COAGULATION ANO FILTPATION OF IINSfREAn C~MREi[E SMOKE IH THE TOBACCO ROD OF A CIGARmE David IJ, Boidridge and Bradley 3, Inge:;ethsen Research and Oeveiownent Department Dowman Gray Tec~nnical tenter R, S, Reynolds Tobacco Comoany Winston·Salea, North Carolina 27102 U,S,A, KEYCIORDS: Smoke, Filtration, Coagulation I, INfROOUCTION Mainstream cigarette smoke is a highly concentrated aerosol that undergoes both coaoulatton and filtration as it travels through a tobacco rod. Measurements of tie particle size distribution and number concentration of mainstream smoke exiting the n~authend of the cigarette during a puff were made for a series of tobacco rod lengths. A model has been developed which predicts the chances in the n~ber concentration and site distribution of the smoke aerosol as it travels through the tobacco rod. The model is based on simple equations for filtration and coagulation of aerosoTs. The predicted values of the mean mars diameter, number density, and total aerosol mass are shown to be In good ·agreement with experimentally determined values, II. EXPERIMENTAL The experimental apparatus used in this work measures the particle site distribution of a fresh, undiluted sample of mainstream smoke by ensemble light scattering (ingebrethsan). Cigarettes were smoked in the reverse puff apparatus shown in Figure i. During the puff, the lit cigarette is enclosed in the retractable glass chamber and the syringe pump forces air into the chamber and through the cigarette to produce a 2·second, 35·ce sine wave·rhaped puff of smoke, A 0.8·lmn i,d, stainless steel tube selects a small portion ei the center of the smoke stream and passes this portion through the smoking nacnlne face plate, G vacuum pomp is used to aerodynamicaily focus the smoke stream issuing from the capillary tube and draw the stream through the observation region. Smoke which does not pass through the capillary is removed by a small vacuum pump prior to the succeeding puff, A vertically polarized argon ion laser beam (514 am) intersects the ~ smoke stream in the observation region. Scattered iioht is detected by in array or shielded !hotodioder (iigure 2]. placed at 106 600, 9001 iZ00, ind f~ ]500 tell:ive to ::,2 interroelting laser Learn. Lioht intensities ire ic~u~rd O every 20 a durinf the puii 5~ a multipr3oramner Pewlett·Packard 3odii 6942$) Clit; PDF -!::!!::!!::!.f3 StlC.i: Dill and stored on 1 microcomputer (Hewie!:-P;ckard model 9920). This produces a file of 100 sets of scattering intensities during the puff. Each set represents the scattering from the smoke crossing the observation region during the corresponding 20 ms time period, Ratios of the experimentally determined intensities (relarive to the intensit]( it 900) are compared with tabulations of theoretical scatterino intensity ratios calculated far log·nonally distributed aerosols with ; refractive index of 1,5J·0.0i (Ishilu, 1978). The geometric mean diameter and geogetric standard deviation Which give the best fit between the theoretical and axperieentally observed intensity ratios define the log·no~fmal distribution which is taken to be an accurate description of the smoke's particle site distribution, ihe final data canris: of 100 sets of geometric mean diameter and geometric standard deviation values which correspond to the 100 experl~ental detelminatians of the rc;tte~ng intensity. The set of data describe the smoke aerosol distribution is a function of time during the puff, The number density is caicul ated as a function of time from the gravitet.rlEaily determined total plr:iaiate mass of the aerosoi, the scatterine intensity at ~00, the theoretical scattering efficiency, and the geometric standard deviation and geometric cleat dianeters, Oata were acquired for the second puff on tobacco rods of 49-, 59·, 69·, 89·, and 119·raa length. Total mass deliveries for the second puff were determined by a separate gravinetric procedure. III. MOOE! A theoretical model has been constructed to predict the evolution of the size distribution and number density of the cigarette smoke aerosol as it undergoes filtration and coacuiation in the tobacco rod. the model is based on the following concept: T~e smoke aerosol exiting the mouthend of a given length of tobacco rod is assumed to relresznt the aerosol which has traveled through that length of tobacco rod. For longer tobacco rods, this aerosol will undergo subsequent coagulation and filtration before exiting the tobacco rod. By determining the aerosoi site distribution and number density issuing from the shortest (49 ~R) tobacco rod in the series and considering filtration and coagulation in an additional arbitrary length of tobacco rod, we can predict the observed aerosol for a tobacco rod of this total length (Ilinimul length ~ arbitrary increment). The physical characteristics of the tobacco rod (eross·sectional area, void volume, tobacco rod length, and effective fiber diameter) and the fl3w profile o: the puff are taken as input data. The effective fiber diameter us determined from the relationship of pressure drop to flow rate. The model uses the site distributions and number densities obtained from the shortest ((9 mn) tobacco rod as an initial data set. Each experimentally determined site distribution and number density is treated as an independent parcel of vrosai. For a given parcel of lerosol, the corresponding flow rate will determine the smoke velocity, residence time, and time of appearance at the end of an additional length of tobacco rdd, The filtration of this parcel of aerolol was predicted from single fiber filtration equations [Hindr, 1982). Particles were apportioned into 100 site bins from 0.01 to 1,00 micron in diaeeter according to the ~ag nPrmai dirtribution, Filtration of each bin was O considered leparately, ed the resultinp leraral distribution war described by ~ new log·50rmal distribution. Co~~ulii;on or this new log·normal aerosol distribution was predicted with analytical expressions for ·;· Clit; PD F---!:!r:!r:!·fil st~l@r;:*n'ln 8rownian coaculation in the continuum regime (Lee, 1983). For those cases where the residence time in the tobacco rod exceeded 20 ms, the :iltrationl coagulation process war repeated in tb·ns intervals until the snore reiched the end of the tobacco rod. The result of these calculations is a predicted set of 100 size distributions and number densities which describe the expected smoke aerosoi for a given tobacco rod length; IV RESULTS ANO OTSCUSSI0H Figure 3 shows the variation of the experimentally de!em;inee. mean mass diameter during the second puff from a 49·(Rm tobacco rod (dashed line) and the calculated and observed mean mass diameter profiles from a tobacco rod of 119·nm length (dotted and solid liner, respectively)~ The obs2rie~ change in the middle of the puii is Ipprauimateil and the mPCei predicts aoprnvilnately PO"n of this change. In addition, the initial appearsnce of the smoke in the longer rod is delayed by approximately 200 ms relative to the shorter rod. This delay is well reproduced by the model, Figure shows the variation in the experimentally determined number density durine the second puff from a 49-rm tobacco rod length (duhed line) and the calculated and observed number density profiles for a tobacco rod of !1P·an length (dotted and solid liner, respectively). The exesrimentally determined perk number density decreases by a factor of aoFroxi;lately 2.5. The ~odel predicts a change in the peak number density of 2.6, Figure 3 shows the calculated and observed mean mass diameter profiles for a tobacco rod of 89·lmn length (dotted and solid lines, respectively), The , calculated and experimentally observed curves are in good a~reenent~ The calalatian slightly underestimates the change in the mean mass diameter at the start of the puff. Figure 6 shows the calculated and observed number density profiles for a tobacco rod of 89-lmn length (dotted and solid lines, respectively). The calculated number density slightly overestimates the reduction in:he maximum of the number concentration. The accuracy of the theoretical prediction of total mass removal by the tobacco rods is illustrated in Figure 7. The experimental (rectangles) and predicted (circles) total aeroso1 mass are presented as a percentage of the total mass observed for the shortest (highest delivery) tobacco rod length. The agreement here is excellent, showing that the single fiber filtration equations succeed in de prediction of the major nars removal meeianisa~s in the tobacco rod. Small deviations are observed between the theoretical and experimental number concentrations and mean mass diameters, particularly at the start and end of the puff. There deviations pmbabiy arise from the c~oiee of the coaaulation equations used in the model. The equations used here (lee) were derived for continuum regime aerosols in a static gas. ~he cigarette smoke u6sdl is actually too small to be accurately described by the continuum regime lodels, and the tlow of the sPoke stream maker it subject to shear coaaulation, in spite of the manic necessary rimptii~ing assuwtions, the ptedic:ed results are in good general agreement vith the ex;erinental observations, ·j. Clit; PDF -!::!!::!!::!.f3 StlC.i: 0~1P V. CONCLUSIONS Ue have experimentally determined the number density and mean mass diameter of cigarette smoke during a puff at a series of tobacco rod lengths. Both the number density and mean mass diameter show significant changes due to filtration and coagulation as the snake passes through the tabwco red. The simple theoretical IwdeI presented here is capable of predicting these changes with reasonable accuracy. Clil;PDF - !::!!::!!?.f35tlc.;:nlil Y~ REFERENCES 1, Hinds, W. C, Aerosol iecinoiogy. John Wiley and Sons, New York. (1982), t. Ingebrethsen, 8. 3. "E~oludoo of the Particle Sire Distribution of Mainstream Cigarette Smoke During a Puff". Aerosol Sci. Tech. In press, 3~ tshitu, Y and Okada, T. 'Oetemination of Particle Sie Oirtributian of Small Aerosol Particles of Unknown Refractive Index by a Light Scattering Method", J. Con. Interface Sci. 66:234 (1978). a. Lee, K, W. 'Chlnge of Particle Size Distribution During Brownian Coagulation". J. Colt. Inte~ace Sci. 92:315 (1983). .rr. · 1114 Y, I ij Dx *I ->3* r 0 o 'E ~ c 0 I I ' I 3 3 vr X x u I 0 1 r I r I r k L L c '1 0 00 O rO I OZ 3 C u i-0 O U O 0 00 U~ c3 D e a ot ro O N Es O C I I b P D F--!;! N ~~d Sf ID~:~~1 C C A" "g c n v B noa 2. Schcatie diagram ot tbs Icattarcd light dctsetot head. A) rotatioo stage, a) vacuum pert, C) shielded photod!odc~, D) lalar bsP. CllbPB~- v~~~astio.soni ol~ v3 rB I, ~V ~rp r I :PGI Cu~ 1 I : PI la00u rEO~e ~*Out PIP ~~ ePOO1 ~XYY ~e ea 10 Iv W I ~un~~ + a OIPO WI 0~0(~ h CI r OIE~O ND IEO I e n "O P c 6e~iur Eixai 81 Ijd ~ V IE~ (11~1 SP ir it o* gui r C U 'UC I QO I~ o~Ew~ t I ~~U~UC O ( Y~Ye I u0110~ W d I 1 E bP I (plX a r IOil 1 I ~PQ 'O i FQ I~Q U~COC I (ukO IEe u < ~E~lO 8 IIE~PE n~ cE 8 ee1141 I~uCO0 I yOO~u~ n 8 Y i d i 1 (Ilu*l, 1·ull a o N 9 U N ClibPDF - v~~fastio.soni E IYu e r we o~O =0_c~i~ B Pie'~ iEmO I I ~u o~lrro ~~mO*iO uacv·lo x u m DsaZFC e OV / rigC~J / t k~~ii~i cu~roE / m ~ uaci I e~ Z .'I I" ri~OC ~ / O YO)~ O aeoEo H deE~O~ f ~·rltPk~ F· Plrl ID E; (r n I LCOC n! pOX(1u / Oa~uoOui NP a~ 'C aicPP~ Z "O EnO C. O W o pC*IYil'l > O U I I m ~Eo E~ Z 'Y C; Ik·~ uOv i~ O W Ocmec ulliPI:~~ U Z d 0dAu ~J n Oe~ E t too oc 03 u OL : 0001 W ~~k=OL~E ID QOPlO~m m kYs,ll~C Z u E j eg c: c 3 P OI t~ eeac~a \ O~Bklm \ Or' C k~ ~10~1 :" EOc~ N kUP13 PEaE808Uu~Ecl :LIDCOI Zu~sOO seseeeem m ~s a m s~ II iii~ ~I .N~m~rN8~~*~ NNN' (BI13~33/UnN) NIIIP~lm~O~ o N a 03 \o ~Ei w Clil;PDF~:f;i~~fl~cT;i3~iil Es OE [P s~~rl f*fyi I~ i"O I; 00 I :Oupl 1110d \ t ~~$o~S \ 111 -4*1 ri LX p0 :~~PIO 12: cC W dl·rpl \ t eJr4u 41 W CPOvI r: H ;"IE~i a: (U O ;YO c ELl~e H "O O~~~*I O "'' C''' O C QcO CO :~~up H ~~OePI L~ v ~4mu lr) j;p~E~ (r I W ·~ ·IOp r ~CL~L100 Z ~J W OP1 k I m ~'X10 Z OIIPI~IU L3m QO i IP CILvOLt91 W I 3 r i " d~SY~~ i ~O*PO 'P eO'O~OE :: PE i 1~ O~aal ~sIEYh a E·rl OD m ei ;r tu ~E 1k111 m e~pl~OlE aourias o u P10 L eo EjptlOu V) LI1 ~ m m ru m a (~U0~311) 113L31b~In 4 O h) 03 W YIJ Z uu I t :C 0 EQE'OI PESIE 8 I~l~uO 1~*0u y ug*udO C IC ex~F 1 eo O c I, ~sO~F~ / OuuOLt r C : O illEde Z O~Yl~u O I- 0 dCoo / OOP~O~ n · 1 liO Ik: u t I+ II Iee / O~~CUQ OX 0, I ,/ , (OulgC I rb 10 OLEIIO YO C, oCa Z ' C IO~uOL Y i 'P CLl O ) 0 I liOF V II 'Y I C k i I! 2 ElOe V u~)C~ O W O~OluF U Oi I \ r 000 eu Oo000e C iO *ii uO~iu: W 4 llQ040 ~eai~e m u~l ul r i " OEC r \\ ol 1001 \ eioocpo \ Oa1kOuk 0=0 0 8 *O 10 ~sPPr~o I 1 ~01 0 N fj:df:: Zu~uE~P7r10j10 B i 9~808~0~8~889 *#lmO~~Q00*N~ LI [01~t33/YnM) HOII~~H~03 tI:BWTM tl O ClibPDF - v~~fastio.soni ioo i- o 8 O 6 i 40 L 0 20 10 60 80 100 fiO TDBICCO ROO 1ENG~1 (mm) PIGIIRE 1. Percent mass dsliverl relative to that of the shortest tobacco rod. O - fxpcrimentally observed deliveries O · Calculated deliveries Clit; PD F~'