Tobacco Documents Online

BROWN & WILLIAMSON TOBACCO CORPORATION PROCESS DEPARTMENT RESTRICTED PGS-B072-d4 PET PROCESSING OF AMMONIUM HYDROXIDE TREATED TOBACCO Hr. E. E. Kohnhorst Nr. M. L. Reynolds Dr. J. N. Jewell Mr. K. E. Philpot Dr. J. G. Esterle Mr. R. F. Denier August 17, 1984 Project 525 DISTRIBUTION Dr. B. B. Chakraborty Hr. T. F. Riehl Mr. R. H. Roneycutt Mr. D. L. Gordon Mr. P. R. Fisher ABSTRACT Dr. D. A. Colby Dr. E. F. Litzinger Dr. L. C. Chao Mr. R. H. Marshall Library (g) Drying of flue cured grade CZFC having three different levels of ammonium hydroxide applied was pursued using the PET dryer. This note presents the results of this study and provide~ direction for future experimentation with the use of ammonia. Specific results from this study show: PET drying of a~onium hydroxide treated tobacco, at S0 psig, 2D ~econds residence time shows favorable decreases in reducing sugar and specific sugar levels i~dicatlng Occurrence of the Maillard reaction. Unlike ammonia gas impregnation, OC P&T analysis indicates PET processlng with ammoni~un hydroxide at the levels used in this test does not increase acetic acidj free nicotine, or acetamide, all of which are indicative of Marlboro-type products. The N~4OH/PET treated samples were perceived as having more tobacco taste wit~ lower impact and ircltatlon to a non-tobacco control by expert smoker aaalysis. Operating PET at 50 psig generated alkaloid losses of 50-60% which may have influenced these results. Future work with PET processing will use reduced dryer residence time to linimize alkalold losses on AMELIA all flue-cured lamina while sustaining the NK~/sugar reaction. NH4OHIPET and NB3(g)IPOP processlng of all flue-cured SDS samples is also planned to try and take advantage of the high sugar levels of all flue-cured stems. GZ00'71 94

INTRODUCTION A previous study* using ammonia gas has shown that pressurization with steam, when applied after a~monla gas impregnation, will increase the conversion percentage of reducing sugars to Maillard-type reaction products. It is hypothesized that Maillard-type reaction products favorably improve clgarette flavor, while reducing impact and irritation, [t was theorized that if a favorable reaction chemistry is achieved with a short reaction time medlumsteam pressure condition, drying in a PET dryer using a steam pressure of 50 psig may generate ~imilar types of reaction products. Small-scale testing on the POP unit using a~m~onia gas impregnation followed by pressurization with steam has ~hown that both reducing sugar, and specific sugar levels decrease with increasing steam pressure. Reducing sugar losses are a measure of the extent to which favorable ammonla/sugar reaclion products are being generated. Lengthening reaction time increases the percentage change in reducing sugars; however, a blackening Of the tobacco occurs rendering the product undesirable for cigarette manufacture. Short reaction times of less than 2 minutes in duration, appear to favorably influence reducing sugar conversion while minimizing color changes. Organelepticafly. higher steam pressures appear to improve tobacco flavor, but increase impact and irritation, while medium pressures have a greater influence on reducing impact and irritation, but minimizing flavor enhanc~men£. ~oth short reac£ion or resldence time and pressures upward of 50 psig are inherent to a PET (Pressure Enhanced Tobacco) drying system; therefore, a logical extension to our a~m.onia studies is to investigate ammonia/sugar reactions in a PET dryer. EXPERIMENTAL WORK Direct injection of a~s~onia gas into a PET dryer will probably result in a release Of armmonia fumes into the area surrounding the dryer. Consequently, it was decided to try to apply the ammonia to the £o5acco in a form that will allow iL Lo be liberated inside the dryer where i~ can react with the tobacco sugars, and gas escape is minimal or no~existen£. Approximately ten-2 lb. samples of a cu~ CZFC tobacco was conditioned to 13~moisture followed by spraying with mixtures OE concentrated ~onium hydroxide and water. Levels of i0, 20 and 30~ ammonia "relative to" reducing sugar levels were utilized with inlet dryer moistures targeted at 21~ after ammomia/water addition. On a wet tobacco weight basis these levels of ammonia corresponded to 5. I0 and 15Z levels of ammonium hydroxide (28~ ammonia content). These three levels ~f ammonium hydroxide were chosen based on concurrent experimentation being performed by Dr. E° P. Litzinger in the area of 6Z00713S, i{ -- g -

am~onla/sugar reaction mixtures. Samples were bulked for three hours prior to drying to insure impregna£1on. A strong ammonia smell was present after bulking and prior to drylng, but no ammonia smell was present upon discharge from the dryer. In addition to drying the ammonium hydroxide sRmples at 50 psig on the PET dryer, two additional samples, with water only, were dried at pressures of 50 pslg and 0 psig on the PET dryer. R conditioned sample of untreated C2FC wag used as a control for comparison purposes. A dryer configuration consisting of six heat exchangers was used foc all PET processing. Shell-side pressure on the heat excha,gers was maintained at 60 psig with superheat temperatures of 385°F bein3 used for the 50 pslg samples, and 385°F being used for the 0 psig samples. Supply steam pressures of 120 psl3 and ~0 psig were used for the 50 pslS and 0 psig sRmples~ respectively. These are identical conditions used during normal, non-a~monia PET drying of lamina. Analysis of samples was per[ormed on an "as is', hasis with results being corrected to a common m~isture content of i~. RESULTS Table I presents chemlcal~ pH and free ammonia results for the C2FC tobacco processed in this study. Ammonia concentration effects are il]ustrated in Appendix Figures A-I and A-3. TABLE 1 PRODUCT PROPERTIES @ 14% MOISTURE AFTER ND4OH/PET TREATMENT A~0NIA STEAM REDUCING CONCENTRATION PRESSURE ALKALOIDS 3UGARS FRUCTOSE GLUCOSE ~HH A~ONIA (~) (psig) (%) (%) (%) (%) (%) UNTREATED NO STEAM 1.86 9.9 3.0 1.8 5.5 .04 NONE (WATER ADDED) 0 1.50 9.6 2.7 1.6 5.& N/A NONE (WATER ADDED) 50 .94 i0.0 3.3 1.7 5.2 .05 I0 50 .RO 9.9 2.4 l.l 5.3 .09 20 50 .75 9.5 Z.& 1.2 5.3 .13 30 50 .73 8.0 2.0 1.0 5.3 .13 6ZOO71 -- 3 -

As shown in Table i, and Figure A I, higher concentrations of ammonium hydroxide show a favorable depletion of reducing sugars indicating occurrence of ammonia/sugar reactions~ Unfortunately, according to GC Purge and Trap analysis, formtlon of increased levels of acetic acid, free nicotine, a,d acetamlde did not occur with the NHgOB/PET process like it did with ammonia gas/POP. The noticeable decrease in free nicotine may be attributed to the heavy alkaloid losses associated with PET processing during this series of tests. There is a need to try to restore higher alkaloid levels in order to fully evaluate the effects on ~ree nicotine. Methods of restoring higher alkaloid levels are addressed in the Future Work section. A greenish color appeared on the tobacco prior to drying with its intensity increasing with increasing levels of applied ammonium hydroxide. This greenish color completely disappeared upon processing through the PET dryer. It is not understood what compounds are being formed to give these observed color changes. Table 2 shows a comparison of sugar and alkaloid losses for ammonia gas impregnation on the POP unit versus am~oniLun hydroxide liquid processes with the PET dryer. "(NHagas) POP" data has been reported previously. TABLE 2 AMMONIA GAg IMPREGNATION VERSUS AMMONIUM HYDROXIDE PROCESSES REDUCING A/~ONIA STEAM ALKALOID SUGARS GLUCOSE FRUCTOSE SUCROSE PROCESS CONCENTRATION PRESSURE LOSS LOSS LOgS LOSS LOSS (%) (psig) (%) (X) (%) (1) (%) 0 21 + 9 + 7 + 16 0 30 23 ii gl 18 0 60 15 33 36 45 0 0 50 49 0 6 + I0 0 I0 50 57 0 39 20 0 20 50 60 4 33 20 0 30 50 61 19 4& 33 . 0 POP/GAS/HED POP/GASIBHD POP/GASIHHD LIQUID/PET LIQUID/PET LIQUID/PET LI~ID/PET As shown in Table 2, alkaloid losses at equivalent steam pressures are m~¢h higher with PET processing. These higher alkaloid losses may result from the relatively long residence time that the tobacco remained in the dryer compared to high humidity drying of POP samples on the Minicon. 620071397

It is known that higher tobacco temperatures increase the amount of alkaloids which are volatilized, and in both the PET dryer and the "POP" impregnator, steam is used under pressure to raise tobacco to the adiabatic saturation temperature of £he steam (-31D~F @ 60 pslg). In a drying system, the gas surrounding a tobacco particle is always changing as the particle is carried through the dryer, thus there is always a concentration gradient for alkaloid removal. At a constmnt tobacco temperature, the longer a particle remains in the dryer, the greater the alkaloid loss. In ~onia gas imgregnatlon~ the tobacco is surrounded by stagnant pressurized steam in which an equilibrium is finally reached where alkaloid concentration in the steam is sufficient to slow further movement of alkaloids into this gas. In order to limit alkaloid losses in a drying system, shorter residence times have been shown to be beneficial. Future PET trials will focus on the use of a short residence time to reduce alkaloid losses. Organoleptioally~ LAgEDO-made cigarettes were evaluated by an expert smoker who detected a difference between cigarettes treated with NH40H and those without. (An expert smoker was used primarily because of the noticeable alkaloid loss (see Table i) which occurred with these samples during drying. It was felt that the difference in alkaloid levels would probably bias organoleptie evaluations.) An untreated control and a s&mple without mmaDnluan hydroxide dried at 0 psig were found to have the most impact and irritation with very little flavor. Different ammonixun hydroxide levels could not be differentiated. This may have been caused by the low alkaloid levels of the 50 psig tobacco which would cause a lowering of impact, irritation, and flavor. Alkaloid levels need to be returned to normal to evaluate the impact of the a[mmonium hydroxide concentration levels. Although all 50 pslg samples were lower in irritation than the two samples just previously mentioned, no distinct differences between samples was noticed for any of the tobacco dried at 50 psig. FUTURE WORK Based on the results of this study, it appears that a shorter residence time is required in order to reduce alkaloid losses while sustaining the armlonia/sugar reactions. Since the majority of residence time is occurring from the inlet and exit airlocks, provisions have been made to increase airlock speed from 7 to g5 RPM which will shorten dryer residence time from 20seconds to approximately 7 seconds. Plans are also being formulated to go directly into the cyclone separator from the inlet airlock (i.e.~ no ehamher) which should shorten the residence time to less than ~ seconds including airlock time. -5-

As part of the ammonia program, an all flue-cured SDS sample will be processed with the NH~OH/PET process and N~(g)/POP process. During testing of SDS, the following evaluations will be performed. High and low residence times on the PET dryer to determine influence on alkaloid losses. 30 and 60~ levels of ammonium hydroxide addition in order to distinguish between amounts of sugar conversion from the Maillard reaction. O, 30, 60 psig steam pressures on the POP unit. In addition, it is planned to evaluate both the flue-cured and hurley sides of the AMELIA blend. These evaluations will be performed on both the PET unit and the POP unit. The flue cured side of the blend will be used to evaluate if the Maillard reaction is occurring predominatefy from natural occurring sugars in the tobacco, while the burley portion will be used to evaluate if the reaction is occurring from artificially applied sugars which Ooeur in the easing. L~EDD cigarettes smoked by experts will be used to deLermlne effects from various combinations of hurley and flue-cured tobaccos at several levels of inclusion. R. F. D. RFD/la 0~4gs Attachments 1. REFERENCE Denier, R. F., "Pressure/Time Effects On Ammonia Impregnated Tobacco", B&W Research PGS B0gT-g&, (5/4/8~). -6-

APPENDIX A 6Z0071400

FIGURE A-I At~ONIA CONCENTRATION EFFECTS IN PET PROCESSINC m r~ NEt/REDUCING SUGAR PERCENTAGE (I.) - A1 - 6Z0071401 ~48s

FIGURE A-2 ~ONIA CONCENTRATION EFFECTS IN PET YROCESSING ~- 50 PS~G A- UNTREATED C2FC 0- 0 PSIC ]liEij ............. ~!!IIIIi ~llIIr~ D ill]i IIJFIIF ..~]T]lIIl )illii i]11ill IIIII~ • i ill 2 ir rlE ~II [I]II ~ll;llJ , ,ITr ,,~,. t I I~'....~ ~ I ~ I ~ ~ I i ~ : ~ ~ ~ ~ i ; i] I , ~,,,,,,v~ I !fill Illl] FRUCTOSE ll[Irl~lPill] lil III]lgiJilllililJll]illl~., ;, Iri'g~'~:"~l~ ]~'IIIIIIIII]FII 11111111111 ~ Tllllll}llllLIllJllfJJlJllllllll]irll JlllllllJII J! i 0 I0 ~0 30 I~IRED~ClNG SUGAR ~ERCENTAgE (If - A2 - o,.,.8. 6~007140Z

FIGURE A-3 /~ONIA CONCENTRATION EFFECTS IN PET PROCESSING 0 r~ 2.0 il.5 "~1.0 i~ 0 I0 EO N~3/REDCCING SUGAR PERCENTAGE (%) ~- 50 PSIG ~- UNTREATED C2FC ~- 0 PEIC 3O -A3- 620071 03 D448s