Tobacco Documents Online

-10- - o t;he Camel blends may therefore reflect hasher air-cured content of the lamina porClon. Rowever, mmaoni~ salts may have been leached from tobaccos dur~ns the paper-type reconstltution process, and consequently it cannot be inferred from t;he low a~-mcmia nlt;rosen content o£ the reconstituted material chat it is flue-cured in orlK~n. 5. $~OKE ~.AS~NTS 5.1 Part;iculate Phase Deliveries (Table 5) The TPH and nicotine deliveries lie vlt;hLn the rs~nKes 20-28 mK and 1.0-1.5 mg respectively. These ranse8 are very similar Co those previously est;ablished (2) for Philip Morris brands, namely, 18-26 mE and 1.O-1.6 mK. However, whereas Philip Morris brands were 8enerally sunilar in t;erms of delivery per puff (2.6-2.9 ~ ~H/puff and 0.I&-O.17 mE nicotine/p-ff) the same yes not; true of the Reynolds brands (2.1-3.0 q TPM/puf£ and 0.10-0.17 mg nicotine/puff). The low nicotine delivery of Winston (Mexico) is in line v£Ch the blend measurements; however, the hiKher TPM and nicotine deliveries of Winston (HONK F~nK, F~aic, Spaln) wlch respect; to Winston (U.S.A.) cannoc be accounted for, since all four cisarat;ces vere sim£1ar physically and blends were similar in chem£stry and construction. Ou the ocher hand, the relatively low TPM delivery of Camel (Holland and West Germany) is due to ~he b/Kh paper porosity used; smoke ~/cocine has been ma£nCained in Holland by the use of e h~sher n~cotine blend. The proportion of nLcoClne whlch is extract;able lles in t;he ranie 18-28Z. Hizhest values are not;ed for Camel (U.S.A. and France) whilst; an unusually low value was noted for Camel (Wast Germany). The proportion of extractable nicotine may be related to the reconstituted tobacco BAT Co LTD - MINNESOTA TOBACCO LITIGATION m

-II- content or even the puffed tobacco content of the blend (Yisure 3); in the Philip Morris survey the proportion was hisher for PCL-type containln8 brands end this may reflect the addition of s predominantly elf-cured reconstLtuted material to the blend, or be related to the addition of chemicals. 5.2 Vapour Phase Deliveries (Table 6) The delivery of carbon monoxide (14.6-18.2 mE) is remarkably steady, considering the rmaKe of paper porosities employed. Winston (Mexico) is outside this reuse (23.3 IS) but in this case the delivery can be at least partly accounted for by the very low paper porosity (74 WTU) used. The delivery of s~nonia is in Keneral elishtly hi6her for Camel versions than for Winston versions; the delivery reflects the ammonia nitrogen content of cbe blend (Fisure &). The delivery of formaldehyde - s suspected irritant in ciBaretce smoke - was found in the previous study of Philip Morris brands (2) to be related inversely no reconstituted tobacco content, specifically PCL-type. Since no relationship with PET-type was found, the diammonium phosphate present on the PCL-type may have been responsible for reduced formaldehyde delivery. Zn the present study of ReDolds brands, formaldehyde delive~ declined with increased proportions of either PCL or PET types in the blend (Fisure 5). For French, Swiss and German versions of Camil, the deliveries were even lower than anEiclpated on this basis. Consequently, if the effect is due to the presence of added chemical(s) e.K. chlorides on the PET-type reconstituted tobacco, a Steerer quantitymay have been added to the reconstituted tobaccos in chess three counnries. BAT Co LTD - MINNESOTA TOBACCO LITIGATION D C CT- m,C C

-12- \ Previous studies on amelioration have shown (6, 7) that ammonium chloride and diaunonium phosphate are parnicularly effective in reducinE formaldehyde delivery and subjective irritation levels. In the Philip Morris survey (2) a plot of s~-onia and formaldehyde smoke deliveries 'mapped' Che brands in the same fashion as reducing sulars/a:~onia ninroEan blend contents. A similmz finding is obtained in nhm present study (Fiance 6), the map again reflecting blend type and reconstituted tobacco effects. The position of VICEROY with respect no the ocher brands has however chanted. Two consciCuenns of Che vapour phase scan, aceCon£crile and 2,5-dimethylfuran, have been extracted from the dana and plotted (FiEure 7). Dana is also included for three European air-cur~ ciSaretues and three European flue--cured cigarettes. The three mlr-cured clearances deliver hish levels of acatonltrile and low levels of 2,5-dlmethylfuran; the reverse is true for the flue-cured ciKaretces. The Reynolds brands are contained within Chess extremes, but 8 partial separation is obtained which supports prmvious data; ~,e. versions of Camel are ~enerally closer to the a/r-anted extreme whilst Winston versions tend more towards the flue-cured llne. Some deEree of overlap is stain apparent, /~dlcaning chat the blend tTpe difference Is not marked. 5.3 011/Water Partition Analysis (O/WP) (Table 7) O/WP analysis was cazTied out on the plain tobacco rods (SI) as well as the complete cisaretces ($2). The relative contributions of tobacco blend and filter characteristics to O/WP ch~acry can thus be determined. "The dana obtained was 'normalised' and 'chenically Srouped' - in this way the smoke quality of the cigarettes can be BAT Co LTD - MINNESOTA TOBACCO LITIGATION O", k.,,'-I r.,o

-13- intercompared on the basis of various chemical classes in smoke for which caste and flavour attributes have been tentatively assiEned (8). Fyridines and C~clic Aldeh~des (Figure fl) SI and S2 daze has bee~ pleiad, toKether with reference SI data for e 1OOZ flue-cured (PLAYERS ex Holland) and • leer air-cured (St. Michel ex Holland) =iEarette. The unf£1terea smoke from the flue-cured oisarecte is low in pyzid£nes and hieh in cyclic aldehydes; the reverse is true for the air-cured cigarette. The reynolds branas (SI data) lie between these two extremes, as does VZCEEOY. Marlboro lies above the li~e; the pyridlnes enhancement is due Co the addition of diammonium phosphate (2). Zn seneral, the various versions of Camel lie closer than the WlnsCon versions to the elf-cured marker and Co Marlboro, whLlst the Winston versions lie c~oser to the flue-cured marker and VICEROY. A elm/far b=and pattern emerses in the $2 dace, supportinE the earlimr conclusions concerning the blends. The proportion of ~ cyclic aldehydes is much lower than the proportion; this reduction is caused by filter plasticlser. It is interesting to note that, on the SI basis, Camel (West Germany) contained the highest level of cyclic aldehydes. The unusually hish level of triacetin in this filter has resulted however in 1:he lowest S2 cyclic aldehydes. On the basis of S2 pyr£d£nes/cycl£c aldehydes data, Camel (U.S.A. and France) ere seed matches co Marlboro (U.S.A.). BAT Co LTD - MINNESOTA TOBACCO LITIGATION

-14- /-- Volatile Carbon~ls and Nicotine (Figure 9) These are the dlmens~ons adopted in a recant examination of U.S.A. brands (9). It is noted that the U.S.A. brands (S2 data) show the same disposition on these axes as in the earlier study. VICEROY smoke conCalns the h£shest proportion of volatile cmrbonyls; Camel the lowest. Both S] and S2 data demonstrate thac the level of volatile carbonyls in smoke is significantly lower in the Reynolds brands than in VICEROY and Marlboro (U.S.A. versions). On the basis of this dace, Marlboro and Camel are not matched. The remaining Eeynolds brands lie between Winston (U.S.A.) and Camel (U.S.A.) in a random manner; for example, Camel (West Germany) resembles Winston (U.S.A.) whilst Winston (Spain) resembles Camel (U.S.A.). The resemblance in terms of these chemical groups is apparent on both S1 and S2 bases sines acetate filters and plaetic£sers show no selective effects on these O/~P constituents. Zndividual PTridinss In the exmmination of Philip Morris brands (2) the level of peak 41 was shown to be much enhanced by nhe presence of dlm~onium phosphate; in the present study, the hiKher leve~s of ~nia nitrosan in the versions of Camel than ~n the versions of Winston, have resultad in an enhancement of this .peak (Fisure I0). (This peak has hitherto been assisted the identity 2,5-dimethylpyridine; recent work using a nitrogen deT.ector suggests the presence of a pyrazlne in this peak, and this may be the predominating component.) A resemblance is asain noted between Winston versions and VICEROY (U.S.A.), whilst Camel versions are intermediate asain between Winston versions and Marlboro (U.S.&.). The anhancemenc in peak 41 can be caused by chemical addition (as in Marlboro) BAT Co LTD - MINNESOTA TOBACCO LITIGATION C

-15- or through the use of hish a.~onia nitrogen a~r-cured tobaccos (as in Geuloises). Burley nips contain a high level of a~onla nitrosen, and could explain the Camel characteristics. The use of dLammonium phosphate on Philip Morris PC'L-type reconstLtuted tobacco was shown (2) to result in an enhancment in peak At wlth respect to peak 47. (The recent work using both f.iod, and nitrogen detec¢or systems may m11ow this to be interpreted as an increase in pyrazine/pyridine ratio.) ~ Figure II, it can be seen that versions of Winston and VZCEEOY (low ammonia nitrogen) have similar 41/47 ratios whilst Camel and Marlboro (high ~onia nitrogen) hate a hither ratio. High a~nonia nitrogen content results in an Incrlase in both peaks 41 and 47, but the increase in peak 41 is the greater. In contrast (Figure IO] the propornion of peak 53 (3-vinyl pyrldine) is llttle affected by the b~end J,mmonia nitrogen level, and it has been conjectured (2, lO) that this pyridine results from pyrolysis of nicotine. Non-Tobacco Constituents Menthol (peak 6A) was not detected in other than zrace quantities on any of the brands; on the other hand, a coumarin peak (peak I02) was noted on three versions of Camel (Swiss, Dutch and French) and a miler amount on one version of WLnston (Mexican). 6. SUBJECTIVE ASSESSPENTS 6.1 ~pact and Zrr£tatlon (Table 8) The level of impact was highest for the ~ U.S.A. products; the remalninE versions of Winston and Camel had slightly mower impacts with the exception of Wins=on (Mexico) which had a significantly lower impact. BAT Co LTD - MINNESOTA TOBACCO LITIGATION m O (..j'l r'O

-16- The level of irritation yea slightly lower for Camel than Winston. Since Winston (Mexico) showed similar levels of irritation to other versions of Winston, thle brand had a 8isnificantly hiKher irzltation:&mpac= ratio. Impact scores did not co=relate well with either extractable nicotine delivery per puff or O/WP nicotine delivery per puff. The hiKh impact score of Winston (U.S.A.) cannot be explained by the smoke chemistry data. The high irritation:impact ratio of Winston (Mexico) may, howewer, be explicable from the smoke cheuListr7 data. This brand delivered vary low amounue of extractable and O/WP nicotine per puff (related uo impact) but high levels of formaldehyde (possibly related to irritation). 6.2 Flavour Profile (Table 9) The brands were assessed with respect to their U.S.A. counterparCs; the assessment wee direcc£onal. The basic conclusion is chat differences between these brands, in flavour profile terms, are Im~ll. ExamininK the non-U.S.A, versions of Winston it can be noted tha~ increased scores for 'flue-cured', 'musty' and 'burnt' are Kenezally obtained with respect Co the U.S.&. vers~8ome increase in 'air--cured' characuer is also noted. On the oth~ hand, 'blended' and 'roasted, toasted' characters have declined in some ~nscances. The overall flavour ampllcude is the same in each case, including the U.S.A. version. An almost identical pattern is obtained in the case of CameZ; increases in 'flue-cured' score and decreases in 'roasted, toasted' charecuer (with respect co the U.S.A. version) beLtS parr/cule=ly notable. Again, the overall flavour amplitude was identical for all five versions BAT Co LTD - MINNESOTA TOBACCO LITIGATION m 0 Cm.

-17- of the brand. The tflue-cured' and 'roasted, toasted' scores are explicable in terms of blend au~onia nitroKen content, since both U.5.A. versions are the most elf-cured on this basis. 6.3 Pack Aroma Whi£st similar flavour elements are used in all varsions of Winston and Camel, the balance between the elements seems to vary more than was the case with Philip Morris brands (2) or versions of Marlboro (i). For example, Camel (U.S.A.) has hiKh carlmll =haracter, Winston (Mexico) has low spice character and Winston (Spain) has a hLKhar emphasis on fruit. The "followinE comments relate to the individual brands; comparative cou~ents referring to the U.S.A. version. Winston. (U.S.A.) - Moderate/hash pack aroma amplitude. Spice (cinnamon/mace type) and herbaceous with liKhc fruit and caramel aspects toKethar with a maple nuance. Little contribution from toasted and cocoa characters. Low Burley character, less heav£1y cased than traditional 'U.S.-blended'. Some green notes detected, more typical of Virslnia types. Winston (Mexico) - S/JnLla~ or slightly Kreater amplitude than U.S.A. version. Fruit/caramel aspects are more apparent, whilst spice/herbaceous aspects contribute less. More Burley notes than the U.S.A. version with sums l~Sen and musty overtones. A cloy~nl balsamic mouth 'feel' was x~ted on lmokinK, possibly due to ethTl cinnamate. W~nston (Honk KOnE) - SLmilar or 81ilh~ly Ir~ster ampILtude than U.S.A. version. AlthouEh spice/herbaceous ollments are present, stabley notes and a more cocoa backEround were detected. The product is more typical of traditional 'U.S.-b~ended' products. BAT Co LTD - MINNESOTA TOBACCO LITIGATION r,,o

-18- Winston (F~wait) - Slightly Steerer amplitude than U.S.A. vers£on. Spice and herbaceous elements are again more apparent; fruit aspect is less apparent and of lower pitch. Caramel note was less apparent due to increased spice element. Winston (Spain) - Sllghtly lower amplitude than U.S.A. version J A high fruit amplitude was noted on some packets. Lass spice but more herbaceous elements of a tobacco compatible type were detected. Caramel notes ware detected. Camel (U.S.A.) - Hodexate pack aroma amplitude. Ven£11ic/caramel character w~h fruit, nut, herbaceous and l£$ht spice nuances. Little or no cocoa aspects and some evidence of a low level of pepperm~nn. Camel (France) - Lower amplitude than U.S.A.. Similar character to the Dutch product (see below) althoush el~ghtly lo~er pitch; this may be a reflection of the blend. No flavour note of the st&ble type was detected. Camel.(Sw.ltserland) - Sim£1ar 8mplltude to U.S.A. A£so s~lar £n character to ~he Dutch product (see below) but nora Burley-llke character with stish~ly stable-like aspect reduclng aroma pitch. Spice and venill£c/caramellic elements were not as apparent with respect to Dutch and U.S.A. versions. C~mel (HoLland) - Fosslbly Ereater In a~pLitude than U.S.A. vers£on. Vanillic components are Less noticeable whilst fruit character is of higher pitch. Spice/herbaceous elements were more detectable, herbaceous elements heine of the flours, 8enet type. The overall impression was of • more tobacco compat£ble character with leas interest notes. BAT Co LTD - MINNESOTA TOBACCO LITIGATION O",.

-19- Camel (West Germany~- Creater ampli:ude than U.S.A, version. Less c&za:el aspect than UoS.&., but more than other v~re£ons. Here £rui¢ and spice aspects were noted than e~l othe= versions, the spice element beinK reminiscent o£ SinKer. A ~nt note was detected initially. 7 • CONCLUSIONS The analyses per£o~d rank the brands ('In tezms og blend aiz~-cured content) in the sequence:- Marlboro (U.S.A.) > Camel > VTCZ~0Y (U.S.A.) • W~neton. Produc~ features co~n to the Reynolds brands examined ~ ~h~s report ares- (a) FAlter lenKth and triaceti~ content (with the exception of Camel ex Vest Germany). (b) Overall cigarette pressure drop. (c) Use of l£quo~cice and cocoa in casings. (d) Use of P~T-type in blends (with the ~cep~.on o~ W~nston ex Mexico). (e) The percenteSe of Pwr-type in Wlnltou blends (all stated to be ~an~actuzed in U.S.A.). (£) Use o£ puf£ed tobaccos (~W~_l~ the e~eeption8 of Winston ex Hexlco end Camel am West Gez~any). (g) Delivery of carbon monoxide and n£r~£c oxlde. (h) Low delivery of 0/WF volatile cazbonyls, (i) I, eveL o£ subjective i~z£t&t£on. Level of flavouz mnpl£tude. BAT Co LTD - MINNESOTA TOBACCO LITIGATION