March, 19661 MOLD, PEYTON, MEANS AND WALKER 189 Determination of Catechol in Cigarette Smoke BY J. D. MOLD, M. P. PEYTON, R. E. MEANS AND T. B. WALKER (Research Department, Liggett and Myers Tobacco Company, Durham, North Carolina, U.S.A .) A procedure has been devised for the specific determination of catechol in cigarette-smoke condensates. This procedure should also be applicable to the determination of catechol in other materials resulting from pyrolytic or combustion processes. As the catechol is isolated without recourse to the formation of a deriva- tive, no interference is encountered as a result of the presence of guaiacol or similar compounds. Avoidance of the use of alkaline conditions throughout the procedure has permitted reproducible and high recoveries. THE presence of catechol in cigarette smoke has been known for many years without any reliable measurements having been reported concerning the amount present. This is princi- pally due to the fact that catechol is rather unstable in alkaline solution, and procedures previously used for its determination in cigarette smoke have utilised an alkaline extraction step to concentrate the acidic materials and remove contaminants.The simple monohydroxyphenolic materials can be determined reliably by gas - liquid chromatography. Application of gas - liquid chromatography to the quantitative deter- mination of catechol was reported by Janak and K0rners.l Recoveries of approximately 70 per cent. were achieved. More recently, von Rudloff2 reported the qualitative separation of catechol from other phenols by using several different gas - liquid chromatographic substrates.Difficulties were noted due to “tailing” or poor elution. Methylation has been used to furnish derivatives of phenols in cigarette smoke for subsequent separation and determinati~n,~ y4 but this is not a quantitative process for catechol, and furthermore the veratrole measured can be derived from guaiacol, which is also present in the smoke. Other workers5 have chromatographed dihydroxybenzene compounds by liquid - liquid partition on silicic acid with cyclohexane as the developing solvent. The amounts of individual phenols were determined by measurement of the optical density at 2800 A. This procedure gave good resolution and appeared well suited for the routine determination of phenolic constituents in samples of similar composition.Disadvantages were the time required and the need for making optical-density measurements on a large number of fractions. The successful use of polyamide surfaces for the qualitative separation of phenolic materials, including cate~hols,~ ?7s8 and their suggested utility for quantitative analysiss prompted us to adapt this technique for the analysis of catechol in cigarette smoke. The procedure that we have developed was designed to avoid exposure of the catechol to alkali. The smoke condensate is dissolved in aqueous acid and ether. Ether-soluble acidic and neutral components are further fractionated by extraction into aqueous borate at pH 7. Catechol and similar vicinal dihydroxy-compounds are selectively complexed by the borate and made soluble in the aqueous phase.Acidification of the aqueous extract permits recovery of these compounds by extraction with ether. Catechol is then separated from other components of this mixture by chromatography on thin layers of polyamide powder with an acidic developing solvent. Since catechol is determined in our procedure without the formation of a derivative such as the methyl ether, no interference from guaiacol is experienced. The method has given reproducible results when applied to cigarette smoke and near theoretical recovery of added known amounts of catechol. EXPERIMENTAL Catechol was unstable when stored in redistilled analytical-reagent grade ether con- taining no preservative, but no loss was observed over a 3-week period when trace amounts of sodium diethyldithiocarbamate were added to the ether.Considerable loss of catechol190 XOLD, PEYTON, MEANS AND WALKER: [Aqzalyst, Vol. 91 did occur if it was extracted from ether into an aqueous alkaline solution and was then recovered from the acidified extract by re-extraction into ether. The losses sustained in this treatment were not as great in the presence of other smoke components. The recovery of catechol was 97 per cent. when subjected to the initial ether extraction used in the present procedure. Similarly, the recovery of catechol in the borate extraction step was 98 per cent. KO significant loss was observed in 0.1 hf sodium borate at pH 7 for periods of time well in excess of that required for the borate extraction.Thin-layer chromatography on polyamide powder from another source gave quite different RF values and less desirable behaviour than did the \Yoelm polyamide powder. The initial use of the upper layer of a benzene - acetic acid - water (6 + 7 + 3) solvent for the thin-layer chromatography gave unsatisfactory results, apparently due to variations in composition of the solvent. BettslO discusses the superior stability of one-phase solvent systems over systems prepared by equilibration of two phases. The solvent mixture that was chosen (100 + 25 + 1) is a single phase and has given reproducible results. Initially, the recovery of catechol from the thin-layer chromatograms was found to vary from 78 to 96 per cent. of the theoretical value. The extent of loss was usually similar for plates prepared at the same time.Volatilisation of catechol from the polyamide film did not appear to be a source of loss, as the yield was essentially the same whether the spot was eluted immediately, or after one hour, after the drying of the plate. Also, when spots of catechol were applied diagonally across the plate and developed in the usual way, recoveries varied from 92 to 96 per cent., with no inhcation of a trend related to the extent of travel of the catechol across the plate. Consistently higher recoveries, 94 to 100 per cent., have been obtained by washing the polyamide powder with ethanol, prior to preparing the slurry. Furthermore, the ethanol- washed polyamide provides layers with less tendency to crack on drying.The recovery of catechol, added to cigarette-smoke condensate and carried through the various steps of the procedure as outlined below, was 95 to 100 per cent. The catechol spot removed from the thin-layer chromatogram of a typical cigarette- smoke extract was found to have ultraviolet absorption identical to that for an authentic sample of catechol, with little background absorption. I t had the same R, value as authentic catechol when chromatographed on borate-treated silica gel with butanol, saturated with 0.1 M sodium borate a t pH 7 , and on Whatman Xo. 1 filter-paper with chloroform (containing 1 per cent. ethanol) which had been equilibrated with methanol - water - formic acid (25 + 24 + 1) as proposed by Reio.ll The same yellow colour was obtained for both known and unknown by spraying with diazotised 9-nitroaniline. The colorimetric procedure used for the quantitative determination of catechol is a modification of the test reported by Mitchell.12 The test is given, in general, by all di- hydric phenols and it is therefore necessary to achieve an adequate separation of catechol from other compounds of this class prior to applying this test.As catechol is the pre- dominant compound of this class, present in the borate extract of cigarette smoke, the separation achieved by this method is usually adequate. Smoke from cigarettes made entirely with flue-cured tobaccos, contains some interfering substances in concentrations that could warrant extending the development of the chromatogram on longer plates. I t is possible to obtain a measure of the total dihydric phenols in the ether extract from the acidified borate solution.The intensity of the colour is comparable to that obtained with catechol for many of these compounds, such as 3-methylcatechol, 4-methylcatechol and protocatechuic acid. Pyrogallol and caffeic acid give 1.3 times the colour intensity observed for catechol. Protocatechualdehyde, coumarin-diols and naphthalene-diols give an orange colour with absorption maximum at about 450mp. These latter compounds, if present, would also contribute to the total colour intensity at 515 to 530mp. Phenol, resorcinol, guaiacol and hydroquinone do not give significant colour production at 515 to 530 mp. Results were erratic when an aliquot of the ether solution containing dihydric phenols was evaporated to dryness before mixing with the colorimetric reagent.This was remedied by adding a portion of the reagent before evaporating off the ether. APPARATUS- METHOD Smoking a$$aratus-That described by Keith and Newsome13 was used. Smoke-coZZection $asks-- JF-6910 (Scientific Glass Apparatus Co. Inc., Bloomfield, N. J.) .March, 19661 DETERMINATION OF CATECHOL I N CIGARETTE SMOKE 191 Apparatus for thin-layer chromatografhy-Desaga-Brinkmann model S-1 1 (Brinkmann Spectrophotometer-Perkin-Elmer model 350 (Perkin-Elmer Corp., Nonvalk, Conn.) . Instruments Inc., Great Neck, N.Y .). REAGENTS- a-CeZlzdose-Solka Floc (Brown Co., Berlin, New Hampshire). Ether-Mallinckrodt anhydrous analytical-reagent grade is redistilled and stabilised by the addition of 0.05 p.p.m.of sodium diethyldithiocarbamate (Distillation Products Industries). Woelm Polyamide powder-(Alupharm Chemicals, Elmont, Long Island, New York). Catechol-Re-sublimed (Aldrich Chemical Co., Milwaukee 10 Wis.) . Ferrous salt solzdion-Dissolve 0.1 g of hydrated ferrous sulphate, FeS0,.7H20, and 0.5 g of sodium potassium tartrate, NaKC,H40,.4H20 (Rochelle salt), in 100 ml of distilled water. This solution is stable for approximately 3 days. Colorimetric reagent-Mix two parts of the ferrous salt solution with five parts of 10 per cent. w/v ammonium acetate and two parts of 0.25 N ammonium hydroxide. This reagent must be freshly prepared for each set of determinations. Cigarettes-Brands A, B, C and D are non-filter-tipped cigarettes (commercially available in the U.S.A.) which contain blended flue-cured (Bright), air-cured (Burley and Maryland) and sun-cured (Turkish) cigarette tobaccos.Brands E, F, G and H are filter-tipped cigarettes (commercially available in the U.S.A.). These cigarettes also contain blends of the three types of tobaccos listed above. Cigarettes I, J and K were prepared to contain a blend of a single type of tobacco, flue-cured, sun-cured or air-cured, typical of that particular component of a commercial blend. The latter cigarettes were cased with a glycol - sugar mixture. SMOKING TECHNIQUE- Cigarettes equilibrated at 60 per cent. relative humidity and 25" C are selected to within k30 mg of the average weight desired. These are smoked mechanically in a room controlled at 25" C and 60 per cent. relative humidity.Forty-ml puffs of 2 seconds duration are taken at intervals of 60 seconds until an average butt of 30 mm remains. The smoke, produced from the sequential smoking of ten cigarettes, is condensed in a trap containing a-cellulose, which is partially immersed in a solid carbon dioxide - acetone coolant. The weight of smoke condensate is determined by the difference in weight of the cellulose trap, at 25" C, before and after smoke collection. A correction is applied for the amount of moisture introduced from the atmosphere during the puffing. To obtain the yield of catechol from the filter-tipped cigarettes without filtration, the filter materials are removed from the mouth-piece before smoking and the cigarettes are smoked to a butt length of 30mm, including the mouthpiece.EXTRACTION OF THE CATECHOL FRACTION- The smoke condensate is removed from the cellulose traps by washing with several portions of ether, 0.1 N hydrochloric acid and acetone, totalling 105, 50 and 5 ml, respectively. These extractants are then combined and equilibrated, the layers are allowed to separate, and the aqueous layer is washed five times with equal volumes of ether. The ether extracts are combined and concentrated at 30" to 34" C in a nitrogen stream to a volume of 25 ml. The ether concentrate is then extracted four times with 25-ml volumes of 0.1 M sodium borate, which has been adjusted to pH 7 with hydrochloric acid. If necessary, the pH of the first borate layer is re-adjusted to pH 7 with cold, dilute sodium hydroxide after contact with the ether concentrate.The borate extracts are cooled and acidified to pH 1 immediately upon collection. Recovery of the catechol from the acidified borate solution is accomplished by extracting four times with 100-ml volumes of ether. After the ether extract has been concentrated to 50 ml in a nitrogen stream a t 30" to 34"C, it is washed twice with 1 ml of water, concen- trated further, and dried with sodium sulphate, in the cold, overnight. The sodium sulphate is filtered from the solution, and the filtrates and rinsings are carefully concentrated in a nitrogen stream to exactly 2 ml. Losses of catechol are observed with complete evaporation of the solvent. The extraction of larger samples of cigarette smoke is accomplished, similarly, by scaling up the extracting volumes.192 MOLD, PEYTON, MEANS AND WALKER: [,4naZyst, 1701, 91 DETERMINATION OF THE TOTAL 0-DIHYDRIC PHENOLS- An aliquot of the smoke fraction containing the o-dihydric phenols, usually representing one cigarette, is mixed with 5 ml of the colorimetric reagent, the vessel is shaken thoroughly, and the ether evaporated in a nitrogen stream.The coloured solution is diluted to 10 ml with the colorimetric reagent and the visible spectrum is obtained with the colorimetric reagent in the reference cell. The absorption maximum, between 515 and 530 mp, is calculated from catechol standardisation data, and the amount of o-dihydric phenols is calculated as if they were catechol. THIN-LAYER CHROMATOGRAPHIC SEPARATIOK OF CATECHOL- Chromatographic plates (20 x 20 cm) are coated with 0.35 mm of Woelm polyamide powder.amide powder in 45 ml of methanol - water (3 + 1). A narrow area along each edge is scraped free of polyamide to promote uniformity of the solvent flow. An aliquot (200 pl) of the catechol extract, equivalent to the smoke condensate from one cigarette, is applied to the lower left corner of each of two plates. A sample of 110 pg of pure catechol is similarly applied to the middle of the lower side of the two plates. After the application solvent has evaporated, development is carried out by ascending chromatography in benzene - acetic acid - water (100 -+ 25 -t- 1). The solvent front usually reaches the top of the plate within an hour. The plates are then removed from the chromatography chamber and the solvents are allowed to evaporate until the plates appear barely dry: extended drying will cause the powder layer to crack.Chromatography is continued by using distilled water to develop the plates in a direction perpendicular to that of the first chromatography. This usually requires about 45 minutes. The plates are partially dried in air, sprayed lightly with the ferrous salt colorimetric reagent, and again partially dried in air. Exposing the damp plate briefly to ammonia vapour will aid in the development of the characteristic purple colour. Both the smoke - catechol spot and the control - catechol spot are marked before the plate has dried completely. After drying thoroughly in air, each spot is removed by suction on to a sintered-glass filter14 and washed into a 10-ml vessel with the ferrous salt colorimetric reagent.COLORIMETRIC DETERMINATION OF CATECHOL- The visible spectrum is measured within 2 hours with the colorimetric reagent in the reference cell. The absorbance is measured a t the maximum, which occurs between 515 and 530 mp. The concentration of catechol in the sample is read off from the curve obtained, by measuring this absorbance for several concentrations of standard catechol. The standardi- sation curve is re-determined for each colorimetric reagent preparation and, at concentrations between 50 and 300 pg per 10 ml, obeys Beer’s law. The value obtained for the smoke - catechol content is corrected for the recovery achieved for control samples of catechol treated in the same manner.Five plates are prepared a t one time by spreading a slurry of 5 g of the poly- RESULTS AND DISCUSSION Catechol analyses for smoke from several nun-filter-tipped cigarettes are given in Table I. The results are expressed both in terms of pg of catechol per cigarette, and as the catechol percentage of condensate. The latter values tend to compensate for variations in the weights of smoke produced due to differences in length or burning characteristics of the cigarettes. The standard deviation calculated for the results presented in Table I is +0.024 per cent. The level of catechol in smoke from cigarettes made with a blend of air-cured (Burley) tobaccos was considerably lower than for smoke from cigarettes made with either flue-cured or sun-cured tobaccos. To evaluate the effect of filtration on the concentration of catechol in cigarette smoke, it is necessary to correct for the moisture content of the smoke.This is required as the filters remove a disproportionate amount of water. When this correction was applied to the analyses performed on smoke from several different brands and types of filter-tipped cigarettes (E, I;, G and H, Table 11), the results indicated no appreciable effect of the filter on the concentration of catechol in the smoke, i.e., the catechol is removed by these filters to an extent equivalent to the removal of the total dry smoke. This would suggest that negligible The reasons for this difference are being investigated.March, 19661 193 DETERMINATION OF CATECHOL IN CIGARETTE SMOKE TABLE I CATECHOL IN SMOKE FROM KON-FILTER-TIPPED CIGARETTES Smoke Catechol content- condensate, r A > Ing per t% Per Type of cigarette cigarette cigarette Cigarette A, 85-mm blended flue-cured, air-cured 48 145 and sun-cured tobaccos 161 49 127 130 Cigarette B, 85-mm blended flue-cured, air-cured 37 113 and sun-cured tobaccos 131 36 118 130 Average .. Average . . Cigarette C, 70-mm blended flue-cured, air-cured 33 92 and sun-cured tobaccos 93 35 102 102 Cigarette D, 70-mm blended flue-cured, air-cured 35 76 and sun-cured tobaccos 95 35 97 96 94 100 53 260 180 54 228 244 65 193 198 52 168 186 39 50 50 39 46 43 Average . . Average . . Average . . Average . . Average . . * Omitted from calculation of average and standard deviation. Cigarette I, 85-mm blend of flue-cured tobacco Cigarette J, 83-mm blend of sun-cured tobacco Cigarette K, 85-mm blend of air-cured tobacco percentage of condensate 0.30 0-33 0.26 0.27 .. 0.29 0.3 1 0.36 0.33 0.36 . . 0.34 0.28 0.28 0.29 0.29 . . 0.28 0.22 0.27 0.28 0.28 0.27 0.29 . . 0.27 0.49 0.34* 0.42 0-45 . . 0.45 0.35 0.36 0.33 0.36 . . 0.35 0.13 0.13 0.12 0.11 . . 0.12 amounts of catechol are present in the vapour phase of the smoke, otherwise its selective removal by the cellulose acetate filters, as is noted for the more volatile monohydric phenols, would be anticipated. TABLE I1 CATECHOL IN SMOKE FROM 85-mm FILTER-TIPPED CIGARETTES Filter intact Filter removed f Dry r smoke, * mg per Brand mm Filter-tipped type cigarette E 17 acetate 26 F 20 acetate 20 G 20 acetate + charcoal 29 H 20 acetate -+ charcoal 21 Catechol 1 percentage of dry PLg smoke 103 0.39 97 0.49 118 0.4 1 92 0.44 r Dry 7 smoke, * ing per cigarette 42 31 44 26 1 Catechol -7 percentage of dry tLg smoke 146 0.35 155 0.50 176 0-40 116 0.45 * The smoke weights are corrected to a dry basis by using moisture values from duplicate smoke collections determined by Karl Fischer titrations.194 MOLD, PEYTON, MEANS AND WALKER [Analyst, Vol.91 The colorimetric test for o-dihydric phenols, applied to the borate-extractable fraction of cigarette smoke, prior to chromatography on polyamide, and calculated as catechol, indicated a total content for these materials about twice the value of that measured for catechol (see Table 111). Smoke samples used for these analyses were prepared under the supervision of D. H. Woods. TABLE I11 TOTAL O-DIHYDRIC PHENOLS IN SMOKE FROM NON-FILTER-TIPPED CIGARETTES Studies are in progress to identify these materials. Smoke o-Dihydric phenols (as catechol) fng Per r-lg Per percentage of condensate, r h \ cigarette cigarette condensate Cigarette A, 85-mm blended flue-cured, air-cured and sun-cured tobaccos . . . . . . 49 336 0.69 Cigarette J, 85-mm blend of sun-cured tobacco 53 286 0.54 Cigarette K, 85-mm blend of air-cured tobacco 39 107 0.27 Cigarette I, 85-mm blend of flue-cured tobacco 54 526 0.98 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. REFERENCES JanAk, J., and Komers, R., Colln Czech. Chem. Cornmun., Engl. Edn, 1959, 24, 1960. von Rudloff, E., J . Gas Chromat., 1964, 2, 89. Commins, B. T., and Lindsey, A. J., Analytica Chim. Acta, 1956, 15, 557. Carruthers, W., and Johnstone, R. A. W., Nature, 1960, 185, 762. Young, J. H., Analyst, 1961, 86, 520. Endres, H., and Hormann, H., Angew. Chern., Int. Edn, 1963, 2, 254. Martin, W. N., and Husband, R. M., Analyt. Chem., 1961, 33, 840. Halmekoski, J., and Hannikainen, H., Suornen. Kern., 1963, B36, 24. Gasparic, J., Petranek, J., and Borecky, J., J . Chromat., 1961, 5, 408. Betts, T. J., J . Pharm. Sci., 1964, 53, 794. Reio, L., J . Chrornat., 1968, 1, 338. Mitchell, C. A., Analyst, 1923, 48, 2. Keith, C. H., and Newsome, J. R., Tob. Sci., 1957, 1, 51. Matthews, J. S., Pereda, V., A. L., and Aguilera, P., A., J. Chrornat., 1962, 9, 331. Received A$ril 12th, 1965