442 BENTLEY AND BURGAN : POLYNUCLEAR HYDROCARBOBS IN [Vol. 83 Polynuclear Hydrocarbons in Tobacco and Tobacco Smoke Part I. 3 : 4 - Benzopyrene BY H. R. BENTLEY AND J. G. BURGAN (The Imperial Tobacco Comeany Ltd., Research Department, Raleigh Road, Bristol, 3) A method is described for determining the concentration of 3 : 4-benzo- pyrene in tobacco and in tobacco-smoke condensate by using the fluorescence spectrum of the hydrocarbon. IT has been suspected for many years, largely by analogy with coal tar, that the water- insoluble “tar” forming the condensable fraction of tobacco smoke would be found to contain polynuclear aromatic hydrocarbons. An essential difference, however, between normal tobacco smoke and coal tar is that, whereas the latter is the by-product of a carbonisation process occurring in the absence of air, tobacco smoke contains the products of combustion of leaf constituents and of distillation of volatile substances in the presence of air.Considerable amounts of leaf constituents are also carried over into the smoke by a process of steam-distillation and entrainment, and, since these substances are removed rapidly from the burning zone, they are to a large extent chemically unaltered. Polynuclear aromatic hydrocarbons are thought to be formed during pyrolysis of many substances in a restricted supply of air between about 750” and 1600” C by a process involving first a breakdown into methylene radicles and hydr0gen.l The methylene radicles dimerise t o ethylene, which breaks down further to hydrogen and “nascent acetylene.” This partly decomposes to carbon, hydrogen and methane and partly polymerises to polynuclear aromatic hydrocarbons with further loss of hydrogen. A mechanism of this sort may explain the formation of carcinogenic tars during the carbonisation in the absence of air of the organic materials that were tested by K e n n a ~ a y ~ ? ~ and Kennaway and Samp~on.~ These carbonisa- tion conditions are, however, not representative of those in the burning zone of a cigarette, and normal tobacco smoke is, a priori, unlikely to contain appreciable amounts of polynuclear aromatic hydrocarbons.This may not be so in tars produced by heating tobacco in closed ~ t i l l s , ~ ~ 6 ~ 7 ~ 8 ~ ~ which may be expected to resemble carbonisation tars more closely, and this essential difference was noted by some of the earlier workerslOJ1 and also in more recent times.12 Because of its important biological effects, and also because of the comparative ease with which it can be recognised in mixtures, attention has been directed first to 3:4-benzo- pyrene.However, many workers who analysed tobacco tar by methods applicable to coal tar failed to detect the h y d r o c a r b o n . l l ~ ~ ~ J ~ ~ ~ ~ An early report by Roffos that tobacco tar contained 3 : 4-benzopyrene arose by a confusion of nomenclature. The hydrocarbon for which spectroscopic evidence was obtained by Roffo was, in fact, 1 : 2-benzopyrene, which he believed t o be the highly carcinogenic isomer. It is now evident that one of the reasons for the earlier failures to detect 3 : 4-benzopyrene in tobacco tar was that the order of magnitude of the concentration present was not appre- ciated.It has recently become clear that the smoke from cigarettes smoked under conditions simulating the human habit does in fact contain small amounts of 3 : 4-benzopyrene. There is, however, little agreement among workers in this field as to the amounts present. For example, Cooper and Lindseyx6 give a figure of about 1 pg per 100 cigarettes smoked, equivalent to 0.2 p.p.m. by weight of condensable material; Wynderl’ gives the concentration as 2 p.p.m. by weight of condensable matter; Alvord and Cardonls found a range of 8 to 18 pg per 100 cigarettes smoked, equivalent to 1.6 to 3.6 p.p.m. by weight of condensable matter; Latarjet, Cuzin, Hubert-Habart, Muel and Royerl9 found 1.2 pg per 100 cigarettes smoked, equivalent to 0.2 p.p.m.by weight of condensable matter: and Bonnet and Neukomm20 found 2-2 pg per 100 cigarettes smoked, equivalent to about 0.4 p.p.m. by weight of condensable matter. Differences in smoking technique are probably responsible for some of these variations. In the work now reported, cigarettes have been smoked under conditions that are thought to resemble most closely those of the human habit.2IAugust, 19581 TOBACCO AND TOBACCO SMOKE. PART I 443 In our experience, which we believe to be general, it has not yet been possible to isolate pure polynuclear hydrocarbons from cigarette-smoke condensate. Even after extensive chromatographic separation, the 3 : 4-benzopyrene-containing fractions contain a large excess of extraneous material, which contributes an intense background absorption in the 300 to 400-mp region and an intense fluorescence a t a somewhat longer wavelength.The typical appearance of the absorption spectrum of a purified 3 : 4-benzopyrene-containing fraction is shown in Fig. 1. In this, the absorption peaks a t 365 and 385 mp, which are characteristic of the hydrocarbon, can be seen as small inflexions against a generalised background absorp- tion. A method for determining substances by the heights of absorption peaks under these conditions is that of Morton and Stubbs,22 which has been applied to the determination of vitamin A in fish oils and of anthracene in petroleum. The small and indeterminate peaks obtained from purified cigarette-smoke fractions, however, makeit impossible to use this method for 3 : 4-benzopyrene.For example, a requirement of the method of Morton and Stubbs is that linear irrelevant absorption must be assumed to be present over the region of the peak used in the determination. This assumption cannot be made for certain fractions of tobacco tar. As will be shown later, the analytical method finally adopted depends on the use of fractions that show the characteristic banded fluorescence spectrum of 3 : 4-benzopyrene, and these fractions invariably also show small inflexions in the absorption spectrum charac- teristic of 3 : 4-benzopyrene. On the other hand, fractions closely adjacent to these on the chromatogram have been found to show apparently typical 3 : 4-benzopyrene inflexions without displaying the characteristic fluorescence bands of the hydrocarbon.The possibility that the absence of typical fluorescence in these fractions might be due to quenching was excluded by the addition of small amounts of pure 3 : 4-benzopyrene; the characteristic bands then appeared with the correct intensity. It cannot, therefore, be assumed that the back- ground absorption is linear for the purpose of measuring peak heights. I I 350 400 Wavelength, mp Fig. 1. Absorption spectrum of purified 3: 4- The values reported hitherto in the literature for the 3 : 4-benzopyrene content of tobacco smoke have all been found by the absorption method. Because they show considerable discrepancies, it was thought that it would be useful to develop, for comparison, an analytical method based on an entirely different principle.The proposed method, based on fluorescence, has therefore been devised. The application of the method to some problems of current interest is shown by the results in Table I. For mixed cigarettes representative of those that have a large sale in the United Kingdom, a large number of replicate determinations permits the reproducibility of the method to be assessed. For the twenty-five results listed in the first section of Table I, the mean maximum 3 : 4-benzopyrene content per 500 g of cigarettes is 4.9 pg, with a range 1.5 to 8.0 pg. This concentration is equivalent to about 0.2 p.p.m. by weight of condensable matter and agrees with the results of Cooper and Lindsey,16 Waller15 and Latarjet, Cuzin, Hubert-Habart, Muel and Royer.19 The concentrations found by other workers, which are considerably higher than this, are therefore not typical of the cigarettes on sale in the United Kingdom smoked under our conditions.benzopyrene-containing fraction444 BEiVTLEY AND BURGAN : POLYNCCLEAR HYDROCARBONS IN [Vol. 83 The results listed in Table I also show that there is no difference in the 3 : 4-benzopyrene content of the smoke from cigarettes containing only American or Rhodesian tobacco, although the chemical composition of these two types of leaf differs appreciably. Also, there is no more 3 : 4-benzopyrene in the smoke from cigarettes made entirely of cut tobacco stems, which have a very high cellulose and lignin content and are practically free from alkaloids.TABLE I DETERMINATION OF 3 : 4-BENZOPYRENE IN TOBACCO SMOKE, LEAF AND STEM Source Naterial analysed Mixed cigarettes typical of current United Tobacco . . production American cigarettes . . . . .. Rhodesian cigarettes . . . . . . . . Cigarettes made from tobacco stems Smoke . . Tobacco . . Smoke .. Stem . . * . Maximum 3 : 4-benzopyrene content per 500 g of 6.0 3.0 8.0 1.5 5.0 2.0 4.0 1.5 3.0 6.0 8.0 7.0 5.5 3.0 4.0 2.7 7.0 3.0 5.0 3.0 6.0 7.0 7.0 8.0 7.0 2.0 2.75 5.0 5.0 2.5 3.0 2.0 source, pg 3.0 3.5 6.0 2.5 2.0 3.0 2.5 5.0 4.0 4.0 5.0 Not 0.2 found The figures for unburnt tobacco and tobacco stem show that after normal curing and manufacture these materials, as might be expected, have become contaminated with 3 : 4- benzopyrene by contact with atmospheric dust and soot.The low figure for stem as compared with lamina is presumably due to the much lower surface area - weight ratio of the former. XETHOD REAGESTS- Light petroleum-Light petroleum, boiling range 40" to 60" C, free from aromatic hydro- carbons, is percolated through chromatographic alumina, distilled and stored over sodium wire. Benzene-Benzene (crystallisable) is washed three times with concentrated analytical- reagent grade sulphuric acid, once with water, twice with 2 N sodium hydroxide solution and then again with water until the washings are neutral. I t is then dried over anhydrous sodium sulphate, distilled repeatedly until the residue (50 ml from 2 litres) does not fluoresce and stored over sodium wire. Diefhyl ether-Anaesthetic ether B.P.is dried over sodium wire. Acetone-Laboratory-reagent grade acetone is used without further purification, Alumina-Woelm alumina, neutral grade, activity 1, is used. 3 : 4-Benzopyrene solzttion-A solution of the commercial material in light petroleum is percolated through alumina (Woelm neutral grade), and the principal fluorescent band is eluted with 40 per cent. v/v of benzene in light petroleum. The hydrocarbon is recovered from the eluate by evaporation and then recrystallised from light petroleum and dried in vucuo. The product has log(e382mp) = 4.46 (literature 4.47). PROCEDURE FOR RECORDING FLUORESCENCE SPECTRA- Fluorescence spectra are obtained with a Hilger medium-quartz spectrograph and photographed. In the examination of chromatographic fractions for the presence of 3 : 4- benzopyrene, Ilford HPS plates (5 inches x 4 inches) are used with an exposure time of 1 to 4 minutes.In the determination of 3:4-benzopyrene, Ilford HP3 plates (5 inches x 4 inches) are used with an exposure time of 2 to 12 minutes. The optical system used is show~n in Fig. 2.August, 19581 TOBACCO AND TOBACCO SMOKE. PART I 445 The source of exciting radiation, A, is a “black-glass” 125-watt mercury-arc lamp used in conjunction with a polished Wood’s glass filter, C. This transmits a group of lines in the 365-mp region, which excites the visible fluorescence of many polycyclic hydrocarbons. By means of a large glass condensing lens, B, and a concave mirror, F, arranged as shown, the exciting radiation is focused on the sample cell, E. This is a I-cm quartz cell, with lid, turned slightly off axis so as to deflect as much stray mercury light as possible from the spectrograph slit.To reduce the amount of stray mercury light still further, a matt-black shield, D, is placed in front of the cell, ” G Silt B C F A = 125-watt “black-glass” mercury-arc lamp (G.E.C. type MBW/U) B = Condensing lens C = Wood’s glass filter (Chance 0x1, 2 mm thick) D = Matt-black shield E = Quartz sample cell F = Concave mirror G = Lens of short focal length Fig. 2. Optical system used for recording fluorescence spectra The fluorescent light emitted from the cell passes through an aperture in the centre of the concave mirror and is focused on the spectrograph slit by means of a lens of short focal length, G.The whole of the optical system is conveniently mounted on the optical bench of a spectrograph. PROCEDURE FOR DETERMINING 3 : 4-BENZOPYREIiE IN CIGARETTE SMOKE- The cigarettes used in this work are conditioned a t 60 per cent. relative humidity and 70” F before smoking, and are smoked under the standard conditions described elsewhere.= The smoke is collected by electrostatic precipitation, use being made of the automatic smoking machine described elsewhere.21 The precipitated smoke solids (about 25 g) from 500 g of cigarettes are extracted from the autosmoker glass tubes with a mixture of equal volumes of diethyl ether and 2 N hydrochloric acid. The combined ether extracts are separated and washed, successively, with four 100-ml portions of 2 N hydrochloric acid, 100 ml of water, four 100-ml portions of 2 N sodium hydroxide and three 100-ml portions of water.The solution of the neutral fraction of smoke condensate in ether is then dried over anhydrous sodium sulphate, filtered, and evaporated on a steam-bath. The residue is recovered by evaporation three times from successive small volumes of light petroleum; it is then dissolved in 50 ml of light petroleum and transferred to a column of 115 g of alumina in a glass tube of 28 mm diameter, protected from direct sunlight. The chromatogram is developed with 500 ml of light petroleum and then, successively, with 250-ml portions of 10, 20 and 30 per cent. v/v benzene - light petroleum mixtures and finally with sufficient 100-ml portions of 40 per cent. v/v benzene - light petroleum mixture to elute all the 3:4-benzopyrene from the column.Each fraction is evaporated to dryness, care being taken to remove all traces of benzene, and re-dissolved in 5 ml of light petroleum. The fluorescence spectrum of each fraction is recorded, and those fractions containing 3 : 4-benzopyrene are combined. The combined fractions are concentrated to small volume and transferred to a column of 10 g of alumina in a glass tube of 16 mm diameter, protected from direct sunlight. The chromatogram is developed with 50 ml of light petroleum, successive 25-ml portions of 10, 20 and 30 per cent. v/v benzene-light petroleum mixtures and finally sufficient 25-ml portions of 40 per cent. v/v benzene - light petroleum mixture to elute all the 3 : 4-benzopyrene.The fractions are evaporated to dryness, care again being taken to remove all benzene, re- dissolved in 5 ml of light petroleum and examined for the presence of 3 : 4-benzopyrene as before. Fractions containing 3 : 4-benzopyrene are combined and evaporated on a steam-bath.446 BENTLEY AND BURGAN : POLYNUCLEAR HYDROCARBONS IN [Vol. 83 This volume depends on the amount of 3 : 4-benzopyrene present, but the concentration of 3 : 4-benzopyrene in it must be such that the following requirements are met- (i) The bands in the fluorescence spectrum a t 403, 408 and 427 m p are visible. (ii) The fluorescence due to the maximum amounts of 3:4-benzopyrene to be subse- quently added to aliquots for the purpose of determination must not be appreciably quenched in the solution.When necessary, this requirement is checked by means of a microphotometer, use being made of the strong 403-mp band. The concentration of 3 : 4-benzopyrene in this solution is determined by the comparison on a single photographic plate of the fluorescence spectra of the unknown solution, of standard solutions of 3:4-benzopyrene in light petroleuin and of aliquots of the unknown solution containing the same added concentrations of 3 : 4-benzopyrene as the standard solutions. For a photographic plate in the normal exposure ranges, the plate density is directly proportional to log(1ight intensity). With the low light intensities involved in this work, however, the plates are very much underexpcsed and the increase in plate density with increasing light intensity, and hence with increasing concentration of 3 : 4-benzopyrene, is found to be nearly linear for solutions of the pure hydrocarbon in light petroleum.Hence, by visual comparison of the spectra of the unknown solution and the solutions containing added 3 : 4-benzopyrene, it is possible to determine the level of added 3 : 4-benzopyrene a t which the concentration of the hydrocarbon in the unltnown solution has been doubled, and hence what this concentration is. With the proposed procedure, the final recovery of pure 3 : 4-benzopyrene added to solutions of smoke condensate in diethyl ether before the initial extraction with hydrochloric acid is 85 to 90 per cent., which provides a correction factor for determinations on normal smoke condensates. The residue is dissolved in a suitable known volume of light petroleum.PROCEDURE FOR DETERMINING 3 : 4-BENZOPYRENE I N LEAF AND CIGARETTE TOBACCO- The leaf is stemmed and cut before extraction. Manufactured cigarette tobacco is extracted without further preparation. Cut leaf or cigarette tobacco (200 g) is extracted with acetone for 4 hours in a Soxhlet extractor. The extract is evaporated under reduced pressure on a steam-bath, and the residue is hydrolysed by boiling under reflux for 2 hours with 100 ml of 10 per cent. w/v ethanolic potassium hydroxide. The resulting: solution is concentrated under reduced pressure on a steam-bath, diluted with water and repeatedly extracted with diethyl ether. The combined ether extracts are washed, successively, with three 50-ml portions of 2 N hydrochloric acid, 50 ml of water, three 501-ml portions of 2 N sodium hydroxide and three 50-ml portions of water; they are then dried with anhydrous sodium sulphate and evaporated under reduced pressure on a steam-bath. The residue is recovered by evaporation three times from successive small volumes of light petroleum, dissolved in 50 ml of light petroleum and transferred to a column of 115 6; of alumina, exactly as for the analysis of smoke condensate. With leaf extracts it is usually unnecessary to re-chromatograph the combined 3 : 4-benzopyrene-containing fractions.The 3 : 4-benzopyrene content of the com- bined fractions can then frequently be determined by direct comparison on a photographic plate of the fluorescence spectrum of the unknown solution with the spectra of standard solutions of the hydrocarbon; otherwise the determination is carried out by the method described for smoke condensate.For pure 3 : 4-benzopyrene added to tobacco before extrac- tion and hydrolysis of the extract, the recovery from the final chromatogram fractions is 75 per cent., which provides a correction factor for determinations on normal tobaccos. We thank the Directors of The Imperial Tobacco Company (of Great Britain and Ireland) Ltd. for permission to publish this paper. REFERE~CES 1. 2. 3. 4. 6, R~~~~ F 2 m r t ~ n n m i . A daaa,,*, r r-..... inn- 4 1 1 - 7 - Falk, H. L , and Steiaer, P. E., Cancer Res., 1952, 12, 30. Kennaway, E. L., Brzt. Med. J., 1925, 2 , 1. -, Bzochem. J . , 1930, 24, 497. Kennaway, E. L., and Sampson, B., J . Path. Btzct., 1928, 31, 609.447 August , 19581 TOBACCO AND TOBACCO SMOKE. PART I 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. Sugiura, K., Amer. J . Cancer, 1940, 48, 41. Flory, C. M., Cancer Res., 1941, 1, 262. Campbell, J. M., Brit. J . Exp. Path., 1939, 20, 122. Schurch, O., and Winterstein, A., 2. Krebsforsch., 1935, 42, 76. Wynder, E. L., Graham, E. A., and Croninger, A. B., Cancer Res., 1953, 13, 855. Cooper, E. A., Lamb, F. W. M., Sanders, E., and Hirst, E. L., J . Hyg., 1932, 32, 293. Wynder, E. L., Graham, E. A., and Croninger, A. B., Cancer Res., 1953, 13, 855. Waller, R. E., Brit. J . Cancer, 1952, 6, 8. Cooper, R. L., and Lindsey, A. J., Ibid., 1955, 9, 304. Wynder, E. L., Brit. Med. J . , 1957, i, 1. Alvord, E. T., and Cardon, S. Z., Brit. J . Cancer, 1956, 10, 498. Latarjet, R., Cuzin, J., Hubert-Habert, M., Muel, B., and Royer, R., Bull. Cancer, 1956, 43, 180. Bonnet, J., and Neukomm, S., Helv. Chim. Acta, 1956, 39, 1724. Sharman, C. F., and Iles, W. G., J . Appl. Chem., 1957, 7, 384. Morton, R. A., and Stubbs, A. L., Analyst, 1946, 71, 348. Received September 24th, 1957