Simple Flow Injection Analysis System for Simultaneous

Anal. Chem. 2010, 82, 8658–8663

Simple Flow Injection Analysis System for Simultaneous Determination of Phenolic Antioxidants with Multiple Pulse Amperometric Detection at a Boron-Doped Diamond Electrode Roberta Antigo Medeiros, Bruna Cláudia Lourenc¸a˜o, Romeu Cardozo Rocha-Filho, and Orlando Fatibello-Filho* Departamento de Quı´mica, Universidade Federal de Sa˜o Carlos, C.P. 676, 13560-970 Sa˜o Carlos - SP, Brazil A method for simultaneous determination of butylated hydroxyanisole (BHA) and butylated hydroxytoluene (BHT) in food was developed that uses multiple pulse amperometry (MPA) with flow injection analysis (FIA). Determination of these phenolic antioxidants was carried out with a cathodically pretreated boron-doped diamond electrode and an aqueous ethanolic (30% ethanol, v/v) 10 mmol L-1 KNO3 solution (pHcond ) 1.5) as supporting electrolyte. A dual-potential waveform, at Edet1 ) 850 mV/200 ms and Edet2 ) 1150 mV/200 ms versus Ag/AgCl (3.0 mol L-1 KCl), was employed. The use of Edet1 or Edet2 caused the oxidation of BHA or of BHA and BHT, respectively; hence, concentration subtraction could be used to determine both species. The respective analytical curves presented good linearity in the investigated concentration range (0.050-3.0 µmol L-1 for BHA and 0.70-70 µmol L-1 for BHT), and the detection limits were 0.030 µmol L-1 for BHA and 0.40 µmol L-1 for BHT. The proposed method, which is simple, quick, and presents good precision and accuracy, was successfully applied in the simultaneous determination of BHA and BHT in commercial mayonnaise samples, with results similar to those obtained by HPLC, at a 95% confidence level. Synthetic phenolic antioxidants (SPAs) are widely used in food systems to prevent lipid oxidation during processing and storage. The phenolic compounds butylated hydroxyanisole (BHA) and butylated hydroxytoluene (BHT)ssee Figure 1sare among the primary synthetic antioxidants widely used to interrupt the chain of free radicals involved in the autoxidation process that constitutes the most common form of deterioration of fats used in the food industry. They have been used both alone and as mixtures in oils, margarine, and mayonnaise,1,2 but their use is not a straightforward solution. Because BHA and BHT are suspected of being responsible for liver damage and carcinogenesis in laboratory animals, their potential harmful effects on health have been * Corresponding author. Phone: +55-16-33518098. Fax: +55-16-33518350. E-mail: [email protected]. (1) Delgado-Zamarreno, M. M.; Gonzalez-Maza, I.; Sanchez-Perez, A.; Martinez, R. C. Food Chem. 2007, 100, 1722–1727. (2) Diaz, T. G.; Cabanillas, A. G.; Franco, M. F. A.; Salinas, F.; Vire, J. C. Electroanalysis 1998, 10, 497–505.

8658

Analytical Chemistry, Vol. 82, No. 20, October 15, 2010

Figure 1. Chemical structures of BHA and BHT.

extensively discussed and studied. Therefore, in several countries the use of these additives is subject to regulations, which define specific approved antioxidants, establish permitted use levels, and include labeling requirements; thus, the determination of SPAs in foods is necessary to ensure the fulfillment of legal requirements as well as quality-control procedures in the food industry. However, there are differences among the individual countries, that is, antioxidants permitted in one country may be prohibited in another. Internationally, the JECFA (Joint FAO/WHO Expert Committee on Food Additives) periodically considers food additives, including SPAs, on the basis of all available scientific data, to establish specifications of identity and purity as well as acceptable daily intake levels.3,4 In the European Union, for example, the amount of synthetic antioxidants in food is limited to 0.01% (100 mg/kg) for each antioxidant if used individually and to 0.02% as total fraction if the antioxidants are used in mixtures.1 In Brazil, the use of these antioxidants is controlled by The National Health Surveillance Agency (ANVISA), which limits the content to 200 mg/kg for BHA and 100 mg/kg for BHT.5 In the United States, limits for antioxidants (expressed as percentage on the basis of total weight), singly or in combination, vary for different food products; for instance, in rendered fat, the limits are 0.01% singly or 0.02% in combination, while for margarine the limit is 0.02% (singly or in combination).6 Many methods for determining BHA and BHT individually or simultaneously have (3) Food and Agriculture Organization of the United Nations and World Health Organization, 1995. http://www.who.int/ipcs/food/jecfa/en/index.html. (4) Guan, Y. Q.; Chu, Q. C.; Fu, L.; Wu, T.; Ye, J. N. Food Chem. 2006, 94, 157–162. (5) National Health Surveillance Agency Brazil (ANVISA), 2005. www. anvisa.gov.br. (6) U.S. Department of Agriculture Food Safety and Inspection Service, Processing Inspectors’ Calculation Handbook, 1995; p 53. 10.1021/ac101921f  2010 American Chemical Society Published on Web 09/21/2010

been recently reported, based on spectrophotometry,7 liquid and gas chromatography,8-10 and micellar electrokinetic chromatography.1,4 Electrochemical techniques, such as voltammetry and amperometry, are a promising alternative to classical approaches due to their relatively low operational cost, good miniaturization potential, and rapid and sensitive detection procedures, which are suitable for faster analyses. Some methods for determining BHA and BHT by voltammetric2,11-19 and amperometric20-22 techniques were already reported. Diaz et al.2 studied the voltammetric behavior of propyl gallate (PG), BHA, and BHT at a glassy carbon electrode (static and rotating) in an acetonitrile-water medium; they used a chemometric procedure for the determination of these antioxidants in different spiked samples of packet soup. Ni et al.17 studied the voltammetric behavior of BHA, BHT, PG, and tertbutylhydroquinone at a glassy carbon electrode in a 0.1 mol L-1 perchloric acid solution containing 1% (v/v) methanol, using chemometric approaches such as classical least-squares, principal-component regression, and partial least-squares regression analyses. Linear calibration plots were obtained in the concentration ranges 2.8-83 µmol L-1 for BHA and 2.8-36 µmol L-1 for BHT, with detection limits of 1.0 and 0.68 µmol L-1, respectively. The method was applied to determine the four antioxidants in a set of synthetic mixtures as well as in several commercial food samples. Ceballos and Fernandez12 used square-wave voltammetry (SWV) with carbon-disk ultramicroelectrodes to determine BHA and BHT in vegetable oils. The determinations were carried out directly in a benzene/ethanol/ H2SO4 solution or in acetonitrile after an extractive procedure, with better results in the latter case. Agu ¨´ı et al.11 used cylindrical carbon-fiber microelectrodes in the simultaneous determination of BHA and BHT by SWV, obtaining detection limits of 4.0 µmol L-1 for BHA and 0.37 µmol L-1 for BHT; however, the supporting electrolyte used contained acetonitrile, a high-cost and moderate toxicity reagent. Recently, our group developed a method for the simultaneous determination of BHA and BHT in food samples using SWV.19 The determinations were carried out with a cathodically pretreated boron-doped diamond electrode (BDD) and an aqueous ethanolic (30% (7) Capitan-Vallvey, L. F.; Valencia, M. C.; Nicolas, E. A. Anal. Chim. Acta 2004, 503, 179–186. (8) Guo, L.; Xie, M. Y.; Yan, A. P.; Wan, Y. Q.; Wu, Y. M. Anal. Bioanal. Chem. 2006, 386, 1881–1887. (9) Perrin, C.; Meyer, L. Food Chem. 2002, 77, 93–100. (10) Saad, B.; Sing, Y. Y.; Nawi, M. A.; Hashim, N.; Ali, A. S. M.; Saleh, M. I.; Sulaiman, S. F.; Talib, K. M.; Ahmad, K. Food Chem. 2007, 105, 389–394. (11) Agu ¨´ı, M. L.; Reviejo, A. J.; Ya´n ˜ez-Seden ˜o, P.; Pingarro´n, J. M. Anal. Chem. 1995, 67, 2195–2200. (12) Ceballos, C.; Fernandez, H. Food Res. Int. 2000, 33, 357–365. (13) Ceballos, C.; Fernandez, H. JAOCS 2000, 77, 731–735. (14) Ceballos, C. D.; Zo´n, M. A.; Fernandez, H. J. Chem. Educ. 2006, 83, 1394– 1352. (15) De la Fuente, C.; Acuna, J. A.; Vazquez, M. D.; Tascon, M. L.; Batanero, P. S. Talanta 1999, 49, 441–452. (16) Kumar, S. S.; Narayanan, S. S. Talanta 2008, 76, 54–59. (17) Ni, Y. N.; Wang, L.; Kokot, S. Anal. Chim. Acta 2000, 412, 185–193. (18) Raymundo, M. D.; Paula, M. M. D.; Franco, C.; Fett, R. LWTsFood Sci. Technol. 2007, 40, 1133–1139. (19) Medeiros, R. A.; Rocha-Filho, R. C.; Fatibello-Filho, O. Food Chem. 2010, 123, 886–891. (20) Luque, M.; Rios, A.; Valcarcel, M. Anal. Chim. Acta 1999, 395, 217–223. (21) Riber, J.; de la Fuente, C.; Vazquez, M. D.; Tascon, M. L.; Batanero, P. S. Talanta 2000, 52, 241–252. (22) Ruiz, M. A.; Garcia-Moreno, E.; Barbas, C.; Pingarron, J. M. Electroanalysis 1999, 11, 470–474.

ethanol, v/v) 10 mmol L-1 KNO3 solution (pHcond ) 1.5 adjusted with concentrated HNO3) as supporting electrolyte. The oxidation peak potentials for BHA and BHT present in binary mixtures were separated by about 0.3 V. The attained detection limits for the simultaneous determination of BHA and BHT, 0.14 and 0.25 µmol L-1, respectively, are lower than the ones obtained by voltammetric techniques reported in the literature. Among the amperometric methods, two of them involved the use of HPLC to separate the antioxidants before their amperometric detection. Riber at al.21 used a polypyrrole electrode modified with a tetrasulfonated nickel(II) phthalocyanine complex as amperometric electrochemical detector in a flow injection system after an HPLC preseparation. The antioxidants analyzed were tert-butylhydroquinone (TBHQ), BHA, and PG, whose detection limits were 0.26, 0.38, and 0.38 µmol L-1, respectively. Ruiz et al.22 developed a nickel phthalocyanine polymer-coated glassy carbon electrode for the amperometric detection of BHT, BHA, PG, and TBHQ after their separation by HPLC under gradient elution conditions, using an applied potential of 0.70 V (vs Ag/AgCl). With TBHQ as internal standard, detection limits of 0.11, 0.15, and 0.6 mg L-1 were obtained for BHA, PG, and BHT, respectively. Luque et al.20 described a new poly(vinyl chloride) (PVC) composite electrode as an amperometric detector in a flow-injection system using a wall-jet flow cell for the determination of PG, octyl gallate (OG), and BHA, separately. Detection limits of 0.23, 0.32, and 0.61 mg L-1 were obtained for PG, OG, and BHA, respectively. However, as far as we know, there are no methods reported in the literature for the simultaneous determination to BHA and BHT in foods by flow injection analysis (FIA) and amperometric detection, without a previous HPLC separation of these antioxidants. FIA with amperometric detection, which is easy to handle and shows excellent sensitivity, has been widely used for the determination of various electroactive species at very low concentrations.23-25 FIA methods with amperometric detection of a single analyte have been commonly reported in the literature; however, the development of flow techniques for multicomponent analysis has been rarely described.26-29 The most common approach is the use of HPLC separation of the analytes before their quantification. Another strategy is the use of two or more sensors with the application of different constant potentials at each sensor, whose resulting signals are analyzed with a multivariate calibration method. On the other hand, few studies employing multiple pulse amperometry (MPA) and FIA were found for simultaneous determinations.30-32 In MPA, an appropriate waveform of potential pulses is applied to a single working electrode (23) Wangfuengkanagul, N.; Chailapakul, O. J. Pharm. Biomed. Anal. 2002, 28, 841–847. (24) Suryanarayanan, V.; Zhang, Y.; Yoshihara, S.; Shirakashi, T. Sens. Actuators, B 2004, 102, 169–173. (25) Siangproh, W.; Ngamukot, P.; Chailapakul, O. Sens. Actuators, B 2003, 91, 60–66. (26) Terashima, C.; Rao, T. N.; Sarada, B. V.; Tryk, D. A.; Fujishima, A. Anal. Chem. 2002, 74, 895–902. (27) Marzouk, S. A. M.; Sayour, H. E. M.; Ragab, A. M.; Cascio, W. E.; Hassan, S. S. M. Electroanalysis 2000, 12, 1304–1311. (28) Ivandini, T. A.; Honda, K.; Rao, T. N.; Fujishima, A.; Einaga, Y. Talanta 2007, 71, 648–655. (29) Matos, R. C.; Augelli, M. A.; Lago, C. L.; Angnes, L. Anal. Chim. Acta 2000, 404, 151–157. (30) Surareungchai, W.; Deepunya, W.; Tasakorn, P. Anal. Chim. Acta 2001, 448, 215–220.

Analytical Chemistry, Vol. 82, No. 20, October 15, 2010

8659

and the resolution of the analyte mixture is obtained without chemical pretreatment of the sample or electrode modification or even the use of chemometric techniques for data analysis.31 Surareungchai et al.30 developed a method for the simultaneous determination of glucose and fructose by flow injection analysis using a quadruple-potential waveform (Edet1 ) -0.5 V/240 ms, Edet2 ) 0.2 V/180 ms, Eoxd ) 1.0 V/180 ms, and Ered ) -0.8 V/300 ms) and a Nafion-coated Au electrode. Edet1 or Edet2 caused the oxidation of glucose or glucose and fructose, respectively. Hence, concentration subtraction could be used to determine both species. More recently, Santos et al.31 described a simple and quick strategy for the simultaneous determination of paracetamol (PC) and ascorbic acid (AA) in pharmaceutical formulations by a flow injection method with MPA detection. The compounds were detected by applying a sequence of four potential pulses (waveform) as a function of time to a three-electrode amperometric system that uses a walljet cell with a gold disk as working electrode and Ag/AgCl (saturated KCl) as reference electrode. AA was directly detected at 0.40 V and PC was indirectly detected at 0.0 V by the reduction (desorption) of the oxidation product (N-acetylp-benzoquinonimine) electrochemically generated at 0.65 V. The fourth potential pulse (-0.05 V) was applied for the complete regeneration (cleaning) of the gold electrode surface. Gimenes et al.32 also described the application of three sequential potential pulses to an unmodified glassy carbon working electrode positioned in a wall-jet cell for the fast and indirect quantification of dopamine (DA) in the presence of a large excess of AA [an Ag/AgCl (saturated KCl) was also used as reference electrode]. DA was indirectly detected at 0.35 V through the reduction of the oxidation product (o-dopaminoquinone) electrochemically generated at 0.80 V. The third potential pulse (0.00 V) was applied for the regeneration (cleaning) of the working electrode. Thin films of BDD have emerged as excellent electrode materials for several electrochemical applications, especially electroanalytical ones, mainly due to such properties as a wide potential window in aqueous solutions (up to 3 V), low background currents, long-term stability, low adsorption, and low sensitivity to dissolved oxygen.33-37 The properties of BDD are commonly affected by several factors, such as their electronic properties and surface termination.38 Suffredini et al.39 called to attention that cathodic pretreatment of a BDD electrode dramatically increased (31) Santos, W. T. P.; Almeida, E. G. N.; Ferreira, H. E. A.; Gimenes, D. T.; Richter, E. M. Electroanalysis 2008, 20, 1878–1883. (32) Gimenes, D. T.; Santos, W. T. P.; Tormin, T. P.; Munoz, R. A. A.; Richter, E. M. Electroanalysis 2010, 20, 74–78. (33) Hupert, M.; Muck, A.; Wang, R.; Stotter, J.; Cvackova, Z.; Haymond, S.; Show, Y.; Swain, G. M. Diamond Relat. Mater. 2003, 12, 1940–1949. (34) Swain, G. M. In Electroanalytical Chemistry; Bard, A. J., Rubinstein, I., Eds.; Marcel Dekker: New York, 2004; Vol. 22, pp 182-278. (35) Sarada, B. V.; Terashima, C.; Ivandini, T. A.; Rao, T. N.; Fujishima, A. In Diamond Electrochemistry; Fujishima, A., Einaga, Y., Rao, T. N., Tryk, D. A., Eds.; Elsevier: Tokyo, 2005; pp 261-286. (36) Peckova´, K.; Musilova´, J.; Barek, J. Crit. Rev. Anal. Chem. 2009, 39, 148– 172. (37) Jolley, S.; Koppang, M.; Jackson, T.; Swain, G. M. Anal. Chem. 1997, 69, 4099–4107. (38) Granger, M. C.; Witek, M.; Xu, J. S.; Wang, J.; Hupert, M.; Hanks, A.; Koppang, M. D.; Butler, J. E.; Lucazeau, G.; Mermoux, M.; Strojek, J. W.; Swain, G. M. Anal. Chem. 2000, 72, 3793–3804. (39) Suffredini, H. B.; Pedrosa, V. A.; Codognoto, L.; Machado, S. A. S.; RochaFilho, R. C.; Avaca, L. A. Electrochim. Acta 2004, 49, 4021–4026.

8660

Analytical Chemistry, Vol. 82, No. 20, October 15, 2010

the electroanalytical detection limit for chlorophenols, reinforcing the notion that the analytical performance of BDD electrodes greatly depends on their surface termination, that is, whether they are hydrogen- or oxygen-terminated. Recently, in our research group, this effect of cathodic pretreatment also was observed in the determination of additives in food products and drugs in pharmaceutical formulations.19,40-46 Considering the above, the aim of this work is to couple MPA and FIA with the unique properties of the BDD electrode for the development and optimization of a simple, low-cost, and rapid method for the simultaneous determination of BHA and BHT in food products (commercial mayonnaise) without a previous separation step. This method is based on a single-line flow injection system and an appropriate waveform of potential pulses that is applied to a cathodically pretreated BDD electrode. EXPERIMENTAL SECTION Apparatus. Multiple pulse amperometric flow measurements were performed in a homemade three-electrode wall-jet electrochemical cell (made of two acrylic blocks), which is described in detail elsewhere.47,48 The working electrode was a BDD film (8000 ppm boron) on a silicon wafer from the Centre Suisse de Electronique et de Microtechnique SA (CSEM), Neuchateˆl, Switzerland. Detailed information on this BDD electrode was reported elsewhere,49 while Gandini et al.50 reported details on the preparation of these diamond films. Prior to the experiments for the simultaneous analysis of BHA and BHT, the BDD electrode (0.64 cm2) was cathodically pretreated in 0.5 mol L-1 H2SO4 by applying -1.0 A cm-2 for 120 s;19 thus, the BDD surface was made predominantly hydrogen-terminated.39,49 As described previously,47,48 a stainless steel tube was used as counterelectrode, along with a miniaturized Ag/AgCl (3.0 mol L-1 KCl) reference electrode. The voltammetric measurements were carried out by use of an Autolab PGSTAT-30 (Ecochemie) potentiostat/ galvanostat controlled with GPES 4.0 software. The BHA and BHT determinations by an HPLC comparative method were carried out in an LC-10 AT Shimadzu system, with an UV-vis detector (SPD-M10-AVP) set at 290 nm (for BHA) or 278 nm (for BHT). A Shim-Pack CLC-ODS (6.0 mm × 250 mm, 5 µm) chromatographic column was used. The mobile phase was an acetonitrile/methanol mixture (50/50, v/v) at a flow rate of 1.0 mL min-1, while the injection volume was 30 µL.9 (40) Lourenc¸˜ao, B. C.; Medeiros, R. A.; Rocha-Filho, R. C.; Mazo, L. H.; FatibelloFilho, O. Talanta 2009, 78, 748–752. (41) Medeiros, R. A.; de Carvalho, A. E.; Rocha-Filho, R. C.; Fatibello-Filho, O. Anal. Lett. 2007, 40, 3195–3207. (42) Medeiros, R. A.; de Carvalho, A. E.; Rocha-Filho, R. C.; Fatibello-Filho, O. Talanta 2008, 76, 685–689. (43) Medeiros, R. A.; de Carvalho, A. E.; Rocha-Filho, R. C.; Fatibello-Filho, O. Quim. Nova 2008, 31, 1405–1409. (44) Sartori, E. R.; Medeiros, R. A.; Rocha-Filho, R. C.; Fatibello-Filho, O. J. Braz. Chem. Soc. 2009, 20, 360–366. (45) Andrade, L. S.; Rocha-Filho, R. C.; Cass, Q. B.; Fatibello-Filho, O. Electroanalysis 2009, 21, 1475–1480. (46) Batista, E. F.; Sartori, E. R.; Medeiros, R. A.; Rocha, R. C.; Fatibello-Filho, O. Anal. Lett. 2010, 43, 1046–1054. (47) Andrade, L. S.; Moraes, M. C.; Rocha-Filho, R. C.; Fatibello-Filho, O.; Cass, Q. B. Anal. Chim. Acta 2009, 654, 127–132. (48) Andrade, L. S.; Rocha-Filho, R. C.; Fatibello-Filho, O.; Cass, Q. B. Anal. Methods 2010, 2, 402–407. (49) Salazar-Banda, G. R.; Andrade, L. S.; Nascente, P. A. P.; Pizani, P. S.; RochaFilho, R. C.; Avaca, L. A. Electrochim. Acta 2006, 51, 4612–4619. (50) Gandini, D.; Michaud, P. A.; Duo, I.; Mahe, E.; Haenni, W.; Perret, A.; Comninellis, C. New Diamond Front. Carbon Technol. 1999, 9, 303–316.

Figure 2. Cyclic voltammograms in an aqueous ethanolic (30% ethanol, v/v) 10 mmol L-1 KNO3 solution (pHcond ) 1.5 adjusted with concentrated HNO3) with a cathodically pretreated BDD electrode: (---) without and (s) with the addition of (A) 0.10 mmol L-1 BHA and (B) 0.20 mmol L-1 BHT. Scan rate was 50 mV s-1.

Reagents, Supporting Electrolyte, and Standards. All reagents were of analytical grade: BHA and BHT (Sigma), KNO3 (Aldrich), and ethanol (Quemis, Brazil). An aqueous ethanolic (30% ethanol, v/v) 10 mmol L-1 KNO3 solution (pHcond ) 1.5 adjusted with concentrated HNO3) was used as supporting electrolyte, in which standard 1.0 mmol L-1 BHA and BHT solutions were prepared. All solutions were prepared with ultrapurified water supplied by a Milli-Q system (Millipore) with a resistivity greater than 18 MΩ cm. Measurement Procedures. Cyclic and hydrodynamic voltammograms were obtained for each compound before the MPAFIA determinations. The electrode potential applied in these analyses was selected with the limiting current range in the hydrodynamic voltammograms taken into account. After optimization of the experimental parameters for the proposed method, analytical curves were constructed by adding small volumes of the concentrated standard solutions of the two analytes to the supporting electrolyte in order to have the following concentrations: 0.050, 0.10, 0.50, 0.70, 1.0, 2.0, and 3.0 µmol L-1 for BHA and 0.70, 3.0, 6.0, 10.0, 30.0, 60.0, and 70.0 µmol L-1 for BHT. The limit of detection (LOD) was obtained as the concentration whose associated amperometric response is equal to 3 times the average (n ) 10) voltammetric response for the blank solution. Treatment of Commercial Mayonnaise Samples. A procedure similar to those proposed by Luque et al.20 and Raymundo et al.18 was followed for the determination of BHA and BHT in commercial mayonnaise samples purchased from a local supermarket. A mayonnaise sample of about 1 g was dissolved in 2 mL of ethanol contained in a large test tube. After being shaken for 5 min, this mixture was centrifuged at 5000 rpm for 10 min. This extraction procedure was repeated twice; the extracts were collected and then diluted to 5 mL with the supporting electrolyte. Previous to injection into the electrochemical cell, a 500 µL aliquot was further diluted to 10 mL with the supporting electrolyte. For the HPLC measurements, after the extraction procedure, the extracts were diluted in 5 mL of mobile phase, of which a 500 µL aliquot was further diluted to 10 mL with mobile phase before injection into the chromatographic column. RESULTS AND DISCUSSION Investigation of Electrochemical Behavior of BHA and BHT and Flow Injection Parameters. First, cyclic voltammograms of BHA and BHT on the cathodically pretreated BDD

Figure 3. Hydrodynamic voltammograms obtained for (b) 0.10 mmol L-1 BHA and (9) 0.10 mmol L-1 BHT by use of a cathodically pretreated BDD electrode. Supporting electrolyte: aqueous ethanolic (30% ethanol, v/v) 10 mmol L-1 KNO3 (pHcond ) 1.5 adjusted with concentrated HNO3); injected volume ) 250 µL; flow rate ) 2.4 mL min-1.

electrode in the aqueous ethanolic (30% ethanol, v/v) 10 mmol L-1 KNO3 solution (pHcond ) 1.5 adjusted with concentrated HNO3) were obtained (Figure 2). Both BHA and BHT exhibited welldefined irreversible oxidation peaks around 660 and 970 mV, respectively. According to the literature,14,15,19 the electrochemical oxidation of BHA or BHT occurs by a two-electron process. Hydrodynamic voltammograms were obtained to investigate the electrochemical response of the cathodically pretreated BDD electrode under dynamic conditions and to optimize its operating potentials for the MPA-FIA measurements. In these investigations, the potential of the BDD electrode was increased in steps, and the resulting steady-state currents were measured and plotted as a function of the applied potential (see Figure 3). For BHA, the current starts to increase at about 600 mV and approaches a maximum value in the potential range of 850-950 mV. For BHT, the current starts to increase at about 850 mV and approaches a maximum value in the potential range of 1150-1250 mV. Hence, a potential equal to or greater than 850 or 1150 mV can be applied for the electrochemical oxidation of BHA or BHT, respectively, in flow systems. Moreover, the influence of parameters such as flow rate and injected sample volume on the multiple pulse amperometric response was evaluated. The flow rate was studied in the range 0.65-2.75 mL min-1, by use of a solution containing both analytes at a concentration of 0.10 mmol L-1 and applied potentials of 850 and 1150 mV. The current increased with flow rate up to 2.4 mL min-1 and remained constant for higher values; thus, this flow rate was selected for the most favorable response and Analytical Chemistry, Vol. 82, No. 20, October 15, 2010

8661

Figure 4. (A) MPA waveform applied to the cathodically pretreated BDD working electrode as a function of time. (B) Flow-injection pulse amperometric responses in triplicate for solutions containing 50 µmol L-1 BHA or BHT or both analytes simultaneously at this concentration. Supporting electrolyte: aqueous ethanolic (30% ethanol, v/v) 10 mmol L-1 KNO3 solution (pHcond ) 1.5 adjusted with concentrated HNO3); flow rate 2.4 mL min-1; injected volume 250 µL.

corresponds to a satisfactory sampling rate (about 63 per hour). The effect of injected sample volume was evaluated by use of volumes in the range of 50-600 µL. The amperometric signal increased linearly with sample volume up to 250 µL; thus, a volume of 250 µL was adopted. Multiple Pulse Amperometric Simultaneous Determination of BHA and BHT. If the traditional chronoamperometry technique were used, indeed BHT could not be quantified in the presence of BHA because the oxidation potential of BHA (850 mV) is smaller than the oxidation potential of BHT (1150 mV). However, the simultaneous quantification of this couple is possible in the MPA technique, with the application of two potential pulses. Figure 4A shows a double-potential waveform for the simultaneous determination of BHA and BHT by flow injection analysis. In this waveform, Edet1 ) 850 mV/200 ms corresponds to the oxidation of BHA without any BHT interference, while Edet2 ) 1150 mV/ 200 ms corresponds to the oxidation of both BHA and BHT. MPA usage in the literature commonly involves the use of another potential pulse to clean the electrode surface, but in this work this third potential pulse was not necessary because adsorption onto the BDD electrode is not common and the electrode presented good repeatability. Figure 4B shows the pulse amperograms obtained at 850 mV (Edet1) and 1150 mV (Edet2) for the injection of three different solutions. The first solution contains only 50 µmol L-1 BHA; the second, only 50 µmol L-1 BHT; the third, both analytes simultaneously at this concentration. As can be seen, at 850 mV only BHA is oxidized, while at 1150 V both BHA and BHT are oxidized. Then, the proportionality between current and BHA concentration is obtained at Edet1, whereas the proportionality between current and BHT concentration is obtained through adequate current subtractions: IBHT ) IEdet2 - IEdet1. Therefore, it is feasible to quantify both analytes simultaneously, without interference from BHA on BHT as happens when the traditional chronoamperometry technique is used. In order to establish the linear working range, a series of experiments was performed with standard solutions of BHA and BHT; both antioxidants were determined by simultaneously 8662

Analytical Chemistry, Vol. 82, No. 20, October 15, 2010

Figure 5. MPA responses for a cathodically pretreated BDD electrode, after injections of solutions containing BHA (0.050-3.0 µmol L-1) and BHT (0.70-70 µmol L-1) simultaneously or different samples of mayonnaise (A-D). Supporting electrolyte: aqueous ethanolic (30% ethanol, v/v) 10 mmol L-1 KNO3 solution (pHcond ) 1.5 adjusted with concentrated HNO3); flow rate 2.4 mL min-1; injected volume 250 µL.

changing their concentrations (see Figure 5). The respective analytical curves presented good linearity in the investigated concentration range (0.050-3.0 µmol L-1 for BHA and 0.70-70 µmol L-1 for BHT), with the following calibration equations: I ) 0.006 19 + 0.0559c I ) 0.144 + 0.0503c

r ) 0.998

(I ) IEdet1 for BHA)

r ) 0.999 (I ) IEdet2 - IEdet1 for BHT)

where I is given in microamps and c is given in micromoles per liter. The estimated LOD values (3S/N) were 0.030 µmol L-1 for BHA and 0.40 µmol L-1 for BHT. The LOD value for BHA is excellent, since it is about a fifth the value previously reported by our laboratory,19 which was already the lowest value in the literature. On the other hand, the LOD value for BHT is 60% higher than the one previously reported by our laboratory19 but still similar to another value in the literature.11 The intraday and interday repeatabilities were determined by successive injections (n ) 10) of BHA and BHT at different concentrations; the obtained relative standard deviation (RSD) values were, for BHA, intraday 3.0% and interday 4.1%; and for BHT, intraday 1.9% and interday 3.4%. The selectivity of the proposed method was evaluated by the addition of possible interferents [NaCl, soluble starch, ethylenediaminetetraacetic acid (EDTA), citric acid, and acetic acid] in a standard solution containing 10 µmol L-1 BHA and 10 µmol L-1 BHT, in the concentration ratios (standard solution:interferent) 10:1, 1:1, and 1:10; the obtained current signals were compared with those in the absence of any interferent. Analysis of the obtained responses allowed us to conclude that these compounds do not significantly interfere with the method proposed here. The highest interference (RSD 4.8%) was found for EDTA at the concentration ratio 1:10, but this ratio was not present in the analyzed samples. Finally, the method developed for simultaneous analysis of BHA and BHT was tested with four samples of mayonnaise. Figure 5 presents also the transient signals in triplicate obtained from injections of solutions of these samples of mayonnaise. Table 1 presents the results obtained for the analysis of these samples:

Table 1. Results Obtained in the Simultaneous Determination of BHA and BHT in Mayonnaise Samples by HPLC and the Proposed Method (MPA) BHA (mg/100 g)

BHT (mg/100 g)

mayonnaise sample

HPLCa

MPAa

HPLCa

MPAa

1 2 3 4

2.0 ± 0.1 1.7 ± 0.2 2.3 ± 0.2 2.3 ± 0.1

2.1 ± 0.2 1.6 ± 0.1 2.2 ± 0.2 2.1 ± 0.3

1.1 ± 0.1 1.3 ± 0.2 1.8 ± 0.2 1.4 ± 0.2

1.2 ± 0.3 1.3 ± 0.2 1.8 ± 0.3 1.3 ± 0.1

BHA BHT errorb errorb (%) (%) 5.0 -5.9 -4.3 -8.7

9.1 0.0 0.0 -7.1

a Average of three measurements. b Error (%) ) 100 × (MPA value - HPLC value)/HPLC value.

average of three determinations for each sample and the corresponding standard deviations. The results obtained through the proposed method were also compared with those obtained by an HPLC method.9 The paired t-test was applied to the results obtained by HPLC and the proposed method, and the resulting t values (1.54 for BHA and 1.66 for BHT) are smaller than the critical one (3.18, R ) 0.05), indicating that in both cases there is no difference between the obtained results at a confidence level of 95%.

CONCLUSIONS The results here reported allow us to conclude that the use of MPA and FIA along with a cathodically pretreated BDD electrode for simultaneous determination of BHA and BHT in commercial mayonnaise is possible. Indeed, very low detection limits were obtained in the simultaneous determination of BHA (0.03 µmol L-1) and BHT (0.40 µmol L-1). Furthermore, addition and recovery studies allowed us to conclude that the matrix effect did not present any significant interference. Additionally, the concentration values obtained for BHA and BHT are similar to those obtained by an HPLC method. The proposed methodology is simple, quick, and can be carried out with good precision and accuracy. Finally, it should be noted that this methodology is readily applicable to other food products. ACKNOWLEDGMENT We gratefully acknowledge the Brazilian funding agencies FAPESP, CNPq, and CAPES for financial support and scholarships.

Received for review July 20, 2010. Accepted September 6, 2010. AC101921F

Analytical Chemistry, Vol. 82, No. 20, October 15, 2010

8663