Tobacco Documents Online

THE EFFECT OF THE HRI[MECTANTS GLYCEROL AND PROPYLENE'CLYCOL ON l+IAINSTREAM AND S.IDESTREAM SMOKE DELIVERIES OF' ACROy.EIN, FORMALDEHYDE, ACETALDEHYDE, ACETONE AND PROPIONALDE$YDE Project lea~dier : BLAKE-C. Technical assistance : Posset-L. Bo:sch-V. Droz-A. PPiiILIP MORRIS EUROPE February 1987 R&D Department Neuchatel/Switzerland C87-05284

CONTENT S PAGE Introduction 3 Cigarettes Analytical Nfetliods 5 - 6 Results and Discussion. Fteferernces 10 Conclusions 11 Tables 12 - 21 Figures 22 ' Appendix : Cigarette specifications, filler andimainstream smoke data

- 3 - I NT RpD'UCT I ON Commer-Cial cigarettes often containlglycerol and/or propylene glycol as humectants to maintain moisture at al desired level duringi storage. It is permitted to add up to 5% total humectants by-the German tobacco regulations (Tabakverordnung). 011 O'H OH OH OH ( l~ 1 CH2 C'H Cii2 CH3 CH Glycerol Propylene glycol The formation of acrolein from glycerol during smoking was described! in an, experiment with labelled glycerol (i1) . It was found that glycerol, addled at a level of 3% to the filler, contributed about 8.5% of the total mainstream smoke (MS) acrolein. A more recent investigiation (2) showed that after spiking cigarettes with glycerol, a small increase in MS acroleim delivery was noted, h~owever a significant increase in sidestream smoke (SS) delivery was found. In order to confirm these findings a more extensive project was commissioned Wroject T'EAPt) . Cigarettes were prepared from a commercial blend (BPt-0105)~ containlingi either propylene glycol or glycerol or mixtures of both humectants. Correlations have been made between amounts of glycerol and/or propylene glycol in the filler and deliveries of MS and SS acrolein. in addition, correlations have been made between amcounts of glycerol and/or propyliene glycol and MS and SS deliveries for formaldehyde, acetaildehyde, acetone and propionaldehyde.

- 4 - CIGARETTEB'. Cigarettes were prepared fromia commercial blend (BR-005), containing : (a)' No humectant (reference cig,arette) (b) 2.0, 3.4 or 5.4% glycerol, respectively (c) 1.2, 2.7'or 4.1% propylene glycol, respectively (d) 0.6% propylene glycol and 1.8% glycerol (e) 0.6% propylene glycol and 4.5% glycerol. The humectant contents were determined~by the. PNiE-QA (Table 1). Cigarette specifications andIMS analytical data generated by the PME-QA are given in the Appendix.. Before smoking, the.cigarettes were conditioned at 60s RH and 22 2°C for at least 72 hours.

- 5 - ANALYTTCAL METHODS Stnoking and aldehyde trapping conditions. For MS generation a Borgwaildt RM 20/'C'S smoking machine was u~sed. Fi~Te cigarettes were smoked for each of three determinations according to:DIN 10240/1 (60% RH, 22°C, 2 sec, puff every 60 sec., 35 ml puff volume). The MS aldehydies were trapped in two Dreschel type traps ('100 ml)) each containing 75 ml of 01.5% dinitrophenyl hydrazine (DNPH) and 0.015% phosphoric acid in acetonitrile. A Cambridge filter (0 4.4 cm) was connected ini series with the traps. For SS generationi, a single cigarette smoking set-up was used in which cigarettes were smoked in a pear-shaped glass chamber of one-liter volume (3). Three cigarettes were smoked for each of three determinations (2 sec. puff'every 60 sec., 35 ml puff volume). The SS aldehydes were collected in two Dreschel type traps (100 ml) each containing 75 ml of 0.5% DNPH and 0.1% phosphoric acid in acetonitrile. A Cambridge filter (014.4 cm) was.connected in series with the traps. After collection, the DNPH solutions were coanbinedi, diluted 1:3 with acetonitrile.and then left for at least 2 hours to complete the dieriwatization in-situ. The DNPH-aildeMyde derivatives were stable for 15' hours under these conditions. The Cambridge filters (which trapped fo maldehyde which broke through!the traps) were cut up into small pieces and extracted with 10 ml of DNPH solution, as used in the Dreschel traps for MS and SS respectively.

HPLC'analysis The trappingisolutions and Cambridge filter extracts were combined and analysed by HPLC under the followingicondiitions : A'Hewlett Packard model 1084 A liquid chromatograph, equipped with a variable wavelength detector sample injector was used.. (model 798'75A) and automatic Columns : 25 cm x 4.6 mm (i.d.), packed~with Zorbax OD'S. (5 u) and 25 cm x 4.6 mm(i.d.) packed with Supelco LC'-18 (5 u) connected in series Detector : UV at 365 nm Injection volume : 10 ul Oven temperature : 35oC Solvent : acetonitrile + water (55' + 45) Flow : 1.5 ml/min Concentrations of DNPH-al:dehydes were calculated by the external standard method~usingipeak area measurements. Calibration solutions were prepared by dissolving, previousiy synthesized DNP'H-alidehydes in acetonitrile.

RESULTS AND DISCUSSION Effect of glycerol and propylene glycol on mainstream smokee parameters A'detailed list of cigarette specifications, filler composition and MS data is.givern inithe Appemdix. S'ome of the more important information regarding the effects of the humectants oniMS components are shownlin,Tables 2, 3 and4I. It can be seen that glycerol and propylene glycol have no significant effect on puff number, carbonimonoxide, nitric oxidie, nicotine andi DPM deliveries. It appears tha~t TPM,(DPM + water) values increase slightly with increasing humectant concentratilons but this is due, partly to an increase inithe water delivery. A mass balance study of the fate of'the humectant glycerol during smoking was made (3) using C'14' labelled glycerol. The gilyceroll content of the filler wa~s 3%. It was calculated that 20$' of the MS TPM is derived froQn~ the humecta~nt glycerol. Effect of glycerol and~propy,lene glycol on MS and SS acrolein deliveries For MS the aidditionlof glycerol andl/or propylene glycol to the filler had no significant effect on the acrolein delivery. In fact, acrolein deliveries at high humectant content (, 5%) were lower than for the reference cigarette. This is probably due to ai lower weight of tobacco inithe fi ller (Table 5; Figs 1, 2 and 3). For SS the addition of glycerol to the filler had a 5ignificantt effect on the acrolein delivery e.g. 5.4% glycerol gave a 50%

increase in acrolein delivery (Table 5; Fig. 1!). The acroleinn does not appear in the MS smoke in significant quantities perhaps due to-filtration by the filler rod or the filter itself. Approximately 1%' of the available glycerol was converted to acrolein during smoking. OH OH OH I I `' CH2 CH CH2 - 2H20 -~ CH2 = CH-CHO Glycerol A~crolein Propylene glycol had no effect on SS delivery of acrolein (Table 5; Fig. 2). The cigarettes containing both humectants (8P and 9P) also showed increased acrolein deliveries which were directly proportional to the glycerol content of the filler (Table 5; Fig. 3). Effect of glycerol andipropylene.glycol on MS and SS formaldehyde acetaldehyde, acetone andipropionaldehyde deliveries The addition of glycerol and!/or propylene glycol to the filler had no significant effect on the dieliveries of MS formaldehyde,, acetaldehyde, acetone or propionaldehyde (Tables 6, 7, 8 and 9; Figs. 4, 5:and 6). For SS the presence of glycerol in the filler markedly increased formaldehyde deliveries e.g. 5.4% glycerol gave a 3.5 fo1d:! increase in formaldehyde. This may be due to direct pyrolysis of glycerol to form formaldehyde (Table 6; Fig. 4!)1. S'lig,htly

increased SS deliveries of acetaldehyde were also observed (Table 7') whereas acetone and propionaL'dehyd~e deliveries were not significantly affected (Tables 8' and 9).. The addition of propylene glycol to thie filler had no effect on SS*deliveries of formaldehyde, acetone and propionaldehyde. (Tables 6, 8 and 9; Fig. 5), although acetald'ehyde deliveries were slightly increased (',Table 7). Cigarettes containing both humectants also showed increasedd formaldemyde!and acetaldehyde deliveries which1were directly related to the glycerol content of'the filler (,Tables 6:and 7'; Fig,. 6)

REFERENCES (1) Ga-ger-F'. L., C'. E. Brunot, Pt'. D. Carpenter, F. E'. Re snik and C.J. Varsel, "Cigarette.smoke precursors 1. The contr'ibution of the humectant glycerol to the acrolein content of cigarette smoke", PM Research Center Richmond, Virginia (USA), April 14, 1961. (2)1 Blake-C.J., "'The contribution of the humec:tant glycerol to the acrolein content of mainstreamiand sidestream smoke. A preliminary investigation", PME-Neucha'tel, May 29, 1986. (3) Piad~e-J.-J., PME Analytical Method "Single cigarette determination of carbon monoxide in mainstream and sidiestream smoke". (4) Holmes-J'.C., M.B. Bennett and B.T. Oakley, "A study of the distribution of the humectant during the smoking process using C14 glycerol as a tracer", unpublished, Philip Morris Inc.