Tobacco Documents Online

j 10 ) A P t a -3 a 3 Some previous studies of human "tar" deliveries based on analysis of retained nicotine have assumed that the filter efficiency for nicotine removal and the delivered "tar"-to-nicotine ratio were totally independent of flow rate and therefore remained con- stant from machine smoking to human smoking. This assumption did not significantly distort the results -obtained for undiluted cigarettes, in which the flow rate through the filter is a relatively high value for machine smoking (1050 cc/min.) and an even higher value for human smoking. 'However, filter efficiencies can change rapidly as flow rates become small Ji.e., every filter is 100% efficient at zero flow rate). Therefore, for highly diluted cigarettes, where flow rates may be well below 1050 cc/min., there can be large differences in efficiency and "tar"-to-nicotine 5/ ratio as between machine and human smoking. 5/ In general, the filter efficiency for total par- ticulate matter (TPM), nicotine and water is a func- tion of the flow rate through the filter, since the longer the residence time of the smoke the more probable that the smoke components are removed. The [Footnote continued on next page.]

1 ) ; 11 Therefore, techniques that ignore the variable nature of filter efficiency and "tar"-to-nicotine ratio will lead to large errors when applied to cigarettes with substantial dilution. Experimental Method ) f 3 i D In addition to Barclay KSSP, two cigarettes were used in the study for comparison. One, Cambridge RSSP, has a measured FTC method "tar" delivery of 1 mg. and an unlit machine dilution of about 78%. The second,'Merit KSSP, has a measured FTC method "tar" delivery of 7 mg. and an unlit machine dilution of [Footnote continued from-previous page] delivered "tar"-to-nicotine ratio is also a function of flow rate since filtration of "tar" increases more rapidly than filtration of nicotine at reduced flow rates. This is because the nicotine and "tar" removal processes in cigarette filters differ. Nicotine removal is gener- ally by particulate filtration. However, much of the "tar" produced by a cigarette puff is volatile material in the vapor phase that can be removed by the very effi- cient process of condensation. As the flow rate is re- duced, the nicotine removal increases hut the "tar" re- moval increases at a greater rate because of the very efficient condensation mechanism coupled with the normal particulate removal. Therefore, the delivered "tar"-to- nicotine ratio increases greatly with increasing flow rate through the cigarette and filter.

12 ) I ) V j y 28%. Cambridge was selected because it is similar to Barclay in~"tar" delivery and dilution when tested in smoking machines. Merit was selected because, in other tests, it has been shown to be similar to Bar- clay in "tar".delivery and dilution when smoked by human smokers. A panel of 13 regular smokers was assembled for this study. included were men and women with high, medium and low flow rates. Each panelist was given four cigarettes of each brand and asked to smoke them at their leisure, saving the butts. They were also asked to puff each cigarette, lit and unlit, on Philip Morris' on-line monitoring equipment (Figure 1). Mea- surements were made of the rod flow rate for each smoker for each brand. From these measurements, the retained nicotine and the individual smoking charac- teristics of-the subjects were determined. Tables 1-3 show both the individual and average values for Bar- clay, Cambridge, and Merit. ~ ;:

~ : ~ r 13 Filter efficiencies and delivered "tar"-to- nicotine ratios were determined for e4ch brand from ) ) machine smoking at several dilution levels. Measure- ments were made of TPM, nicotine and water, both on the filter and on the collecting pad. The delivery and retention data were fitted to a power function, as described below, and this function was used to calculate the component efficiencies as a function of lit rod flow rate: Filter Efficiency = Amount on Filter x 100 (5) Amount on Filter + Amount on Machine Pad Experimental Results a ~D- The standard method of measuring cigarette dilu- tion is to use unlit cigarettes. However, since the results reported in this document involve the smoking of lit cigarettes, dilution values on lit cigarettes are used to calculate percent dilution -- i.e., the split of the total puff flow between tobacco rod and dilution holes. The percent dilution measured on a lit Z-)

14 • I D cigarette is greater than that on an unlit cigarette because the hot coal (900° C) increases the viscosity of the air being pulled through it. This increase in viscosity leads to an increase in resistance to flow at the cigarette coal. Since there is then an added resistance to flow down the rod of a lit cigarette, more of the flow is drawn through the dilution holes on the filter giving a higher percent dilution on lit cigarettes. Figure 2 shows the lit dilution values as a function of the unlit dilution values. These data were derived from the on-line dilution measuring equip- ment when both lit and unlit dilution values were mea- sured for each cigarette over a wide range of flow rates. A nonlinear, least-squares fit was made to the function: Power Lit Dilution = Constant x(IInlit Dilution) (6) 9 in order to facilitate applying the flow rate correc- N O N r* Z; tions. Figure 2 shows very little scatter in the data r CA points, and this is taken to show that the same correction ~

I D , 3k 15 is applicable for each of the three cigarettes for any flow rate. Using the data reflected in Figure 2, the unlit dilutions derived from standard machine measurements were transformed to lit dilutions so as to plot work- ing curves in terms of lit.rod flow rates. Once this correction is applied, the component values retained in the filter and those delivered to the smoking machine pad were fitted to the following general equa-• tion: Amount = Constant x (Lit Rod Flow Rate) Power Figure 3 and Figure 4 show the best fitted lines and the experimental points for the delivery to collect- ing pads and for the material retained in filters for Barclay. In Figure 5 and Figure 6 are corresponding values for Cambridge and in Figure 7 and Figure 8 for Merit. Table 4 shows complete equations derived for all components and cigarettes. Once the best curves have been derived by non- linear, least-squares analysis of the data, nicotine (7)

) I ) ) - D .t 3 3 16 filtration efficiency versus flow rate can be calcu- lated from the retained nicotine and the delivered nicotine data using Equation 5. The results are shown in Figure 9. Note that the efficiency curves for each cigarette become asymptotic to 100 percent efficiency as the rod flow rate approaches zero. Also observe that the Cambridge filter is more efficient at every flow rate than the.Barclay filter. This is a highly significant result. It means that even if the same amount of nicotine were retained in the Cambridge and Barclay filters, the nicotine delivered by Barclay would be much greater than that delivered by Cambridge because of the greater efficiency of Cambridge's filter. In fact, as our studies show It~. (Table,6rj, Cambridge's retained nicotine was substan- tially lower than Barclay's, indicating an even lower nicotine delivery.by Cambridge relative to Barclay. Discussion of Results The equations have been developed for describ- ing both the component delivered and the component

: ) I ! I 3 , V 17 retained on the filters as a function of flow rate for the cigarettes. Curves of nicotine efficiency versus flow rate and "tar"-to-nicotine ratio versus flow rate• have been constructed from the data. (Figures 9 and 10.) This information can now be used with the smoker's flow rate, dilution-in smoking, and nicotine retained on the smoker's filter to compare the rela- tive "tar" and nicotine deliveries of the three brands to each smoker. The results of this analysis are shown in Tables 5 and 6. Note that the relative "tar" deliveries of Merit and Barclay are basically identi- cal in Table 5, while Barclay's nicotine delivery ex- ceeds Merit's by approximately 50 percent (Table 6). This is due to the substantially higher nicotine availability in the Barclay blend (2.8%) than in either the Merit blend (1.5%) or the Cambridge (2.0$). Similarly, Figure 10 reflects these differences. The consistency of these results for each of the panelists, with very different individual smoking characteristics, is remarkable. These tables confirm 2

18 a . I ) "4 .4 that Barclay, when.smoked in the lips by human smokers, is equivalent to a cigarette in the range of 7 to 8 mg. "tar" by FTC method, and far above the 1 mg. range.

1 ,.. ... sp 0% o*`. ~ •aft