Tobacco Documents Online

C 1000359943 q

Accession No. 75-247 Copy No. Issued To PHILIP MORRIS U..S.A. R E S E A R C H' C E N T E R CHARGE NO. 8 TITLE: 2305 - Flavor Development TYPE REPORT: OANNUAL OSEMIANNUAL OCOMPLETION OSPECIIAL PERIOD COVERED: July, 1974 - July, 1975 CONFIDENTIAL ES CASING STUDY - PROPYLENE GLYCOL W!R I TTEN BYt ~* J. W. Swain Mr. F. E. Resnik Mr. A. B. Clarke Mr. R. N. Thomson Mr. H. L. Spielberg Dr. W. F. Gannon Dr. R. M. Ikeda Mr. K. S. Burns Central Files Mr. H. G. Daniels Day Files (2) Mr. J. D. Hind Indexer Mr. Ji. W. Madures Mr. E. A. Christopher ~ KEYWORDS: ES Casin~, Propylene G1yco1' Elimination, Physical Data, Subjective Evaliuations, Superwet Casing Applications, Casing, After Expansion ©

I F r F I i I I I I. II. . TABLE OF CONTENTS Page Introduction and Summary ........................ 1 Experimental Procedure, Results and Discussion . B. .. 2 Casing of ES after Expansion ................ 2 Processing Results ...................... 2 . Run 438 versus 439 ................... ................ 2 . Subjective Results ........... ............ 3 . Run 438 versus 439 .................. 3 Casing at Superwet Cylinder ................. 4 1. Processing Results ....................... 4 a. Run 6281-A verus 6282-A ............. 4 b. Run 6790 versus 6793 ................ 6 2. Subjective Results ...................... 7 a. Run 6281-A versus 6282-A; 6790 versus C. . a. Run 495 versus 464 ................... 7 7 7 7 b. Run 7400 verus 7401 ................. 8 2. Subjective and Smoke Results ............ ]0 a. Run 495 versus 464 .................. 10 I b. Run 74i00 versus 7401 ................ 10 3. Recommendation of PG Elimination ........ 11 L Appendix ......................................... ................ 15 IV. References 6793 ........................... Elimination of PG from Casing .......... 1. Processing Results ................. ...................................... 28 L

r r F F F F r i I I- I L L I. Introduction and Summary ~ ,.Analysesof casing components on Bright expanded stems . (ES) have shown 60-80% loss of propylene glycol during Pilot Plant processing. The volatility of PG has been shown in other studies of humectants on blends as well as expanded materials (Hind, 1975, and Crayton, 1975). Most of the PG'lost from ES during processing escapes into the tower gases during,ex- pansion; therefore, alternatives to current processing were proposed to avoid high temperature exposure for PG. Eliminating these losses during exapnsion could save approxi- mately $100,000 per year. More savings could be gained by the total elimination of'PG from ES casing. Alternative methods of casing to reduce PG losses or eliminate PG have as follows: A. Casing of ES after expansion B. Casing at superwet cylinder C. Elimination of PG from casing The thirdalternative was tested'most extensively, but each was analyzed for casing recoveries in addition to physical and subjective effects. Attributes monitored to avoid alternations of the final ES product are listed below: L

Page 2 F U F I i i I I L 1. Flavor differences (Subjective): 2. Equilibrium moistures 3. Cylinder volumes Firmness data Sieve fractions 6. Casing recoveries Alternative methods of casing to reduce PG losses were evaluated prior to tests of PG elimination, and the main objective was to assure equality of acceptance through testing. In summary, the most feasible solution was the elimination of PG from ES casing.. Test results which support this conclusion are as follows: 1. PG elimination did not significantly affect the O!.V., 2. C.V., sieve fractions, nor firmness of ES product. PG elimination did not significantly affect flavor of ES in Marlboro cigarettes. Casing after expansion produced clumping problems in ES product. The cost savings are best for PG elimination com- pared to other alternatives tested. II. Experimental Procedure, Results and Discussion A. Casing of ES after Expansion 1. Processing Results a. Run 4~~3'8, versus~, 43'9! i

Page 3 F I I- r Pilot Plant test run 438 was designed to produce uncased ES which was cased with the normal level of casing upon exiting the 8"' tower. Casing recoveries were closer to theoretical than for the control run 439 which was cased prior to rolling and cutting._ PG was detected at 1.7% on samples cased after expansion compared to only 0.5°l on control cased ES. Observations'during processing indicated that the ES I i I I 1 I I ~ I i ~ L cased after expansion was slightly sticky and tended~to clump worse than normal. Since ES is usually bulked or stored before blending, the clumping and possible moisture problems would need to be resolved. 2. Subj ective Results a. Run 438 versus 43'9' Flavor Group screening indicated no obvious differences between cigarettes with 10% control ES and!10%o ES cased after .expansion. Booth testing of these candidates showed only a slight tendency for the experimental to be more spicy (Attach- ment I). Further testing on Rl-3515 matched control ES at 6.75% in the Marlboro ESS-2 blend to the same level of ES cased~after expansion. Analytical smoke data (Attachment II) showed no significant differences. Subjective responses were similar except the combined panel showed a tendency to judge the control more like the control.

r U U Ii f I L L ~ L L L L x Page 4 These subjective responses and!the clumping problem were probably the consequence of casing at the full level after expansion.'Spraying of ES at a lower component level based on normalYrecoveries was proposed, but not tested due to processing problems. B. Casing at Superwet Cylinder 1. ~ Processing Results Earlier tests had shown that location of casing applica- tion did not affect degree of expansion~nor level of casing on ES (Turano, 1972). However, it seemed logical that casing, losses would be higher for stems cased before rolling and cutting since cased~stems lost prior to superwetting would contribute to overall losses. Manufacturing Service&initiated testing of casing application at the superwet cylinder in 1974 (Madures). Flavor .Development andISEF testing of the experimental ES showed no obvious flavor differences. . Extended runs at the Louisville ES facility of super applications have shown several advantages (Madures, 1975). In~addition to processing and quality improvements, a casing savings of 8-10% was reported with respect to PG recoveries. a. Run 6281-A versus 6282-A The effects of the point of application for the experimental casing replacing PG wex e studied on Pilot Plant and Pilot L

r I U I r I I I.. i L L i L i L x Pa ge 5 Primary modified equipment. Casing before rolling was com- pared to superwet application in respectively. Blended stems were cased at with experimental casing, bulked the Pilot Plant (6282-A). In the superwet application additional spray nozzel was used test 6281-A. tests 6282-A and 6281-A, the conditioning,cylinder overnight, and rolled~in of experimental casing an and the water content of the casing was considered in calculating the superwetting spray rate. Stems were sprayed with~water in the conditioning cylinder and other steps were the same as the control in D ifferent oven volatile levels were obtained on the superwet RKS (Table 1). The superwet application was on target (45% O.U.) while the control cased was around 40%Q O.V.; however, the effects on expansion were minimal since the corrected cylinder volumes were slightly higher (50.9 versus 49.1 cc/10 g)) for ES cased at the conditioning cylinder. Slightly lower glycerine and reducing sugar recoveries were observed on the application before rolling possibly indicating that the spray rate and/or absorption were lowPr. The variable feed rate into the expansion tower (due to hand feeding) and the resulting product O.V.'s caused concern, but overall, no~signficant physical differences could be i

r r F U F F I I L ~ It attributed~ to the Page 6 point of experimental casing application. Flavor Group screening of 100% ES cigarettes from these processing tests showed~no obvious flavor differences. Therefore,'subsequent processing tests of casings were planned~ with applications during superwetting. 'b. Run 6790 versus 6793 PG elimination was tested in the next series of casing tests (6790 and 6793) in combination with different DAP levels. Casings and DAP in the superwetting solutions were sprayed separately. Data in Table 1 (Attachment III) shows that the superwet RKS was not significantly different in "a&is"'nor equilibrated O.V. for the tests with and without PG within the 0, 2.0, and 3.5% DAP levels. Cylinder volumes on equilibrated feed samples were 32.2 for RKS with PG and 32.6 cc/10 g for samples without PG.. Equilibrated expanded samples gave similar values for O.V., C.V. and sieve fractions whether cased with or without PG. Differences in O.V. and C.V. were only observed between the 0, 2.0, and 3.5% DAP levels in Table III andlIV (Attachment Analytical recoveries showed the control ES with 1.8% L L glycerine, 0.7% PG and 13.8% reducing sugars compared to experimental with 2.1%, <0.1%, and 13.9%, respectively. Losses of PG in the tower amounted to around 50% of that applied to the control. Values for casing components were questionable L

V iT r F ( I t i ~ I L l . Page 7 when comparisons were made between superwet RKS and ES. Reordering of superwet RKS down to around 15% O.V. in blow boxes resulted in losses of PG from control samples. Casing recovery studies were initiated at that time to resolve these problems and to define the effects of moisture level, matrix, and conditioning procedures. These studies are continuing ; in the'Analytical Services GC section. 2. Subjective Results a. Run6281-A versus 6282-A; 6790 versus 6793 These ES sample pairs were compared~at 6.75 and 10.0% levels in 1975 and 1976 Marlboro blends. Flavor Group evaluations showed no obviou~s differences in flavor due to elimination of PG from the casing. Larger scale testing waa planned for the next phase which~involved more extended runs on superwet applications of casing without PG. C. Elimination of PG from Casing 1. Processing Results a. Run 495 versus 464 Conventional casing procedures were followed for these runs which were cased in~the conditioning cylinder prior to rolling. Blended Bright stems were used £or both runs, but a different blend of stems was probably inadvertently processed on Run 464. Sieve fractions showed 12% higher longs and a:proportionately lower per cent mediums on the ES cased'without PG. These values were similar to the values from I