Tobacco Documents Online

Accession number 2176 Lorillard Research Center Greensboro Tobacco Improvement - Progress Report - A Comprehensive Study of the Chemical and Organoleptic Properties of Four Burley Tobacco Grades and Their Relationship to Taste Differences Submitted by: J. P. Morgan Report number: ' Date: June 17. 1985 Summary or Abstract: Three offshore HL burley grades from-Malawi, Chile, and Argentina and a domestic HL grade were evaluated chemically and organoleptically to determine reasons for differences in taste of cigarettes made with these tobaccos. Chemical analyses, used for this investigation, were leaf analysis, organic acid analysis, ether extractions, sequential soxhlet extractions, smoke.analyses, gas and vapor phase analyses, pyridine analysis, and organoleptic evaluations. Chemical and organoleptic differences were seen among these samples. The majority of the differences appeared to be quantititave rather than qualitative. No correlations were seen between the chemical analyses and organoleptic evaluations. /1p:1 Xc: J. W. H. F. H. C. J. J. Bell Crouse Minnemeyer Schultz (b ~ Library ~ O N GD a

Introduction It has been demonstrated that marginal offshore burley tobaccos can be improved with chemical treatment; however, the character- istic offshore cigar taste still cannot be completely removed or covered in the total cigarette blend (2). One of the major objectives of this investigation was to organo- leptically and chemically identify the components in the tobacco which generate these offtastes or differences in taste character. By establishing a correlation between these two evaluations, one may better predict the treatment necessary to improve the usabil- ity of marginal tobaccos. Experimentation and Discussion of Results Four different HL burley grades were selected for this investi- gation. Two of these grades, C-HL-A-83/84 (Chilean) and HL-K-80 (Domestic) were acceptable as domestic HL grades. Two of the grades, A-HL-D-83 (Argentine) and.MB-HL-U-82 (Malawian) were unacceptable as domestic grades in all blends. These grades were chosen for three reasons: 1. They were readily available in Danville. 2. They were rated as HL grades. 3. They were representative of acceptable & unacceptable burleys.. The investigation was divided into two parts - First, the tobac- cos were analyzed for differences in their chemical composition and secondly, cigarettes were machine made with each tobacco grade and their smoke analyzed chemically and organoleptically. I. Tobacco Analysis A. Each of the burley grades were submitted to the Leaf Laboratory for complete tobacco analyses. The results are listed in Table I. There were significant differ- ences among the four grades: 1. The total nitrogens of the Domestic and Malawian tobaccos were 4.56 and 4.44% respectively. They were approximately 17% greater than the two remaining grades. 2. The total volatile bases of the Argentine tobacco was 0.81%. This was approximately 20% less than the three remaining grades.

3. The Domestic and Chilean tobaccos had the highest nicotine content, 3.78 and 3.51% respectively. The nicotine contents of the two remaining grades were approximately 30% less. 4. The specific volume of the Malawian tobacco was 210. This was approximately 16% less than the remaining three grades. 5. There were no significant differences in the total reducing substances. 6. The nitrate of the Domestic tobacco was signifi- cantly greater than the remaining three grades. It was 113.0% greater than the Argentine tobacco and approximately 309.0% greater than the remain- ing two grades. 7. The Argentine tobacco had the highest pH. It was 6.20 versus a mean of 5.27 for the remaining three grades. 8. The Domestic tobacco had the greatest chloride ion content, 0.48%. The remaining grades ranged from 0.20 to 0.34%. 9. Thee chemical oxygen demand (COD) was greatest for the Chilean tobacco. The remaining three grades ranged from 1.50-1.54. 10. The Malawian tobacco had the lowest ash (15.2%). The remaining grades ranged from 18.6 to 19.5%. B. In investigating the reason for differences in the aroma of these tobaccos, an ether extraction similar to one by Hajime Kaneko and Kiyoko Ijichi was used (4). A sample of each burley grade was dried overnight at 105°C, ground.in the Wiley Mill in the Pilot Plant (PP), and screened. The tobacco remaining on the 40 mesh U.S. Standard Series Screen was retained for analysis. One gram of each was weighed into a 125 ml round bottom flask for extraction. The samples were extracted twice with diethyl ether (C H ) O. The extract was decanted through a grade 102 iilter into another 125 ml round bottom flask. The extracts were evaporated to dryness. The remaining residues were transferred to a 1 ml volumetic tube with methylene chloride (CH C1 ), and diluted to exactly 1 ml with CH C1 . They wire then analyzed on a Model 5880 He~leit Packard Gas Chromatograph. The conditions used for the analysis are listed in Table II. - 2 -

Chromatograms were recorded for each sample. (Figures I through IV) A comparison of these showed no significant differences in the components present in the four tobacco samples. C. A more extensive extraction adapted from one by John Steel (5) was used to further evaluate any chemical differences which could help explain the aroma and organoleptical differences in the tobaccos. A sequential soxhlet extraction was employed using hexane, methylene chloride, and 95% ethanol. Each tobacco sample was extracted in triplicate and the extracted residues weighed and recorded (Table X). Each sample was dried overnight at 105°C, ground in the Wiley Mill, and screened. The tobacco remaining on the 40 mesh U.S. Standard Series Screen was retained for the analyses. One gram of tobacco was placed in each glass thimble and covered with glass wool. The thimble was placed in the soxhlet extractor and 200 mis of solvent added. The extraction time for each solvent was ten hours. The recycle times for each extractor was set at approximately seven minutes. The first extraction solvent was hexane. Following the extrac- tion, the extract was transferred to a tared flask, evaporated to dryness, and dried overnight at 105°C. The same tobacco samples were used for the methylene chloride extraction, followed by the 95% ethanol extraction, aHd the same residue determination proce- dure followed. There was good agreement within extrac- tions for each sample with several exceptions. A review of these results showed: 1. The Chilean hexane residue was approximately 25% less than the other three tobacco grades which ranged from 6.11 to 6.61%. 2. The Chilean methylene Chloride residue was also the lowest, 0.40%. The Domestic sample had the greatest CH2CL residue at 1.68% followed by the Malawian (0.991) and Argentine (0.83%). Due to the variation within the CH2CL2 extractions, a sample of each was analyzed on the Hewlett Packard 5880 Gas Chromatograph using the conditions listed in Table II. Chromatograms were recorded for each sample (Figures VIII through XI). From a review of these, it appeared that the same components were present in the residues of each sample, with greater quantities of some of these in the Domestic and Malawian samples. - 3 -

3. The ethanol residue of the Domestic sample was the greatest (16.44%). This was approximately 15% greater than the three remaining samples. D. Organic acids significantly effect smoke strength and taste character (3). Burley tobaccos contain large quantities of oxalic, malic, and citric acids. The five burleys were analyzed for organic acid content using the Gas'Chromatographic method determination as their methyl esters (1). The results are listed in Table IX. A review of the results showed that the major organic acids were approximately 60% greater in the Chilean and Domestic tobaccos than in the Argenine and Malawian tobaccos. The citric acid content of the Malawian tobacco was only 0.6%. The lower acid content in the tobacco could effect smoke pH and strength, allowing more free nicotine and a more basic smoke to be trans- ferred upon pyrolysis. This may contribute to a residual cigar note in the Argentine tobacco as previ- ously reported by the expert taste panel. II. Smoke Analyses The four burley grades were spray cased in the P.P. with 2% glycerine and the moisture adjusted to 19%. After approxi- mately eight hours equilibration, they were cut 32 cuts per inch and dried back to 14.5% moisture. One tray of 85mm non-filter cigarette were machine made in the P.P. The cigarettes were supposed to be made at the same weight; however, due to differences in tobacco volume, moisture, and firmness the weights were adjusted to give a more uniform looking cigarette. Cigarette samples were stored in the conditioning room at 75°F and 60% relative humidity until all investigations were completed. Samples were submitted for complete smoke, organoleptic, and vapor phase analyses. In addition a pyridine analysis method was developed. A. For the smoke analysis it was requested that all samples be picked within the same weight range and cigarettes smoked to a 23.0 mm butt length. The results are listed in Tables III-VI. The weight differences were due to variation of the mean weights among the samples. The tobacco weights of the Argentine and Domestic tobaccos were less which was somewhat predictable since they had greater specific volumes. Their dry particulate matter, corrected particulate matters, and puff counts were also less. This could be explained by their reduced tobacco weight as well as their significantly higher nitrate content. The nicotine deliveries were indicative of that found in the tobacco analyses. The carbon monoxide results were all within the same range. - 4 -

B. Permanent gas phase results are listed in Table VII. A review of these showed that the nitric oxide content was significantly greater on the Domestic and Argentine samples. This was related to the higher nitrate content of these tobaccos. The hydrogen cyanide (HCN) contents were significantly greater in the Domestic and Malawian samples. Since HCN is a degration product of tobacco proteins, this could indicate that the levels of proteins were greater in these samples. In the organic vapor phase analyses the isoprene level was significantly greater in the Malawian samples, indicating there may be more essential oils in this tobacco or a greater lamina to stem ratio. The acetalde- hyde, acetone, and acrolein levels, all related to the cellulosic nature of the tobacco, were lower in the Malawian sample. This, along with the higher isoprene level above and lower ash content reported earlier for the Malawian tobacco, verify a higher lamina to stem ratio in this sample. The delivery of benzene was significantly lower in the Domestic sample. The deliveries of acetonitrile were significantly greater in the Chilean and Malawian samples. This may be indicative of greater levels of n-heterocyclics, such as pyridines, picolines, or methyl pyrazine. C. For n-heterocyclics in the smoke of these samples the following procedure was developed: 1.' Five cigarettes for each sample were smoked to a 23.0 mm butt length on an 8-port Filtrona Smoking Machine. The smoke condensate was collected on a Cambridge filter pad and the vapor phase collected in two 50 ml impingers containing 25 mis of 0.2N HC1. 2. The Cambridge pad was combined with the acid solution from the impingers in a 125 ml round bottom flask. Approximately 25-30 mis of 0.2N HC1 was used to wash the glass Lubing connecting the impingers to the smoking machine. This solution was shaken for 30 minutes. 3. The extract was filtered through glass wool and extracted with 4 x 40 mis of methylene chloride (CH2C12). 4. The CH C1- was back washed with 2 x 25 mis of 0.2N HC1 ang tAe CH2C12 discarded. - 5 -

5. The acid solutions were combined and the pH made basic with lON NaOH. 6. This solution was extracted with 4 x 30 mis of CH2C12 and the base discarded. 7. The CH Cl was dried over Na2SO4 and filtered througA giass wool. 8. The CH C1 was evaporated carefully to 2-3 mis and transfirrid to a 5 ml volumetric flask and diluted to the mark with CH2C12. 9. The samples were analyzed on the Model 5880 Gas Chromatograph. The conditions used for the analysis are listed in Table VIII. Chromatograms were recorded for the Domestic, Argentine, and Chilean samples (Figures V-VII). There were no significant differences in the n-heterocytic components present in the three samples. When time permits, the major components should be identified and quantitated. D. Cigarette samples of the four tobaccos were submitted to the expert Product Development taste panel for a blind or.ganoleptic evaluation. It was requested that they evaluate these samples with respect to: 1. Character 2. Strength 3. Offtaste They were also asked to make comments concerning unusual offtaste including possible cigar tastes and to rank according to preference as a burley tobacco. The results are listed in Tables XI & XII. The samples were submitted two times for the blind organoleptic evaluation. In the first evaluation, the the order of their pre- ference :aas : Grade 1. A-HL-D-83 Argentine 2. K-HL-80 Domestic 3. C-HL-A-83/84 Chilean 4. MB-HL-U-82 Malawian The Argentine sample had no offtaste with good charac- ter and strength. The Domestic sample was very close with an offtaste of 0.25, good character, and the - 6 -

most strength. The Chilean and Malawian samples followed with low character, medium strength, and high offtaste. In the second evaluation of the same samples the order of preference-was: Grade 1. MB-HL-U-82 Malawian 2. A-HL-D-83 Argentine 3. C-HL-A-83/84 Chilean 4. HL-K-80 Domestic The Malawian sample, rated the worst in the previous evaluation, was rated the best. Its offtaste was the lowest, 0.25, and its character the best. The strength was good. The Argentine sample, rated best previously, was rated second. It had low character, good strength, and medium offtaste. The Chilean sample was again rated third with low character, good strength and high offtaste. The Domestic sample, rated second previ- ously, was surprisingly rated last. It had low charac- ter, good strength, and the highest offtaste. A review of the two organoleptic evaluations showed that the.Argentine and Malawian samples, overall, received the best ratings. The two grades which are currently used as domestic grades were rated below the two offshore grades which are not currently used as domestic grades. Conclusion Significant differences were noted among these four burley grades in the various tobacco analyses used in this evaluation; however it appeared that the major components were very similar. The tobaccos vary primarily in the quantity of the components pre- sent. The Domestic tobacco was much higher in nitrogenous components, due probably to the greater use and availability of nitrate fertilizers. The organic acid and nicotine content of the Domestic and Chilean tobaccos were much higher also. There were also significant differences found in the evaluation of smoke from cigarettes made with these tobaccos. As in the tobacco analyses no different components were detected which might lead to an explantation of differences in taste. The taste differences were certainly related to the quantity of the compo- nents delivered in the smoke. This delivery is dependent upon many uncontrollable and some controllable variables. Due to the inconsistency of the organoleptic evaluations, no correlations could be drawn between chemical differences and taste differences. - 7 -

Even though two of these burley tobaccos have previously been rated as unexceptable for domestic use, this investigation didn't support that. It found that they were different in the quantity of their components but yet so similar in their smoking quality that they couldn't be consistently identified This evaluation did uncover significant differences among these four burley grades. With Gas Chromatograph/Mass Spectrophoto- meter analyses to facilitate more qualitative and quantitative investigations, more differences could be uncovered and identi- fied. To utilize these data, a means to correlate them with the organoleptic evaluations must be established or all of our analyses are in vain. References 1. Rose A. Shoffner & J. H. Bell, "An improved Gas Chromato- graphic Method for the Determination of Oxalic, Malic, and Citric Acid in Tobacco As Their Methyl Esters", (June 15, 1978). 2. J. P. Morgan, "An Investagation of pH Adjustment on Offshore Burley Tobaccos - Progress Report One; (April 29, 1983). 3. W. E. Crouse, "A Comprehensive Study of the Chemical and Physical Properties of B3FR Burley Tobacco and Their Relation- ship to Smoke Strength, (April 8, 1981). 4. Hajime Kaneko & Kiyoko Ijichi, Agr. Biol. Chem., Vol. 32, No. 11, p. 1337-1340 (1968). 5. John A. Steele; JAOAC, 57, p. 1379 (1974).

Table I Tobacco Analysis of Five Burley Grades Sample Nit. TVB Nic. Vol. Ash TRS N0 pH C1 COD Argentine 3.84 0.81 2.56 262 18.6 0.8 3 1.23 6.20 0.20 1.51 Chilean 3.81 0.98 3.51 240 19.2 1.2 0.75 5.20 0.34 1.63 Domestic 4.56 1.01 3.78 249 19.5 0.8 2.62 5.33 0.48 1.50 Malawian 4.44 0.98 2.60 210 15.2 1.2 0.52 5.29 0.25 1.54 ezosos9e