ANALYST, MAY 1992, VOL. 117 857 Simple Voltammetric Method for the Determination of p-Carotene in Brine and Soya Oil Samples at Mercury and Glassy Carbon Electrodes B. Valentin Pfund and Alan M. Bond Department of Chemistry, La Trobe University, Bundoora, 3083, Victoria, Australia Terence C. Hughes* Denehurst Ltd., 961 Glenhuntly Road, Caulfield South, 3162, Victoria, Australia (3-Carotene is a naturally occurring yellow-orange pigment, which can be derived from saline micro-algae marine phytoplankton and some plant-derived natural oils. In this work, a simple method for the determination of (3-carotene, involving only solvent extraction from brine (or soya oil) samples into dichloromethane, followed by addition of electrolyte and direct measurement of the differential-pulse polarogram at mercury electrodes or differential-pulse voltammograms at a glassy carbon electrode, is described, based on the extremely well-defined two-electron oxidation process that occurs in non-aqueous solvents.The method has been applied t o soya oil and brine reference concentrates and t o feed and effluent samples associated with the production of 6-carotene via marine micro-algae. Excellent agreement with a well-established spectrophotometric method has been obtained, confirming that the simple voltammetric method should be a useful addition t o the analytical methodology available for monitoring the production of (3-carotene concentrates. Keywords: Voltammetry; analytical determination; (3-carotene (3-Carotene is a precursor of vitamin A and contributes to the colour of many plants.It is particularly well known as a yellow-orange pigment found in carrots. (3-Carotene contains 11 carbon-carbon double bonds in conjugation (Fig. 1) and owes its colour to absorption at the violet end of the visible spectrum (A = 451 nm).] In view of its colour, it is not surprising that spectrophotometric methods for the deter- mination of p-carotene have been widely used.2 In addition to being present in many plants, carotenoids, including p-carotene, are present at relatively high concentra- tions in saline micro-algae marine phytoplankton and other marine matter. For the determination of (3-carotene in marine samples, high-performance liquid chromatographic separa- tion procedures, coupled with spectrophotometric detec- tion3.4 and a very sensitive and specific resonance Raman method, have been described.5-6 In view of the presence of @-carotene in many natural products, it is not surprising that commercial products are usually derived from sources such as micro-algae. In the production of (3-carotene from natural sources, it is necessary to have quality control at every stage of the plant production process.Simple methods for the rapid determination of b-carotene are, therefore, required as an alternative or in addition to the commonly used spectrophotometrie methods,*-h which generally require relatively time-consum- ing separation procedures to achieve adequate selectivity. Despite the fact that the extensive series of conjugated double bonds that are present in the structure of (3-carotene should indicate that carotene is likely to be electroactive, the analytical use of voltammetric oxidation and/or reduction processes, known to occur on a range of electrode surfaces,7-13 has been rather limited.However, the amperometric detec- tion of (3-carotene in irradiated fruits after chromatographic separation has been described in detail.14 In this method, carotene esters and other carotene compounds are hydrolysed by addition of KOH, mixed with ethanol containing pyrogal- late and extracted into light petroleum for injection on to the chromatographic column. This application refers to the determination at naturally occurring trace levels, where interference from many other related electroactive com- pounds is expected to occur without inclusion of a chromato- graphic or other form of separation into the analytical methodology.In this work, we have investigated the possibil- ity of developing a more direct voltammetric method in which the @-carotene present in relatively high concentrations in the commercial concentrate is simply extracted into dichloro- methane containing an electrolyte, and a direct voltammetric determination is then undertaken in the non-aqueous solvent. Results from this very simple procedure are then compared with a spectrophotometric method to confirm the validity of the voltammetric method. Experimental A 30% (3-carotene sample (Hoffman-La Roche, Basle, Switzerland) was used to prepare a 10-2 mol dm-3 standard solution in dichloromethane (electrolyte). The reference material samples (technical-grade quality), provided by Betatene (Melbourne, Australia), were as follows: (i) 30% (3-carotene in soya oil, and (ii) 1.5% fi-carotene in brine.Samples obtained from various stages of the production of (3-carotene, and examined in this work, were also supplied by Fig. 1 Structure of &carotene * Present address: Unichema Australia Pty Ltd., 164 Ingles Street, Port Melbourne, 3207, Victoria, Australia.858 ANALYST, MAY 1992, VOL. 117 Betatene. The reference method for the determination of (3-carotene was a spectrophotometric procedure, which is approved by the AOAC.15 It consists of a sample homogeniza- tion step, separation and purification of the (3-carotene by quantitative solvent extraction into hexane, and a spectropho- tometric determination under standardized conditions at 436 nm.Before use, analytical-reagent grade dichloromethane was distilled with use of a 30 cm Vigreux column. The electrolytes were electrometric grade (G. Frederick Smith Chemicals, Columbus, OH) tetrabutylammonium perchlorate (Bu4N- C104) or tetrabutylammonium tetrafluoroborate (Bu4NBF4), used at a concentration of 0.1 mol dm-3 in distilled di- chloromethane. Initial vol tamme tric (polarographic) investigations to con- firm the mechanism of the electrode processes and to establish optimal conditions for the analytical procedures were under- taken with a PAR Model 174A polarographic analyser, equipped with a dropping-mercury or platinum-disc working electrode, a platinum-wire auxiliary electrode and an Ag- AgCl (dichloromethane; saturated LiCI) reference electrode. Analytical determinations of (3-carotene were carried out with a Metrohm Model 646 VA processor and Model 647 VA stand, with use of a multi-mode mercury working electrode operated in the dropping-mercury mode, or a glassy carbon working electrode, with a glassy carbon auxiliary electrode and the same Ag-AgC1 reference electrode as above.All experiments were undertaken at 20 k 1 "C and where necessary (for reduction studies) solutions were de-gassed with high-purity nitrogen to remove oxygen before comment- ing a voltammetric experiment. Results and Discussion Details of the Electrode Processes in Dichloromethane Fig. 2 shows a differential-pulse polarogram for a 1 mmol dm-3 solution of 0-carotene in dichloromethane (0.1 rnol dm-3 Bu4NC104).At a (peak) potential (Ep) of -1.70 V versus the Ag-AgC1, a well-defined reduction wave is observed, and at +0.59 V versus the Ag-AgC1, a narrower and larger oxidation process is observed. As also shown in Fig. 2, the reference compound ferrocene (Fc) exhibits a reversible one-electron oxidation process (Fc # Fc+ + e-) with a peak potential of 0.50 V versus the Ag-AgC1 under the same conditions. At a mercury electrode, no other waves were observed prior to the solvent limit (negative potential limit) or mercury electrode oxidation (positive potential limit). Mairanovsky et a1 .I3 report a reversible one-electron reduction with a half-wave potential (E; value) of -1.68 V versus the SCE in non-aqueous solvents to produce a L 1 I 1 I I 0.50 0 -0.50 -1.00 -1.50 -2.00 €N versus Ag-AgCI-LiCI Fig. 2 Differential-pulse polarogram (drop time = 1 s, pulse amplitude = 50 mV) for reduction and oxidation of p-carotene in dichloromethane (0.1 mol dm-3 Bu4NC104). 1, Solvent (baseline); 2, solvent and 1 X lo-' rnol dm-3 6-carotene; and 3, solvent with 1 x 10-3 mol dm-3 /3-carotene and 1 X 10-3 rnol dm-3 ferrocene. A.6-Carotene and B, ferrocene moderately stable anion radical (and other reduction processes at a more negative potential, which are outside the dichloromethane solvent range) and a two-electron oxidation process with an Eh value of +0.61 V versus SCE. The oxidation process is a reversible two-electron charge transfer with an irreversible chemical step, following charge transfer being observed with long-term domain experiments.The separation between the E,-values in this work and the El values reported in ref. 13 agree completely, confirming that the solvent is not particularly important in determining the potential for reduction or oxidation nor probably the mechan- ism for either process. In agreement with Mairanovsky et al., we find that the reduction process is a one-electron step and the oxidation a two-electron step, which disagrees with Takakachi and Tachi,8.9 who described the reduction as a four-electron process. The reduction process is chemically and electrochemically reversible in dichloromethane under conditions of cyclic voltammetry in the sense that it has a AE, value at both platinum and mercury electrodes for separation of reduction and oxidation components identical to that obtained for the known reversible one-electron oxidation of Fc.The product of the reduction in dichloromethane can, therefore, be postu- lated to be the anion radical as is the case in other non-aqueous solvents.13 As required for the proposed mechanism, the direct-current polarographic limiting current for the reduction step is one-half that of the oxidation process (ignoring sign differences), and, in agreement with Mairanovsky et al. ,I3 the oxidation process, while having electrochemically reversible two-electron characteristics with respect to charge transfer, exhibits some degree of chemical irreversibility at both platinum and mercury electrodes at a scan rate of 500 mV s- 1 (Fig. 3). Oxidation processes at more positive potentials are not discussed in this paper.The mechanism for oxidation proposed by Mairanovsky et a1.13 involves the initial formation of a dication by a two-electron charge transfer process followed, in longer time domain experiments, by proton loss to form a monocation and the one-electron reduction of the monocation to a neutral, but unstable, radical is then detected on the reverse scan of cyclic voltammograms (Fig. 3). The oxidation mechanism in di- chloromethane, while not studied in detail in this work, therefore appears to be the same as in other solvents examined by Mairanovsky et al. 13 In dichloromethane, the above data demonstrate that, in principle, either of two processes could be used for analytical purposes. However, the oxidation process is a more sensitive two-electron step, has a narrower half-width, is better resolved from the solvent limit process, and because it is an I I 0 0.2 0.4 0.6 0.8 1.0 E N versus Ag-AgCI Fig.3 Cyclic voltammograms (scan rate = 500 mV s-I) obtained at 1, a slowly growing mercury electrode and 2, at a platinum disc electrode for oxidation of 1 x 10-3 mol dm-3 (%carotene in dichlo- romethane (0.1 rnol dm-3 Bu4NCI04)ANALYST, MAY 1992, VOL. 117 859 oxidation occurring at relatively positive potentials, rather than a reduction step, its use does not require the removal of oxygen. This process was, therefore, employed in subsequent analytical studies. Differential-pulse Polarography at the Dropping-mercury Electrode Calibration With a drop time of 1.0 s and a pulse amplitude of 50 mV, a plot of differential-pulse peak height for the oxidation of (3-carotene in dichloromethane (0.1 rnol dm-3 Bu4NC104) was linear over the concentration range of 2 x to 10-3 rnol dm-3 with a correlation coefficient of 0.9997 (slope 27.5 pA/pmoi dm-3, intercept 0.006 PA).At concentrations above 2 x 10-3 rnol dm-3, non-linearity was observed in the calibration curve, which may have been a result of Ohmic iR drop. Consequently, determinations were confined to concen- trations up to 10-3 rnol dm-3 and were undertaken by the method of standard additions to avoid matrix effects. Determination of [%carotene in a reference soya oil sample A 100 mg sample of (3-carotene in soya oil (30%) (Betatene) was weighed into a 100 ml calibrated flask containing distilled dichloromethane (0.1 rnol dm-3 Bu4NCI04).After dilution to 100 ml with dichloromethane, the differential-pulse polaro- gram (drop time = 1.0 s, AE = 50 mV) was recorded for the solution over the range +0.20 to +0.85 V versus the Ag- AgCl. A well-defined differential-pulse peak corresponding to oxidation of 0-carotene was observed at +0.55 V versus the Ag-AgCI [ Fig.4(a)]. The standard-additions method was used to determine the @carotene in the sample with a value of 34.7 -t 0.8% of 0-carotene being obtained from four determina- tions, which is in satisfactory agreement with the manufac- turer’s nominal value of 30%. Recoveries of 100 k 3% were obtained for soya oil samples spiked with known amounts of fi-carotene, which also suggests that a valid procedure has been developed for the voltammetric determination of p- carotene in soya oil.10 pA I I 0.80 0.50 0.20 0.6 0.4 0.2 EN versus Ag-AgCI Fig. 4 Differential-pulse polarograms (drop time = 1 s, pulse amplitude = SO mV) obtained for the dctcrmination of (3-carotenc in ( u ) soya oil rcfcrcncc sample: 1. sample; 2, after addition of 5 x IO-’ rnol dm - j 13-carotene; and ( h ) brinc reference sample: 1. sample; 2. after addition of S x 10Y rnol dm-3 @-carotene; and 3 , after addition of 1 x 10-3 rnol dm-3 p-carotcnc. For details see text Determination of (3-carotene in brine A 2.50 mg sample of (3-carotene concentrate in brine (1.5%) (Betatene) was extracted with distilled dichloromethane (4 x 10 ml) and, after addition of 3.4 g of Bu4NC104, the extract was diluted to 100 ml with distilled dichloromethane. Dif- ferential-pulse polarograms [Fig.4(b)] obtained for the solu- tion and analysed by the standard-additions method, as for the soya oil sample, gave a value of 1.9 k 0.1% of (3-carotene, based on four determinations, which again is in satisfactory agreement with the manufacturer’s nominal value of 1.5%. As was the case with the soya oil sample, recoveries of 100 2 3% were obtained for samples of brine spiked with known amounts of @carotene. Table 1 Data for the differential-pulse polarographic determination of B-carotene in brine samples obtained at various stages of production from marine micro-algae (details of the method used are given in the text) Results (@-carotene concentration) Voltammetry Sample origin Mass used/g ( n = 4) Product 2 0.625 3.06 f 0.07 Product 3 50.0 37 f 3 Product 1 510 1.47 k 0.02 Effluent 1 46 1 0.06 k 0.02 Effluent 2 509 <0.05 (g kg- (mg kg-’) (mg kg- 1 (mg kg- 1 (mg kg- l ) t Y E 2 u 0.8 0.6 0.4 0.2 0.8 0.6 0.4 0.2 EN versus Ag-AgCI Fig.5 Differential-pulse polarograms (drop timc = 1 s, pulse amplitude = 50 mV) for the determination of (S-carotene in brinc plant feed samples. ( a ) product 2 sample: 1, sample; 2, sample plus 200 yl of 1 x 10-2 rnol dm-3 8-carotene standard; 3 , sample plus 400yl of 1 x 10-2mol dm-3 /3-carotene standard; ( h ) product 1 sample: 1, sample; 2, sample plus 100 yI of 1 x 10-2 rnol dm-3 B-carotene standard; 3 , sample plus 200 yl of 1 x lo-’ rnol dm-3 @-carotene standard. For details see Table 1 and textANALYST, MAY 1992, VOL.117 0.8 0.6 0.4 0.2 0.8 0.6 0.4 0.2 E N versus Ag-AgCI Fig. 6 Differential-pulse polarogram for the determination of (3-carotene in brine plant effluent samples. (a) Effluent 1 sample; (b) Effluent 2 sample. Curve 1 is the sample and curves 2 and 3 correspond to addition of 100 and 2001.~1 of 1 X 10-2moldm-3 (3-carotene standard, respectively. For details see Table 1 and text Determination of (3-carotene in brine during various stages of production In the commercial production of (3-carotene from marine micro-algae, a substantial number of determinations are required at each of various stages of the manufacturing process. Results for the determination of (3-carotene in the production samples and effluent samples are in excellent agreement with values determined spectrophotometrically, as indicated in Table 1.For the brine samples cited in Table 1, an aliquot of each sample was extracted with distilled dichloromethane (4 X 5 ml) with use of a centrifuge to speed up the separation of the aqueous and dichloromethane phases. After addition of 0.85 g of Bu4NC104, the extract was diluted to 25 ml with dichloro- methane, and a differential-pulse polarogram was recorded. In the ‘product 2’ sample, where (3-carotene concentrations are very high, the polarograms are well defined and equivalent to that in Fig. 3(b). In the ‘product 3’ and ‘product I’ feed samples, resolution from neighbouring peaks can be achieved and the expected peak from oxidation of (3-carotene is observed at +0.5-+0.6 V versus the Ag-AgCI (Fig.5 ) . In the example of the ‘effluent 1’ sample [Fig. 6(a], the (3-carotene levels are low and near the detection limit, whereas for the ‘effluent 2’ sample, the @-carotene concentration is below the detection limit [Fig. 6(b)]. However, in all instances, excellent agreement is obtained with the spectrophotometric method. Consequently, it is confirmed that the simple and direct method can be used for the determination of (3-carotene in these commercially important samples without the need for separation procedures other than those introduced via the solvent-extraction step. Of course, if samples at naturally occurring (3-carotene levels were being examined, where many compounds at much higher concentrations than (3-carotene would be present, chromatographic methods with amperometric detection, as described in ref.14, would almost certainly be required to 12 8 Q, .k 4 0 t a 7 6 5 0.3 0.6 0.9 0.3 0.6 0.9 E N versus Ag-AgCI 0.3 0.6 0.9 0 20 40 60 80 100 [p-Carotene]/mg I-’ Fig. 7 (a)-+) Differential-pulse voltammograms and ( d ) calibration graph for oxidation of (3-carotene at a glassy carbon electrode in 25 ml of dichloromethane (0.1 mol dm-3 Bu4NC104) after addition of an aliquot of a 1.00 g 1-1 @-carotene standard solution. Curve 1 is for a sample containing no (3-carotene, curves 2-5 are for samples containing 4 x 100 p1 additions of the @-carotene standard, and curves 6-10 are for samples containing 5 x 400 p1 additions of the (3-carotene standard. Duration between pulses = 0.8s. Pulse amplitude = 50 mV. For details see text achieve adequate resolution from overlapping peaks that could arise from the presence of compounds having similar oxidation (reduction) potentials to (3-carotene.In contrast to naturally occurring matrices, the sample we have been interested in, (3-carotene, is present at very elevated levels and is a major constituent. In this sense, it may, therefore, be surprising that no significant interference has been encountered in the analytical procedure. Differential-pulse Voltammetry at a Glassy Carbon Electrode and with Bu4NBF4 Electrolyte In a production plant environment, analytical procedures requiring mercury, as used with the technique of polar- ography, may not be considered to be suitable from the aspect of occupational health. Consequently, differential-pulse vol- tammetry at a glassy carbon electrode was examined as an alternative to polarography at a dropping-mercury electrode.As shown in Fig. 7, well-defined curves and linear calibration curves can be obtained at the glassy carbon electrode, and data essentially indistinguishable from those obtained at the mercury electrode are observed for all samples. Use of a glassy carbon electrode for asymmetric detection after chromato- graphic separation obviously also would be preferred in a liquid chromatography-electrochemical detection method. Similarly, replacement of a 0.1 mol dm-3 Bu4NC104 by 0.1 mol dm-3 Bu4NBF4 was examined and also found to be equally suitable for the determination of P-carotene. Perchlor- ate electrolytes are often regarded as potentially explosive materials, and the use of tetrafluoroborate electrolyte as anANALYST, MAY 1992, VOL.117 861 alternative may, therefore, be preferred. Fortunately, the peak potentials and characteristics of the voltammetric curves obtained at carbon, platinum or mercury electrodes with either perchlorate or tetrafluoroborate electrolytes are essen- tially the same as described in detail for those at the dropping-mercury electrode, with 0.1 mol dm-3 Bu4NC104 as the electrolyte, so that the occupationally safer alternatives are also viable. Conclusions A simple voltammetric procedure for determining 6-carotene, involving extraction from brine into dichloromethane, fol- lowed by direct determination in dichloromethane with 0.1 mol dm-3 Bu4NCI04 or 0.1 mol dm-3 Bu4NBF4 as the electrolyte, has been shown to be applicable to solutions relevant to the production of 6-carotene from marine micro- algae and phytoplankton.Equivalent results to those of methods based on the well-established spectrophotometric procedure are obtained, and the method should be useful for monitoring important stages of plant production of p-caro- tene. Equivalent and equally useful voltammetric methods could also probably be developed for other components in the increasingly important field of the determination of caro- tenoids and porphyrins in foods and natural products. We thank Betatene Ltd. for the supply of the S-carotene samples. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 References Morrison, R. T., and Boyd, R. N., Organic Chemistry, Allyn and Bacon, Boston, MA, 4th edn., 1983, p. 1153. Walton, H. F., and Reyes, J., Modern Chemical Analysis and Instrumentation. Marcel Dekker. New York, 1973. Crompton, T. R., Determination of Organic Substances in Water, Wiley Interscience, Chichester, UK, 1985, vol. 2, p. 498. Abaychi, J. K.. and Riley. J. P., Anal. Chim. Acta, 1979,107, 1. Beyermann, K., Organic Trace Analysis, Ellis Honvood. Chichester, UK, 1985. p. 235. Hoskins, L. C., and Alexander, V., Anal. Chem., 1977,49,695. Takahashi, R.. Rev. Polarogr., 1961, 9. 247. Takahashi, R., and Tachi, 1.. Agric. Biol. Chern., 1962,26,771, 777. Takahashi, R., and Tachi, I . , Abh. Dtsch. Akad. Wiss. Berlin, Kl. Med., 1966, 589. Kuta, E. J., Science, 1964, 144, 1130. Kuta, E. J., and Yu, M., Lipids, 1967, 2, 411. Mairanovsky, V. G., Engovatov, A. A., and Samokhvalov, G. I., Zh. Org. Khirn., 1970,6, 632. Mairanovsky, V. G., Engovatov, A. A., Ioffe, N. T., and Samokhvalov. G. I.. J. Electroanal. Chem., 1975, 66, 123. Argneessens, R., Nangniot, P., Lacroix, J. P., and Muri, D., Bull. Rech. Agron. Gembloux, 1989, 24, 85; Chem. Abstr., 1989, 111, 1327012. Official Methods of Analysis of the Association of Official Analytical Chemists, ed. Williams, S . , AOAC, Washington, DC, 14th edn., 1984, Nos. 4301443023. p. 834. Paper I fO3684J Received July 19, I991 Accepted November 11, I991