Analyst, August, 1979, Vol. 104, pp. 723-729 723 Direct Differential - pu Ise Pola rsg ra p h ic Determination of Mixtures of the Food Colouring Matters Tartrazine - Sunset Yellow FCF, Tartrazine - Green S and Amaranth - Green S in Soft Drinks A. G. Fogg and K. S. Yo0 Chemistry Department, Loughborough University of Technology, Loughborough, Leicestershire, LE 11 3T U Tartrazine and Sunset Yellow FCF can be determined directly in orangeade by differential-pulse (d.p.) polarography on the addition of pH 9 Britton - Robinson buffer and tetraphenylphosphonium chloride. The tetraphenyl- phosphonium chloride removes the large polarographic maximum obtained with tartrazine at pH > 4 and causes the d.p. polarographic peaks of the two colouring matters to be separated. Tartrazine in limeade can be determined in a similar supporting electrolyte but these conditions are not suitable for the determination of Green S, which is usually present at a low concentration relative to the tartrazine and for which the d.p.polarographic peak is depressed by the addition of tetra- phenylphosphonium chloride. Green S can be determined after adding pH 4 Britton - Robinson buffer and tetramethylammonium chloride to the limeade : the addition of tetramethylammonium chloride gives a better base line in the presence of tartrazine. The solution is then re-adjusted to pH 9 and tetraphenylphosphonium chloride is added in order to determine the tartrazine. At pH 4 the sugar present in blackcurrant syrup gives a small d.p. polaro- graphic peak at the same potential as Green S.At pH > 6 the peak of the sugar disappears but amaranth gives a broad polarographic maximum. This maximum is suppressed at pH 7.8 by the addition of tetramethylammonium chloride. Under these conditions the Green S peak is separate but the small concentrations of Green S normally present in blackcurrant drinks can only just be detected. The procedures have been tested on soft drinks prepared with known concentrations of colouring matter. Keywoevds : Diffeevential-pulse polarography ; food colowing mattevs ; tartrazine ; Green S ; amaranth The UK Statutory Instrument concerned with colouring matter in food1 lists permitted colouring matters and gives specifications for purity of colouring matter samples; colour is to some extent self-limiting and no limits are set on the amounts of these colouring matters that can be added to foods. Most analytical publications on food colouring matters are concerned with the identification of colouring matters, usually by thin-layer chromatographic method^.^-^ Spectrophotometric quantification can be applied after thin-layer chromato- graphic ~eparation.~ The continuing need for the use of food colouring matters, together with some concern about their widespread use, has aroused interest in their determination in foods.Food colouring matters and their intermediates have been determined by liquid - solid, ion- exchange and steric-exclusion forms of high-performance liquid chromatography (HPLC) ,* but more recently better results have been obtained by using the newly introduced paired- ion chromatographic m ~ d e .~ . ~ For thin-layer chromatographic - ultraviolet spectrometric or HPLC determination, the food colouring matters are first extracted from the foodstuff .3 With beverages and water-soluble foods the colouring matters are first adsorbed on wool or polyamide and are then re-extracted into an organic solvent. Colouring matters in more intractable foods undergo a more rigorous extraction with an organic solvent or liquid ion- exchange resin, before being adsorbed on wool or polyamide.724 FOGG AND YO0 : DIRECT DIFFERENTIAL-PULSE POLAROGRAPHIC Analyst, VOZ. 104 This work was undertaken in order to assess the value of the differential-pulse polarographic method for the determination of food colouring matters. Most colouring matters are reducible at the dropping-mercury electrode and give distinct polarographic waves.' Clearly, samples of colouring matters obtained using the clean-up procedures described above could be determined by using a diff erential-pulse polarographic finish.The peak potential observed could be used as partial confirmation of the identity of the colouring matter. Colouring matters with the same reducible group, e.g., the azo group, do tend to be reduced at similar potentials, however, so that identification cannot usually be made unequivocally and the analysis of mixtures of colouring matters can be difficult without prior separation. The ion-pair extraction method has been used extensively in the determination of ionisable pharmaceutical compounds,* and is now being adapted increasingly to paired-ion HPLC.9 We decided to try the ion-pair extraction approach with food colouring matters, and found that the acidic food colouring matters tartrazine.and Sunset Yellow FCF could be extracted into chloroform from orange squash using tetraphenylphosphonium chloride. After evaporating the chloroform and dissolving the colouring matters in pH 9 Britton - Robinson buffer, two distinct polarographic waves were obtained. Subsequently, the extraction step was found to be unnecessary; the addition of tetraphenylphosphonium chloride to orange squash buffered at pH 9 alters the potential at which tartrazine is reduced. Whereas in the absence of tetraphenylphosphonium chloride the two colouring matters are reduced at similar potentials (EP values: Sunset Yellow FCF -0.64 V; and tartrazine -0.73 V), in its presence two separate waves are obtained. Direct differential-pulse polarographic procedures for the determination of three mixtures of colouring matters have been developed.Experimental and Results Apparatus Polarographic measurements were made with a PAR 174 polarographic analyser (Princeton Applied Research). For differential-pulse operation, the forced drop time was 1 s, the pulse height 50 mV and the scan rate 2 mV s-l. Two-electrode operation was used with a dropping-mercury electrode and a saturated calomel reference electrode. A water-jacketed polarographic cell was used and the temperature was maintained at 25 "C. Reagents and Samples Britton - Robinson buffer (PH 1.9; 0.04 M in each constituent) was prepared by dissolving 2.47 g of boric acid in 500 ml of distilled water containing 2.3 ml of glacial acetic acid and then adding 2.7 ml of orthophosphoric acid and diluting to 1 1 with water.The pH of the buffer was adjusted as required by means of 0.2 or 4 M sodium hydroxide solution. Tetraphenylphosphonium chloride (0.01 M) and tetramethylammonium chloride (1 M) solutions were prepared from laboratory-grade reagents. Samples of amaranth, Sunset Yellow FCF, tartrazine and Green S, and of a blackcurrant health drink syrup and a basic syrup, were kindly provided by Beecham Products Ltd. The concentrations of colouring matters quoted in this paper assume that the samples of colouring matters are pure, i.e., that they contain 100% m/m of the colouring matter.This assumption was adequate for carrying out the recovery tests made here, but for normal routine analytical use calibration ' graphs should be obtained by using assayed samples of the colouring matters. Determination of Sunset Yellow FCF and Tartrazine in Sparkling Orangeade An orangeade containing no colouring matter was prepared by diluting 15 ml of the basic syrup to 100ml with distilled water that had previously been carbonated by using dry-ice. A blank polarographic solution was prepared by mixing 5ml of this orangeade, 5ml of 0.01 M tetraphenylphosphonium chloride and 20 ml of Britton - Robinson buffer (pH 1.9), adjusting the pH to 9 with sodium hydroxide solution and diluting to 50ml. An aliquot (20 d) of this solution was pipetted into the polarographic cell, deoxygenated for 10 min and polarographed.When carrying out development work, successive aliquots of con- centrated solutions of Sunset Yellow FCF and tartrazine were added to the blank solution by means of a 100-p1 syringe, the solution being deoxygenated for 1 min after each addition.Aztgust, 1979 DETERMINATION OF FOOD COLOURS IN SOFT DRINKS 725 a -? E Y C 3 0 t 0.5 PA I 11 PA L I I I I -0.90 -0.50 -0.90 -0.50 PotentialN Fig. 1. Effect of tetraphenylphosphonium chloride on the d.p. polarogram of a mixture of Sunset Yellow FCF and tartrazine using the recommended procedure (A) with the addition of tetraphenylphosphonium chloride and (B) without its addition. The effect of adding the tetraphenylphosphonium chloride on the polarograms of a mixture of Sunset Yellow FCF and tartrazine is shown in Fig.1. The effect on the peaks of the individual colouring matters is illustrated by the results given in Table I. On the addition of small amounts of tetraphenylphosphonium chloride, the peak potential of tartrazine is shifted to a more negative potential and the peak current is increased by 140y0. The peak potential of Sunset Yellow FCF remains unchanged but the peak current is decreased by 50%. TABLE I EFFECT OF TETRAPHENYLPHOSPHONIUM CHLORIDE (TPPC) CONCENTRATION ON THE D.P. POLAROGRAPHIC PEAKS OF TARTRAZINE AND SUNSET YELLOW FCF Tartrazine concentration = 0.4 p.p.m. ; Sunset Yellow FCF concentration = 1.84 p.p.m. TPPC concentration, p.p.m. . . . . 0 9 20 40 60 90 100 380* 900 1200 i , ~ (Sunset Yellow F c F ~ ~ ~ A .. 5.4 4.5 3.20 1.60 2.10 2.50 - 2.60 - 2.56 i,? (tartrazine) /PA . . . . 0.70 - 0.67 1.49 1.65 - 1.65 1.68 1.67 - * TPPC concentration used in recommended procedure. 1 E, = - 0.65 V (in presence or absence of TPPC). Calibration graphs for both colouring matters deviate slightly from rectilinearity at higher concentrations owing to adsorption effects at the dropping-mercury electrode but are not affected by the presence of the other colouring matter in the proportions normally found in soft drinks. The latter effect is illustrated in Fig. 2, which shows typical polarograms obtained to produce a calibration graph for tartrazine in the presence of Sunset Yellow FCF. The recommended procedure for the determination of Sunset Yellow FCF and tartrazine in sparkling orangeade is as follows.Pipette 10 ml of 0.01 M tetraphenylphosphonium chloride solution into a 50-ml beaker. Add 5 ml of 0.01 M tetraphenylphosphonium chloride solution and 20 ml of pH 1.9 Britton - Robinson buffer. Adjust the pH to 9 with sodium hydroxide solution and dilute to 50 ml in a calibrated flask. Deoxygenate a portion of this solution in a polarographic cell and polarograph it between -0.3 and -1.0 V. E, = -0.73 V (in absence of TPPC) or -0.80 V (at all TPPC concentrations > O p.p.m.1.726 FOGG AND YO0 : DIRECT DIFFERENTIAL-PULSE POLAROGRAPHIC Analyst, VOZ. 104 The procedure was tested by using a sparkling orangeade (42 p.p.m. of Sunset Yellow FCF and 20 p.p.m. of tartrazine) prepared from the basic syrup and the samples of colouring matter.The result obtained for ten determinations was 41.4 p.p.m. of Sunset Yellow FCF (coefficient of variation = 1.4%) and 20.5 p.p.m. of tartrazine (coefficient of variation = . ~. 1 .O%). I c C B A!. -0.80 -0.50 PotentialN Fig. 2. D.p. polarograms produced in obtaining a cali- bration graph for tartrazine in the presence of Sunset Yellow FCF (1.2 p.p.m. in measured solution). Tartra- zine concentration in mea- sured solution: (A) 0, (B) 1.0, (C) 2.9, (D) 5.1 and (E) 7.5 p.p.m. Determination of Tartrazine and Green S in Sparkling Limeade Tartrazine is best determined under the solution conditions above for sparkling orangeade, viz., at pH 9 in the presence of tetraphenylphosphonium chloride. Under these conditions Green S is reduced at a similar potential to tartrazine but fortunately, because only a small proportion of Green S relative to tartrazine is normally present in sparkling limeade and because the Green S wave is depressed in the presence of the phosphonium salt, interference of Green S in the determination of tartrazine is negligible.Green S is best determined at pH 4 and the addition of tetramethylammonium chloride was found to improve the base line considerably. This improvement seems to arise owing to a partial suppression of the polarographic maximum of the tartrazine. Calibration graphs for Green S in the presence of tartrazine are rectilinear. Tartrazine can be determined after re-adjusting the pH to 9 and adding tetraphenylphosphonium chloride. The recommended procedure for the determination of Green S and tartrazine in sparkling limeade is as follows.Add 5 ml of 1 M tetramethylammonium chloride and 10 ml of pH 1.9 Britton - Robinson buffer and adjust the pH to 4 with sodium hydroxide solution. Dilute to 50 ml in a calibrated flask. Deoxygenate a portion of this solution in a polarographic cell and polarograph it between -0.45 and -0.85 V. Pipette 10 ml of sparkling limeade into a 50-ml beaker.A %gust, 1979 727 Carefully re-adjust the pH of the solution in the cell to 9 with several drops of 4 M sodium hydroxide solution and add 10 mg of tetraphenylphosphonium chloride. Pass nitrogen through the solution to aid dissolution of the solid and to deoxygenate the solution, and polarograph the solution between -0.6 and -1.0 V. The procedure was tested using a sparkling limeade (20 p.p.m.of tartrazine and 2 p.p.m. of Green S) prepared from the basic syrup and the samples of colouring matter. Typical polarograms are shown in Fig. 3, DETERMINATION OF FOOD COLOURS I N SOFT DRINKS 1 / 0.2 1-1 A Green S 1 B 0.5 p A IL I I I I -0.60 -0.20 -0.90 -0.50 PotentialIV Fig. 3. Typical d.p. polarograms of Green S and tartrazine in limeade: (A) pH 4.1 and (B) pH 9.2. The result of ten determinations was 2.1 p.p.m. of Green S (coefficient of variation = 3.4%) and 19.6 p.p.m. of tartrazine (coefficient of variation = 1.9%). At pH 9 the Green S gives a small peak a t -0.75 V that is masked by the tartrazine peak. At the levels of Green S in the limeade this Green S peak is negligible; at very high concentrations of Green S a shoulder appears on the tartrazine peak.Determination of Amaranth and Green S in Blackcurrant Health Drink The sugar in the blackcurrant health drink syrup gives a small peak at pH 4 at the same potential as Green S. At pH > 6 this peak is absent. Amaranth gives a broad polaro- graphic maximum at pH > 6 but this can be suppressed by the addition of tetramethyl- ammonium chloride; tetraphenylphosphonium chloride, tetraethylammonium chloride and Triton X-100 do not suppress the maximum as effectively. The recommended procedure for the determination of amaranth and Green S in black- currant health drink is as follows. Pipette 5 ml of blackcurrant health drink into a 50-ml beaker. Add 5 ml of 1 M tetramethylammonium chloride solution and 10 ml of pH 1.9 Britton - Robinson buffer.Adjust the pH to 7.8 with sodium hydroxide solution and dilute to 50 ml in a calibrated flask. Transfer a portion of this solution to a polarographic cell, deoxygenate it for 10 min and polarograph the solution between -0.30 and -0.80 V. The procedure was tested using a blackcurrant health drink (250 p.p.m. of amaranth and 4 p.p.m. of Green S) prepared from blackcurrant health drink syrup and the samples of colouring matter. Clearly, the Green S is being polarographed near its detection limit for the amount of amaranth present but the procedure can be used as a limit test for this colouring matter. The result of ten determinations for amaranth was 245 p.p.m. with a coefficient of variation of 1.3%. A typical polarogram is shown in Fig.4.728 FOGG AND YOO: DIRECT DIFFERENTIAL-PULSE POLAROGRAPHIC Artalyst, VoZ. 104 A P- 0.2 PA Green S -0.80 -U.30 PotentiaVV Fig. 4. (A) Typical d.p. polaro- gram for Green S and amaranth in blackcurrant health drink. (B) Polarographic peak of 0.6 p.p.m. of Green S in the measured solution in the presence of a smaller pro- portion of amaranth (amaranth concentration = 2.5 p.p.m.). Discussion The established method of identifying food colouring matters is thin-layer chromato- graphy. This excellent method has the advantage of being inexpensive and the disadvantage of being difficult to quantify. Although relatively costly equipment is required, HPLC is now the nearest approach to an ideal method for the identification and determination of food colouring matters, combining efficient separations with precise quantification. Both thin- layer chromatographic and HPLC methods, however, generally require separation of the food colouring matters from even simple food matrices before application to the plate or column.The diff erential-pulse polarographic procedures described here are applied directly to the soft drinks without the need for prior separation of the colouring matters on wool or poly- amide. The differential-pulse polarographic peaks are sharp and afford some measure of identification. A 600-mV scan takes 5 min at 2 mV s-l but a scan rate of 5 mV s-1 might be acceptable for routine use; deoxygenation takes about 10min but can be effected on a batch basis in advance. Therefore, the procedures can be considered for routine use. Micro- processor-controlled polarographs are available that have full sample and data handling facilities after the solution-preparation step. The authors thank Mr. B. A. Saturley of Beechams Products for the provision of samples They also thank Dr. N. T. Crosby of the Laboratory of the and for helpful discussions. Government Chemist and Dr. W. P. Hayes for helpful discussions.August, 1979 DETERMINATION OF FOOD COLOURS IN SOFT DRINKS 729 References 1. 2. 3. 4. 5. 6. 7. 8. 9. “Food and Drugs: Composition and Labelling: The Colouring Matter in Food Regulations 1973,” Macek, K., Editor, “Pharmaceutical Applications of Thin-layer and Paper Chromatography,” Pearson, D., “The Chemical Analysis of Foods,” Seventh Edition, Churchill Livingstone, Edinburgh, Venkataraman, K., Editor, “The Analytical Chemistry of Synthetic Dyes,” John Wiley, New York, Wittmer, D. P., Nuessle, N. O., and Haney, W. G., Jr., Analyt. Chem., 1975, 47, 1422. Knox, J. H., and Laird, G. R., J . Chromat., 1976, 122, 17. Zuman, P., “Organic Polarographic Analysis,” Pergamon Press, Oxford, 1964. Schill, G., Ion Exchange Solvent Extraction, 1974, 6 , 1. Johansson, I. M., Wahlund, K. G., and Schill, G., J . Chromat., 1978, 149, 281. SI 1973 No. 1340, HM Stationery Office, London. Elsevier, Amsterdam, 1972, p. 618. 1976, p. 50. 1977, p. 466. Received November 29th, 1978 Accepted March 23rd, 1979