Physics, Chemistry, and Dyftarnics of Interplanetary Dust ASP Conference Series, VoL J04t 1996 BQA> St Gustafion and Marthti 5. Flamcr On the crystalline silicate component of cornet dust 1 1 2 J, Mayo Greenberg , Algen Li , Tctsuo Yamamoto & Takashi Kozasa 3 Laboratory Astrophysics, University of Leiden, The Netherlands; 2. Department of Earth and Planetary Sciences, Hokkaido Univfirsity} Japa n; 3, Department of Earth and Planetary Sciences, Kobe University, Japan Abstract, The effects of p re aggregation silicate crystallinity and post aggregation silicate crystallinity on the 11.3 ftin structure in cornet dust emission have been calculated* Of the order of 10 - 20% crystallinity provides the best agreement to date with observation but the fits arc not yet fully satisfactory. Introduction The 10 /im feature in the emission by comet dust whkh was first detected and identified by Ney (Maas, Ney & Woolf 1970) has now been observed in four comets to consist of two main features: one at about 9.7 firn and the other at about 11.3 firn (see Manner et al 1994a for a review). The former is generally attributed to the amorphous form of olivine and the latter to the crystalline form of olivine. The latter identification is not as secure as the first and has also been variously attributed to a PAH (Polycyclic Aromatic Hydrocarbon) emission or SiC, If comets were formed out of the cold interstellar dust as proposed by Greenberg and others (Greenberg 1982; Yamatnoto 1995) and the interstellar dust contained only amorphous silicates as indicated by its featureless 10 fun absorption, what is the source of crystallinity? If there is some crystalline fraction in the silicates, there are essentially two possibilities for its origin: 1) before the comet formed, some of the interstellar dust was heated to the point at which conversion from amorphous to crystalline silicate occurred; 2) the conversion from amorphous to crystalline occurs after the dust comes off the comet. Since amorphous silicates crystalline at 875 K in 105 hr (Koike & Tsuchiyaraa 1992) while at 1000 K the process is essentially instantaneous (Day Sc Donn 1978), it is vinderstaadable that the interstellar dust could have been heated to the required temperature in the protosolax nebuia. On the other hand the comet dust could never have achieved such high temperatures at the solar distances at whkh the 1L3 firn structure is already observed- For comet Mueller 1993a (Manner et al 1994b) the il.3/£m feature appears at a distance of 2 AU* At this distance^ the temperature of the fluffy comet dust in a McDonnell size distris bution which could provide silicate emission excesses (m < 10~ fl) would be in the range 320 < Td < 500K (Greenberg & Hage 1990; Grcenberg & Li 1996). How the silicate in the comet dust can be partially metamorphosed turns out 497 Downloaded from https://www.cambridge.org/core. IP address: 107.172.128.61, on 12 Jun 2019 at 11:14:14, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0252921100502103 498 to be a somewhat subtler mechanism than direct total hecitiiig but on the other hand, it provides a possibility for arriving at a picture more consistent with the cold intnrstidlar dust aggregation of comets whereas the preheating of the dust leads to some inconsistencies with current understanding of the romet nucleus composition wit it regard to both volatiles and degree of ice crystallinity, Tn section 2 we present a summary of the chemical and morphological structure of interstellar dust based on ivccnt observations and interpretations. The third section is a brief discussion of the consequences to the cornet nucleus of PL 1 V -L ^ J. catino; and rcconuensaAjori or a sigmiic&nt fraction of the interstellar dust in the prolosolar nebula* In section 4 we present some introductory remarks on a. possible mechanism for converting; a fraction of the interstellar silicate to c:rystrillinity by healing to temperature far bolow 800 K. Section 5 contains some calculations of the spectral emission by comet dnsl consisting of flu fly aggregates of partially modi lied interstellar dust minus their volatile fomponents. 2, Interstellar dust components The major emphasis in the paper is lo follow the refractory component of inter stellar dust and hi comets. 2*1. Interstellar silicate The most clearly demonstrated ubiquitous solid material ia interstellar space is concluded from its 10 pm Si-0 stretchh absorption to be n silicate. The optical constants of crystalline olivine — or any other crystalline silicate — provide much sharper features than tliose observed in tlio interslollar medium so that it is generally accepted that the interslollar silicates are amorphous- The most cloarlv defined prop^rtieh^ of tlio interstellar silicates can he. deduced better from polarization than from absorption because the latlor requires a knowledge of the background a.s well as possible silicate emission. The polarization observed for the Beckliri-ISTeugebauer object a.t both f).7 fim and L8 i-trn provides the beat basis for comparison with dust modols* It ca,n bo .shown that no pure silicates whicli have bec^n proposed provide a fully suit-able match to the observation fGreenbere fe Li 1995)- However. if an "astronomical silicate" is defined by tlio measured properties of an amor]>lious (glassy) silicate core with organic refractory mantle (see section 2.2) a remarkably good fit with the observations is possible (Greeuhurg & Li 1995). Since the 3.4 pm feature of organ ics is extremely well correlated witJj llie 9.7 fim absorption (Pendletou et al 1994) it appears that the silicates are indeed coated by an organic refractory mantle so that no (or little) purely silicate particles exist in the interstellar medium. Thus the 1 term "astronomical silicate" is taken to refer to a silicate? core -organic refractory mantle particle. The optical constants of the amorphous silicate MgFeSiO* are ven in Dorsdincret al (1095). 2*2, Organic refractory I he prcHfmcu of a carbonacooii^ component, m interstellar <lust that; exhibits the 3,4 pin features characteristic of C1I stroiches in CII2 ^tu3 CII3 groups is now establishetl (Butchart et al 19S6; Sandford vtt al 1991)- Although many Downloaded from https://www.cambridge.org/core. IP address: 107.172.128.61, on 12 Jun 2019 at 11:14:14, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0252921100502103 499 carbon compounds exhibit features similar to the Interstellar one, a recent survey (Pcndleton et al 19!) I) shows that no previously produrod laboratory analog candidate absorbs exactly like the interstellar dust, A simulation of cyclic interstellar dust evolution which occurs sequentially in molecular and diffuse clouds is accomplished in two stages. The analog of the cyclic ultraviolet processing of interstellar dust described in the following provides a remarkably close match to the observed 3.4 f.tm absorption. In the first stage we produce laboratory residues, which result from ultraviolet photo processing of analog grain mantles that start out as simple ices (H 2 0, CO, CH4> CH3OH, NH3, etc.) at 10 K (Grecnberg 1978). These "first generation " or^anics are then exposed to longterm ultraviolet radiation by exposure to the Sun for the full time available on the EURECA satellite ERA platform (Innocent! & Mesland 1995). Tlie evolution of interstellar dust organics resulting from extensive exposure in diffuse clouds to ultraviolet irradiation of organic mantles produced in interstellar clouds is well simulated by the long-term ultraviolet irradiation of laboratory organic residues in the solar environment. The resulting 3.4 firn infrared feature of the solar irradiated laboratory organic residues is much closet to that of diffuse cloud interstellar dust than that of a wide variety of other suggested sources of organics (Greenberg et al 1995), ui r i g J we present a full infrared spectrum of the solar irradiated organics. Note particularly the absorption at A > Sum which plays an important role in modifying the 9,7 um feature of pure silicates (Greenberg & Li 1995), For the interstellar case, the absorptivity beyond 5 //?n is probably smaller than that for the ERA sample, but for comet organics it could be as high or higher than the Ell A values because the organics made in the molecular cloud contraction leading to comets and solar system have not been exposed to the degree of ultraviolet radiation of the organic refractory mantle as in the diffuse cloud stage. The absorptivities of the organic refractory mantle in the visual and near ultraviolet are responsible for the heating of the comet dust particles as originally noted by Greenberg & Hage (1990). In the current calculation we use m = 0*15 in the visual as in the earlier work (Greenberg & Hage 19J)0)* We note that this is probably an underestimate for processed organics (Jenniskens 1993), 3- Heating of interstellar dust in t h e protusolar nebula Let us assume that the material out of which the cornet nuclei formed was initiaJJy distributed throughout the protosoW nebula, JJI other words some of the interstellar dust was modified by heating to various degrees. As an approximation we shall assume that the early Sun was similar in its radiation output to the current one so that the temperature of the dust at various solar distances can be calculated assuming the present Sun radiation, Using the silicate core-organic refractory mantle model of prcsolar interstellar dust <\s the basic refractory component the temperature of ^ 875 K at which amorphous silicates are crystallized implies that the organic refractory component is at least partially evaporated and the volatile (ices) components are fully evaporated. It will turn out that,: at least 15% of the? silicates in comet dust must be crystallized in order to account for the 11*3 ftm feature. Tf this crystalline silicate is to be recirculated out to the region where comet formation takes Downloaded from https://www.cambridge.org/core. IP address: 107.172.128.61, on 12 Jun 2019 at 11:14:14, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0252921100502103 500 I .r- a 1 ure 1, Comparison of the absorb an re of three solar irradiated residues on tne EURECA satellite ERA plat.form (Grepnberg et al 1995) with the Galactic center GC IRS 6E (Tondletnn fit al 1994), place two factors must be taken into account in the composition of the nucleus aggregation: 1) recondensation of volatile^ on the silicates: 2) chemical processes in the protosolar nebula. We first note that, if even only 15% of the silicate achieves a temperature of 875 K a much lamer fraction of the dust niusl have been raised to temperatures tit which all the ices have evaporated and even some organic; mantles as well. In order to estimate what fraction of ices have been evaporated, we "first assume a radial distribution of the dust and let the number of dust particles which have achieved T > 875 A' ht> 15% nf the total. We then conservatively let. the fractional number of dust particles with T > 300/if be those for which not only the H^O but all volatile^ have been evaporated. The H^O by itself would subliinale at a much lower temperature of ~ 1S0K (Hanner 191 so that th« following estimate of the dcerce of recondensation is a. lower limit. If the total ensemble of dust grains then a.ggregatos to form the cornet nucleus the fraction of recuiidensed to initial ice is greater than llio ratio of the numbers with r T > 300 A to those with T < ,100 A". The radial distribution of temperatures of the dust has heen calculated for individual core-mantle par tides. Tf the ra.dia,l a then the number distribution of initial interstellar dust; particles is n(r) ~ r ratio of the number with fully evaporated volaliks to the number of crystalline PS is fassumine; a spherical disl-TibuticjiO l r (T > 300) a. where I'c is the assumed ininiinuin crysta-llinity temperature?; picv is the fractional nuinber of dust grains with fully evaporated ice and t3c ]b tho fractional number K of crystallized silicates. Using 2, T AU, xAU and /?c — 0-15 the amount of re::on(ienscrt to initial amorphous ice is > 90%. Since IT^O which is recondensed a,t any region in the pi-nsnlar nebula is crystalline ( _ _crystalline ice! 90% Kouchi eX al 1994) the comet nucleus contains at. of pure ainorplicniK ire is Whether or not tho interior of comet nuclei consists Downloaded from https://www.cambridge.org/core. IP address: 107.172.128.61, on 12 Jun 2019 at 11:14:14, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0252921100502103 501 not fully provable but thu general evolutionary properties of comet nuclei appear to be best followed from the idea that it is initially amorphous (Klinger 1080; ania et al 1993; Tancredi et aJ 1994). In addition the volatile chemical distribution appears to be closer to the interstellar dust than to the protosolar cheniicai composition (Mumma et al 1903). In particular the CO/CHj ratio would appear to be incompatible with 90% of the chemical processes having occurred in the solar nebula4, Silicate m e t a m o r p h o s i s in c o m e t d u s t Yamamoto (Yamamoto et al 1996) has proposed a novel way to produce a level of crystallinity in the silicate in moderately heated comet dust- Even if one considers that the presolar systein dust lias never been heated beyond about 30 K, the radicals remaining in the icy mantles which have been photoprocessed at T < ISA' wilt have been totally annealed away in times much, much smaller than the aggregation time scale (Greenberg et a! 1993). However, any remaining radicals in the organic refractory mantles may well p m i s t not nnly during aggregation but also in the billions of years since the birth of the solar system. Because of exponential reduction in diffusion rates with well depth, a relatively moderate increase in radical storage depths is all that is required. However, when the comet dust come off the nucleus its temperature, even at 2 AU from the Sun can reach temperatures of 300—500 K. This is in itself insufficient to crystallize the silicates but, if radical recombination occurs In the organic refractory mantle, radical-radical combination energy of ^ 1 - 2 eV per reaction can be released at the core-mantle interface, A preliminary calculation by mamoto et ill (1996) suggests that to a depth of at least several molecules this can lead to enough local heating to produce cryslallinity* Given a mean silicate radius of 0.07 urn* and a mean molecule diameter of 5 A. a 20% fraction of crystallinity corresponds to a depth of 10 molecules. Whether the process is sufficient to penetrate to this depth is currently under investigation, If it is possible, the presence of significant crystal]] iiily in comet dust does not require the precometary interstellar <Iust ever to have been heated beyond 30 K. Thus the 11.3 /im emission component of comet dust could be consistent with cold n of interstellar dust. m 5, r r 10 /jm emission by comet dust In view of the fact that only small (— \ftm) size silicate particles can reproduce the observed width of the 10 firn emission of comet dust it is tempting but inconsistent to consider them as the direct source of the spectral feature. Using the scattering/absorption properties of only such small particles ignores the observed comet dust mass distribution- A consistent way of calculating the spectral energy distribution of comet dust was described by Greenberg and Hagc 1090) in terms of aggregation of submicron silicate core-organic refractory man1 tle interstellar dust particles* We shall follow the same procedure here without discussing the details of tho scattering calculation in which Mie theory was used for the spherical aggregates, Wo note that for high porosity aggregates the opti- Downloaded from https://www.cambridge.org/core. IP address: 107.172.128.61, on 12 Jun 2019 at 11:14:14, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0252921100502103 502 cal resonance by individual particle is diluted by the presence of a large vacuum fraction. 12 CJ B 10 CM 8 - x E 6 ~ 8 10 X (/urn) 12 14 8 10 X 12 14 1 =*-0.8 " B 0.6 CM 7 x 0.4 ~ CO 0.2 0 Figure 2. a) The observational and theoretical spectra of comet Halley at 0.79 AU: points — observational data (Campins & Ryan 1989); solid line — model result assuming 6% pre-aggregation crystallinity; dotted line — the same as solid line but with the 9.7 fim absorptivity of crystalline silicate reduced to 1/2; thin solid line — the assumed dust thermal emission, b) The same as a but for the silicate excess emission (i.e., with the dust thermal emission subtracted). We represent comet dust as spherical aggregates of core-mantle particles with uniform porosity P and size distribution as given by McDonnell et al (1989). The size distribution undoubtedly depends both on the comet solar distance and on the radial distance of the dust from the nucleus. There is abundant evidence for some fragmentation with distance. There also appear differences in the spectral emissivity of different comets with varying 11.3 fim structure (Hanner et al 1994a). However our first aim will be to determine whether, within some framework, varying mainly the degree of crystallinity may lead to some of the main features of comet dust emissivity in the 10 fim region. We have considered porosities from 0.93 to 0.975 but the following results are limited to P = 0.95. Downloaded from https://www.cambridge.org/core. IP address: 107.172.128.61, on 12 Jun 2019 at 11:14:14, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0252921100502103 503 5.1. Pre-aggregation crystallization We model the dust as consisting of a mixture of pure crystalline silicates (as defined by Mukai & Koike 1990) mixed in with amorphous silicate core-organic refractory mantle particles with the mantles equal in mass to the cores. In Fig.2 the percentage of crystalline to amorphous silicate is 6%. Fig.2a shows the spectral energy distribution and Fig.2b shows the net emission at 10 fim after dust thermal continuum subtraction. As already noted earlier (Greenberg & Li 1995) the short wavelength side of the 10 fim emission is dominated by the presence of the organic refractory mantle absorptivity. Without such mantles the short wavelength side would be substantially too low compared with the comet emission. It is obvious that with only 6% crystallinity the 11.3 fim feature is not produced. In view of uncertainties in the spectral properties derived by Mukai Sz Koike (1990) we have tried to estimate what effect there would be if the short wavelength portion of their crystalline absorptivity were reduced, thus enhancing the relative effect of the 11.3 fim feature. This effect occurs but the combined optical properties in the ensemble (by M ax well- Game tt (1904) effective medium theory) appear to result in a shift of the crystalline feature to significantly shorter wavelength. What prescription one requires to provide the observed peak wavelength is not clear. ^0.8 - a 0.6 - 0.4 x a 0.2 o 8 10 A 12 14 Figure 3. The observational and theoretical excess emission of comet Halley at 0.79 AU: points observational data (Campins & Ryan 1989); solid line — model result assuming 15% pre-aggregation crystallinity; dotted line — the same as solid line but with the 9.7 fim absorptivity of crystalline silicate reduced to 1/2; dashed line — the same as solid line but with the 9.7 fim absorptivity of crystalline silicate reduced to 0. Since 6% crystallinity seemed inadequate we have tried the same model assuming 15% crystallinity. The series of curves shown in Fig.3 indicate what happens using not only the full Mukai & Koike crystallinity silicate structure but also using arbitrarily reduced 9.7 fim by a factor of 1/2 or 0. The effects are rather striking in that not only is the "11.3" fim feature enhanced it is also made much sharper. This is a property produced again by using Kramers-Kronig relation on the fluffy aggregates optical properties. Downloaded from https://www.cambridge.org/core. IP address: 107.172.128.61, on 12 Jun 2019 at 11:14:14, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0252921100502103 504 n the final analysis it appears that one requires about 15% crystallmity to produce the required extra feature "strength" hut we, hesitate excellent match with the observation because this extra structure is in position and modified in shape. We have also considered the dust Tt is worth noting that the calculated crystalline feature is shifted observed peak position if one uses elongated comet dust particles. to claim an both shifted shape effect. dose to the 1 0.8 5 0.G 0.4 0.2 o 8 10 12 The theoretical excess dust emission of comet flalley by Figure 4 assuming 20% post-aggregation silicate crystallinity. Post-aggregation crystallization We have chosen to present the result only for the case where the silicate core is crystallized at its surface to 20% of its volume. The result shown in Fig.4 is not very encouraging. The strength of the crystalline feature for this morphology is lamely hidden, and furthermore, even the short wavelength side of the 10 m band appears to be reduced. We reiterate that this must be considered a very preliminary result until further examination can be made of such effects as produced by varying size distribution of the comet, dust, for example. oncluding remarks We have considered two possible ways ol obtaining a degree of silicate crystal! inity in comet dust. The one which assumes that interstellar dust is preheated in the solar nebula leads to a model of comet dust which gives a rather good representation of the 10 txm structured emission feature with the assumption of a,l>o\il 15% cry at a.U in it v. However, tlie problem I hen is t,ha,t t.hu comet does not act as if it contains 85% unmodified interstellar dust but rather as if it contains less than 10% unmodified interstellar dust. If this is the case there would be substantial inconsistencies with current ideas fibout the nucleus conslating of predominantly amorphous ice and current interprelation of the coma material composition with respect to interstellar dust ice mantles. On the other is a surface phenomenon on the amorphous silicate core,, jj if crystaJlinity y y p p he (Just optical pioportjes do not provide a very good match tu the observed 10 firn structure. It is our suggestion that, in spite of the lack of good agreement produced by surface crystallinity, its effects should be studied in greater depth. Downloaded from https://www.cambridge.org/core. IP address: 107.172.128.61, on 12 Jun 2019 at 11:14:14, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0252921100502103 505 Perhaps a mixture of the two processes sliuuld also be considered though it Is important to bear in mind that any large de^r^e of pra-comet silicate crystallinity Implies a much larger degree of change in the comet nucleus composition and morphology. A c k n o w l e d g m e n t s . We are grateful for the support by NASA grant NGit 33-018-148 and by a grant from the Netherlands Organization for Space Research (SRON) for research on the organic refractories in space. We would like to thank Dr. M, Haimef and Dr. H, Campins for sending us tho comet Halley silicate emission data* We also thank Dr. M. Hanner for some helpful suggestions. One of us (AL) wishes to thank the World Laboratory for a fellowship. Refe r e n ccs Hutcliart, I., McFadzean, A,D., WliiUet, D.C.B., (iGballo, T.R. &; Greonberg, J.M., 1086, A&A 154, ho Campins, IT. & Ryan, E., 1989, ApJ 341 ? 1009 Day, K.L. & Donn, B., 1978, ApJ 222, L45 Dorsclmcr, J., Degemaiin, 11., Henningj Tli., .lager, C. & Mutschke, II., 1995T A&At 300, 503 , J.M. 1978, in : Cosmic Dust, edL McDonnell, J.A.M., Wiley, P. 187 Groonbcrg, J.M., ]f)82, in : Comots, e.d. VVilkcning, L.L., University of Arizona Press, P. 131 Grceubcrg, J.M. k llago, J.T., 1990, ApJ 361, 2G0 Green berg, J.M, Mend oza- Gomez, C.X., dc Groot, M.S. & Breukere, R., 1993, in: Dust and Chemistry in Astronomy, etls. Millar, T.J. & Williams, D.A,,TOL Publ. T Green berg, J.M. & Li, A., 1995, A&A, iu press Grocnborg, J.M., Li, A., Menrioza-Gomoz, C.X., Schutlc, W.A., Gerakines, P.A. dc Groot, M., 1995, ApJ 455, LI77 ^, J.M. & Li, A., 1996* submitted to Planetary & Space Science IIaimer, M.S.» 1981, Icarus 47, 342 Han nor, M.S., Lynch, D.K., & Russell, Xt.W., Ifl94a, ApJ 425, 274 Han net, M.S., Hack well, J.A., Russell, Jl.W. k Lynch, D.IC, 1994b, Icarus 112, 490 Haruyama, J-, Yamamoto^T,, Mizutani, H. &: Greenberg, ,KM, 199'3, Journal of Geophysical Research, Vol.98, N0.E8, 1M5079 Innocvnti, L., Meslaud, D.A.M. eds EURIvCA Scientific liesuIts, Adv. Sp. Res, 15, # 8 . ISSN 0273 117 Jenniskons, P., Baratta, G.A., Kouchi, A., de Groot, M.S., Gr^enbcrg, J.M. & Strazzulia,, G. 1993, A&A 273,583 Jenniskons, P., 199:!, AfeA 274, 653 Klingor, J., 1980, Science 209, Koike, C. k Tsuchiyama, A., 1992; MKRAS 255, 248 • 1 • 1 S Downloaded from https://www.cambridge.org/core. IP address: 107.172.128.61, on 12 Jun 2019 at 11:14:14, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0252921100502103 506 Konchi, A., Yamamolu, T., Knzasa,, T., Kuroda, T. & Green berg, J.M., 1994, A&A 2f)0, 1009 Maas, H.W., Ney, E.l\ & Woolf, N.F., 1970, ApJ 160, L101 McLxwcll-Garnett, J.C-, 1904, Phil/Iran s.R.Soc, London, 203A, 385 1 McDonnell, J,, Panlriewicz, G.? Birchley, I ., Green, S. & Perry, C ? 1089, in Proc. Workshop on Analysis of Returned Comet Nucleus Samples, Milpitas, Muiai, T. k Koike. C , 1990, Icarus 87, 1H0 Murnina, M-? Weissman, P. & AlansterTi? 1993, in: ProtostEirs and Planets III,* l . L i JT h MS Ui Levy EH E.H.. L Limiiie J,T. & M Mathews M.S., T Tucson.: Univ. Arizona ^ss, P. 1177 Peiidleton, Y.J., Sandfnrd, S.A., Allamandola, L.J., Tielens, A.G.G.M. & Sellgren, K. 1994, ApJ, 437,683 Sandford, S.A., Allama.ndola, L.J., Ticlcns, A.G.G.M., Sollgron, K., Tapia., M. & Pcndlcton. Y.J. 1991, ApJ. 371, 607 Tancrcdi, G-, Rkkin^n s II, & Grcenberg, J.M., 1994, A&A 286, 659 ,o. T., 1995, in; The Cosmic Du^t Cuiuicction, cd. Grccivbcrg, Kluwcr, Dordrecht, in press ;o? T., Kozasa, T M Shirono, S., Green berg, J.M, fe Li? A.. 199C, in preparation ? Downloaded from https://www.cambridge.org/core. IP address: 107.172.128.61, on 12 Jun 2019 at 11:14:14, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0252921100502103 7 J ' *