Jump to:

Product Design

[Acid Treated Cigarette Filters]

Date: 16 Oct 1992
Length: 37 pages
509687543 -7579
Jump To Images
snapshot_rjr 509687543-509687579

Abstract

Consists of multiple documents: (1) Description of exploratory study of acidic filters and "the mechanism of mainstream smoothing with acid treated filters"; and (2) various memos regarding trials of "Acid blends on PET web". Includes, in study report, sections entitled: "Objective; Summary; Status; Keywords; Motivation; Introduction; Preparation of filters/cigarettes; Analytical procedures; Results; Discussion; Standard analyses; Special analyses; Filter length studies; [and] Conclusions". Describes new "procedure for manufacturing acid filters" resulting in "smoothing effects on low T/N smoke that correlate with substantial reductions in 'volatile' nicotine but only slightly reduced total nicotine".

User-Contributed Notes

Fields

Author
Blakely, D.B.
Ingebrethsen, Bradley James (RJR Scientist)
Lyman, C.S.
Perfetti, Thomas Albert, Ph.D. (RJR Flavorist)
Saintsing, B.L.
Recipient
Blixt, C.A.
Burger, W.M.
Ehmann, Carl W
Hildebolt, William M.
Defense
Hardin, Robert V. (RJR R&D Technical Support Staff VP 1994)
Defense
Hayes, Andrew Wallace (RJR, VP Biochemical/Bio)
A. W. Hayes was Director of Biological Research in 1987. VP of Biochemical/Biobehavioral Research & Development 1990.
Hein, Carl C., III (RJR VP & General Manager 1994)
Defense
Lloyd, Robert A. Jr., Ph.D. (RJR New Business Development VP)
Defense
Pugh, Dan R. (RJR Process Tech & Dev. Director, 1987)
Defense
Simmons, William Samuel, Ph.D. (RJR Smoking & Health Director)
Stowe, Mary Evelyn, Ph.D. (RJR Applied R&D Director)
Defense
Willard, Ron L. (RJR New Business Development VP, 1994)
Defense
Hypothesis
Free Nicotine
FTC machine testing and ratings
Design changes to achieve altered FTC smoke machine tar and nicotine ratings, with or without measured changes in human intake.
Use of filters, paper, and ventilation
Modification of tobacco products through use of filters, paper, and ventilation, and measuring effects on dependence, behavior, and toxicity.
Nicotine transport, transfer, and uptake
Design changes which alter nicotine delivery or effect how the product causes and maintains dependence, including transfer of nicotine from tobacco to smoke, and uptake into the body.
Mainstream constituent yields
Modification of selected mainstream smoke constituents in response to health concerns.
Measuring human smoking behavior
Measuring the effects of changes in human smoking behavior on intake of nicotine and smoke constituents.
Sensory effects
Technologies used to measure, control, or alter sensory effects
Keyword
Adsorption
Substance held inside another by physical bonds
Volatile nicotine
Total particulate matter (TPM or Tar)
WTPM
Additive
Acids
Citric acid
glycerol
Lactic acid (Lactic Acid and dl-Lactic Acid)
Levulinic acid
Polyethylene teraphthalate
Potassium hydrogen phthalate
Phosphoric acid
Sulfuric acid
Smoke Constituent
Nicotine
Design Component
Acid treated filters
Web acid filters
Burley tobacco
Nicotine transfer efficiency (NTE)
G7-1
Top dressing (Top flavor)
Cellulose acetate filter (CA filter, Conventional filter)
B-3
PET filters
Operation/Project
Project XB
Project designed to test acceptability and smoke contents of experimental low-tar cigarettes
Project 113
Named Organization
Tennessee Eastman
Federal Trade Commission (Enforcement agency for laws against deceptive advertising)
Enforces laws against false and deceptive advertising, including ads for tobacco products. Ensures proper display of health warnings in ads and on tobacco products;collects and reports to Congress information concerning cigarette and smokeless tobacco advertising, sales expenditures, and the tar, nicotine, and carbon monoxide content of cigarettes.
Technology/Method
STT tobacco extract
FTIR
Gas chromatography
Subject
aerosol (technology)
Effects—Smoking Behavior (Effects)
Filters (Design)
Length (Design)
Levulinic Acid (Additives)
reduces the harshness of cigarettes
pH Manipulation (Technology)
Puff Count (Measures)
Sensory Effects—Taste (Effects)
Smoke Deposition (Measures)
Smoothness/Harshness (Effects)
T/N Ratios (Measures)
Tar (Measures)
Test/Consumer Preference (Testing)
Test/Smoke Machine (Testing)
Test/Toxicity (Testing)
Transfer to Smoke (Measures)
Ventilation (Design)
Particle Size (Technology)

Document Images

Text Control

Highlight Text:

OCR Text Alignment:

Image Control

Image Rotation:

Image Size:

Page 1: dha73d00
AUTHORS: Bradley J. Ingebreth Cynthia _ S,. Lyman Barry L. Saintsing Thomas A. Perfetti Dallas B. Blakley DATE: October 16, 1992 DEPARTMENT:Biochemical/Biobehavioral (771) Brands R&D (782) Materials Development (773) Packaging Technology (783) DIVISION: Biobehavioral (555) Brands (582) Tobacco Ingredients (572) Packing Applications Development (570) PROJ. NO./NAME: 113/XB NOTEBOOK PAGES: NONE REPORT NO.: BIOBMM 92-020 PREVIOUS REPORTS: BIOBMM 91-013 NO. OF PAGES: 38 91-014 91-015 ACID TREATED CIGAREITE FILTERS OBJECTIVE: To describe the motivation for investigating acidic filters, the exploratory development work, and the research toward understanding the mechanism of mainstream smoothing observed with acid treated filters. SUMMARY: Studies of 'volatile' nicotine suggest that acidic filters may be a means to smooth low 'tar' to nicotine, T/N, smoke that avoids some of the difficulties that have been encountered with blend acid addition. We have developed a procedure for manufacturing acidic filters involving the coating of a chemically resistant fiber web with acids and acid solutions and the making of filters from the coated webs. Coatings of citric, lactic, and levulinic acids and solutions of these acids in glycerol have been found to be the most promising. Informal smokings of cigarettes made with these filters indicate smoothing effects on low T/N smoke that correlate with substantial reductions in 'volatile' nicotine but only slightly reduced total nicotine. A variety of standard and special analyses were made on acidic filter prototypes and a working hypothesis for the mechanism of 'volatile' nicotine reduction has been developed and tested against the measurements. We believe that all the acidic filters adsorb vapor nicotine and that this 1
Page 2: dha73d00
results in reduced 'volatile' nicotine. The reasonableness of this hypothesis is discussed in detail. Additionally we. believe that the volatile acids tested, lactic and levulinic,. elute acid into the'smoke, fiirtlier reducing 'volatile' "nicotine: VVe -found 'that the addition of glycerol to the filter coatings reduces the transfer of the volatile acids and prevents the bulk dispersion of citric acid powder. The citric acid coating provides a smoothing effect without addition of acid to the smoke. Measurements with acidic filters of varying length revealed an interesting leveling off of T/N at long filter lengths that may lead to a means to quantify total adsorbable nicotine in smoke. STATUS: This project is ongoing. SED evaluations of several acidic filters vs. low T/N prototypes made with blend acid additions are pending. KEYWORDS: ACID FILTERS, ACIDIC FILTERS, WEB ACID FILTERS, WA, XB, CITRIC, LACTIC, LEVULINIC, VOLATILE NICOTINE, SMOOTHNESS, TAR TO NICOTINE RATIO, T/N, VAPOR NICOTINE, NICOTINE ADSORPTION, SMOKE pH, GLYCEROL, POLYETHYLENE TERAPHTHALATE, VAPOR NICOTINE, NICOTINE ADSORPTION
Page 3: dha73d00
MOTIVATION: The XB project objectives are to provide a good-tasting cigarette with maintained nicotine and reduced 'tar', i.e. a lowered 'tar' to nicotine ratio, T/N. The harshness associated with low T/N smoke in prior XB prototypes has been smoothed by the addition of organic acids, mainly levulinic acid, to the blend. The latter approach has not resulted in an outstanding product in part due to the overall lowering of tobacco taste and the introduction of off tastes resulting from the addition of levulinic acid to the blend. Our work on evaporative deposition suggests that a smoothing of low T/N smoke may be achieved without addition of acid to the smoke by stripping nicotine vapor in the filter with an acidic adsorbent. The elimination or reduction of acid addition to smoke may offer advantages in the form of reduction of off tastes and/or simplification of toxicological concerns. INTRODUCTION: We have prepared filters and configured cigarettes in analogous fashion to the STT dual segment filter design (report/patent titled "Cigarette and Cigarette Filter Element Therefor", Docket #BB-118-R&D) using acids and acid solutions in place of the STT tobacco extract. Eastman PET (polyethylene teraphthalate) 4SW non-woven web is used as the base material which allows substantial loadings of relatively strong acids to be applied due to the unique pore structure of the 4sw fibers and their chemical resistivity. Several types of acid coated filters have been found to have significant smoothing effects on low T/N smoke. While exploring acid type and loading levels in a development'mode, we have at the same time attempted to understand the mechanism of smoothing through a series of experiments. This report largely deals with the latter efforts. PREPARATION OF FILTERS/CIGARETTES: Three coating runs will be described, each of which yielded several filter types. The PET web samples were coated with acid solutions in Plant #641-1 on a Faustel laminator. The solution coating levels were determined by laminator cylinder groove size, the solution acid concentration, and the affinity of the solution for the fibers. Tables 1-3 list the coating weights achieved in the three runs.' In all Tables of this type the column labeled "Solution" lists the coating solution description and in some cases the acid and glycerol concentrations of the solutions. The column marked "Weight" lists the gravimetrically determined mass % loading of solution on web. In Table 1 the solution concentrations were 7% by weight for potassium hydrogen phthalate and 29% by weight for all others. In Table 2 the solution concentrations are as indicated in weight %. In Table 3 the concentrations are the indicated mass amount of solute in 13 kg. of solution. 3
Page 4: dha73d00
Filter rods were made from 9 inch wide rolls of web on the gathered web filter maker in the R&D pilot plant. Thee rods were made with a non-porous.-plugwrap. _.. Cigarettes'were hand made by reinbving tlie filteis froai made cigarettes, insettirig the acid coated filter segment (usually 12 mm length) next to the tobacco rod, and placing a segment of CA filter at the mouth end. A variety of tobacco rods were made in this way for informal smoking evaluation, but all the analytical data reported here were obtained with tobacco rods of a single type described in Table 4 and made in a 0% filter dilution configuration. ANALYTICAL PROCEDURES: In addition to standard FTC 'tar' and nicotine, smoke pH, smoke levulinic acid, and 'volatile' nicotine, two specialized measurements were developed for this project. Nicotine vapor adsorption: This procedure provides a comparison among filters of the efficiency of nicotine vapor adsorption from nitrogen passed through filters. This measurement does not directly predict nicotine adsorption rates during smoking because of the presence of the aerosol during smoking vs. nicotine in nitrogen in this measurement. Even so, the comparison provided by this measurement is useful since the factors that determine the vapor nicotine adsorption, affinity of the coated filter surface for nicotine and its capacity to hold nicotine, also are important during smoking. A schematic representation of the nicotine vapor adsorption measurement procedure is shown in Figure 1. Nitrogen was passed for one minute through a 21 mm segment of acid treated PET web filter followed by three cambridge pads in disposable holders. The flow rate was set at one liter per minute. The nitrogen flow was first allowed to equilibrate for four minutes by passing through a 50 ml vacuum flask holding 10 grams of nicotine thermostated at 25°C in a water bath for four minutes. Nitrogen plus nicotine vapor was then routed through the filter and the pads for a set exposure time. Figure 1 Initially, an FT-IR was used to determine at what point the nicotine vapor broke through the third pad. On this basis exposure time was chosen that allowed nicotine to pass through an untreated filter and two cambridge pads, but did not allow enough time for the vapor to pass through the third pad. After exposure, the filter and pads were submitted individually for low level GC nicotine analysis. -J Ln 4 0
Page 5: dha73d00
Acid elution: This procedure provides a comparison among filters of the rate of acid transpQ.r.t: fromm filters to air passed thrQugh the. filter,s. A.gairt, this.measurement, does not. . directly predict the rate of transfer of acid to smoke during smoking, but does depend on the same basic property that determines the rate of transfer during smoking, i.e. the vapor pressure of the acid over the filter fiber surfaces. A schematic representation of the acid elution procedure is shown in Figure 2. With a flow rate of one liter per minute, nitrogen was allowed to flow through a 21 mm segment of CA web filter and into 40 grams of pure water for five minutes and 10 seconds, during which time pH measurements were taken and recorded by a Mettler DL70 Titrator. Each water sample was purged with nitrogen for one minute prior to exposure to the filter. The final pH reading, in conjunction with the total mass of the water and the acid molecular weight and pKa, was used to calculate the actual mass of eluted acid by means of the following equations: K = [H+] [A ] a [HA] Figure 2 (1) Ca=[A _] +[HA] (5) A-+OH-=H+ (6) [~l - ([H+]2-Kw) (7) K a K*[HAl [A _l = a (8) [H+] Mass=(YolH2O) * (Ca) *Ma (9) a 5 ~ ~
Page 6: dha73d00
Where K~ is the acid dissociation constant, [H+] is the hydrogen ion concentration, [A-] is the anion concentration, [HA] is the acid concentration, K,,, is the dissociation constant of water, [OH.] is the hydroxide ion concentration, C, is the analytical concentration of acid; Mass is ihee mass -in. grams of-eluted acid, Vol HZO is the volume of water,, and. M,: is the molecular weight of acid. RESULTS: Table 5 presents standard analytical results for cigarettes prepared as described above (Table 4, 0% dilution) with filters from Runs 1, 2, and 3 corresponding to Tables 1, 2, and 3, respectively. KHP is potassium hydrogen phthalate. The percentages listed under filter type are in mass percent on the coated filter segment of the acid, glycerol, or acid/glycerol solution as indicated. Two nicotine per puff values are given; one from the FTC smokings and the other from the 'volatile' nicotine measurement, as indicated. Percent 'volatile' nicotine, %VNIC, is the fraction of the total nicotine, 'volatile' plus 'non-volatile' (that which passes the bubbler), represented by 'volatile' nicotine. 'Tar' to nicotine ratios, T/N, are calculated from the FTC data. pH is smoke pH determined by the standard Analytical procedure. Table 6 lists the results of the filter only measurements, nicotine vapor adsorption and acid elution, along with smoke levulinic acid and glycerol for selected samples. The percentage listed as nicotine adsorbed is the fraction of the total nicotine presented that was found on the filter. The mass of acid eluted is calculated from the change in pH of the downstream trap and the equilibrium ionization equations for the acids. The mass shown for the control and glycerol only filters is calculated as citric acid in order to show the noise level for zero mass of acid eluted. The mass of levulinic acid in smoke per cigarette is shown in the last column for Run 2 and in the second column for Run 3. The average starting and ending pH values for each filter type from which the mass acid eluted was calculated are listed in Table 7. As a way of further investigating the function of acid filters and possibly of gaining insight to the properties of mainstream smoke, a series of yield measurements were made as a function -of filter length. In the first set of measurements, a control and three acid filters (Table 3) of lengths 10, 30, and 49 mm were studied on our standard high nicotine blend tobacco rod (Table 4). The results of these measurements are presented in Table 8. Also listed in Table 8 are the results of regression analyses of the logarithm of the indicated quantity ('tar', nicotine, and 'volatile' nicotine) vs. the filter length. The slope from the latter analysis is a removal rate coefficient for the various quantities and is listed in the right most column of Table 8. A second more extensive study of filter length was performed with the filters described in Table 9 and an untreated control. Filter lengths of 10 to 90 mm in 20 mm increments were used with the same tobacco rod as above (Table 4). Table 10 lists the results showing 'volatile' nicotine, total nicotine, 'tar', percent 'volatile' nicotine, and 'tar' to nicotine ratio for each filter type at each length. Plots of the same data are shown in Figures 3-8 for 'tar' per puff, total nicotine per puff, 'volatile' nicotine per puff, 'tar' to nicotine ratio, percent 'volatile' nicotine, and 'volatile' nicotine to 'tar' ratio, respectively. 6
Page 7: dha73d00
DISCUSSION: : We laave. draw.n. severa.l - general-conclusions from. the .results of. this. work, which include: * Acid filters of several types smooth low T/N smoke by the reduction of 'volatile' nicotine with a small but significant impact on total nicotine. The smoothing effects are presently verified only by informal smoking evaluations with SED tests pending. * Two mechanisms for 'volatile' nicotine reduction can be active with acid filters: vapor nicotine adsorption and addition of acid to smoke. * A smoothing effect can be achieved with citric acid glycerol coated filters without any addition of acid to the smoke. * Glycerol can be used to control the rate of acid transfer to smoke. A more detailed discussion of the results follows and is divided into three sections by analysis type: standard analyses, special analyses, and filter length studies. STANDARD ANALYSES: 'Tar' filtration was unaffected by any of the coatings as is clear from the constancy of 'tar' per puff in Table 5 for each of the Runs. This observation is not unexpected and indicates that the macroscopic fiber structure of the web filters is not significantly'altered by the presence of the coatings. Total nicotine, measured by both the FTC smokings and the 'volatile' nicotine procedure, was affected by the acid filters. The effect is somewhat inconsistent in Run 1 of Table 5, but in Runs 2 and 3 all the acid coated filters showed a reduction of total nicotine vs. the control ranging from about 15 to 30 ug. of nicotine per pufii approximately 9% to 18% of the control value. Given the comparative constancy of the 'tar' yields, these total nicotine yields indicate slight selective filtration of nicotine by the filters. The combined effects of constant 'tar' and slight total nicotine reduction result in a small elevation of T/N for the acid filters vs. the control as indicated in Table 5. As will be discussed further we believe that the total nicotine reduction results from adsorption of vapor phase nicotine by the acidic surfaces. 'Volatile' nicotine was reduced for the acid filters vs. the control by greater proportions than total nicotine as is shown in Table 5. The reductions ranged from insignificant up to almost 50%. It should be emphasized here that 'volatile' nicotine is an empirical measure and is most useful for comparisons among cigarettes. Particularly, the 'volatile' nicotine number should not be interpreted as vapor phase nicotine concentration in absolute units. The preferred physical interpretation of the 'volatile' nicotine value is as a measurement of the mass exchange rate of nicotine from the smoke to a surface. Run 1 indicates that the reduction of 'volatile' nicotine is both acid type 7
Page 8: dha73d00
and loading level dependent with lactic being more effective than citric which is in turn more effective than KHP. Run 2 shows levulinic acid to be comparable in effect to lactic acid with a substantial reduction (30%) in 'volatile' nicotine achieved with the relatively low.filter loading.of.3.2% by weight. Run 3 shows an unusual.effect for.. glycerol loaded filters. While-tlie total nicotine is ieduced for the glycerol filters by similar amounts compared to the acid filters, the 'volatile' nicotine is virtually unaffected. The other data for Run 3 show that 'volatile' nicotine reductions can be maintained with 50/50 glycerol/acid mixtures for lactic and levulinic but are reduced for citric acid relative to the 46% citric acid loading of Run 1. The indications that led to the use of glycerol are discussed in the next section on special analyses. Smoke pH showed a general reduction with decreasing 'volatile' nicotine. The two lowest values of smoke pH, 4.11 (Run 1) and 4.98 (Run 2), both correspond to higher filter loadings of lactic acid only. The glycerol only filters in Run 3 again show a contrary pattern with a slight increase in smoke pH despite a reduction in total nicotine. The pH changes for the acid coated filters will be elucidated in the next section. Overall, the standard analyses indicate that the 'volatile' nicotine can be reduced by acid coatings with a much smaller proportional reduction in total nicotine and no effect on 'tar'. Also, acids differ in their 'volatile' nicotine reduction efficiency which is somewhat affected by the addition of glycerol. Smoke pH reductions generally accompany the 'volatile' nicotine reductions. SPECIAL ANALYSES: Table 6 lists the results of the special analyses described in detail in the analytical procedures section. The columns labeled "% NIC ADSORBED" show the vapor nicotine adsorption measurements. While these measurements were made with nicotine vapor in nitrogen and not smoke we believe that they are an indicator of 1) the existence of a vapor nicotine adsorption capability of the filters and 2) relative adsorption efficiency of the different f lters. In Run 1 we see only lactic acid substantially exceeding the control in adsorption efficiency with citric acid and KHP only marginally higher than the control at the higher loading levels. The 100%. adsorption value for the high level of lactic acid corresponds to the largest value for 'volatile' nicotine reduction in Table 5. In Run 2 we see 100% adsorption rates for two coating levels of both lactic and levulinic acid. The % 'volatile' nicotine reductions for all four of the cigarettes made with these ~e filters are similar as shown in Table 5. Run 3 shows that the removal rates remain near 100% for lactic and levulinic with the addition of glycerol to the coating and the fractional 'volatile' nicotine reductions remain high. The 50/50 citric acid/glycerol coated filters showed intermediate vapor nicotine adsorption which correspond to intermediate 'volatile' nicotine reductions in Table 5. The glycerol only coated filters again show a selectivity for nicotine with intermediate vapor nicotine adsorption which corresponds to a reduction of total nicotine but no reduction in 'volatile' nicotine. It should be noted that informal smokings of the glycerol only products vs. the other acid filter products showed minimal reductions in harshness. From the vapor nicotine adsorption measurements we conclude that, except for glycerol only filters, vapor nicotine adsorption efficiency as determined by this method positively correlates in a general way with 8
Page 9: dha73d00
'volatile' nicotine reductions and is coating type dependent. We further hypothesize that vapor nicotine adsorption in the acid treated filters may play a role in 'volatile' nicotine reduction. This latter point requires additional discussion given the form of the . ~evaporative transfer..hypothesis. . We will suggest later, as an overall conclusion of these studies, that there are two mechanisms of 'volatile' nicotine reduction with acid treated filters: addition of acid to the particles, for volatile acids only, and vapor nicotine stripping, for both volatile and non-volatile acids. So while it could be argued that acid addition to the smoke could be the cause of 'volatile' nicotine reduction with volatile acids, for the non-volatile case, citric/glycerol, vapor adsorption only must be considered as the mechanism of 'volatile' nicotine reduction and that is where the conflict with the evaporative deposition hypothesis arises. The evaporative deposition hypothesis suggests that as vapor nicotine is stripped from smoke during inhalation, by adsorption onto surfaces, additional nicotine evaporates from the particles, driven by the newly established concentration gradient, and becomes available for adsorption. This should also apply for other adsorbable semi-volatile smoke components. The question that arises is: Why is nicotine not driven to evaporate from the particles as vapor is adsorbed in the filter, re-establishing the initial conditions and yielding an unchanged (from non-acid filter) 'volatile' nicotine measurement? A possible explanation for the effect of nicotine vapor removal by acidic filters on 'volatile' nicotine level is a kinetic limitation on the rate of vapor nicotine replenishment by evaporation from particles. This description of the system would say that there is nicotine in the particles that is free to evaporate but that it does so so slowly that it is not collected in the 'volatile' nicotine bubbler trap. This model is equivalent to saying that only nicotine present in the vapor phase when the smoke exits the cigar6tte contributes to 'volatile' nicotine since nicotine that evaporates too slowly on the time scale of the measurement, which is the same as that for inhalation, can be considered to be bound in the particles. In evaluating the latter possibility several interesting and relevant observations can be made, unfortunately none of these conclusively resolves the matter but they do merit presentation: * 'Volatile' nicotine values as high as 85 micrograms per cigarette (ug/cig) have been measured for blended cigarettes and values exceeding 230 ug/cig were measured for an all burley cigarette. The room temperature saturation vapor pressure of nicotine (R&DM, 1988, No. 94) predicts a maximum vapor concentration in a 35 cm about 6-8 ug. Thus, measured masses of 'volatile' nicotine appear'to exceed in some cases the theoretical limit at vapor saturation suggesting that 'volatile' nicotine does not reflect only initial vapor phase nicotine. This argument, however, carries an important caveat. Some small but un-quantified amount of particulate nicotine is collected and reported as 'volatile' nicotine in the standard procedure. We know that the amount of particulate nicotine collected is a small fraction of the total since the 'tar' determined by the 'volatile' nicotine procedure agree very well with standard FTC 'tar', but even a small amount could have some impact on the 'volatile' nicotine value which typically is 5-15% of the total nicotine yield. So the fact that 'volatile' nicotine exceeds, in some cases, the expected maximum vapor nicotine concentration suggests, but does not prove, that 'volatile' nicotine includes evaporated nicotine in addition to initial vapor 9
Page 10: dha73d00
nicotine; this due to the empirical nature of the 'volatile' nicotine method. * All attempts to measure mainstream vapor phase nicotine as that which passes through a Cambridge filter pad have yielded a null result, i.e., no measurable nicQtine._ passes the filter. If Cambridge pads were ideal particle filters, these observations would resolve the matter. Unfortunately, Cambridge pad filters and similar glass fiber filters have been shown to adsorb vapor phase nicotine in ETS studies (see Ogden et al. TCRC 1992). These data are not sufficient to determine whether 'volatile' nicotine is vapor phase nicotine only and is removed by filter adsorption in standard FTC smoking. * 'Volatile' nicotine values below the predicted saturation vapor value are observed even though total nicotine is in excess of that needed for vapor saturation. If 'volatile' nicotine represented only initial vapor nicotine this observation might be explainable on the basis of a nicotine vapor pressure in smoke below the saturation level and tied to the nicotine mole fraction in the particles according to Henry's law. However, we have estimated the fraction saturation of nicotine vapor in smoke and found it, according to certain assumptions and approximations, to be very small, possibly 1-5% (BIOBMM 91-013). At this fraction of saturation not even the lowest values of 'volatile' nicotine can be accounted for if initial vapor phase equates with 'volatile' nicotine. * The fraction of total nicotine that is measured as 'volatile' nicotine increases as smoke concentration decreases. This is an effect that can not be totally accounted for by the slight decrease in T/N that accompanies increases in ventilation level, as it is also observed to some extent when smoke concentration is reduced by increasing filtration efficiency. We think that this may be related to an impedance to mass transport of vapor nicotine by particles that have lost nicotine by evaporation and act as sinks when ' ,• diffusion of nicotine creates vapor concentrations greater than that at the particle surface (see BIOBMM 91-014 for other evidence of this effect). An alternative explanation for the observed 'volatile' nicotine reduction for filters that only remove nicotine vapor, that is consistent with the data, is as follows. The vapor concentration of nicotine in the fresh smoke exiting the tobacco rod is much lower than the saturation value and the 'volatile' nicotine concentration. The nicotine vapor pressure is determined by the thermodynamic activity of free base nicotine in the particulate solution, and changes as the free base concentration in the particles changes. In the acid filters nicotine vapor is adsorbed and replenished to the gas phase by evaporation from the particles on a very rapid time scale consistent with the high surface area of the aerosol and its intimate mixing with the adsorbing filter fibers. As free base nicotine evaporates from the particles its concentration in the particulate solution decreases, its activity decreases, and a lower vapor pressure surrounding the particles is established. The net effect on the smoke as it exits the filter is a lowered nicotine vapor concentration and a lowered particulate free base nicotine activity which together yield a lower mass transfer rate of nicotine, predominantly due to a reduced particle nicotine evaporation rate, from the smoke in the 'volatility' measurement and, by extrapolation, during smoking. This model is equally consistent with the data vs. the vapor phase only scenario and avoids some of the difficulties listed above. Neither hypothesis is yet proven unambiguously. 10
Page 11: dha73d00
The column labeled "MASS ACID ELUTED" in Table 6 lists the results of the pH based determination of acid vaporization from filter to flowing nitrogen stream described in the analytical methods section. In Run 2 we observe a greater mass of lactic transfers than levulinic. and that. the mass. transferred is. indepen.dent,. for. a:given acid, of : the loading level for the range measured. Levulinic acid in smoke values appear in the forth column of the Run 2 section and are comparable to those typically found for levulinic acid in blend cigarettes. A slightly higher mass of lactic acid in the smoke can be implied if one accepts an extrapolation from the filter elution measurements. Run 3 shows that the mass of lactic and levulinic acid eluted from filters loaded with a 50/50 acid/glycerol solution is substantially reduced vs. the neat filter loadings of Run 2. The levulinic acid in smoke value for 50/50 glycerol/acid at 33% by weight on the filter (16% acid) is about the same as that found in Run 2 for 3.2% neat acid on the filter. Measurements on the 50/50 glycerol/citric acid filter indicate no acid elution yielding the same background pH change as observed for the uncoated control and glycerol only filters. Initial measurements by the acid elution method on filters with citric acid (a room temperature powder) only, showed sharp, irregular changes in pH in contrast to the smooth, continuous change in pH observed for the lactic and levulinic acid (room temperature volatile liquids) coated filters. We believe the irregular changes for citric acid filters were the result of bulk powder dispersion from the filters and that this could be prevented by coating with a solution of citric acid in glycerol. When it became apparent that the volatile acids transferred to the smoke, glycerol solutions of these acids were used to reduce the transfer by lowering the vapor pressure of the acids over the solutions vs. that over the neat liquids. The desired effects were achieved in both cases. No transfer of glycerol to smoke was observed for any filter as indicated in Table 6, Run 3 by the unchanging glycerol in smoke measurement listed under "GLY/PUF". In summary, the analyses presented in this section are consistent with a two mechanism model of 'volatile' nicotine reduction by acid coated filters. Where volatile acids are used, acid is transferred to the smoke and reduces 'volatile' nicotine by altering the particle composition in similar fashion to what takes place when certain acids are added to the blend. The acid elution and levulinic acid in smoke results support this hypothesis. A second mechanism appears to be the adsorption of nicotine vapor by the acid filters. This effect is probably active for both volatile and non volatile acids and is the only mechanism active for citric acid/glycerol coatings. The nicotine vapor adsorption measurements lend support to this hypothesis. The physical reasonableness of vapor adsorption reducing 'volatile' nicotine is best supported by a model in which the adsorption of vapor nicotine gradually reduces the evaporative transfer rate by reducing the free base nicotine concentration in the particles as discussed above. The results of this section also indicate that coating filters with glycerol solutions of acids is an effective means to control acid transfer to smoke. 11
Page 12: dha73d00
FILTER LENGTH STUDIES: The results of the two filter length studies were similar and only the second more more comprehensive set.-of data, Table 10, will be discussed here. .1Vleasurements with filters of different lengths were undertaken to explore the mass exchange aspects of the acid filter mainstream smoke system. The two mechanisms hypothesized for 'volatile' nicotine reduction by acid filters are both mass exchange processes, either vapor nicotine from the smoke to the substrate or volatile acid from the substrate to the smoke. The use of different length filters simply controls the time scale of the mass exchange processes and, if the hypothesized mechanisms are on target, has predictable effects on smoke yields as a function of filter length. In contrast, a variable filter loading approach as an alternative means to manipulate the mass exchange processes is less straightforward from a basic understanding point of view in part because the surface area of available acid may or may not change with increased loading while the mass of acid available for elution and interaction with deposited WTPM will increase. Finally, we thought that by varying the substrate part of the system in a controlled fashion certain basic smoke properties might become manifest through the performance of the whole system i.e. the cigarette yields. In Figure 3 we see that the 'tar' per puff vs. filter length curves are essentially unchanged for.the treated filters vs. the control, indicating that no significant changes to fiber size or web structure resulted from the coating process. In Figure 4 total nicotine per puff is plotted vs. filter length. Here we see a nearly equal reduction, vs. the control, for all the acid treatments at all lengths. Remembering that of the two mechanisms of 'volatile' nicotine reduction only vapor adsorption results in total nicotine reduction, we find here a result consistent with vapor nicotine adsorption at nearly equal rates taking place for each filter, regardless of acid type. In examining the 'volatile' nicotine per puff vs. filter length data plotted in Figure 5, it must be kept in mind that 'volatile' nicotine is not a simple yield quantity like total nicotine and 'tar', but is rather a measurement of the mass exchange rate of nicotine from smoke, its magnitude influenced by other smoke properties. Again we see the three acid filter curves grouped together all at lower levels of 'volatile' nicotine per puff than the control. It should be noted that the citric acid/glycerol coating in this case is at a 6:1 acid to glycerol ratio rather than the 1:1 level for the data in Tables 5 and 6. The higher acid to glycerol ratio used here is more efficient at reducing:'volatile' nicotine than the lower ratio used earlier as indicated by the comparable 'volatile' nicotine values for citric vs. the other acids in Figure 5, particularly at the shorter filter lengths. The citric curve does differ from the lactic and levulinic curves as will be revealed in the logarithmic analysis to be presented later. Qualitatively the reduction in 'volatile' nicotine for citric is not as large at longer filter lengths as it is for the volatile acids. This latter effect may be due to the depletion of adsorbable nicotine from the smoke by the filter in the case of the citric acid coating for which we hypothesize only the vapor adsorption mechanism being relevant. The volatile acid coatings may continue to reduce 'volatile' nicotine at the longer lengths due to the continued addition of acid to the smoke i.e. by the second mechanism. This explanation raises another question of theoretical consistency. Why do we observe any 'volatile' nicotine from the longest citric acid filters if the limit of vapor 12
Page 13: dha73d00
and evaporated nicotine adsorption has been reached in the filter? This same question is raised by the T/N data and will be discussed next. . In Figure. 6. we see that the: T/N.variation with:filter length ..is very similar- for. the -. - three acid filters, increasing from 10 to 50 mm filter lengths and reaching a plateau value of about 12.0 to 12.5 for lengths 50 to 90 mm. The control filter does not exhibit this pattern and has lower T/N values at all filter lengths. Again, a hypothesis that would be consistent with this data would be that there is a limit to the amount of vapor and evaporated nicotine that can be adsorbed by the filters that is reached at the 50 mm length. Why then if a limit has been reached do we continue to see measurable 'volatile' nicotine? The unknown, but probably small, fraction of particulate nicotine that is included as 'volatile' nicotine, as discussed above, may be a contributor but is probably not a full explanation (note that 'volatile' nicotine for citric exceeds that for lactic and levulinic at long filter lengths in Figure 5 while for equal T/N at these lengths particulate deposition would be predicted to yield equal 'volatile' nicotine). Another possibility is a kinetic limitation on the nicotine mass exchange from the smoke controlled by the residence time of the smoke in the filter which is smaller than the residence time of the smoke in the 'volatile' nicotine bubblers. However, the fact some nicotine is selectively removed up to 50 mm filter length, as indicated by the rising T/N, but stops being selectively removed at longer lengths, flat T/N, does not suggest a kinetic limitation since increased residence times for longer filters should result in increased removal if the rate of removal does not change. Yet another possibility to reconcile the T/N data, that suggests a limit on adsorbable smoke nicotine is reached in the longest filters, with the non-zero 'volatile' nicotine measurements, that suggest additional adsorbable nicotine, is a hypothesized effect of particulate water concentrations. The particulate water of conventional smoke in a filter is expected to be low compared to that of the smoke in the 'volatile' nicotine bubblers or in the throat. The smoke in the acid filters may'be dried even more than conventional smoke due to water adsorption by the hygroscopic acid/glycerol filter coatings. If particulate water promotes nicotine evaporation from the particles (solubilizing nicotine from the organic phase and/or increasing particle surface area or some other mechanism) then we might speculate that a nicotine removal limit is reached in the filters at a certain water concentration but that this limit changes when particulate water is added to the smoke in the 'volatile' nicotine measurement device and during puffing and inhalation. Considerable experimentation would be needed to decide anything definitive about the effect of water on particulate nicotine evaporation, but that work may be warranted given the additional implications such an effect would have with regard to nicotine exchange throughout the lung during smoke inhalation. The fact that the plateau T/N value is so similar for the three acid filter types suggests that there may be an intrinsic smoke property controlling the effect i.e. some absolute fraction of total nicotine that is adsorbable nicotine under these conditions. With a couple of assumptions about the nature of the smoke and the measured T/N's we can estimate a fraction adsorbable nicotine. We assume the mass of total nicotine in smoke, Nt, is composed of adsorbable nicotine, Na, and non-adsorbable nicotine, Nn. We take the 10 mm control filter T/N as approximating the fresh smoke with all adsorbable and non-adsorbable nicotine still present and take the average plateau T/N to represent smoke with all adsorbable nicotine removed. We further assume that 13
Page 14: dha73d00
the 'tar' to non-adsorbable nicotine ratio is the same for the two cases i.e. non-adsorbable nicotine is particle bound as is 'tar' and both are removed with the same efficiency by the filters. This yields: . .. . .... .. (Na+Nn)/T1 = Na/T1 + Nn/T 1 = 1/9.5 = 0.1053 (fresh smoke) Nn/T2 = 1/12.4= 0.0806 (Na removed,long acid filter) since we assume Nn/T1=Nn/T2 Nn/T1 = 0.0806 and NalT1 = 0.1053-0.806 = 0.0246 from which the fraction of total nicotine that is adsorbable nicotine in the fresh smoke is (Nafl'1)/(Na/T1+Nn/I'1) = Na/(Na+Nn) = Na/Nt = 0.2336 or about 23%. This calculation should be viewed within the context of the assumptions made and the prior discussion regarding 'volatile' nicotine, but with further work this approach may offer a route to the measurement of a more basic smoke property than 'volatile' nicotine. Interestingly, no measured fraction 'volatile' nicotine of total nicotine value has exceeded 23% although some have approached it at 17%. In Figure 7 the % 'volatile' nicotine of total nicotine is plotted vs. filter length. We see an increase in fraction volatile nicotine with increased filter length for the control cigarette. We believe that this increase is caused by the decrease in smoke particle- ' concentration with increased filter length which in turn causes an increase in nicotine transfer efficiency from the smoke to surfaces, the 'volatile' nicotine bubbler in this case. We see this increase in % 'volatile' nicotine with decreased smoke particle concentration routinely, but frequently as the result of increased ventilation level. Since increased ventilation level is known to be associated with decreased T/N, which would also be expected to increase % 'volatile' nicotine, we have examined past 'volatile' nicotine data and found numerous cases where % 'volatile' nicotine increased with decreasing smoke particle concentration without an associated decrease in T(N. The present data for the control filter shows a case where T/N actually increased slightly with decreased smoke concentration but % 'volatile' still increased, reinforcing our; belief in the hypothesized impedance of nicotine mass transfer efficiency by increased particle concentration. We also see in Figure 7 that the citric acid filter counteracts the increase in % 'volatile' nicotine but not as effectively as the volatile acid coated filters. This again is consistent with the two mechanism 'volatile' nicotine reduction hypothesis discussed above. The same trends are exhibited in the 'volatile' nicotine to 'tar' ratio shown in Figure 8. The final analysis applied to the filter length data was a classical filtration theory analysis. For filters with a constant filtration efficiency per unit length with respect to a given smoke component, a plot of the logarithm of the smoke component yield vs. filter length yields a straight line the slope of which is a filtration coefficient. This analysis was 14
Page 15: dha73d00
applied to 'tar', total nicotine, and 'volatile' nicotine through linear regression and the regression results are listed in Table 11 for each filter type. Again we note that 'volatile' nicotine technically does not fit the criteria for this type analysis because it is not a simple smoke component, nevertheless we found that the regression fits were as good -for 'volatile' nicotine as for the other smoke components. The regression fits were good overall with a minimum r2 of about 0.968. Plots of the best fit lines are given in Figures 9-11 for 'tar', total nicotine, and 'volatile' nicotine all per puff, respectively. We see in Figure 9, as we did in Figure 3, virtually identical 'tar' removal efficiencies for the four filters indicating no alteration of mechanical particulate filtration efficiencies due to the coatings. In Figure 10 we see similar total nicotine removal slopes for all four filters due to the domination of total nicotine removal by particulate nicotine filtration. The three acid filter curves are reduced from the control indicating nearly equal selective nicotine removal, we believe by vapor nicotine adsorption. The best fit lines for 'volatile' nicotine vs. filter length in Figure 11 show all the acid filters having lower magnitudes and more negative slopes, higher removal coefficients, than the control. The same conclusions are drawn as in the discussion of Figure 5. The less effective 'volatile' nicotine reduction efficiency for citric vs. the other acids is accentuated here as the average performance over all lengths is included in the determination of the regression line. CONCLUSIONS: Restating the conclusions given at the beginning of the discussion section: * Acid filters of several types smooth low T/N smoke by the reduction of 'volatile' nicotine with a small but significant impact on total nicotine. The smoothing effects are presently verified only by informal smoking evaluations with SED tests pending. * Two mechanisms for 'volatile' nicotine reduction can be active with acid filters: vapor nicotine adsorption and addition of acid to smoke. * A smoothing effect can be achieved with citric acid glycerol coated filters without any addition of acid to the smoke. * Glycerol can be used to control the rate of acid transfer to smoke. We add: * 6:1 citric acid/glycerol coatings reduce 'volatile' nicotine by comparable amounts to the volatile acids at filter lengths useful in products without the addition of acid to the smoke. * The most reasonable physical scenario for 'volatile' nicotine reduction through vapor nicotine adsorption in the filter appears to be a continuous reduction in particle free base nicotine activity which reduces the evaporation rate of nicotine and the vapor nicotine concentration which results in a reduction of the rate of nicotine transfer from smoke to surface. 15
Page 16: dha73d00
* As acid filter length increases, T/N initially increases but levels off at the same value for three different acid type filters for a given tobacco rod, suggesting that under the conditions extant in the filter there is a limit to the amount of nicotine that can be removed from the smoke by vapor adsorption and particulate nicotine evaporation. 16
Page 17: dha73d00
Table 1 Winston-Salem, N.C. 27102 919-741-5000 INTER-OFFICE MEMORANDUM SUBJECT: Acid Blends on PET Web DATE: December 12, 1991 Trial #2674101 TO: Mr. Barry L Saintsing FROM: Mr. Dallas B. Blakley, Jr. Product Development Division Packaging Technology R&D Brands R&D On December 9 and 10, 1991, eleven samples were produced on the Faustel laminator at Plant #641-1. Two different cylinders, 110 Quad and 45 T.H., were used to apply acid solution to a 1.0 ounce per square yard nonwoven PET web. Samples #1 through #6 were produced with the 110 Quad and Samples #7 through #11 with the 45 T.H. A description of the samples is: . .........n nv.::. •.. . v.....{ : v..{ .{ {. . . . • 1 Phosphoric Acid 2.0 - 2.5 .43 -.47 2 Citric Acid 3.1 - 4.0 .71 -.82 3 Sulfuric Acid 2.1 .46 4 Lactic Acid 1.4 - 1.8 .32 -.37 5 Potassium Hydrogen Phthalate 1.4 .28 -.32 6 Water --- --- 7 Water --- 8 Phosphoric Acid 27.2 - 32.0 5.8 - 6.4 9 Citric Acid 45.7 - 47.0 9.6 - 10.0 10 Potassium Hydrogen Phthalate 9.3 - 11.8 2.2 11 Lactic Acid 42.0 - 58.0 8.9 - 12.7 This material was trimmed to 9" widths and shipped to Plant #611-5 on December 11. Dallas B. Blakley, Jr. DBBjr/clm xc: Mr. D. L. Carespodi .'`Ne work for smokers: '
Page 18: dha73d00
Table 2 Winston-Salem, N.C. 27102 919-741-5000 INTER-OFFICE MEMORANDUM SUBJECT: Acid Blends on PET Web DATE: February 3, 1992 Trial #2678601 TO: Mr. Barry L Saintsing FROM: Mr. Dallas B. Blakley, Jr. Product Development Division Packaging Technology R&D Brands R&D .On January 28, 1992, five samples were produced on the Faustel laminator at Plant #641-1. An 85 T.H. cylinder was used to apply acid solution to a 1.0 ounce per square yard non-woven PET web. A description of the samples is: 1 2 3 5 Lactic Acid - 14% in H20 pH - 2.57 Lactic Acid - 29% in H20 pH - 2.38 Levulinic Acid - 14% in H20 pH - 3.07 Levulinic Acid - 29% in H20 pH - 3.05 Lactic Acid 14% } in H20 G7-1 SDE 20% pH - 3.67 7 5 - 15 0 . . 17.7 - 18.6 3.6 2.9 - 3.5 0.6 - 0.7 21.0 - 24.3 4.5 - 5.2 28.0 - 35.0 6.2 - 7.7 These samples were trimmed to 9" widths and shipped to Plant #611-5 on January 30, 1992. Dallas B. Blakley, Jr. DBBjr/clm xc: Mr. D. L. Carespodi . 'We worK for smokers."
Page 19: dha73d00
Table 3 wnston-Salem,..N C. 27102 919-741-5000 INTER-OFFICE MEMORANDUM SUBJECT: Acid Blends on PET Web Trial #2681501 ~ DATE: March -3,1992 ~' TO: Mr. Barry L Saintsing Product Development Division Brands R&D FROM: Mr. Dallas B_Blakley, :Jr..-~.., Packaging Technoiogy,R&D..~ On March 3, 1992, six samples were produced on the Faustel Iaminator-at=Plan"641-1.=._- An 85 T.H. cylinder was used to apply acid solution to a 1.0 ounce per square yard_norrviroven PET web. A description of the samples produced is: ~ ~' 1 Glycerol - 2M Gms. 3.4 - 4.6 . 0.7 - 0.9 ~ 2 Glycerol - 4M Gms. 6.7 - 7.9 1.4-1.6; 3 Glycerol - 2M Gms. Citric Acid - 2M Gms. 11.0 - 13.0 2.4 - 2.6: 4 Glycerol - 4M Gms. Citric Acid - 4M Gms. 17.6 - 26.0 3.6 - 5.3 Glycerol - 3M Gms. Levulinic Acid - 3M Gms. 30.2 - 36.8 6.6-7.1 6 Glycerol - 3M Gms. Lactic Acid - 3M Gms. 24.5 - 26.0 5.0 - 5.8.. This material will be trimmed to 9" varide and delivered to Plant #611-5 on March 5:-,. Dallas B..BIakley,:Jre-. DBBjr/c1m xc: Mr. D. L Carespodi "We work for smokers."
Page 20: dha73d00
Table 14 Tobacco Rod Specifications The tobacco rod used.in...the.acid fi.lter analyses came..from thee banded product CT 2437 AB which had the following specifications: KG2 (cased with 16% B-26) 24.4% 2% B 25-W CG3 53.2% casing for TB 22.4% cutting Entire blend top dressed with S-182 (0.5%) Paper I.D. 1005471 Tipping I.D. AR0541 (non-porous) Cigarettes were hand-made using first a 12mm segment of the acid coated polyester web filter next to the tobacco rod described above, then a second segment of cellulose acetate (3.0/35,000) at the mouthend.
Page 21: dha73d00
50968 7563 FTC, "VOLATILE" NICOTINE, pH Table 5 ACID FILTERS RUN #1 FT % reduced VNI % reduced % reduced FILTER TAR/PUF NIC/PUF vs. con. NIC/PUF vs. con.. VNIC/PUF vs. con. %VNI. T/ m m m u % CONTROL 2.06 0.180 0.0 0.179 0.0 7.64 0.0 4.27 11.44 6.55 LACTIC 1.6° 2.10 0.181 -0.6 0,172 4.0 6.38 16.4 3.72 11.60 LACTIC 50% 2.02 0.171 5.0 0.177 1.2 4.05 47.0 2.29 11.81 4.11 CITRIC 3.5°/ 2.27 0.197 -9.4 0.167 6.7 6.59 13.6 3.95 11.52 CITRIC 46% 2.03 0.171 5.0 0.167 6.8 5.20 31.9 3.12 11.87 6.02 KH P 1.4% 2.29 0.200 -11.1 0.180 -0.5 7.40 3.1 4.12 11.45 KHP 10.5°l° 2.10 0.184 -2.2 0.162 9.7 6.20 18.8 3.84 11.41 6.38 ACID FILTERS RUN#2 FT % reduced VNI % reduced % reduced FILTER TAR/PUF NIC/PUF vs. con. NIC/PUF vs. con. VNIC/PUF vs. con. %VNI T/ m m m u °/ CONTROL 2.27 0.207 0.0 0.193 0.0 7.46 0.0 3.86 10.97 6.15 LACTIC 11°/ 1.97 0.169 18.4 0.160 17.2 4.75 36.3 2.97 11.66 5.43 LACTIC 18% 2.18 0.188 9.2 0.166 14.0 4.02 46.1 2.42 11.60 4.98 LEV 3.2% ~ 1.99 0.172 16.9 0.170 12.2 5.26 29.5 3.10 11.57 5.47 LEV 23% . 2.18 0.188 9.2 0.178 8.0 4.75 36.4 2.67 11.60 5.19 ACID FILTERS RUN#3 FT % reduced VNI % reduced % reduced FILTER TAR/PUF NIC/PUF vs. con. NIC/PUF vs. con. VNIC/PUF vs. con. %VNI T/ m m m u °/ CONTROL 2.21 0.199 0.0 0.194 0.0 8.97 0.0 4.63 11.11 6.15 GLY 4% 2.13 0.175 12.1 0.170 12.4 8.66 3.5 5.10 12.17 6.36 GLY 7.3% 2.08 0.175 12.1 0.168 13.5 8.70 3.0 5.19 11.89 6.36 CIT/G 12% 2.17 0.182 8.5 0.161 16.8 7.80 13.0 4.84 11.92 5.91 CIT/G 22%° 2.15 0.181 9.0 0.166 14.2 7.48 16.6 4.50 11.88 5.92 LEV/G 33% 2.05 0.173 13.1 0.159 17.8 5.99 33.2 3.76 11.85 5.71 LAC/G 25% 2.14 0.175 12.1 0.158 18.3 5.86 34.7 3.70 12.23 5.66
Page 22: dha73d00
`-_ a b-_ e o ACID ELUTION, VAPOR NICOTINE ADSORPTION, SMOKE GLYCEROL AND LEVULINIC ACID ACID FILTERS RUN #1 FILTER /° NIC ADSO RBED MASS ACID ELUTED;.. LEV.SMK. 'u m /ci CONTROL 6.8 LACTIC 1.6° 39.8 LACTIC 50% 100.0 CITRIC 3.5% 3.3 CITRIC 46% 17.2 KHP 1.4% 8.5 KHP 10.5% 12.7 ACID FILTERS RUN#2 FILTER /° NIC ADSO RBED MASS ACID ELUTED LEV.SMK. 'u m /ci CONTROL 6.3 *5. LACTIC 11 % 100.0 120 LACTIC 18% 100.0 110 LEV 3.2% 100.0 73 0.19 LEV 23% 99.1 73 0.38 ACID FILTERS RUN#3 FILTER /° NIC ADSO RBED MASS ACID ELUTED GLY/PUF 'ug m CONTROL 4.8 *5. 0.064 GLY 4% 69.0 *5. 0.058 GLY 7.3% 77.0 *5. 0.058 CIT/G 12% 54.3 4.4 0.060 CIT/G 22% 67.6 4.9 0.058 LEV/G 33% LEV 0.17 m 100.0 34.0 0.065 LAC/G 25% 97.7 33.0 "' 0.067 *calculated as citric acid
Page 23: dha73d00
Table 7 Filter materials were coated on the Faustel laminator 12-12-91, 1-28-92;:.and.-3-3-9.2:. See attached data sheEts for.additional information on actual levels of loading per filter and calculation methods. Filters used were 12mm in length. Flow = 1LPM Duration = 5min. 10 sec. Gas Purge = Nitrogen (Faustel trial numbers: 2674101, 2678601, 2681501) Avg Mass Eluted (g) Avg Start Avg End pH pH Low Lactic 1.2E-04 6.222 4.516 High Lactic 1.1E-04 6.008 4.533 B-3/Lactic 3.3E-05 6.156 5.045 Low Levulinic 7.3E-05 6.012 4.969 High Levulinic . 7.3E-05 6.051 4.967 B-3/Levulinic 3.4E-05 6.160 5.235 Lt Citric 7.8E-06 6.063 5.995 Heavy Citric 4.8E-05 5.924 5.215 Low B-3/Low Citric 4.4E-06 6.288 6.246 High B-3/High Citric 4.9E-06 6.278 6.199 Low B-3 5.5E-06 6.191 6.155 High B-3 5.5E-06 6.196 6.141 Heavy Water 5.3E-06 6.475 6.428
Page 24: dha73d00
Table 8 EFFECT OF FILTER LENGTH ON YIELDS 50968 7566 citric/gly L mm vnic/ tnic/ %vni tar/ t/ t/v intercept remv. coeff. 10 8.61 228.95 3.76 24U7.89 10.52 279.52 tar 2678.92 -0.01342 30 5.45 147.44 3.69 1692.31 11.48 310.73 nic 267.6309 -0.01814 49 4.48 112.99 3.97 1428.57 12.64 318.75 vn i c 9.79078 - 0. 01681 , citric/ I L mm vnic/ tnic/ %vni tar/ t/ t/v 10 8.45 225.68 3.74 2486.49 11.02 294.26 tar 2811. 655 - 0. 0142 30 5.67 149.38 3.79 1765.43 11.82 311.48 nic 264.8595 -0.01784' 49 4.50 112.66 4.00 1430.38 12.70 317.59 vnic 9.688654 -0.01617 water L mm vnic/ tnic/ %vni tar/ t/ t/v 10 10.39 242.67 4.28 2546.67 10.49 245.22, tar 2857.92 - 0. 01293 30 8.41 182.05 4.62 1884.62 10.35 224.15 nic 270.9668 -0.01235 49 7.35 150.00 4.90 1539.47 10.26 209.56 vnic 11.22764 -0.00889 ' levulinic L mm vnic/ tnic/ %vni tar/ t/ t/v 10 6.46 215.07 3.00 2369.86 11.02 366.84 tar 2726.715 -0.0133 30 4.50 155.84 2.89 1857.14 11.92 412.70 n i c 256.9435 - 0. 01712 49 3.34 110.26 3.03 1410.26 12.79 422.10 vnic 7.59308 -0.01692 lactic L mm vnic/ tnic/ %vni tar/ t/ t/v 10 5.96 227.63 2.62 2421.05 10.64 406.45 tar 2682.429 -0.01262 30 3.13 156.25 2.00 1750.00 11.20 558.88 n ic 255. 3081 - 0. 01438 49 2.67 130.12 2.05 1481.93 11.39 555.30 vnic 6.795595 -0.02069 L mm - filter length in millimeters vnic/p - volatile nicotine per puff ug tniv/p - total nicotine per puff ug %vnic - percent volatile nicotine tar/p - 'tar' per puff ug t/n - 'tar' to nicotine ratio t/vn - 'tar'to volatile nicotine ratio intercept - zero filter length quantity prediction ug remov. coeff. - filtration removal coefficient, k in CL=CO*exp(k*L)
Page 25: dha73d00
Table 9 Winston-Salem, N.C. 27102 919-741-5000 INTER-OFFICE MEMORANDUM SUBJECT: Acid Solutions on PET Web DATE: April 1, 1992 Trial #2685701 TO: Mr. Barry L. Saintsing FROM: Mr. Dallas B. Blakley, Jr. Product Development Division Packaging Technology R&D Brands R&D On March 30 - 31, 1992, three samples were produced on the Faustel laminator at Plant #641-1. An 85 T.H. cylinder was used to apply acid solution to a 1.0 ounce per square yard nonwoven PET web. A description of the samples is: :............ . {:,.,.} --:r::::::: . ..}. $;,-r::'::r.}:-~::~.~,}:;~.}:.$:::<: $>$:<n{:: :F... ..,.} f.~•:- .•. . {+., y :.X1$,}}?:>::<: v + . -...-.::.:.. aa..r. ..- .,.r.•::.:.......~• k..:::..,. 1 Glycerol - 500 Gms. 5.0 - 14.9 0.9 - 3.1 Citric Acid - 3M Gms. Avg. - 10.3 Avg. 2.2 Q.S. to 30 Lbs. with Water 2 Glycerol - 1 M Gms. 2.6 - 6.8 0.6 - 1.4 Levulinic Acid - 1 M Gms. Avg. - 4.0 Avg. - 0.9 Q.S. to 30 Lbs. with Water 3 Glycerol - 1 M Gms. 5.3 - 7.2 1.3 - 1.5 Lactic Acid - 1 M Gms. Avg.: 6.4 Avg. - 1.4 Q.S. to 30 Lbs. with Water This material will be trimmed to 9.25" wide and delivered to Plant #611-5 on April 1. Dallas B. Blakley, Jr. DBBjr/clm xc: Mr. D. L. Carespodi "We work for smokers."
Page 26: dha73d00
50968 7568 Table 10 CONTROL LENGT PUFF VOLNI TOTNI TA VOLNIC/ TOTNIC/ TAR/ %VOLNI T/ 10 7.8 71.88 1930 19300 9.22 256.7 2474 3.59 9.64 30 8.0 63.95 1590 15700 7.99 206.7 1963 3.87 9.49 50 7.4 51.18 1060 12100 6:92 150.2 1635 4.61 10.89 70 7.4 47.72 880 10200 6.45 125.4 1378 5.14 10.99 90 7.7 40.65 790 8500 5.28 107.9 1104 4.89 10.23 CITRIC ACI /GLYCER L LENGT PUFF VOLNI TOTNI TA VOLNIC/ TOTNIC/ TAR/ %VOLNI T/ 1.0 8.0 64.53 1910 19400 8.07 246.8 2425 3.27 9.83 30 8.1 48.11 1350 14900 5.94 172.6 1840 3.44 10.66 50 7.4 37.05 920 11700 5.01 129.3 1581 3.87 12.23 70 7.2 31.75 760 9900 4.41 110.0 1375 4.01 12.50 90 7.5 26.97 670 8400 3.60 92.9 1120 3.87 12.05 LACTIC AC D/GLYCER L LENGT PUFF VOLNI TOTNI TA VOLNIC/ TOTNIC/ TAR/ %VOLNI T/ 10 8.2 61.13 1760 18100 7.45 222.1 2207 3.36 9.94 30 8.2 40.65 1260 14400 4.96 158.6 1756 3.13 11.07 50 7.0 29.40 900 11600 4.20 132.8 1657 3.16 12.48 70 7.4 24.26 730 9300 3.28 101.9 1257 3.22 12.33 90 7.7 19.49 670 8800 2.53 89.5 1143 2.83 12.76 LEVULINIC ACID/GLYC ROL LENGT PUFF VOLNI TOTNI TA VOLNIC/ TOTNIC/ TAR/ %VOLNI T/ 10 7.8 66.49 1700 17800 8.52 226.5 2282 3.76 10.08 30 7.8 48.41 1260 14600 6.21 167.7 1872 3.70 11.16 50 7.7 33.03 960 12400 4.29 129.0 1610 3.33 12.49 70 7.4 26.77 800 10200 3.62 111.7 1378 3.24 12.34 90 7.3 19.54 570 7300 2.68 80.8 1000 3.31 12.38
Page 27: dha73d00
Figure r3. 3000 2500 1000 500 0 10 20 PILT~`~ LE~GT~ mm'° 80 90 100 + CONTROI,- CITRIC ~ LACTIC ~ LEVULINIC
Page 28: dha73d00
Figure 4 300 so o 10 20 70 FILTNK LENGTH mm 80 90 ~ CONTROI,,- CITRIC ~_ LACTIC ~ LEVULINIC
Page 29: dha73d00
Figure 5 10 8 7 6 5 4 3 2 0 20 FILTffR LENGTfl mm 70 80 90 CONTROI~ CITRIC ~ LACTIC ~ LEVULINIC 10 100
Page 30: dha73d00
Figure 6 13.5 13 12.5 W Z 12 ~--~ ~ 011.5 U ~ 10 9.5 9 0 10 20 ~ILTE"P, LENGTH' mm70 ~° 90 ~ CONTROI,,,-- CITRIC A LACTIC 9 LEVULINIC 100 ..2 v
Page 31: dha73d00
F.Lgure 7 e 5.5 5 3 2.5 20 PIL~ LENGTA mm'° 80 90 CONTROL~ CITRIC ,LACTIC ~ LEVULINIC 0 10 100 ,,)
Page 32: dha73d00
Figure 8 0.5.5 0.5 0.25 0.2 0 z~l e I I I I I I I I I 10 20 I~ILT~k LE~TGT~ mm70 80 90 _, _ CONTROI,,- CITRIC ,LACTIC 3 LEVULINIC
Page 33: dha73d00
Table.: 11 LINEAR REGRESSION ANALYSIS OF LOG(X) VS. FILTER LENGTH CONTROL VOLNIC Regression Output: Constant Std Err of Y Est R Squared No. of Observations Degrees of Freedom TOTNIC Regression Output: 2.284456 Constant 0.031006 Std Err of Y Est 0.983933 R Squared 5 No. of Observations 3 Degrees of Freedom TAR Regression Output: 5.639049 Constant 0.054208 Std Err of Y Est 0.982638 R Squared 5 No. of Observations 3 Degrees of Freedom 7.897987 0.019048 0.997196 5 3 X Coefficient(s) -0,00664 Std Err of Coef. 0,00049 CITRIC ACID/GLYCEROL X Coefficient(s) -0.01117 Std Err of Coef. 0.000857 XCoefficient(s) -0.00984 Std Err of Coef. 0.000301 VOLNIC Regression Output: Constant TOTNIC Regression Output: 2.127134 Constant TAR Regression Output: 5.551922 Constant 7.843812 Std Err of Y Est 0.054499 Std Err of Y Est 0.080205 Std Err of Y Est 0.042197 R Squared 0.976244 R Squared 0.967698 R Squared 0.984402 No. of Observations 5 No. of Observations 5 No. of Observations 5 Degrees of Freedom 3 Degrees of Freedom 3 Degrees of Freedom 3 X Coefficient(s) -0.00957 Std Err of Coef. 0.000862 LACTIC ACID/GLYCEROL VOLNIC Regression Output: Constant X Coefficient(s) -0.01202 Std Err of Coef. 0.001268 TOTNIC Regression Output: 2.075642 Constant X Coefficient(s) -0.00918 Std Err of Coef. 0,000667 TAR Regression Output: 5.460167 Constant .764925 Std Err of Y Est 0.063232 Std Err of Y Est 0.060095 Std Err of Y Est 0.055025 R Squared 0.982216 R Squared 0.979209 R Squared 0.967753 No. of Observations 5 No. of Observations 5 No. of Observations 5 Degrees of Freedom 3 Degrees of Freedom 3 Degrees of Freedom 3 X Coefficient(s) -0.01287 Std Err of Coef. 0.001 X Coefficient(s) -0.01129 Std Err of Coef. 0.00095 X Coefficient(s) -0.00826 Std Err of Coef. 0.00087 LEVULINIC ACID/GLYCEROL VOLNIC Regression Output: Constant TOTNIC Regression Output: 2.253134 Constant TAR Regression Output: 5.519603 Constant 7.846661 Std Err of Y Est 0.055559 Std Err of Y Est 0.049044 Std Err of Y Est 0.055352 R Squared 0.988778 R Squared 0.988299 R Squared 0.976543 No. of Observations 5 No. of Observations 5 No, of Observations 5 Degrees of Freedom 3 Degrees of Freedom 3 Degrees of Freedom 3 X Coefficient(s) -0.01428 X Coefficient(s) -0.01234 X Coefficient(s) -0.00978 Std Err of Coef. 0.000878 Std Err of Coef. 0.000775 Std Err of Coef. 0.000875 SLSL 89603
Page 34: dha73d00
8 7.9 7.8 7.7 7.1 7 6.9 6.8 0 10 20 hILTL~Z LE~tGT~ mm70 Figure 9 80 90 100
Page 35: dha73d00
6 W Z 5.5 ~--4 ~ ~ U z 5 4 0 10 20 fILTffR LENGTA mm70 Figure 10 80 90 100
Page 36: dha73d00
Figure 11 2.5 0.5 L 0 I I 10 20 i PILTffR LE&GTA mm70 i 80 90 100
Page 37: dha73d00
Accepted By Approved 0/ 0 lds Don deBethizy DISTRIBUTION: Cover Page Only_ Mr. C. Blixt Dr. W. M. Burger Dr. C. W. Ehmann Dr. W. M. Hildebolt Mr. B. V. Hardin Dr. A. W. Hayes Mr. C. C. Hein Dr. R. A. Lloyd Mr. D. R. Pugh Dr. W. S. Simmons Dr. M. E. Stowe Mr. R. L. Willard U o ohn H. Reyn Complete Copy To: Library - Original Dr. Bradley J. Ingebrethsen Ms. Cynthia S. Lyman Mr. Barry L. Saintsing Dr. Thomas A. Perfetti Mr. Dallas B. Blakely Dr. J. Don deBethizy Barry A allas B. a aintsing ~li 0 17

Text Control

Highlight Text:

OCR Text Alignment:

Image Control

Image Rotation:

Image Size: