a~ (U1CandExp~rtj limited REsEARUI I DeYILDmCNt eINTRE Wvtl~wlOn U10~0 TO OR. R,R, BAKER FROM OR. N, EGI1MEt RLF GATE. NE/3P/19 t7Uh July 1997 STAWS REVIEW OH AEROSOL RESEARCH (As Part of Chelnosensory Research) ~ynti~etic Aerolols - Effect pf Particle Slie.an senrory Perception The influence of aerosol characteristics on sensory perception has been investigated using synthetic aerosols of menthol. (In Scuthampton) and nicotine tin Louisville), simulating cigarette ake, As cigarette smoke consists of particles in various sites and chemistries In different concentrations, the approach In Southampton was to use fairly monodispersed, sub·micron aerosols which could simulate a fraction of cigarette smoke at any one time so that the Individual effects of different particles on sensory perception could be Identified, Southampton's work has shown that 0.70 pt nedlan·d~arneter particles give significantly more impact than 0.55 P particles, These tro sires are present in tobacco smoke~ Therefore, the implication would be that larger particle fractions In smoke could give more Impact, If they have the right chenistry~ Experiments with varying sass concentrations also confiwed that more impact was perceived with increasing aerosol mass. Loufrvfllelr work har~rhorm that very polyd~sperred aerosolr at varying pHs gave impact sensations which were very much dependent on the puff nurnher, Although the nicotine concentration was similar to the levels in cigarette smoke, the particle size·distributian was very different to smoket therefore, a few very large particles carried the majority of the mass onto the taste buds, more efficiently than the cigarette smoke particles and probably at larger quantities, which caused a saturation of the taste buds, The outcome of the two areas of work was that, although the particle size affects the taste perception, aerosot chemistry, mass concentration and particle size distribution are also very important factors and should be considered simultaneously in perception studies, cigarette Snoke - Effect of Clgaette Parameters an Particle eharacteristfes In the next phase of the aerosoilchemosensory work progranme, cigarette design parameters which influence the smoke particle site have been Investfgated. O Clit; PB F-!:!!:!!::!.t3·Stl 8:EO I·i~ (a) Products with additives which enhance nicotine deliveries It was found by P,C, Bevan that some additives, when applied to tobacco, increased the nicotine and nicotine-to-tar ratios in the cigarette smoke, Of those tried, solanesbl was found to cause a 7", increase in total nlcotf ne del ivered~ particle si ze measuremnts showed a 3", increase in count median and a 6! increase in mass median diaoleters. (b) Extreme filter designs iwo filters were made by J~A, Luke which had fine filaments in high number density and coarse filaments in low number density. They were compared at similar (f10%) TPM deliveries obtained by 25'P laser ventilation of the latter, The coarse filament caused: (i) 6: increase in count median hiameter s (11) tO: " mass (iil) 30: " " nicotine per puff (iv) 19: r II nicotine·to·tar ratio Some of this increase was thought to be due to ventilation, which is also known to increase the particle si2e and nicotine delivery, (F) ultra-slim vs some other low delivery products Comparison of CAPRI with similar delivery products of larger circlrmferenee has shown that 25 mm circumference products gave particles which were: (1) r2: bigger in count median diameter " " mass in comparison to CAPRI, which was 17 nm in ejreumfereneel (d) Swirl vs non-swirl products of ultra·lcw delivery (about 1 mo TPM) ihese products were prepared by M.G, Ouke and had very variable particle characteristics when experiments were repeated with cigarettes of the same kind. This intra-product variation has been observed with some other low delivery products as well and was experienced to be consistent when the deliveries are so IcH~ Nevertheless, on average, the particles from the swirl product were only slightly larger than those from the non-swirl product when fresh smoke characteristics were analysed at the butt end~ However, when the measurements Here repeated Inside an artificial mouth, the particles In the amoke from the non·swfrl product, which was in the form of a thin, dense plume, seem to grow by aging (eoagulation) and slightly exceed those from swirl filters~ From the measurements taken at the back of the mouth, corresponding F~ to the throat location where the impact Is perceived, the most O apparent difference between the two products was observed to be in the number counts~ Here particles were counted for longer periods N C I i I; P D F --~::a::~::l.f 3 Sf I.C...i: 0.111 · of time with the swirl products than with the non-swirl one. The implication would be that, due to a more homogeneous distribution of smoke in the mouth resulting from the swirl effect, the sensory receptors are exposed to a higher amount of smoke for longer periods of time. overall, particle site of swirling smoke did not' change either with aging or location variation in the mouth as the measurements taken in different sampling points indicated, (e) various puff parameters vs aerosol characteristics t:n order to investigate the influence of variation in puffing characteristics on smoke aerosol properties, aerosol measurements were taken under different machine-smoking parameters such as puff volume, puff duration, puff profile, Is well as variations during the puff~ The results can be sunanarised as follows: · As ti~e puff voi~ane ii increased froa~t~ cm3 to. 70 cn3/2 sec~, the count median diameter decreases frcP 0,2rJ 'to 0.16 ~. The mass median diameter is de~j~hfrom 0,33 to 0,23pl, The nu~er concentration increases with increasing puff'vollrne, The Jile dlstrilut!on also changes (69 : l.a to 1,39 respectively), - As the puff duration is increased from 1 to 4 seconds, the count and mass median diameters increase from 0,16 to 0,21JMI and 0123 to 0,34 P, respectively, The number counts decrease with increasing puff duration, Geometric Standard Oeviation is also increased from 1,39 to 1,48, - The measurements taken between 0,2 to 2,0 seconds into the puff indicate that the count median diameter Increases from 0,14 to 0~22pr with increasing time, The mass median diameter is highest (0.41P) at t -- 0.2 sec. it falls to 0.21Con, then Increases up to 0,34 pm at the end of the puff. The number counts reach a peak tlovards the middle of the puff as expected, The geonetric·standard deviation is highest when the number counts are lowest to,: 1,47 · 1l48)t particles are more holaogeneous (~ : 1.43) nhen the counts an high, In the middle of the puff. · The variations of particle characteristics in the middle of the puff (t · 1,0 sec.) have also been investigated as a function of.puff number. the number counts increased from 1.5 to 3,5 x 10y rml3 with increasing puff number, The median particle diameter slightly decreased with increasing puff number~ Conclusions It has been shown that smoke aerosol properties do influence the sensory perceptlon~ It should be renelbered that the term "sake aerorol" refers to the whole of smoke, with particles and vapour phase, Smoke aerosot studies should not be oversimplified ~conldering it solely as particle site measurements, As shown in R~O, snake particle site deer influence sensory perception, bit this effect can only be valid if O the particles carry the right chenistry in the right mass concentration O levels detectable by the sensory reeeptors~ C: I i I; P D F~-~::a::~::l, f3 Sf.l~b~.~: D 1~~ -- - A lot of work has been done in analysing the overall smoke for Its chemical characteri stics, Particulate phase studies have been performed by filtering out the whole smoke onto a Cambridge pad and analysing the filtrate as a whole, however, very little is known about what fractions of sloke are responsible for sensory perception, such as impact or irritation. Knowing that the cigarette smoke is composed of a large number of chemicals, and that some fractions of smoke are transported onto the sensory receptors more efficiently than the others las implied by chemosensory research findings), a lot mon work would be needed to identify what actually contributes to the sensory perception of cigarette smoke. Limited work performed with products or varying designs suggests that smoke particles are slightly larger when (a) additives are used which enhance,nicotine transfer; (bJ the filter is made of coarse filaments of low number density; (C) larger circ~rmference tobacco rods are used. 1 Smoke is delivered to the sensory.receptors ~n·larger concentrations if the filter generates swirling smoke, which is more evenly distributed in the mouth cavityl If all the sire fractions carried the same chemical compounds, in the same quantities, the differences In particle size distributions mentioned above would be of little significance in infiuencf ng sensory perception. However, if different fractions had different chemistries las suggested by some other researchers of tobacco smoke), the increase inthe concentration of a single site fraction would be of importance if it carried, say nicotine, more than the other fractions of smoke, even if the median particle size changed very little, The largest changes In the smoke particle sire distribution have been observed as the puffing parameters were changed. particle sizes increase: (a) with decreasing puff volume; (b) with Increasing puff duration; (e) with time, within a puff. fhe puff number influences particle numbers more than particle site; number concentrations increase with increasing puff number, ~hese results are all In agreement with the external work perfotmed under contract for 8AT Hamburg (see Appendix), the other parameters which can be looked at are the influence of puff profile, ventilation, humidity, delivery levels and blend variation an smoke aerosol properties, The implication of these results with different puffing parameters is that larger changes can be achieved In smoke aerosol characteristics if smoke is somerhat influenced at Its early stages of generation, d Coagulation is the easiest way of achieving big particle sites when highly concentrated lerorols are left to age. ru C: I i I; P B F-r!r:!r:!-f3 SPI'CX: 0'11l --- ---------- Ci~arette wire is bcileved to have 1012 partlclesld at Its falma#on zo~e in the vicinity of the burn line, The n~unber concentration is reduced by 1000·fold as snro~e leaves the butt end and changes little when passed through some extreme fitter designs avaitable, Enc, cc: Or. O.P, Robfnson plE . a i~Cbi· ~~ 13"he~; ,I PF-Sn 21~,02,1986 Determination of the mean oartic!e size in ciParerte smoke in the ori.sinsl smoke Concentration Contents 1, Aim of the project (external contribution) 2 2, Method h 3 3, Investigation of the predictive value of the a~et~od 3,1 Effect of refractive index · ' 3 3,2 Significance of the mean diameter ~ b 4, Measurements k,~ leproducibilft~ b 4,2 Effect of the mount during draught 8 4,3 Climatic effects 9 b,b Effect of the draught number 10 4,5 Effect of the draught volume rl 5, Summary Appendix b: Description of the CQ - Method Appendix B: ~w~difLat:oa of the apparatus Appendix C; Note on t~e statistics for measurements using the D[1- method CO Clit; PDF -!::!!::!!::!.f3 StlC.i: 0111 i, 1IIIOF n~ PROJECT : A DS~S~ICUIT SUB-PP,OBLFM IS NE~SUS~iT OF P~e~TfCP SIZE, SINCE i, THE CICARmTE SIYORE PARTICLES CONSIST IN P~45T OF VOLATILE SUBSTANCES (WHIQ ~A~lj ~I~i~ ~TriODS OF ~IWSJRE~EX sml Pson~TIal AT HIGH DIIOTIONS OR IN ATYPICAL FLOGI CONDITIONS) ?, THE GENER~TION IWD IU~S~T CON3InONS OF TIIE mosE PARTICLES C~NGZ DRZSTICI~WIY DURCIC TiE DR\'1'C~ (~WCII LY'C~UI~TIQ~ rjmt SEliSORY PROPERTIES MAKES FOH I REPLTII(F~II~Sr Mtl VERY RAPID ~IJBPIEaT), A IIEVI~7~ OF THE M~STING CIGbmTE dEBOSOL PARTICLE EIE~SUEF~TS ILSD Tii~IR PSOBLY"IS KILS BEE~I CIYEN BBIEFIY M [1], E AM OF TKE PROJECT IS NObl TO FIND A IYEI~IOD ~I~CH dU~S TIIE MI;IS~G~f~h~T OF aZ EIE!N SWP;E PARTICLE SIZE UNDER TIIS FOUOWI1C ASSUEIPTIONS: 1~ FIRST MEASURE~~I~T H)INTS AFTER TL\LES 6;HICH CORRESPOND IO TIIE P~SShCE TI~IE OF TaTER TIP J~S, 2, VERY SHORT MEASUR@IENT TL\1E, MULTIPLE ~SUIID[ENTS DWIBIC OEIE DIUUGIET 3, ~SUP~E?IEBIS IN IMDUI~I~D SIEIOI(E 4, NECESSARY TO ILIVE ~iO SIGIZGFIGIBT CHANCE IN TrIE nOP VELOCITY B~IUEEN FILTER TIP MID IYF~SUI~E~T LOCATION, -3- P ClilTP'D'F- ~::! .fsal·c~-:i3·lil -- ·- 2, Method A method which promised to fulfil these requirements IS presented In 1980 [2], The assignment proposed by us in 1982 provided for a modification of this apparatus for measurements on cigarettes, h description of the principle and the basic construction is given in appendix ~, the subsequent modifications in appendix 8, Particular Sportance was attached to the equipping of the measurement chambers in respect of optical depth and the avoidance or' vortices, A iur~hcr important point ES Lhe recording and connol of the scnenr of measurement (or the electronics associated with it), 3, investigation of the Predictive value of the method ~1 other optical methods of determination presuppose a knowledge of the refractive inlier, and this a) is not known a priori for the smokes of different types of cigarette, b) fan in principle vary during the period of smoking c) varies for different cigarettes of the same type 13], baking use of the refractive indices from earlier studies [3] (4] [5], the effect of the variation of the real part of the refractive index between 1,1 and 1,6 and the effect of the imaeinary part on the evaluation of the mean particle diameter was calculated (see also appendix a, pp AC,AS), 13hile the imaginarp part for realistic values has only a negligeable effect, the increase in the real part tin a realistic framework) signifies a ndn-negligeable decrease in the mean diameter, ,'II ----~ ·· · ."'."'." ··--·.~.,,.,,. Examples : Assumed refractive index : n. 1,15. 1,54 1,59 reau!ranr mean diawter d [~I] . - far given size distribution type ~ : 0,37 0,36 0,33 type 2 : 0.52 Or~5 type 3 1 0.63 ~~ -' · 0.5 For comoarison: if one assumes the proportion by mass of volatile components in the aeroso! pa~eieles to be 30X, then the [mass averaged] dianetP1$ would 5e reduced by about 11X on complete removal of all volatile coiponelts, This corresponds to the order of magnitude which occurs when the refractive index n changes from 1.15 to 1,54 (for d · 0,37), -5- rrC-"VIL' Cdnsc·ucn~?s i. Attention must be paid When comparing the mean diameter for different smokes to the fact that a reduction of the mean diameter (- Jd) can be a result of anl of the following alternative causes: ad physical chemical ojlenge: reduction~ alh:r phpical \ Napntion & ' d] veri. processes vola~~ie coiaponen:s nl coast, n varies 2. Actual differences in diameter may remain hidden in the presence of appropriate chemical changes Which lead to a change in the retractive index if - the diameter increases and a chemical change occurs which is such that n increases - the diameter decreases and a chemical change occurs such that n decreases, n) n: refractive index ·6- · C: I i I; P D ~r~l(!(!:f 3T t I'CTi"O'l ii 3,2 Significance of the~nean diameter The advantage of a rapid measurement facility in the original concentration of the smoke using the Dp method is obtained together dth the disadvantage that only information on a mean diameter, but not about the particle sire distribution, is obtained, An advantage again is the fact that the method supplies approximate!g the volumetric mean (see appendix C) so that considerations of sass Ire pl.i~le nco vl~h unkaaP I~srti~u;ions. Appendix C gives formulae for the con?ection beken the volumetric and arithede means for the Gaussian and the log normal distributions, 4, ~!easurcnents Systematic measurements for the detenination of special design effects are the first object of the folloving investigations, The measurements recorded here belong essentially to the establishment of the method, S, ileoroducibilitv Tne follovLg table sho!~s the man smoke particle size of the same cigarette type, measured in I different months, Measurements vere made at b different mpnents (el to tl) during the draught, -j- Clil; PDF -!::!!::!!::!.f3 StlC.i: 0111 Table 1 ban Particle i?lapeter d [r.) tloment during the Drau~t tl Standard Cigarette 0,38 0,36 0.36 0,52 ist n)eririon 0,38 0,36 0.11 0ljj 2nd repetition 0,36. 0,35 0,40 0,54 3rd repetition 0,3i 0,36 i 0,39 0,53 It has been found that a RaximuP of 40 cigarettes are sufficient for the achievement of reproducible results, with several draughts being made from each cigarette (see here 1,4), The size distributions shoun in fig, 3 are tVpically obtained, whereby mean~l~l particle we are here concerned with a distribution of the mean size of different'(l) cigarettes of the sane tips, The mean values of the mean particle sizes are found to be very reproducible, where however within a measurement series a considerable number of cigarettes with clearly differing mean particle sizes d oaur (aunlid ample: ah inr~asl In i irm 0.1Y ~o O.ju maaoi a doubling of the mass!) here the possible variability of the refractive index should be stressed (see 3,1), Soundly reproducible mean values of d map mean not only adequate averaging out of physical but also of possible chemical differences, -P- 03 Clil;PDF - !::!!::!!::!fasfle,cnl~;l The order of magnitude of the values is 10 - 2rJ 2 higher than the values - measured on other cigarette types - which can be obtained by other methods wt:ich alloy a rapid measurement during the draught at the original concentration [6] or after fixing of the particles [7], Ibth methods [6] and [7] are expensive on apparatus and method [7] substantially more difficult to exccure than that here described, 4,2 ~!ean oatticle si:e as a friction of the moment of time duriln the ---- drauptt $ Using the quasi-continuous data acauisition·aet~od described in appendix R, the mean particle site d during the draught was determined ton the same standard cigarette Is in table 1). The values agree yell with those in cable 1 for tl · 0.5, t~.1,0, 5 ·1.5, t1 · 2.0 sees. Noticeable is the clear rise touards the end of the draught (vhich, on the assumption of constant dendef and refractive index, represents almost a tripling of the particle mass), The lest three quarters of the draught process agree qualitatively Kith the measurements of [6], .g. C: I i I; P D F·-r!rlr~:f3 St I*CTi"O'iil X pwntitative eqlanatloo of this rise in d towards the end of the draught here must make allowance for the (bell-shaped) draught profile, the temporally varying temperature gradient end absolute temperatures in tile smoke formation zone, the changing shape of the combustion core, reevaporation effects etc, 4,3 Climatic effects The effect of the ambient climate and p~etreatsent o? the measured variables was determined in a rough. preliminary test, "ceasuraent A: Standard cigarette, ftesb'~om the package leasu:epent B: " n , open several daps and placed on a warm substrate hasutenent C: " " smoked in a very moist atmosphere Table 2 mean particle i~ia~ecet d Ir] durina draught Itotn,55 ·· t~l ,s tOt2 h ,37 0,36 0,39 0.53 Ci~~ R ,37 1,36 0,38 0,49 Cia, C 0,116 0,3F · 0,52 0.5? d if the effects were genuine then the reslllts could mean that in case C the Dean particle;dsses at the beginning of the draught are 70X higher than for case A, Here accurate investigations are recluired here, 4,a Effect of the E:et~tnuaSEr Tl~e effect o~ the iraught number, or point bf origin o: t~. smoke and of the subsequent stem length on the particle size was relatively small (rhis is in agreement with [7]),~ clearly less than the effect. of the mamen: during I:suShl [as in [6j). ? 8,5 Effect of the drz~~~t volume ·~ ""' First tests:Jere carried our on a standard cigarette for three different draught ~oluPes (fig.3), L with other methods of measurements (see e,QI [7] - 19]) there is a decrease of eean particle diameter with increasing draught volume. This may involve many processes (see also i,?,), a closer examination not only of the trend hut of the order of magnitude is du:e 5, SuTnzrv i~ oethod for Pe!sa:fnp Ge mean particle site of the ciga:otle aerpsol within a very short space of time during the draught and at the original sno~e concentration has been developed, .O -11· h) Clil;PDF .-i.lrlrl:f;istlC.;:nljl In the interpretation of the nezsured Cia, allowance must be made for the facl that a part of the variation in the values can be attributed not to actual variation in the particle diameter but to ch~i~~as in the refractive index a, determined by changes in the chemical eoa)osfcton of the particle, The repronucibillty of the mean values of the mean particle sizes for the rnoke of different cigarettes of the same type is good, There nay be large differences between cigarette and cigarette, which can LmrLlpad di~dll to a deulli4 tt ~.pat~ide sllc.l) DPrinp 1 Iraugh~ & pr:itL r:1 can lenare bl a faclpr af 9. 11 The differences between the individual draughts (the first and last excepted) are less than within one draught, The familiar phenomenon of decrease of the particle size xiih increasing draught volune Is evident with t~is method, Pr, Schneider n · constant asswed -1?- d Clit; PDF -!::!!::!!::!.f3 StlC.i: 0111 Bibliagraphy [1] J, 4(ijOHSi, pre)ll,nt (1985) 12] S, rjitti8, et al,, CI~F - Keecing (1()80) [31 D.a, ?nrp~e, J, Colloid nnd Barfcee Sei~nc· 87 (19n), 1 [6] 'I, Ishin, T, Okada, J, Colloici and Interface Science 66 (1978), ul - 239 [5] i~, Lrl:er at al,, Ipplieq flptics 17 /11713, 19 [6] I, Ingcbrethszn, TC;IC (1981) [7] C,H, geith, 6eIt;, kba~f, 7 (1973), 95·10b [e2 Y~ IsMzll e~ al,, J. Aerosol Science 9 (197a), 25 - ?9 [91 P,, RicharSscn, Tech. FxESsnee I~eeting, Scuth~pron (1983), -13- O Clit; PDF -!::!!::!!::!.f3 StlC.i: 0111