Tobacco Documents Online

SECTION I VARIOUS ASPECTS OF MENTHOL PRODUCT DEVELOPMENT By: D. V. Cantrell SECTION II MENTHOL MIGRATION STUDY By: Hoechst Celanese Corporation SECTION III EFFECT OF AGE ON TIIEMENTHOLAND TRIACETIN DELIVERY OF MENTHOLATED CIGARETTES By: Eastman Chemical Company, Eastman Kodak Company SECTION IV MECHANISMS O} ~NTHOL DE~Y IN FILTER CIGARETTES By: Eastman Chemical Company, Eastman Kodak Company 5S4 .0G 23

r,9 c) ID l SF~ION I

PRODUCT DEVELOPMY.NT REPORT i'" VARIOUS ASPECTS OF ~q~l~OL PR..ODUCT DEVELOPMENT by : D.V. CANTRELL This document summarizes various issues which are normally considered when conducting menthol cigarette development. Five important issues are discussed, including leaf blending, casing additives, cased leaf drying, desired menthol perception and flavorings, and menthol application systems. While this information represents ~W views, the direction practiced regarding each issue may be dictated by consumer's taste preferences and/or manufacturing limitations. SUMMARY OF U.S. MENTHOL MARKET The U.S. cigarette market volume in 1990 was 522 billion cigarettes. Of this, menthol cigarettes accounted for 26.7% of the market. The major brands which make up the menthol market are Salem (6.1%), KOOL (4.9%), Newport (4.7%), BELAIR (0.6%), and Alpine (0.6%). All are free-standlng menthol brands, i.e., they don't have a non-menthol companion. The remainder of the menthol market consists of mostly menthol companion brands of popular non-menthol brands, such as Marlboro, Benson & Hedges, Virginia Slims, Dotal, CAPRI, and RICHIAND. Over the last 15 years the menthol segment has been declining slowly, from a high of 33% market share in 1975 to the present share of 26.7%. MENTHOL PRODUCT DEVELO~ While there are many issues to address during menthol cigarette developmento usually the same four to five issues will always be examined during a menthol development project. These are: I. Leaf blending II. Casing additives III. Cased leaf drying IV. *Desired menthol perception and flavor character V. Menthol application systems Each issue will be discussed separately, however, items l-IV should be considered collectively since together they act to define a menthol cigarette's taste properties. Item V, menthol application systems, is the only issue which is not dictated by consumer's taste preferences. Rather, it is an issue determined by manufacturing conditions/limitations. To keep this discussion simple, any mention of menthol refers to the "L" isomer. The "D" isomer is not used since it lacks the minty character which U.S. menthol smokers highly desire. S 410012

I. LEAF BLENDING The leaf blend should not only provide the customer's desire for tobacco taste, but should also be synergistic or at least compatible with the clean, fresh taste of menthol. A blend direction with a Flue-cured skew is recommended, eg., a 70/30 to 55/45 flue/hurley lamina blend ratio~ If available, the Flue-cured and Burley Erades of choice and use ranges are shown below in Table I. TABLE i FLUE-CURED AND BUELE¥ CRADES AND EECO]'D, IT..,NDED USE RANGE GRADE "C" "L" "T" FLUE- CU~.ED I~UP,.L-~, RANGE (%) GRADE RANGE (%) 7" 14 "X" O- 2 12-14 "B" 12-18 9-13 "R= 2-4 "~3" 0-3 The use of Oriental tobacco is recommended but should not exceed 15%. A preferable range is 6-10%. For best overall taste in a menthol product, we do not recommend using the best quality oriental grades having the most aromatic character. We find these grades, like CAB, offer greater benefit in a non-menthol blend. The =A= grades frequently used are: YAB, UAB, TAB, and BAB. Since Oriental taste should not be dominant, the use of "B" grade Oriental (eg., YB and TB) can be used to dilute the cost and aromatic quality of the "A" Oriental grades. The practice of blending with medium to high levels of Flue and Burley flavor grades also help dilute the Oriental aromatic taste. The reason for keeping the Oriental level low is that the =woody, cedary" notes are not totally compatible with menthol taste. The two remaining blend components for discussion are the types of by-products, reconstituted tobacco (recon) and stem (WTS), used for menthol product development. Historically, B&W has practiced using PJS or ART (examples of non-ammonlated recons) and~/TS (a blend of flue and burley stems) in their menthol products. The levels of PJS and~T/S in blends have been 10-14% and 12-15%, respectively. While this practice still continues for some menthol brands, our newer products are using around 6% all flue-curedWTS and either P3S/ART or EBR at around 17%. With respect to recon, we have experimented with the use of ammoniated recons (EBR and CPCL) and found them to be beneficial among certain menthol smokers, especially smokers of Philip Morris -2-

products. Relative to PJS, EBR and CPCL offer less irritation but also a slightly dirt-y/ammonia taste. When using EBR/CPCL the type of stem used is flue-cured WTS. Obviously, the type of recon used in development projects will mostly be dictated by the target consumer. An excellent example of this is Philip Morris's recent change in the type of recon used in Alpine and Marlboro Menthol from ammoniated to non-ammoniated. This was done to attract KOOL smokers. If stem is a consideration in rahe blend then, we recommendusing flue-cured stem. It offers a cleaner taste than burley stem. As in B/~W's case, availabillt-ymay dictate the use of both flue and burley stem in certain brands. The last blend topic concerns expanded tobacco (ET). At B&W, we practice expanding a blend of 50/50 flue-cured and burley. Since our expanded leaf is cased only with glycerine, we limit the ET levels to around 10% of the blend for Full Taste and Light~ products. Ultra low tar products usually contain greater than 10% to enhance mouthful of smoke. II. CASING ADDITIVES Casing additives are used for two purposes: 1) as humectants to maintain moisture and help plasticize the leaf to reduce break-up during primary processing, and 2) as taste modifiers to enhance tobacco smoke quality or to ameliorate harshness and irritation. Examples of additives used for these two purposes wlthusage levels are presented in Table 2. TABLE 2 CASING ADDITIVES (level represents % in final cigarette) HUMECTANTS/PLASTICIZERS LEVEL Propylene Glycol I-2 Glycerine 2-3 Sorbltol 0~5-i Sugars: sucrose 1-2 corn syrup 1-2 honey 1-2 invert syrup I-3 (%) TASTE MODIFIERS LEVEL (%) Lactic Acid 0.5-1 Cocoa powder 0.2-0.5 Licorice 0.5-1 Chocolate Liquor 0.2-0.5 - 3 -

The level of casing materials such as sugar depends on the grade types of flue-cured and burley used. With the use of good flavor grades and tip grades the level of sugar applied is low (eg., 25-50 k8/1000 kilos of flue-cured leaf, and 100-150 kg/1000 kilos of burley leaf). For less flavorful burley leaf which is low in quality, higher levels of sugar may be needed. Flue-cured leaf which is low in nicotine, high in sugar, and overall low in quality should not be cased with further sugars. The recommended sugar for flue-cured is corn syrup or honey, while r_he sugars for burley are a combination of sucrose, invert syrup, and/or honey. The additives of cocoa and licorice are used only when the burley ~o~c ~ .~ is to be redrled. While lactic acid can be used to ameliorate / ~ irritation and harshness, it can also be applied in the final flavor step. Chocolate liquor is very high in butterfat (-50%) which is effective in reducing harshness of r_he burley tobaccos and provides body to the smoke. If the hurley is to be redrled, then sugars like sucrose and invert syrup are applied prior to redrying in a ratio of 70/30, respectively, at a combined level of i00 kg/1000 kilos of burley. The redried leaf is then given a final casing which normally contains honey and invert syrup in a ratio of 25/75, respectively, at a level of 25 kg/lO00 kilos of redried burley leaf. The oriental is normally cased with the flue-cured leaf. The amount of oriental is not however considered when calculating the amount of casing additives for the ! flue-cured. So, in effect, the oriental dilutes the amount of casing applied to the flue-cured leaf. This is quite acceptable. If possible, we would recommend casing the recon with either the flue-cured or post redrled burley leaf. An adjustment in casing amounts should be made to account for the dilution effect of the recon. While B&N does not practice post-caslng the ET or NTS, this is certainly acceptable for certain ET or NTS blends. IIl. CASED LEAF DRYING Cased leaf drying refers to redrying r_he burley leaf blend. B&W has menthol brands which have the burley redried and brands which do not. The majority of our menthol product sales do not use redried burley. The reasons we don't redry certain menthol brands are: . . ° 4. Redrying produces Maillard type reactions giving rise to roasted/toasted notes, which cause an incompatible taste character with the menthol. By using good flavor grades of hurley, the amount of burley needed in a blend to give acceptable air-cured taste is less if it isn't redried. Redrying subdues the robust strength of burley which in some brands "robustness" is a desired attribute. Limited redrylng capacity. - 4 - 5841001 S

The brands which do utilize redrled burley are newer, and are targeted at smokers of co~petitive menthol brands which also utilize redried burley. Besides the effects of redrying mentioned above, redrylng is effective in reducing harshness and irritation. However, one must be able to accept the modified tobacco taste as a result of redrylng. This is dlffigult to do when trying to improve an established menthol brand with a large franchise of loyal smokers. The best opportunity to use redrylng is in a new product having no pre-established smoker base. An example of this is KOOL XL for Japan where the menthol market is small but has strong potential for growth. IV. DESIRED MENTHOL PERCEPTION This issue covers two topics, one being the type of menthol taste desired and the other being the level of menthol cooling desired. Type of Menthol Taste; While there are basically two types of menthol, natural and synthetic, we break them down further because of ~helr dlfferen~ taste properties. The natural menthol types available are: I. "Brazilian" 2. Chinese 3. Indian 4. Australian - currently produced predominantly in Paraguay. Each of the above natural menthol types has a distinctly different taste character in cigarettes. These differences are described below: Brazilian Chinese Indian Australian - gives a mint~j, clean, crisp cooling effect. gives a musty, slightly dirty, camphoraceous cooling taste character. gives a variable degree of mintiness and has a green, herbaceous cooling taste character. - gives a sweet, eucalyptol cooling taste character. As one can see, the type of menthol taste can be altered distinctly by using different types of natural menthol. The reason for r_he taste properties is due to the different levels of minute quantities of impurities in each type. Obviously, these chemical differences are a function of the various soll and climatic conditions, cultivation, and distillation techniques practiced in the production of menthol throughout the world. - 5 -

The other type of menthol used within the industry is synthetic. We know of two major suppliers of synthetic menthol. These are Haarmann & Relmer, a German Co. with production facilities in Charleston, South Caroliru~-USA and Holtzmlnden, Germany, and the second being Takasago, a Japanese Co. Currently, B&W purchases synthetic menthol from Haarmann & Reimer. Their product was rated better than Takasago's, and we have a better working relationship with Haarmann & Reimer. The difference between synthetic and natural menthol is that synthetic is essentially pure L-menthol (99.95a). Natural menthol, on the other hand, has mint impurities which produce a distinct taste difference. Natural menthol is usually 99.5% pure L-menthol. In 1987, B&W researched different ratios of natural and synthetic menthol for the purpose of learning which ratio offers the ~est menthol taste properties. The result of the research established that the best tasting menthol smoke quality was between ratios of 80% natural/20% synthetic and 60% natural/40% synthetic. We chose the midpoint of 70% natural/30% synthetic. Upon establishing r_his by subjective panel testing, we theorize that the addition of synthetic menthol acts to dilute the natural menthol impurities to a more optimum level for menthol smoke quality. We also know from years of GC analyses that the natural menthol impurities vary in level from one year to the next. So by using synthetic, the cigarette's menthol taste properties are more consistent because the synthetic acts to moderate the inherent variation in natural menthol. Level of Menthol CoolinK Desired: The level of menthol cooling is predominantly a function of the menthol content within the cigarette. Levels discussed in this section are expressed in %. The percentage value correlates to the measured amount of menthol analyzed within the total cigarette (tobacco, paper, and filter) divided by the amount of tobacco in the cigarette times I00. Factors which affect menthol cooling are: i. Menthol content 2. Blend 3. Cigarette tar and nlcotine delivery 4. Flavor additives 5. Filter plasticizer o 6 -

Menthol Content The most slgniflcant factor affecting menthol cooling is the cigarette's menthol content. Examples of this are shown in Table 3, below. Note also the impact of the cigarette's tar and nicotine delivery upon menthol delivez~. TABLE 3 TAR, NICOTINE AND ~JTTHOL DELIVERY VS. CIG. MENTHOL CONTENT SMOKE MENTHOL TAR NIC MENTHOL # of KOOLKS 0.42 17 1.2 0.50 6.8 K00L X L KS (Japan) 0.34 9 0.8 0.24 7.6 K00L MILDS KS 0.52 12 0.9 0.50 7.3 K00L DELUXE LTS KS 0.70 9 0.8 0.50 7.8 NEWPORT KS 0.34 17 1.3 0.~ 7.2 SALEM KS 0.42 18 1.4 0.58 7.3 SALEM LTS KS 0.58 8 0.7 0.40 6.9 ALPINE KS 0.42 17 I.I 0.48 7.3 N0W MENTHOLKS 1.90 1 0.2 0.15 6.3 For a full flavor product, the menthol contents range from 0.34 to 0.42%. These represent the typical menthol ranges in the U.$. currently. There is a significant difference in the menthol cooling perception between K00L and Newport, with Newport being lower in menthol cooling sensation. For markets desiring to introduce a menthol product where the current smoker dominance is non-menthol, we recommend a =llght" loading of menthol. Examples of "light" loadings would be: • Full Flavor (14+ tar delivery) a level of 0.25 to 0.30%. • Lts. (7-13 tar delivery) a level of 0.35 to 0.40%; eg. KOOLXLJapan. • Ultra (3-6 tar delivery) a level of 0.45 to 0.50%. "Light" menthol loadings are recommended since "high" menthol loadings will usually be quickly rejected by non-menthol smokers. Converting non-menthol smokers to menthol is not easy to accomplish. It is, however made simpler if the product offers a low level of menthol taste and cooling. Regarding menthol loading, typically menthol is increased as car decreases to maintain somewhat of a consistent menthol delivery. A good example is KOOLKS and KOOLMILDS KS. As shown in Table 3, the menthol loading is 0.10% greater in KOOLMILDS KS @ 12 mgs. tar than KOOLKS @ 17 mgs. tar, yet both have the same menthol delivery of 0.50 mgs. menthol/clg. The reason this is done is mostly dictated by consumers; i.e., full taste smokers who move down in tar still want.the same level of menthol cooling. $ - 7

For planning purposes, keep in mind the following information recently learned in a menthol study performed by B&N. It showed that i0 to 20% of U.S. menthol smokers can detect a relative difference of +/- 10% in a cigarette's menthol level. For example, at a level of 0.42% (llke KOOLKS) 10 to 20~ of regular menthol smokers can detect a difference in menthol cooling if its menthol content varies from 0.38 to 0.42% or 0.42 to 0.46%. Therefore, for consumer testing, we recommend that the menthol level be within +/- 5% of target before testing. However, for normal production, this target is too stringent. Blend The type of blend used can also affect menthol delivery. This occurs for two reasons. The first is that different blend components have different affinities for menthol, and the second reason is the blend components have different tar and nicotine deliveries. This latter reason will be discussed in more d~r~ail in the next section. The affinity tobacco has for menthol differs by tobacco type. In this case, "affinity= denotes the activity/ of absorption and desorption of menthol. The general rule is that tobaccos which absorb menthol more rapidly also desorb menthol more rapidly. The factor which controls this activity is the leaf's surface struct~ure. For example, oriental and flue-cured leaf have a reslnous/waxy leaf surface which acts to slow absorption of menthol, and once absorbed it is slow to release menthol. Burley tobacco, recon and stem, having less coating on the surface absorb menthol much faster and also lose menthol more rapidly. While this phenomenon occurs, it is mostly nullified by the tar delivery potential of each blend type. For example, even though flue-cured has a lower affinity for menthol than stem or recon, it's menthol delivery will be greater due to flue-cured tobacco having a greater tar and nlcotlne dellverypotential. Cigarette tar & nicotine delivery Ignoring for a minute a cigarette's design parts (filter, paper, and ventilation) the tar and nicotine deliveries are a function of the blend components used. It is also the tar and nicotine which act as a major driving force for menthol delivery. A good example demonstrating this point is the data in Table 3 for KOOLXLKS (Japan) and Newport KS. Both products have a menthol content of 0.34% but KOOLXL delivers 9 m8. of tar and 0.24 m8. of menthol while Newport delivers 17 mg. tar and 0.44 mg. menthol. The menthol/tar delivery ratio is essentially the same for both products, KOOLXL- 0.24/9 - 0.027 and Newport - 0.44/17 - 0.026. This shows that there is close to a one-to-one reletlonship with menthol delivery and tar delivery. Meaning that with~enthol content held constant, the menthol delivery w~l increase proportionally as ta~ increases or vlce/versa.~While no nicotine data are included in Ithe table, the same relationship holds true for nicotine and men~ol delivery 8