36 Analyst, January, 1967, Vol. 92, $9. 3642 A Thin-layer Chromatographic Method Determination of “Quassin” in Cosmetic for the Preparations BY E. C. HUNT (Ministry of Technology, Laboratory of the Government Chemist, Cornwall House, Stamford Street, London, S.E. 1) A method is described by which “quassin” may be determined when used as a denaturant for the alcohol contained in cosmetic preparations. The two bitter principles, neoquassin and quassin are determined together as quassin, after oxidation with a solution of sodium dichromate in glacial acetic acid. Perfume oils are first removed by solvent extraction and the aqueous phase is then evaporated to dryness and the residue oxidised. Further solvent-extraction steps yield a solution of quassin in chloroform which is applied to a thin-layer chromatographic plate.The spots are compared visually with standards prepared from an oxidised “quassin” solution to permit determinations within the range of 5 to 36 p.p.m. in samples. “QUASSIN,” extracted from the wood of the Simarubaceae family of trees, has been used for many years as a denaturant for the alcohol used in cosmetic preparations. It is odourless and intensely bitter. The concentration required in Great Britain for the partial denaturing of industrial methylated spirit preparations is about 20 to 30 p.p.m. The chemical structures of the two bitter principles of “quassin,” vix., quassin (I) and neoquassin (11), have recently been elucidated,l y 2 93 and are represented below. O.CH, Nothing appears to have been published on the analytical chemistry of “quassin.” The term “quassin” used in this paper refers to the commercial material which generally contains quassin, neoquassin and other substances.The efficacy of “quassin” when used as a denaturant is judged by the extent to which it produces an unpalatable bitterness, but it is desirable that this effect of bitterness should be related to a chemical assessment of the amount of “quassin” present. Two reactions appeared potentially useful. In the first, the lactone ring of quassin opens in alkaline solution4g5 to give a compound no longer extractable by, for example, chloroform, whereas neoquassin is unaffected by alkalis; quassin can be recovered on acidifying the solution. Secondly, neoquassin is oxidised to quassin by a solution of sodium dichromate in glacial acetic acid.496 Thin-layer chromatography was also considered as a possible means of isolating the bitter principles.This technique has been used by other worker^^,^ for the identification of quassin and related substances. Pure specimens of quassin and neoquassin were prepared from commercial “quassin”4Jj and were found to have about the same degree of bitterness. Four “quassin” samples were each shown by thin-layer chromatography to contain a total of about 90 per cent. of bitter material. Three of these samples, which were of British manufacture and presumably ex- tracted from the wood of Picrasma excelsa, as specified in the British Pharmacopoeia (1953), contained a preponderance of neoquassin. The fourth sample was of French manufacture from the wood of Quassia amara and consisted predominantly of quassin.HUNT 37 The method studied for the examination of perfume samples involved extraction of the sample to remove essential oils, and oxidation of the residue obtained by evaporating the aqueous phase, followed by solvent extraction.The extracts were subjected to thin-layer chromatography and the quassin spots from the sample compared visually with those from standard solutions. EXPERIMENTAL Quassin is soluble in chloroform, ethyl acetate and benzene, and sparingly soluble in ether or light petr~leum.~ Aqueous solutions of quassin were extracted separately with carbon tetrachloride, ether and light petroleum (b.p. 40” to 60” C); only light petroleum failed to extract a significant amount of quassin and this solvent was therefore chosen for the extraction of the essential oils.Normally, a 5-ml volume of perfume would be diluted with water before extraction and the aqueous phase afterwards evaporated to dryness. OXIDATION- The oxidant used was a 10 per cent. w/v solution of sodium dichromate in glacial acetic acid,6 with a reaction time of 45 minutes at 25” C, or 1 hour at 20” C. Pure quassin was unaffected and could be recovered completely, but neoquassin invariably suffered a loss of about 15 per cent. This loss was not diminished by reducing the sodium dichromate concen- tration to 2 per cent. w/v, nor by changing the time or temperature of reaction. Oxidation was incomplete when the oxidant was diluted with an equal volume of N sulphuric acid, and when the glacial acetic acid was replaced by 80 per cent.v/v acetic acid. A solution of alkaline hydrogen peroxide in acetone had no effect on neoquassin. Potassium permanganate, whether in acetone solution or in alkaline aqueous solution, left no detectable amounts of either neoquassin or quassin. Solutions containing about 1 per cent. w/v of “quassin” in industrial methylated spirit are available commercially and are similar to those used for denaturing purposes. Five such solutions were each diluted with ethanol to give a concentration of 1 mg of “quassin” per ml (nominal) and an aliquot (0-1 m1) of each solution was evaporated to dryness; each residue was dissolved in 0.25 ml of chloroform and aliquots of the chloroform solutions were chromato- graphed as described in the Method.The spots that developed after spraying with potassium permanganate solution were assessed visually by comparison with spots of quassin and neoquassin standards on the same plate. Further 0-1-ml aliquots of the solutions containing 1 mg per ml of “quassin” were evaporated to dryness and oxidised with sodium dichromate solution as in the Method. Each sample was then diluted with 50 ml of water and extracted with chloroform as described, but omitting the pH adjustments and the preliminary extraction with carbon tetrachloride. Aliquots of the final chloroform solutions were chromatographed as before, and the developed spots compared visually with spots of quassin standards on the same plate. The recoveries were reasonably consistent at 70 per cent.and warranted the use of oxidised “quassin” solution for chromatographic standards to compensate for oxidation losses when the method was applied to cosmetic samples. The oxidation step was valuable in that it decomposed perfume residues remaining after the initial extraction and thus reduced possible interference effects on the chromatogram. Neither the extraction nor the oxidation alone gave an adequate degree of purification for chromatography . EXTRACTION OF QUASSIN- If the oxidised sample is made alkaline with sodium hydroxide solution, quassin is converted to sodium quassinate4 and more impurities may be removed by extraction with carbon tetrachloride. Chloroform was found to extract a portion of the quassin despite the presence of an excess of sodium hydroxide.The addition of dilute acid decomposed the sodium quassinate with the regeneration of quassin, which could then be extracted by chloroform. In some experiments chromatographic spots were found which corresponded with neither neoquassin nor quassin. They were probably caused by hydrolysis products that resulted from temperature rises during pH adjustments; the action of heat on alkaline solutions of quassin is to cause hydrolysis with the formation of a stable a ~ i d . ~ $ ~ s ~ Subse- quently, to minimise temperature rises, acid and alkali solutions were cooled in an ice-bath before use, which eliminated the appearance of spurious spots.38 HUNT : THIN-LAYER CHROMATOGRAPHIC METHOD [Analyst, VOl. 92 CHROMATOGRAPHY- Silica gel used as adsorbent on thin-layer plates was found to give better separations of neoquassin and quassin than alumina or cellulose because of its stronger adsorptive properties.Many solvent combinations were examined for developing the chromatograms on silica gel ; the best of these were found to be- A. Approximate R F values: neo- B. Chloroform - methanol (90 + 10). Approximate RF values: neoquassin 0.65; This solvent has been used in a study of extracts of the wood of Picrasma exceZsa.8 The chromatograms are illustrated in Fig. 1. The order of elution was different in these two systems, a fact that was useful in confirming the identity of doubtful spots from cosmetic samples. Both neoquassin and quassin were strongly adsorbed from chloroform solution a t the point of application on the plate so that the volume of the solution applied was not important unless other strongly adsorbed materials were also present. Solvent mixture A was generally preferred because, with cosmetic samples, most of the foreign matter was carried beyond the quassin spot.With solvent mixture B, streaks were frequently found over most of the chromatogram although the quassin spots were usually clear and capable of assessment. Several spray reagents were examined, but none was found to give a selective reaction with quassin and neoquassin. The simplest method of spot location was to incorporate an inorganic additive in the silica gel, which caused it to fluoresce when irradiated with ultraviolet light (wavelength 254 mp). Neoquassin and quassin each absorb light strongly at this wavelength and show as dark spots.A similar effect may be obtained by spraying an ordinary silica gel plate with rhodamine B solution in ethanol and examining the absorption of fluorescence at 254 mp. Alternatively, the plate can be sprayed with 0.5 per cent. potas- sium permanganate solution, dried at 60” C for 10 minutes, and then the excess permanganate eluted with waterg; neoquassin and quassin leave brown spots on a white to pale buff back- ground. Other spray reagents that reacted with the two bitter materials were concentrated sulphuric acid and a solution of molybdophosphoric acid in ethanol, after heating at 110” C in both instances. Ammoniacal silver nitrate reacted with neoquassin, but not quassin, after heating at 110” C. The fluorescence - absorption method and the potassium perman- ganate method were considered to be best, each giving a well graded series of standard spots from 1 to 5 pg with both neoquassin and quassin; the gradation from 5 to 8 pg was slightly poorer.The fluorescence - absorption method showed less foreign matter from cosmetic samples than the potassium permanganate method, but the latter was considered better for spot matching . PREPARATION OF STANDARDS- Pure neoquassin and quassin were separated from commercial “quassin” and purified by published meth0ds.l 94 Neoquassin was separated by crystallisation from methanolic potassium hydroxide solution diluted with 3 volumes of water. It was purified by treatment with charcoal, followed by a second crystallisation, and finally, by crystallisation from dilute methanol (1 + 3).The product melted at 228” C and showed no impurities by thin-layer chromatography. Quassin was recovered from the mother liquor by saturating it with carbon dioxide, and purified by chromatography on an alumina column with chloroform as eluent. The alumina was mixed with an inorganic fluorescent agent (Woelm, 2 per cent. w/w) and packed in a fused silica tube. The quassin could be seen moving as a dark band in ultraviolet light (254 mp). Subsequent treatment with charcoal and crystallisation from dilute methanol (1 + 3) yielded crystals that showed no impurity by thin-layer chromatography, and that melted at 222” C. Because of the low content of quassin in most commercial “quassins,” a second sample of quassin was prepared, after oxidation of the “quassin” with sodium dichromate in acetic acid solution4 to convert neoquassin to quassin.Preparative-layer chromatography was also found to be feasible on plates coated with a 2-mm layer of fluorescent silica gel (Merck kieselgel HF,,,, without binder). Five successive developments were neces- sary in a solvent mixture of chloroform and methanol (95 + 5), with a loading of about 5 mg of “quassin” per cm width of plate. Standard solutions for chromatography were prepared by Ether - glacial acetic acid - water (65 + 25 + 10). quassin 0.75 ; quassin 0.65. quassin 0-75.I --r - - - 1 (4 (4 Fig. 1. Thin-layer chromatograms run in : (a) and (c), ether - glacial acetic acid -water (65 + 25 + 10 v/v) ; (b) and ( d ) , chloroform - methanol (90 + 10 v/v).Q = quassin, NQ = neoquassin, S, and S2 are 10 and 15 p1, respectively, of a perfume extract. Plates (a) and (b) are photographed in ultraviolet light (254 mp), and plates (c) and (d) after spraying with potassium permanganate solution and eluting the excess of reagent [To face$. 38January, 19671 FOR THE DETERMINATION OF “QUASSIN” 39 dissolving 10mg of pure quassin or neoquassin in 10ml of chloroform. Ethanol was used as solvent in standard solutions to be added to cosmetic samples for recovery tests. Standard quassin solutions for routine analysis are prepared from oxidised “quassin.” The “quassin” should be the commercially pure material which normally contains about 90 per cent. of total quassin and neoquassin, and is white to pale yellow in colour; solutions should be colourless to straw coloured.Some foreign “quassin” powders have been found to be diluted and to contain 10 per cent., or less, of total quassin and neoquassin. A 1 per cent. w/v ethanolic solution of “quassin” is standardised by diluting with water to 10 p.p.m. of “quassin” and measuring the absorption on a spectrophotometer against a 1 per cent. v/v ethanol blank. Correction for absorption by impurities may be made by use of a method similar to that used by Morton and StubbslO in the determination of vitamin A. At the peak wavelength of 257.5 mp which is used in this correction, the optical density of neoquassin is about 4 per cent. higher than that of quassin. This type of correction has been criticised by Wilkiell who proposed measurements based upon the difference in absorption at two wavelengths on the slope of the spectral curve.Neither type of correction shows any marked advantage over the other when applied to “quassin” solutions, so the Morton and Stubbs’ type was adopted as it has been used for several years in conjunction with potability tests. METHOD PREPARATION OF STANDARD QUASSIN SOLUTION Evaporate a 5-ml portion of the “quassin” solution (1 per cent. w/v) just to dryness on a water-bath. Add 5 ml of ethanol and again evaporate to dryness. Dissolve the residue in ethanol to make 50 ml of solution. Dilute 10 ml of this solution to 1 litre with water and measure the absorption at 245, 257.5 and 270 mp in a l-cm cell against a 1 per cent. v/v ethanol solution.Calculate the corrected value for Eiik as follows- E:& at 257.5 mp = a ; at 245 mp = b ; and a t 270 mp = c. The correct Ei& = 2.148 [2a - (b + c)]. E;iL for crystalline “quassin” at 257-5 mp is 302. loo, per cent. corrected E:& 302 “Quassin” content of sample = Calculate the volume of solution required to yield 1 mg of quassin from the known con- centration of “quassin”; make a correction for the loss of 30 per cent. found after oxidation, as shown in Table I. Evaporate this volume to dryness and oxidise the residue with 2 ml of the 10 per cent. w/v sodium dichromate solution as described for the extraction of quassin. Dilute the solution with 50 ml of water and, omitting the pH adjustments and carbon tetra- chloride extraction, extract the quassin with chloroform as described below.Evaporate the chloroform to dryness and dissolve the residue in exactly 1 ml of chloroform to give a quassin concentration of 1 pg per p1. TABLE I RECOVERY OF QUASSIN FROM “QUASSIN” AFTER OXIDATION Quassin Sample found, Pg 1 25 2 75 3 33 4 33 5 25 Neoquassin found, Total, Pg Pg 62 87 41 116 50 83 50 83 58 83 Quassin found after oxidation, Recovery, Pg per cent. 60 69 83 71 58 70 58 70 58 70 EXTRACTION OF QUASSIN REAGENTS- Light petroleum (b.p. 40” to 60” C). Carbon tetrachloride-Use analytical-reagent grade Chloroform-Use analytical-reagent grade.40 HUNT : THIN-LAYER CHROMATOGRAPHIC METHOD [Analyst, VOl. 92 Sodium hydroxide, N. Sulphuric acid, N. Sodium dichromate solution-Dissolve 1 g of sodium dichromate dihydrate in 10 ml of glacial acetic acid.PROCEDURE- Place 50 ml of water in a 100-ml separating funnel, add 5 ml of sample, and mix. Add 20 ml of light petroleum (b.p. 40" to 60" C) and shake the funnel for 1 minute. Transfer the lower aqueous phase to a second 100-ml separating funnel. Wash the walls of the first separating funnel with a few millilitres of water and transfer the washings to the second separating funnel. Shake the aqueous phase again with 20 ml of light petroleum and run the lower layer into a 250-ml beaker. Wash the walls of the separating funnel with a few millilitres of water and run this into the beaker. Evaporate the aqueous solution to dryness on a steam-bath and cool. Add 2 ml of sodium dichromate solution, cover the beaker and allow it to stand for 45 minutes at 25" C, or 1 hour at 20" C.Add 40 ml of N sodium hydroxide solution that has been cooled in anice-bath, to give a yellow solution showing astrongly alkaline reaction to pH test-paper, and transfer the solution to a 100-ml separating funnel, Shake it twice with 20-ml portions of carbon tetrachloride for 1 minute each and discard the lower layers. Add 10ml of N sulphuric acid that has been cooled in an ice-bath, to produce a colour change to the orange dichromate; the solution is buffered at about pH 7 by the sodium acetate present. Shake the solution with three 20-ml portions of chloroform for 1 minute each, and run each lower layer in turn through a dry filter-paper into a 250-ml beaker. Evaporate the combined extracts to dryness, dissolve the residue in 1 ml of chloroform, and transfer to a glass-stoppered test-tube, about 100 mm long and 15 mm in diameter.Wash the beaker successively with two 1-ml portions of chloroform and add them to the test-tube. Evaporate the contents of the tube carefully to dryness on a steam-bath, cool, and dissolve the residue in exactly 0.25 ml of chloroform. This is the sample extract ready for thin-layer chromatography. THIN-LAYER CHROMATOGRAPHY REAGENTS- Silica gel-Kieselgel GF,,, (E. Merck and Co.). Standard solation of quasszn, 1 pg pey p l in chlovofoym-(see "Preparation of standard Spray reagent-Dissolve 0.5 g of potassium permanganate in 100 ml of water. Developiizg solvent-Mix 65 ml of diethyl ether. 25 ml of glacial acetic acid and 10 ml quassin solution"). of water. PROCEDURE- Spread the mixture im- mediately over a row of 5 plates, 20 x 20 cm, to a thickness of 0.25 mm.Allow the plates to dry in air, and activate them by heating in an oven at 120" C for 1 hour. Store them until required over silica gel in a desiccator; plates should be re-activated daily before use. Line the inside walls of a chromatographic tank, about 22 cm long by 9 cm wide and 20 cm high, with a sheet of filter-paper and add 100 ml of developing solvent 30 minutes before use. Apply with a capillary micro-pipette, spots of 10 and 15 pl of the sample extract, in 5-pl aliquots, about 1.5 cm above the bottom of the plate. Allow each aliquot to dry before applying the next. Also apply spots of standard quassin solution to the plate to contain 1, 2, 3, 4 and 5 pg of quassin.Place the plate in the tank and allow the solvent to rise about 10 cm; this requires about 45 minutes. Remove the plate, allow it to dry in air and then in an oven at 120" C for 10 minutes. Allow the plate to cool and compare the sample spots visually with the standards under ultraviolet irradiation at 254 mp. The spots are dark against a green fluorescent background. Glass or plastic goggles, or ordinary spectacles should be worn to protect the eyes from the harmful effects of the ultraviolet light. The spots may be assessed to the nearest 0.5 pg of quassin. Alternatively, spray the dried plate evenly with 0.5 per cent. w/v potassium permanganate solution until visibly damp. Warm the plate in an oven at 60" C until just dry (about 10 minutes), and then place it in a tank containing a 1-cm layer of water.Allow the water to rise until the front has carried the Shake 30g of silica gel with 60ml of water for 2 minutes.January, 19671 FOR THE DETERMINATION OF “QUASSIN” 41 excess of potassium permanganate past the developed spots, and dry the plate in an oven at 120” C. The silica-gel layer can be protected by covering with a clean glass plate. Multiply the assessed quassin value by 1.4 to obtain the equivalent amount of “quassin.” RESULTS AND DISCUSSION “Quassin” is generally added to cosmetic preparations of British manufacture as one of the denaturants when alcoholic formulations are used; the concentration of “quassin” required for denaturing may vary according to the different regulations governing different types of products, and may also vary with the alcohol content.The samples investigated during the development of this method were expected to contain 15 to 20 p.p.m. of bitter material, expressed as quassin after allowance had been made for alcohol content and for the purity of the “quassin.” Commercial “quassins” have usually been found to contain not more than 90 per cent. w/v of total quassin and neoquassin. The method described was applied to some typical cosmetic samples and to perfumes denatured in the laboratory by known additions of “quassin.” The results are shown in Table 11. TABLE I1 DETERMINATION OF “QUASSIN” IN COSMETIC PREPARATIONS “Quassin” present, ‘ Sample p.p.m. Hair lacquer . . . . .. 20 Eau de Cologne I . . . . . . Eau de Cologne 11.. .. .. - - - Toilet water I . . . . . . Toilet water I1 . . .. . . 17 Perfume I . . .. .. . . Perfume 11.. . . .. . . Synthetic perfume I . . .. 15 - - Synthetic perfume I1 . . , . 15 Synthetic perfume I11 . . . . 12 ‘Quassin” found, p.p.m. 22 16 16 1 s 1 s 15 18 16 16 11 The synthetic perfumes were compounded in this laboratory from a sample of mixed perfume oils. The values shown in column 3 are theoretical “quassin” contents based on known alcohol contents in these samples. The other samples would be expected to contain from 15 to 20 p.p.m. of “quassin,” depending upon alcohol content. Some further samples gave results lower than the accepted minimum. These were analysed again after the addition of standard “quassin” solution to test for completeness of recovery.TABLE I11 RECOVERY OF “QUASSIN” FROM COSMETIC SAMPLES “Quassin” found p.p.m. Toilet water . . 12 After-shave lotion I 1 1 , l l After-shave lotion I1 Hair tonic . . . . 1 1 , l l Perfume . . . . 1 1 , l l Sample in sample, ‘ 12, 12, 11 Total “quassin” ‘Quassin” added, iound, p.p.m. p.p.m. 17.5 29 17.5 25 17-5 29 17.5 29 17.5 27 Recovery of added “quassin,” per cent. 9s 88 98 102 95 The recoveries of added “quassin,” shown in Table 111, confirmed that the original samples were low in “quassin” and that no appreciable losses had occurred that could be attributed to failure of the method. However, two cosmetic preparations failed to yield results by the method described. One of these was a beer hair-setting lotion, which left a gummy residue on evaporation of the aqueous solution after the initial extraction with light petroleum.This residue did not dissolve in the oxidising solution and little oxidation occurred. Heavy interference was found on chromatograms in each of the two developing solvents. The other sample was a bath essence that contained 55 per cent. of detergent that was probably anionic. Extractions with both light petroleum and carbon tetrachloride led to emulsions which had to be spun in a centrifuge to recover the aqueous phase. The chloroform extraction also emulsified, the chloroform phase forming a gel.42 HUNT The lower limit of determination by the method described is 1 pg of quassin; this is equivalent to 1.4 pg of the standardised “quassin,” or 5 p.p.m., on a 5-ml sample. Lower concentrations of “quassin” may be determined for most samples by applying larger aliquots to the chromatographic plate; by using less chloroform to dissolve the final extract residue; or by taking a larger initial sample.The accuracy of the method is governed by the accuracy of the spot matching. Spots are usually assessed to the nearest 0.5pg, which limits the maximum error to 10 per cent. on the 5-pug standard and to 50 per cent. on the l-pg standard. Sample aliquots are generally chosen so that matching with the upper end of the standard range is possible. The accuracy of the method was found to be adequate for the purpose of this investigation. “Quassin” solutions may be examined directly by chromatography after evaporation and dissolution of the residue in chloroform; the sample spots are compared with standard ranges of both quassin and neoquassin. The author is indebted to the Government Chemist for permission to publish this paper. The provision of samples of quassin and neoquassin by Dr. J. S. E. Holker of Liverpool University, and of “quassin” by William Ransom and Sons Ltd., and Stafford Allen and Sons Ltd., is gratefully acknowledged. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. REFERENCES Valenta, Z., Papadopoulos, S., and Podesva, C., Tetrahedron, 1961, 15, 100. Valenta, Z., Gray, A. H., Orr, D. E., Papadopoulos, S., and Podesva, C., Ibid., 1962, 18, 1433. Carman, R. M., and Ward, A. D., Aust. J . Chem., 1962, 15, 807. London, E., Robertson, A., and Worthington, H., J . Chew. Soc., 3431, 1950. Hanson, K. R., Jaquiss, D. B., Lamberton, J. A., Robertson, A., and Savige, W. E., Ibid., 4238, Adams, R., and Whaley, W. M., J . Amer. Chem. Soc., 1950, 72, 375. Lavie, D., and Kaye, I. A., J . Chew. Soc., 1963, 5001. Stahl, E., Editor, “Thin-layer Chromatography,” Academic Press Inc., New York and London, Mansfield, R. C., and Locke, J . E., J . Amev. Oil Chem. Soc., 1964, 41, 267. Morton, R. A., and Stubbs, A. L., Analyst, 1946, 71, 348. Wilkie, J. B., Analyt. Chem., 1964, 36, 896. 1954. N.B.-These authors refer to quassin as “isoquassin,” and use the term “quassin” for a mixture of quassin and neoquassin. 1965, p. 387. Received April 29tJa) 1966