separation and detection took place. The software allowed free programming of the order and timing of the sequential procedure. The system can run unattended for long periods during calibration and repeatability studies and, once provided with a sample changer, with real samples.
3 Results and discussion According to the manufacturer, the antiseptic mouthwash R Tartar Control is an aqueous solution containing Listerine www.electrophoresis-journal.com
866
M. S. F. Santos et al.
Figure 2. Electropherograms of anions for a diluted antiseptic mouthwash solution with (b = 1.47 V and c = 1.74 V versus Ag/AgCl, KCl3mol/L ) and without (a) electro-oxidation obtained with the EC-CE-C4 D system showing peaks of 10 mmol/L pentanoate (Pent− —internal standard), 242 mol/L benzoate (Bz− ), 86 mol/L saccharine (Sac− ), 67 mmol/L ethanoate (Et− ), and bicarbonate system peaks (SP). Conditions: electro-oxidation time of 60 s with sample diluted 50-fold; capillary 50 m id with a length of 45 cm (20 cm effective length); separation potential of 25 kV; BGE: 30 mmol/L Tris – 10 mmol/L HCl, pH 8.6.
benzoic acid and sodium benzoate, saccharine, ethanol, methyl salicylate, n-propanol, eucalyptol, menthol, sodium fluoride, sorbitol, sucralose, thymol, flavor, blue dye, and a nonionic surfactant (Poloxamer 407). Concentrations are not provided for the product marketed in Brazil and the formulation may differ from one country to another. The BGE used in this study presents a pH value of 8.6, which allows the ionization of saccharine and benzoic acid as well as its separation and detection by CE-C4 D in counter-flow mode (Fig. 2). Sodium is simple to measure by CE-C4 D and presents higher mobility than its coion Tris+ , thus presenting a positive peak in the electropherograms. Since saccharinate and benzoate have chloride as the higher mobility coion, these anions produce negative peaks (valley profile) [23,24]. Fluoride migrates fast enough to be detected under the applied counter-flow conditions. All other ingredients of the formulation are neutral at pH 8.6, but ethanol and n-propanol can be converted to ethanoate and propanoate by EC derivatization. Ethanoate was effectively separated and detected in this manner (Fig. 2, electropherograms b and c). Propanoate could not be detected even in a less diluted sample (tenfold instead of 50-fold) but, after spiking the 50-fold diluted sample with propanol (10 mmol/L), a well-resolved propanoate peak was observed after EC derivatization. The label of the product mentions propanol as an ingredient, but does not indicate its concentration, and it is probably lower than the range specified in a related patent (0.5 to 0.75% propanol by mass) [25]. Methyl salicylate is another potentially measurable ingredient in the formulation after alkaline hydrolysis with the formation of salicylate. The concentration in the 50-fold diluted mouthwash was below the detection limit, and a well-resolved salicylate peak was observed after spiking with methyl salicylate. Analysis of a less diluted sample (e.g. tenfold) caused a gradual change in EOF over repeated C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Electrophoresis 2014, 35, 864–869
Figure 3. Simultaneous determination of cationic, anionic, and neutral species by EC-CE-C4 D with an evaluation of the derivatization potential (versus Ag/AgCl, KCl3mol/L ) on the signal amplitude. Same conditions as described in Fig. 2.
injections, possibly due to a buildup of adsorbed surfactant in the capillary. This should be avoided because it demands frequent regeneration of the capillary. The remaining neutral ingredients in the formulation were difficult to electro-oxidize to ionizable species at the applied potential. At higher potentials, in addition to O2 evolution, species like methanoate (formiate), CO2 , etc., are produced without a quantitative relationship with the target analyte. Figure 2 presents some results for the separation of sample components in the anionic species region. Without EC derivatization (EC off), the electropherograms present signals only for benzoate and saccharinate anions, while after electro-oxidation, the negative peaks of ethanoate (ethanol oxidation) and pentanoate (n-pentanol oxidation—added as an internal standard) were also observed. The use of a neutral internal standard like n-pentanol has been described before [9] for the correction of eventual variations in the EC oxidation conditions. Pentanol presents the same behavior as ethanol in electroderivatization regarding the formation of carboxylic acid. The small perturbation in the baseline after the acetate peak corresponds to a HCO3 − system peak. A second internal standard for the correction of the signals from saccharine, benzoate, and sodium was not found to be necessary due to the high repeatability of all steps under computer control [9], including the injection into the capillary (a water column of 20 cm was used to adjust and stabilize the pressure). The set of electropherograms in Fig. 3 shows the effect of the derivatization potential on the registered signal. The fraction of ethanol converted into ethanoate increased in the range of 1.40–1.67 V and that of pentanol (internal standard) into pentanoate from 1.40 to 1.60 V. The gain in sensitivity and detection limit for ethanol (and pentanol) at higher potentials was accompanied by peak base broadening, but not to the extent of compromising the pentanoate/benzoate www.electrophoresis-journal.com
Electrophoresis 2014, 35, 864–869
CE and CEC
867
Figure 4. Comparison of electropherograms of 1.0 mmol/L benzoate in 3.0 mmol/L HNO3 and 1.0 mmol/L HCl without EC derivatization (EC Off) and at 1.67 V versus Ag/AgCl, KCl3mol/L after 60 s of electro-oxidation. Insets: (A) variation in the salicylate signal with derivatization potential; (B) magnified salicylate signal; (C) variation in the methanoate signal with derivatization potential. Same conditions as described in Fig. 2.
pair resolution at the given concentrations. At 1.74 V, the electrolysis of water becomes evident, with oxygen evolution, at a rate that may result in bubble formation in the cell, reducing the available electrode area and, at worst, interrupting the liquid film in the channel with a consequent loss of potential control by the potentiostat. The potential of 1.67 V was thus the upper limit, but, due to the considerable concentration of ethanol in the sample, a derivatization potential of 1.50 or 1.54 V was found to be appropriate. In any case, the derivatization potential is an easily adjustable parameter that provides some degree of modulation of the sensitivity. In addition to the effect of the derivatization potential on the ethanoate and pentanoate signals, a methanoate peak was observed, appearing 4 min after injection (Fig. 3). Methanol is obviously not an ingredient or detectable contaminant in this alcoholic mouthwash, due to its toxicity. The formulation contains methyl salicylate, the alkaline hydrolysis of which yields methanol and salicylate. However, the concentration of this ingredient is low, the mouthwash is acidic (pH ca. 3.9), and the electrolyte used during electrolysis is even more acidic (pH ca. 2.5, virtually unchanged by the 50-fold diluted mouthwash) and therefore the hydrolysis rate is negligible. The methanoate seems thus to originate from EC oxidative cleavage of a carbon–carbon bond of a sample ingredient with the formation of methanoate, and eventually into intermediate species like methanol and methanal. Electropherograms of a 1.0 mmol/L benzoate solution presented a methanoate peak at a derivatization potential of 1.47 V or higher, as shown in Fig. 4 inset C. For similar experiments with saccharine, no methanoate was detected. C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
On boron-doped diamond electrodes, traces of salicylic acid, 2,5-hydroxybenzoic acid and hydroquinone have been detected by off-line HPLC analysis of soluble electro-oxidation products of benzoic acid [26]. CO2 was reported as the main volatile product detected by MS [27] for the irreversible oxidation of benzoic acid adsorbed on polycrystalline platinum electrodes in acid medium [28]. Formic acid, although volatile, was not noted, possibly due to negligible production at these conditions. A magnified view of the region around 1.8 min in the electropherogram of Fig. 4 inset B reveals a barely detectable peak after derivatization at 1.74 V that could well be salicylate (based on the electromigration time of injected salicylate), indicating that some salicylic acid is formed not only on diamond but also on platinum electrodes. A conclusive identification of the salicylate peak and an investigation of the electrode reaction mechanism are beyond the scope of this work, however, the initial results confirm the potential of on-line EC-CE-C4 D for such kind of research. Data on the C4 D peak area for the target species in the commercial antiseptic mouthwash sample with n-pentanol added as the internal standard is presented in Fig. 5 as a function of the derivatization potential. The carboxylate signals followed a sigmoidal curve up to a plateau, as expected from voltammetry, all presenting a similar half-height potential. This means that direct voltammetric determination would be unfeasible, as would direct CE-C4 D separation and detection of neutral analytes. Conversely, the peak areas of the nonderivatized ionic species, that is, benzoate, saccharinate, and sodium presented some attenuation while the signal of www.electrophoresis-journal.com
868
M. S. F. Santos et al.
Electrophoresis 2014, 35, 864–869
4 Concluding remarks
Figure 5. Variation of analyte signal (area) with derivatization poR tential for a real sample of Listerine after 50-fold dilution. Same conditions as described in Fig. 2.
the derivatized species increased. A possible explanation for this follows. At sufficiently high-derivatization potentials, benzoate and saccharinate can be oxidized to CO2 and intermediate products, like methanoate and salicylate in the case of benzoate, or other undetected nonionic species. Figure 5 shows that the sodium peak area was nearly unaffected by the derivatization potential, as expected. A closer look reveals, however, an average decrease of about 5% from the region of E ࣘ 1.50 V to the region of E ࣙ 1.55 V, where various species are being electrolyzed. A small change in the peak shape and base line adjustment for Na+ peak integration was experimentally observed when more electrolysis products are present, possibly due to some change in the stacking conditions of the sample plug during the electrophoretic separation. A slight depletion of Na+ near the electrode due to its migration away from the anode during the electrolysis preceding the injection into the capillary could also have some influence, although charge transport in the solution is largely dominated by high-mobility protons at the electrolysis pH of 2.4. Once the derivatization potential for a new sample matrix is chosen and kept constant, the calibration curves for each analyte are little affected, except when a major analyte presents considerable variation, for example, alcoholic versus nonalcoholic mouthwash, a situation where the use of standard addition to the ionic species is advisable, besides to the internal standard (n-pentanol) for the derivatized analytes. Although a full quantitative application with determination of the figures of merit of the analytical method is beyond the scope of this proof-of-concept study, linear analytical curves (r ⬎ 0.995) were obtained for benzoate, saccharinate, and ethanol (using pentanol as internal standard). Determination of alcohol in beer with the EC-CE-C4 D system has been validated in a recent short communication [9] comprising a repeatability evaluation throughout 40 successive automatic determinations with no significant decrease in electro-oxidation. C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
The most remarkable feature of the system and method, as shown in Fig. 3, is the simultaneous determination of anionic, neutral, and cationic analytes in a single 3 min run (methanoate, requiring 4.5 min, does not need to be determined as it derives from the electrode reaction of benzoic acid). The derivatization process is based on the widely studied EC oxidation of alcohols, involving the removal of four electrons for the formation of the respective carboxylates [29]. Due to the design of the thin layer flow cell and the capillary interface, no more than 60 s were necessary for the derivatization of the neutral analytes. The automated management of the full system assists in sample handling and results in high efficiency and analytical throughput. The selection of the derivatization potential, as shown, helps to adjust the relative sensitivity of the eletro-oxidizable analytes with regard to other components determined in a sample. The use of EC cells as an auxiliary tool has become common practice in techniques like MS. Some commercial cells are available for this purpose, and the number of applications is increasing [30]. By indicating the formation of salicylate and methanoate from benzoate oxidation, this application also presents an alternative to the evaluation of EC reactions and the study of oxidation or reduction pathways. With C4 D, any ionic species can be measured at concentration above its LOQ. This system is less expensive and simpler than MS and could help in quantitative and qualitative analysis where evidence of product formation must be obtained. The authors wish to thank CNPq (Conselho Nacional de Desenvolvimento Cient´ıfico e Tecnol´ogico, Brazil) for fellowships and a grant. The authors have declared no conflict of interest.
5 References [1] Drozd, J., J. Chromatogr. 1975, 113, 303–356. [2] Meng, H.-B., Wang, T.-R., Guo, B.-Y., Hashi, Y., Guo, C.-X., Lin, J.-M., Talanta 2008, 76, 241–245. [3] Johns, C., Yang, W., Macka, M., Haddad, P. H., J. Chromatogr. A 2004, 1050, 217–222. [4] Silva, I. S., Vidal, D. T. R., do Lago, C. L., Angnes, L., J. Sep. Sci. 2013, 36, 1405–1409. [5] Mai, T. D., Hauser, P. C., J. Chromatogr. A 2012, 1267, 266–272. [6] Kuban, P., Kuban, P., Kuban, V., Electrophoresis 2002, 23, 3725–3734. [7] da Silva, J. A. F., do Lago, C. L., Electrophoresis 2000, 21, 1405–1408. [8] Underberg, W. J. M., Waterval, J. C. M., Electrophoresis 2002, 23, 3922–3933. [9] Santos, M. S. F., Lopes, F. S., Vidal, D. T. R., do Lago, C. L., Gutz, I. G. R., Anal. Chem. 2012, 84, 7599–7602.
www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 864–869
869
[11] Chen, Y., Edwards, B. L., Dasgupta, P. K., Srinivasan, K., Anal. Chem. 2012, 84, 59–66.
[21] Saito, R. M., Brito-Neto, J. G. A., Lopes, F. S., Blanes, L., Costa, E. T., Vidal, D. T. R., Hotta, G. M., do Lago, C. L., Anal. Methods 2010, 2, 164–170.
[12] Bulska, E., Walcerz, M., Jedral, W., Hulanicki A., Anal. Chim. Acta 1997, 357, 133–140.
[22] Francisco, K. J. M., do Lago, C. L., Electrophoresis 2009, 30, 3458–3464.
[13] Lopes, F. S., Antunes, O., Gutz, I. G. R., Electrochem. Commun. 2010, 12, 1387–1390.
[23] Brito-Neto, J. G. A., da Silva, J. A. F., Blanes, L., do Lago, C. L., Electroanalysis 2005, 17, 1198–1206.
[14] Lopes, F. S., Coelho, L. H. G., Gutz, I. G. R., Electrophoresis 2011, 32, 939–946.
[24] Brito-Neto, J. G. A., da Silva, J. A. F., Blanes, L., do Lago, C. L., Electroanalysis 2005, 17, 1207–1214.
[15] Lopes, F. S., Nogueira, T., do Lago, C. L., Gutz, I. G. R., Electroanalysis 2011, 23, 2516–2519.
[25] Kumar, L. D., Nair, M. K., Pan, P. C., U.S. Patent WO2000000166 A2, 2000.
[16] Matysik, F.-M., 1021–1024.
´ ´ [26] Montilla, F., Michaud, P. A., Morallon, E., Vazquez, J. L., Comninellis, Ch., Electrochim. Acta 2002, 47, 3509–3513.
[10] Schieffer, G. W., Anal. Chem. 1979, 51, 1573–1575.
Electrochem.
Commun.
2003,
5,
[17] He, J.-B., Cui, T., Zhang, W.-W., Deng, N., Anal. Chim. Acta 2013, 786, 159–165. [18] Esaka, Y., Okumura, N., Uno, B., Goto, M., Electrophoresis 2003, 24, 1635–1640. [19] Scholz, R., Matysik, F.-M., Analyst 2011, 136, 1562–1565. [20] Palatzky, P., Matysik, F.-M., Electrophoresis 2012, 33, 2689–2694.
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
´ [27] Souto, R. M., Rodr´ıguez, J. L., Fernandez-M ´ erida, L., Pastor, E., J. Electroanal. Chem. 2000, 494, 127–135. [28] Rabaaoui, N., Allagui, M. S., J. Hazard. Mat. 2012, 243, 187–192. [29] Singh, S., Datta, J., Ionics 2011, 17, 785–798. [30] Jahn, S., Karst, U., J. Chromatogr. A 2012, 1259, 16–49.
www.electrophoresis-journal.com
Received September 19, 2013 Revised November 4, 2013 Accepted November 5, 2013
Electrophoresis 2014, 35, 864–869
Research Article
Electrochemical derivatization-capillary electrophoresis-contactless conductivity detection: A versatile strategy for simultaneous determination of cationic, anionic, and neutral analytes The simultaneous determination of cationic, anionic, and neutral analytes in a real sample was demonstrated by coupling electrochemical (EC) derivatization with counter-EOF CE-C4 D. An EC flow cell was used to oxidize alcohols from an antiseptic mouthwash sample into carboxylic acids at a platinum electrode in acid medium. The carboxylates formed in the derivatization process and other sample ingredients, such as benzoate, saccharinate, and sodium ions, were separated in counter-flow mode and detected in one run in Tris-HCl buffer, pH 8.6. Fewer than 5 min were needed to complete each analysis with the automated flow system comprising solenoid pumps for the management of solutions. Insights into the electrochemistry of benzoic acid, present in the sample matrix, were also gained by EC-CE-C4 D; more specifically, by applying potentials higher than 1.47 V to the platinum electrode, some formiate and minute amounts of salicylate were detected. Keywords: Benzoic acid electrooxidation / EC-CE flow system / EC-CE of neutral species / Electrochemical derivatization DOI 10.1002/elps.201300463
1 Introduction Despite the ability of GC to separate and detect volatile species, thermally unstable compounds and metals are out of the scope of this technique if not chemically derivatized prior to separation [1]. HPLC is complementary to GC and has wide applicability, but presents some limitations; for example, for the simultaneous determination of cationic and anionic species by ion chromatography, two different columns are required [2]. In this context, CE is a versatile separation technique allowing the simultaneous determination of cationic and anionic species (organic or inorganic ones). It can be implemented by separation in counter-EOF mode [3, 4], single injection with dual detectors [5], or by dual opposite end injection with separation and detection performed in one run [6]. The determination of neutral analytes is also performed with CE equipment by operating in micellar electrokinetic chromatography mode [7] or by CE after a chemical derivatization step, prior to or concurrently with separation [8]. Precolumn derivatization has wider applicability than on-line Correspondence: Professor Ivano Gebhardt Rolf Gutz, Instituto de ˜ Paulo, Av. Prof. Lineu Prestes, 748, Qu´ımica, Universidade de Sao ˜ Paulo, SP, Brazil 05508–000, Sao E-mail: [email protected]
Abbreviation: EC, electrochemical C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
derivatization, but it increases the ionic strength and complexity of the matrix and augments the risk of interference. Electrochemical (EC) derivatization is a new alternative to deal with neutral species in CE. It consists of the oxidation or reduction of neutral analytes into charged or ionizable ones at an electrode. The first practical example is the formation of carboxylates from alcohols by oxidation in an EC flow cell directly coupled to the CE separation column, allowing separation and C4 D detection based only on the conventional electromigration mechanism [9]. While the quantification of a single alcohol can be done directly by voltammetry, on the basis of the oxidation wave current, many alcohols, sugars, and aldehydes present similar oxidation potentials, thus compromising selectivity. Therefore, voltammetry and particularly amperometry are best applied as detectors to samples containing such compounds after a separation step, for example, by HPLC or CE. In addition to their role as detectors, EC cells are also frequently used preceding separation [10, 11] and spectrometric [12] techniques, mainly for preparation or accumulation purposes. The coupling of EC cells with the inlet of CE equipment, EC-CE, is quite recent, but its versatility has already allowed the development of strategies for on-line sample preconcentration [13,14], the study of an electro-oxidation process by measuring its products by CE-C4 D [15–17], the
Colour Online: See the article online to view Figs. 2–5 in colour.
www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 864–869
formation and detection of anion radicals [18], and the formation of ionizable species amenable to CE separation [9]. This technique may also help to increase ESI efficiency in mass spectrometric detection [19, 20]. Broadening the scope of the previous conceptual short communication on the EC derivatization of alcohols [9], in this work it was first demonstrated that the EC-CE-C4 D system is suitable for the simultaneous determination of cationic, anionic (detected in counter-flow mode), and neutral species (detected after on-line EC derivatization). An antiseptic mouthwash was selected as a complex real sample for this purpose and an automated EC-CE-C4 D flow system facilitated the selection of conditions for the separation of analytes. It was observed that EC derivatization can also occur with other species in the sample formulation besides the intended neutral ones, for example, formic and salicylic acids were detected and attributed to the derivatization of benzoic acid.
2 Materials and methods 2.1 Chemicals and solutions A BGE, composed of 30 mmol/L Tris and 10 mmol/L HCl solution presenting a pH of 8.6, was used in the CE sepaR Tartar Control, ration. The antiseptic mouthwash Listerine used as the sample, was diluted 50-fold in an electrolyte composed of 3.0 mmol/L HNO3 and 1.0 mmol/L HCl (pH = 2.4). All reagents used in electrolyte preparation or as standard solutions were of analytical grade or better.
2.2 Apparatus The EC derivatization was carried out in a three-electrode EC flow system. The propulsion of solutions was performed by three solenoid pumps (two with an output of 10 L per stroke were used for water and BGE delivery and one with 5 L per stroke was used for sample propulsion). The EC flow cell was composed of a block with dual platinum electrodes (MF1000 block from the ESA Company) and a counter block with a spacing gasket in between; and the blocks were joined with screws. The counter rectangular cuboid, self-made from acrylic, comprised the solution inlet and outlet orifices, a pseudo reference electrode (Ag/AgCl) and an orifice normal to the surface of one of the platinum electrodes for the introduction of one-end tip of the capillary of the electrophoresis equipment (Fig. 1). The capillary tip was slant polished as previously described [13] in order to define an average distance of 50 m from the working electrode to the capillary inlet once its edge touches the metal. The second Pt electrode served as the auxiliary electrode during EC derivatization, while during the separation step, it acted as the grounded anode of the electrophoretic cell, powered by a negative high-voltage supply. The cathodic semicell at the opposite end of the capillary consisted of a Pt wire immersed in a vial half-filled with buffered electrolyte. The potentiostat was disconnected from C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
865
Figure 1. Schematic drawing of the EC flow cell coupled to the electrophoresis system, fully interfaced with a microcomputer (connections not shown for clarity). Flow in: Inlet of solutions displaced by solenoid driven pumps; Out: solution outlet; Pseudoreference electrode: Ag/AgCl wire; S1, S2, and S3: software controlled relays to switch all electrodes of the cell to the potentiostat or only the auxiliary electrode to the electrophoresis equipment.
the cell by relays during the CE run to protect it from accidental high voltage leakages (e.g. caused by air bubbles in the system). The EC-CE-C4 D system, mostly built and assembled in the laboratory from mechanical and electro-electronic components, operated under computer software control and data acquisition in the LabView environment. Data and details have been already published for the CE equipment [21], the detector [22], and the automation of the entire EC-CE-C4 D system [14].
2.3 EC-CE-C4 D procedure Before the derivatization step, the EC cell and electrodes were cleaned by pumping 300 L of BGE through the cell and applying a positive potential to the working electrode. After a volume of 200 L of the sample was pumped through the cell, the flow was stopped and a selected positive potential was applied for up to 60 s for derivatization. Afterwards, the software disconnected the potentiostat and hydrodynamically introduced a minute amount of the electrolyzed sample into the capillary. BGE was pumped again through the cell before separation and detection took place. The software allowed free programming of the order and timing of the sequential procedure. The system can run unattended for long periods during calibration and repeatability studies and, once provided with a sample changer, with real samples.
3 Results and discussion According to the manufacturer, the antiseptic mouthwash R Tartar Control is an aqueous solution containing Listerine www.electrophoresis-journal.com
866
M. S. F. Santos et al.
Figure 2. Electropherograms of anions for a diluted antiseptic mouthwash solution with (b = 1.47 V and c = 1.74 V versus Ag/AgCl, KCl3mol/L ) and without (a) electro-oxidation obtained with the EC-CE-C4 D system showing peaks of 10 mmol/L pentanoate (Pent− —internal standard), 242 mol/L benzoate (Bz− ), 86 mol/L saccharine (Sac− ), 67 mmol/L ethanoate (Et− ), and bicarbonate system peaks (SP). Conditions: electro-oxidation time of 60 s with sample diluted 50-fold; capillary 50 m id with a length of 45 cm (20 cm effective length); separation potential of 25 kV; BGE: 30 mmol/L Tris – 10 mmol/L HCl, pH 8.6.
benzoic acid and sodium benzoate, saccharine, ethanol, methyl salicylate, n-propanol, eucalyptol, menthol, sodium fluoride, sorbitol, sucralose, thymol, flavor, blue dye, and a nonionic surfactant (Poloxamer 407). Concentrations are not provided for the product marketed in Brazil and the formulation may differ from one country to another. The BGE used in this study presents a pH value of 8.6, which allows the ionization of saccharine and benzoic acid as well as its separation and detection by CE-C4 D in counter-flow mode (Fig. 2). Sodium is simple to measure by CE-C4 D and presents higher mobility than its coion Tris+ , thus presenting a positive peak in the electropherograms. Since saccharinate and benzoate have chloride as the higher mobility coion, these anions produce negative peaks (valley profile) [23,24]. Fluoride migrates fast enough to be detected under the applied counter-flow conditions. All other ingredients of the formulation are neutral at pH 8.6, but ethanol and n-propanol can be converted to ethanoate and propanoate by EC derivatization. Ethanoate was effectively separated and detected in this manner (Fig. 2, electropherograms b and c). Propanoate could not be detected even in a less diluted sample (tenfold instead of 50-fold) but, after spiking the 50-fold diluted sample with propanol (10 mmol/L), a well-resolved propanoate peak was observed after EC derivatization. The label of the product mentions propanol as an ingredient, but does not indicate its concentration, and it is probably lower than the range specified in a related patent (0.5 to 0.75% propanol by mass) [25]. Methyl salicylate is another potentially measurable ingredient in the formulation after alkaline hydrolysis with the formation of salicylate. The concentration in the 50-fold diluted mouthwash was below the detection limit, and a well-resolved salicylate peak was observed after spiking with methyl salicylate. Analysis of a less diluted sample (e.g. tenfold) caused a gradual change in EOF over repeated C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Electrophoresis 2014, 35, 864–869
Figure 3. Simultaneous determination of cationic, anionic, and neutral species by EC-CE-C4 D with an evaluation of the derivatization potential (versus Ag/AgCl, KCl3mol/L ) on the signal amplitude. Same conditions as described in Fig. 2.
injections, possibly due to a buildup of adsorbed surfactant in the capillary. This should be avoided because it demands frequent regeneration of the capillary. The remaining neutral ingredients in the formulation were difficult to electro-oxidize to ionizable species at the applied potential. At higher potentials, in addition to O2 evolution, species like methanoate (formiate), CO2 , etc., are produced without a quantitative relationship with the target analyte. Figure 2 presents some results for the separation of sample components in the anionic species region. Without EC derivatization (EC off), the electropherograms present signals only for benzoate and saccharinate anions, while after electro-oxidation, the negative peaks of ethanoate (ethanol oxidation) and pentanoate (n-pentanol oxidation—added as an internal standard) were also observed. The use of a neutral internal standard like n-pentanol has been described before [9] for the correction of eventual variations in the EC oxidation conditions. Pentanol presents the same behavior as ethanol in electroderivatization regarding the formation of carboxylic acid. The small perturbation in the baseline after the acetate peak corresponds to a HCO3 − system peak. A second internal standard for the correction of the signals from saccharine, benzoate, and sodium was not found to be necessary due to the high repeatability of all steps under computer control [9], including the injection into the capillary (a water column of 20 cm was used to adjust and stabilize the pressure). The set of electropherograms in Fig. 3 shows the effect of the derivatization potential on the registered signal. The fraction of ethanol converted into ethanoate increased in the range of 1.40–1.67 V and that of pentanol (internal standard) into pentanoate from 1.40 to 1.60 V. The gain in sensitivity and detection limit for ethanol (and pentanol) at higher potentials was accompanied by peak base broadening, but not to the extent of compromising the pentanoate/benzoate www.electrophoresis-journal.com
Electrophoresis 2014, 35, 864–869
CE and CEC
867
Figure 4. Comparison of electropherograms of 1.0 mmol/L benzoate in 3.0 mmol/L HNO3 and 1.0 mmol/L HCl without EC derivatization (EC Off) and at 1.67 V versus Ag/AgCl, KCl3mol/L after 60 s of electro-oxidation. Insets: (A) variation in the salicylate signal with derivatization potential; (B) magnified salicylate signal; (C) variation in the methanoate signal with derivatization potential. Same conditions as described in Fig. 2.
pair resolution at the given concentrations. At 1.74 V, the electrolysis of water becomes evident, with oxygen evolution, at a rate that may result in bubble formation in the cell, reducing the available electrode area and, at worst, interrupting the liquid film in the channel with a consequent loss of potential control by the potentiostat. The potential of 1.67 V was thus the upper limit, but, due to the considerable concentration of ethanol in the sample, a derivatization potential of 1.50 or 1.54 V was found to be appropriate. In any case, the derivatization potential is an easily adjustable parameter that provides some degree of modulation of the sensitivity. In addition to the effect of the derivatization potential on the ethanoate and pentanoate signals, a methanoate peak was observed, appearing 4 min after injection (Fig. 3). Methanol is obviously not an ingredient or detectable contaminant in this alcoholic mouthwash, due to its toxicity. The formulation contains methyl salicylate, the alkaline hydrolysis of which yields methanol and salicylate. However, the concentration of this ingredient is low, the mouthwash is acidic (pH ca. 3.9), and the electrolyte used during electrolysis is even more acidic (pH ca. 2.5, virtually unchanged by the 50-fold diluted mouthwash) and therefore the hydrolysis rate is negligible. The methanoate seems thus to originate from EC oxidative cleavage of a carbon–carbon bond of a sample ingredient with the formation of methanoate, and eventually into intermediate species like methanol and methanal. Electropherograms of a 1.0 mmol/L benzoate solution presented a methanoate peak at a derivatization potential of 1.47 V or higher, as shown in Fig. 4 inset C. For similar experiments with saccharine, no methanoate was detected. C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
On boron-doped diamond electrodes, traces of salicylic acid, 2,5-hydroxybenzoic acid and hydroquinone have been detected by off-line HPLC analysis of soluble electro-oxidation products of benzoic acid [26]. CO2 was reported as the main volatile product detected by MS [27] for the irreversible oxidation of benzoic acid adsorbed on polycrystalline platinum electrodes in acid medium [28]. Formic acid, although volatile, was not noted, possibly due to negligible production at these conditions. A magnified view of the region around 1.8 min in the electropherogram of Fig. 4 inset B reveals a barely detectable peak after derivatization at 1.74 V that could well be salicylate (based on the electromigration time of injected salicylate), indicating that some salicylic acid is formed not only on diamond but also on platinum electrodes. A conclusive identification of the salicylate peak and an investigation of the electrode reaction mechanism are beyond the scope of this work, however, the initial results confirm the potential of on-line EC-CE-C4 D for such kind of research. Data on the C4 D peak area for the target species in the commercial antiseptic mouthwash sample with n-pentanol added as the internal standard is presented in Fig. 5 as a function of the derivatization potential. The carboxylate signals followed a sigmoidal curve up to a plateau, as expected from voltammetry, all presenting a similar half-height potential. This means that direct voltammetric determination would be unfeasible, as would direct CE-C4 D separation and detection of neutral analytes. Conversely, the peak areas of the nonderivatized ionic species, that is, benzoate, saccharinate, and sodium presented some attenuation while the signal of www.electrophoresis-journal.com
868
M. S. F. Santos et al.
Electrophoresis 2014, 35, 864–869
4 Concluding remarks
Figure 5. Variation of analyte signal (area) with derivatization poR tential for a real sample of Listerine after 50-fold dilution. Same conditions as described in Fig. 2.
the derivatized species increased. A possible explanation for this follows. At sufficiently high-derivatization potentials, benzoate and saccharinate can be oxidized to CO2 and intermediate products, like methanoate and salicylate in the case of benzoate, or other undetected nonionic species. Figure 5 shows that the sodium peak area was nearly unaffected by the derivatization potential, as expected. A closer look reveals, however, an average decrease of about 5% from the region of E ࣘ 1.50 V to the region of E ࣙ 1.55 V, where various species are being electrolyzed. A small change in the peak shape and base line adjustment for Na+ peak integration was experimentally observed when more electrolysis products are present, possibly due to some change in the stacking conditions of the sample plug during the electrophoretic separation. A slight depletion of Na+ near the electrode due to its migration away from the anode during the electrolysis preceding the injection into the capillary could also have some influence, although charge transport in the solution is largely dominated by high-mobility protons at the electrolysis pH of 2.4. Once the derivatization potential for a new sample matrix is chosen and kept constant, the calibration curves for each analyte are little affected, except when a major analyte presents considerable variation, for example, alcoholic versus nonalcoholic mouthwash, a situation where the use of standard addition to the ionic species is advisable, besides to the internal standard (n-pentanol) for the derivatized analytes. Although a full quantitative application with determination of the figures of merit of the analytical method is beyond the scope of this proof-of-concept study, linear analytical curves (r ⬎ 0.995) were obtained for benzoate, saccharinate, and ethanol (using pentanol as internal standard). Determination of alcohol in beer with the EC-CE-C4 D system has been validated in a recent short communication [9] comprising a repeatability evaluation throughout 40 successive automatic determinations with no significant decrease in electro-oxidation. C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
The most remarkable feature of the system and method, as shown in Fig. 3, is the simultaneous determination of anionic, neutral, and cationic analytes in a single 3 min run (methanoate, requiring 4.5 min, does not need to be determined as it derives from the electrode reaction of benzoic acid). The derivatization process is based on the widely studied EC oxidation of alcohols, involving the removal of four electrons for the formation of the respective carboxylates [29]. Due to the design of the thin layer flow cell and the capillary interface, no more than 60 s were necessary for the derivatization of the neutral analytes. The automated management of the full system assists in sample handling and results in high efficiency and analytical throughput. The selection of the derivatization potential, as shown, helps to adjust the relative sensitivity of the eletro-oxidizable analytes with regard to other components determined in a sample. The use of EC cells as an auxiliary tool has become common practice in techniques like MS. Some commercial cells are available for this purpose, and the number of applications is increasing [30]. By indicating the formation of salicylate and methanoate from benzoate oxidation, this application also presents an alternative to the evaluation of EC reactions and the study of oxidation or reduction pathways. With C4 D, any ionic species can be measured at concentration above its LOQ. This system is less expensive and simpler than MS and could help in quantitative and qualitative analysis where evidence of product formation must be obtained. The authors wish to thank CNPq (Conselho Nacional de Desenvolvimento Cient´ıfico e Tecnol´ogico, Brazil) for fellowships and a grant. The authors have declared no conflict of interest.
5 References [1] Drozd, J., J. Chromatogr. 1975, 113, 303–356. [2] Meng, H.-B., Wang, T.-R., Guo, B.-Y., Hashi, Y., Guo, C.-X., Lin, J.-M., Talanta 2008, 76, 241–245. [3] Johns, C., Yang, W., Macka, M., Haddad, P. H., J. Chromatogr. A 2004, 1050, 217–222. [4] Silva, I. S., Vidal, D. T. R., do Lago, C. L., Angnes, L., J. Sep. Sci. 2013, 36, 1405–1409. [5] Mai, T. D., Hauser, P. C., J. Chromatogr. A 2012, 1267, 266–272. [6] Kuban, P., Kuban, P., Kuban, V., Electrophoresis 2002, 23, 3725–3734. [7] da Silva, J. A. F., do Lago, C. L., Electrophoresis 2000, 21, 1405–1408. [8] Underberg, W. J. M., Waterval, J. C. M., Electrophoresis 2002, 23, 3922–3933. [9] Santos, M. S. F., Lopes, F. S., Vidal, D. T. R., do Lago, C. L., Gutz, I. G. R., Anal. Chem. 2012, 84, 7599–7602.
www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 864–869
869
[11] Chen, Y., Edwards, B. L., Dasgupta, P. K., Srinivasan, K., Anal. Chem. 2012, 84, 59–66.
[21] Saito, R. M., Brito-Neto, J. G. A., Lopes, F. S., Blanes, L., Costa, E. T., Vidal, D. T. R., Hotta, G. M., do Lago, C. L., Anal. Methods 2010, 2, 164–170.
[12] Bulska, E., Walcerz, M., Jedral, W., Hulanicki A., Anal. Chim. Acta 1997, 357, 133–140.
[22] Francisco, K. J. M., do Lago, C. L., Electrophoresis 2009, 30, 3458–3464.
[13] Lopes, F. S., Antunes, O., Gutz, I. G. R., Electrochem. Commun. 2010, 12, 1387–1390.
[23] Brito-Neto, J. G. A., da Silva, J. A. F., Blanes, L., do Lago, C. L., Electroanalysis 2005, 17, 1198–1206.
[14] Lopes, F. S., Coelho, L. H. G., Gutz, I. G. R., Electrophoresis 2011, 32, 939–946.
[24] Brito-Neto, J. G. A., da Silva, J. A. F., Blanes, L., do Lago, C. L., Electroanalysis 2005, 17, 1207–1214.
[15] Lopes, F. S., Nogueira, T., do Lago, C. L., Gutz, I. G. R., Electroanalysis 2011, 23, 2516–2519.
[25] Kumar, L. D., Nair, M. K., Pan, P. C., U.S. Patent WO2000000166 A2, 2000.
[16] Matysik, F.-M., 1021–1024.
´ ´ [26] Montilla, F., Michaud, P. A., Morallon, E., Vazquez, J. L., Comninellis, Ch., Electrochim. Acta 2002, 47, 3509–3513.
[10] Schieffer, G. W., Anal. Chem. 1979, 51, 1573–1575.
Electrochem.
Commun.
2003,
5,
[17] He, J.-B., Cui, T., Zhang, W.-W., Deng, N., Anal. Chim. Acta 2013, 786, 159–165. [18] Esaka, Y., Okumura, N., Uno, B., Goto, M., Electrophoresis 2003, 24, 1635–1640. [19] Scholz, R., Matysik, F.-M., Analyst 2011, 136, 1562–1565. [20] Palatzky, P., Matysik, F.-M., Electrophoresis 2012, 33, 2689–2694.
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
´ [27] Souto, R. M., Rodr´ıguez, J. L., Fernandez-M ´ erida, L., Pastor, E., J. Electroanal. Chem. 2000, 494, 127–135. [28] Rabaaoui, N., Allagui, M. S., J. Hazard. Mat. 2012, 243, 187–192. [29] Singh, S., Datta, J., Ionics 2011, 17, 785–798. [30] Jahn, S., Karst, U., J. Chromatogr. A 2012, 1259, 16–49.
www.electrophoresis-journal.com
