•j!/M UNAMBIGUOUS SPECTRAL EVIDENCE FOR HIGH- (AND LOW-) CALCIUM PYROXENE IN ASTEROIDS AND METEORITES. J.M. Sunshine _,S. J.Bus2,T. H. Burbine*, T. J. IvlcCoy;, and R. P. Binzet 4 _Science Applications International Corporation, Suite 400, 4051 Daly Dr., Chanti/ly, VA 20151 ([email protected]), 2Uni'_:ersity of Hawaii, institute for Astronomy, Hi[o, HI 96720, 3Department of Mineral Sci- ences, National Museum of Natural History, Smithsonian Institution, Washington, DC 20560, *Department of Earth, Atmospheric, and Planetary, Sciences, Massachusetts institute of Technology, Cambridge, MA 02139. Introduction: Spectroscopy remains a powerful PYX absorptions and may be similar to ordinau' chon- tool for inferring the modal mineralogy and mafic drites [9]. Finally, as discussed below, some asteroids mineral composition of asteroid surfaces [1,2]. Since are dominated by PYX absorptions. similar measurements can be made on meteorite sam- AIGM Modeling. MGM modeling of 847 Agnia pies, spectroscopy can help link the two populations (Fig. 1) reveals a surface composed of a mixture of and add spatial and geologic context to detailed geo- two pyroxenes. The LCP results in two bands (red), chemical knowledge derived from meteorite samples. while HCP results in three bands (blue). The 1.25 _m For example, analysis of the recent NEAR- band has been interpreted by many as an indication of Shoemaker mission to Eros include detailed study of plagioclase. However, the spectra of PYXs (particu- NIS spectra to assess the affinity of Eros to ordinary larly HCP) contain an absorption in this region due to chondrites [3-5]. As discussed in these studies, pyrox- the presence of Fe'-+ in the M1 crystallographic site ene (PYX) and olivine (OLV) absorption are readily [14]. It is noteworthy that there is no indication of detectable in the spectra. Furthermore, subtleties in OLV in the MGM fit to 847 Agnia. Two other mem- band parameters (position vs. area) suggest the pres- bers of the Agnia family have also been measured with ence of both low- and high-calcium pyroxene (LCP SpeX and have similar spectral properties. and HCP) [3-5], as expected from the petrology of or- The MGM fit to 17 Thetis (Fig. :.2)appears almost dinary chondrites [e.g. 6]. However unambiguous identical to that of 847 Agnia, indicating these two identification and detailed compositional inferences for asteroids have very similar surface compositions. The both LCP and HCP (and OLV) are difficult from band only significant difference between these two spectra is parameters analysis, Inthis study, we examine spectra in the absolute strengths of the absorption bands, with of S-asteroids and meteorites with the Modified Gaus- 17 Thetis having weaker bands. Interestingly, 17 The- sian Model (MGM), an absorption band model [7], to tis plots in the S(VI) field under the Gaffey et al. clas- explore the role of HCP in these silicate-rich spectra. sification [5], which is interpreted to be rich in LCP, Data: Asteroid Spectra. High quality spectra of but with an OLV component. Our analysis shows no 17 S-asteroids have recently been measured using evidence of OLV, but instead, significant HCP. SpeX, a low- to medium-resolution infrared spectro- While visual and MGM analysis of 17 Thetis and graph, newly commissioned at the Infrared Telescope 847 Agnia show them to be spectrally similar, the ra- Facility (IRTF) on Mauna Kea, Hawaii [8-9]. In its tios of the 1- and 2- tam band areas (1.17 and .55) sug- low-resolution mode (R - 100), SpeX can produce gest these two asteroids have very different composi- spectra of faint asteroids from 0.8 to 2.5 gm with S/N tions (i.e. based on Cloutis et al. [15]). The presence comparable to data typically collected with visible of HCP is apparently affecting the continuum points, wavelength CCDs. The SpeX data have been com- which in-turn affects the band ratios. MGM derived bined with visible CCD data measured during the band ratios are more consistent (.92 and .74). Taken at SMASSII survey to produce high S/N spectra from face value these band areas suggest significant OLV 0.44 to 2.5 gm for several asteroids. components (45-70%), yet recall there is no evidence Meteorite Spectra.. Burbine et al. [10] have meas- of OLV absorptions in these spectra. ured >70 spectra of meteorites from the well charac- Interpretation: The inferred mineralogy of these terized Smithsonian's Analyzed Meteorite Powder asteroids, two pyroxenes (LCP:HCP ~60:40; based on collection [ll]. These data and previous studies on LCP to HCP band ratios [16]) and little OLV (<15%) other meteorite types, including primitive achondrites imply that they are evolved igneous bodies. Among the [12] and eucrites [13], are analyzed in parallel for meteorites, such compositions are most similar to the comparison to the asteroid spectra. basaltic achondrites. For comparison, the spectra of Results: The 17 S-asteroids measured span a large basaltic achondrites are also examined with the MGM. range of the S-asteroid class. Some have very strong For example, the MGM fit to the Bouvante basaltic olivine absorptions (OLV>>PYX) and represent a eucrite (Fig. 3), reveals the presence of both LCP and continuum between S- and the olivine dominated A- HCP, and no OLV. These MGM results are very con- asteroids [8]. Others contain strong OLV and strong sistent with the known composition of Bouvante. ItI(III-CAI,CII ;Xl I"_'I),OXENE IN ASI'EROIDS AND METEORITES: J. IM.Sunshine, etal. MGM derived LCP:HCP range from 72:28 (from I gm Figures: MGM model fits to 847 Agnia, 17 The- bands ratios) to 58:42 (2 gm band ratios), while the tis, and the Bouvante eucrite. Plotted in each fig- electron microprobe modal analyses shows LCP:HCP ure is the residual error (purple, offset 10%), the of 68:32 [17]. In addition, the band centers for both individual modified Gaussian distributions (solid pyroxenes in Bouvante are at long wavelengths, which lines), the baseline continuum (dashed), and the is very consistent with petrologic results that show it to modeled spectrum (black line) overlying the be iron-rich (LCP: Wo_En3_Fss8 [18]). Bouvante is measured spectrum (orange). All spectra include also known to have no significant OLV component. visible (gray), LCP (blue) and HCP (red) absorp- Finally, it should be noted that Bouvante contains 45% tions, yet show no evidence of OLV. plagioclase. However, as noted above, ambiguity with Fig. 1: MGM Fit to 847 Agnia M I pyroxene absorption complicates any spectral de- 02 • • termination ofplagioclase content. This analysis gives confidence to the MGM results for asteroid spectra. Compared to Bouvante, the aster- oids are less iron-rich and have more HCP suggesting, on average, a more gabbroic composition. It is likely however, that the hemispherically averaged asteroid -O,4 spectra are mixtures of evolved igneous lithologies. Conclusions: The high quality SpeX spectra in- clude detailed absorption features which are accurately t modeled with the MGM. In several S-asteroids and 'QlJ eucrites, we find evidence for both LCP and HCP, with little to no OLV. Similar modeling efforts with ordi- 0.$ ii1 1,11 _a _l.s nary chondrite spectra, unambiguously require OLV, LCP, and HCP absorptions with inferred modes and Fig. 2: MGM Fit to 17 Thetis compositions that are consistent with known petrogra- phy [10]. It is now clear that HCP is a significant con- tributor to the spectra of S-asteroid and silicate-rich meteorites. We therefore must approach the analysis I° of these spectra not simply as OLV-PYX mixtures, but as ternary mixtures of OLV, LCP, and HCP. •QA ¢ Acknowledgements: Support for this research from NASA's PGG program (NASW-00012 to J.M.S.) .OII is greatly appreciated. All meteorite spectra were col- lected at Brown University's Keck/NASA Reflectance 1 L..C..m .,m--, t Experiment Laboratory (RELAB), a multi-user facility supported by NASA (NAG5-387 l). References: [i] Gaffey M. J. et al. (1993), Meteoritics, ._.Q . L............. ,t............... a................. t......... 0,4, !.0 1,11 3.0 I1.I 28, 161-187. [2] Pieters C. M., and McFadden, L. A. (1994),Rev Earth and Planet Sci, 22, 457-497. [3] McFad- Fig. 3: MGM Fit to Bouvante Eucrite den L. A. et al. (2001) MAPS, 36, 1711-1726. [4] McCoy T. J. et al. (2001) MAPS, 36, 1661-1672. [5] Gaffey M. L et al. (1993) Icarus, 106, 573-602. [6] McSween H. Y. Jr. et al. (1991)Icarus, 90, 107-116. [7] Sunshine J. M. et aL, JGR, | "* L&2--: --- 95, 6955-6966, 1990. [8] Binzel R. P. et al. (2001), MAPS, 36, !167-1172. [9] Bus S. J. et al. (2001), BuIIAAAS/DPS 33, #41.16. [10] Burbine et aL (2002) LPSC, X,Eglll this volume. [11] Jarosewich E. (1990) Meteoritics, 25, 323-337. [12] Burbine T. H. et aL (2001) LPSC, XXXll, #1860. [13] Burbine T. H. et aL (2001) MAPS, 36, 761-781. [14] Burns R. G., (1993) Mineralogical Applications of Crystal Field Theory, 551 pp. [15] Cloutis E. A. et al. (1986), JGR, 91, t_-Qalm • 11,641-11,653. [16] Sunshine J. M. and Pieters C. M. (1993) JGR, 98, 9075-908. [17] Christophe Michel-Levy, M., et. al. <2" t Hl_h_ab=la. P_jr_-- (1987), Bull Mineral 119, 449-458. [18] Delaney, J. S. et al. (1984), LPSCX1; 210-211. aJ 1.0 1.§ 2.11 2.§ SPECTRAL MEASUREMENTS OF METEORITE POWDERS: IMPLICATIONS FOR 433 EROS. T. H. Burbine _,T. J. McCoy _,E. Jarosewich _,and J.M. Sunshine:, _Department of Mineral Sciences, National Museum of Natural History, Smithsonian Institution. Washington, DC 20560-0119, USA ([email protected]), 2Advanced Technology Applications Division, Science Applications International Corporation (SAIC), 4501 Daly Drive, Chantilly, VA 20151. Introduction: One of the goals of the NEAR- mass available. These powders were prepared under Shoemaker mission to 433 Eros was to determine if it clean conditions so that they can be used for trace ele- has a meteoritic analog. The primary means of making ment analyses and other types of chemical studies. such a link are the X-ray/gamma-ray spectrometers [1], The meteorite samples were ground in an agate which measure elemental compositions of the surface, mortar in a hood with positive air pressure and then and the multi-spectral imager (MSI) and near-infrared sieved through a nylon sieve to usually pass through spectrometer (NIS) [2], which measure spectral re- 100 mesh (< 150 _m). A few samples were sieved to flectance. pass through 200 mesh (< 75 _m). This fine-powdered For determining meteoritic analogs using the X- fraction (< 150 _m) consisted of silicates and sulfides ray/gamma-ray spectrometer data, the primary data and contained up to 0.4 wt.% of fine-grained metal. used for comparison is the set of bulk chemical analy- The meteoritic fraction larger than 100 mesh (>150 ses of meteorites done by Jarosewich [3]. These bulk _m) was primarily metal. A detailed description of the chemical analyses were done on samples now found in preparation of powders and the analytical procedures is the Smithsonian's Analyzed Meteorite Powder collec- given in Jarosewich [3]. Normative mineralogies for tion (USNM 7073). For determining meteoritic ana- most of the ordinary chondrites, based on the bulk logs using MSI/NIS spectral data, the primary data chemical analyses, are found in McSween et al. [12]. used for comparison is the set of meteoritic spectra Reflectance spectra at room temperature were ob- compiled by Gaffey [4]. tained on the powders using the bidirectional spec- To expand the set of meteoritic spectra available to trometer at RELAB. The incident angle was 30 de- the scientific community, we have initiated a spectral grees and the emission angle was 0 degrees. The study of over 70 samples (primarily ordinary chon- spectral coverage was 0.32 to 2.55 jam with a sampling drites) found in the Smithsonian's Analyzed Meteorite interval of 0.01 p.m. Almost all samples were also Powder collection and an electron microprobe study of measured out to 26 lam using a Fourier transform infra- their corresponding thin sections. This set of spectral red (FTIR) spectrometer. Selection of the powders for and compositional data should allow for better con- spectral analysis was done after visually examining straints on the distribution of meteorites in plots of each sample to check for possible weathering effects. band area ratios versus Band Icenters [5] and the use- Results: Reflectance spectra and the olivine and fulness of equations for deriving mineralogic compo- pyroxene mineralogies for the -70 meteorites are cur- sitions from band parameters [6]. These spectral data rently being compiled. To determine the Band I cen- can also be combined with previous spectral studies of ter, a linear continuum tangent to Band I is first di- other meteorite types such as the primitive achondrites vided out. We have now measured the band areas and [7], euerites [8], and angrites [9] to determine how band centers of 18 meteorite powders. We are cur- useful the derived band parameters are for differenti- rently estimating error bars for the calculated values. ating between different meteorite classes. These spec- Their values are plotted on a Band Area Ratio plot tral data can also be used for testing the Modified (Figure 1) with the olivine, ordinary chondrite, and Gaussian Model (MGM) [I0,11] for determining mo- basaltic achondrite regions defined by Gaffey et al. [5]. dal abundances and marie mineral chemistries from The average band area and band center for 433 Eros reflectance spectra. determined by NEAR-Shoemaker [6] are also plotted. Meteoritic Powders: As part of the Smithsonian's An interesting preliminary result is that the new or- meteorite research program, approximately 300 mete- dinary chondrite measurements fall in the lower range orites have been analyzed in the past 36 years for bulk of the ordinary chondrite region defined by Gaffey et chemical data [3]. The primary requirement for analy- al. [5]. This offset is due to the calculated Band I cen- sis was to obtain, if possible, sufficient amount mate- ters tending to be at lower wavelengths than those cal- rial to assure a representative sample of the whole me- culated by Gaffey et al. [5] for similar types of ordi- teorite, and to retain some material for future studies. nary chondrites. Measurements of all the ordinary Usually from five to forty grams of a meteorite were chondrite powders should determine if this offset is powdered depending on the type of meteorite and the real. It is important to determine the cause of this off- NPEICRALMEASUREMEONFTMSETEOR[PTOEWDERTS.H: .Burbineetal. setsince433Erosdoesnotyetplotintheregionde- [6] McFadden L. A. et at. (2001) Aleteoritics & tinedbythefirstmeasuremeonfttshesenewordinary. Planetary Science, 36, 1711-1726. [7] Burbine T. H. et chondritpeowders. al. (2001) LPSC, XVXII, #1860. [8] Burbine T. H. et al. Onepossibilityforthisoffsetistheahnosctom- (2001) Meteoritics & Planetary Science. 36, 761-781. pleteremovaolfmetallicironfromthemeasurepdow- [9] Burbine T. H. et al. (2001) LPSC, ,EL\TL #1857. ders.MorozandArnold[13]havefoundthataddition [10] Sunshine J. M. and Pieters C. M. (1993)JGR. 98, of opaques(e.g., pyrrhotite)to an olivine- 9075-908 [11] Sunshine et al. (2002) LPSC. X.VVIII orthopyroxemneixture tends to move the Band Icenter this volume. [12] McSween H. Y. Jr. et al. (1991) Ica- to longer wavelengths. We are currently planning on rus, 90. 107-116. [13] Moroz L. and Arnold G. (1999) re-mixing the metallic iron with the fine-powdered JGR, 104, 14109-14121. [14] Gaffey M. J. (1984) Ica- fraction to determine if the Band Icenter changes with rus, 60, 83-114. [15] Gaffey M. J. and Gilbert S. L. the addition of metallic iron. (1998) Meteoritics & Planetary Science. 33, 1281- Another possibility is that the wavelength calibra- 1296. [16] Pieters C. M. and Pratt S. F. (2001) personal tion correction of 0.025 m that is used [14,15] on the communication. [17] Russell S. S. et al. (1998) Mete- Gaffey [4] meteorite data is too large. This correction oritics & Planeta_ Science, 33, 853-856. [18] Schulze is done because of a later-discovered wavelength offset H. et al. (1994) Meteoritics, 33, 275-286. found between the spectra taken by the spectrometer Acknowledgments: Brown University's that measured the Gaffey [4]_meteorites and more re- Keck/NASA Reflectance Experiment Laboratory cent spectra. By comparing spectra of an Apollo soil (RELAB) is a multi-user facility supported by NASA sample taken by the Gaffey [4] spectrometer and the grant NAG5-3871. RELAB spectrometer, Pieters and Pratt [16] found that 1,15 a non-linear wavelength correction, which has an off- "olivine set of +0.015 _m at a wavelength of 0.92 _m, was bet- 3 ter. We are planning on measuring a number of ordi- r-IC_londnie s • L¢hol_lntQl nary chondrites also measured by Gaffey [4] to deter- _¢_1.o5 11[11tlftlll • LL chondntH mine which wavelength calibration is better. • Bumwlll Of the meteorites measured, two of the most inter- _(_dinary chondritesbasahii"':"h:_rites esting mineralogicaily are the ungrouped chondrite Burnwell [17] and the R-chondrite Rumuruti [18]. 0.95 I Bumwell is more reduced than H chondrites with oli- vine compositions of Falsuo2 and pyroxene composi- 0.9 ........ _1; -w............. 2" i i i i3l tions ofFsu.4,0.7. Burnwell plots amongst the H chon- BandAreaRatio drites in the Band Area Ratio plot (Figure 1). Rumu- ruti is an oxidized (-Fa39), olivine-rich, metal-poor chondrite. As expected, Rumuruti plots within the Figure 1. Plot of Band Area Ratios versus Band I Centers. olivine region (Figure 1). .... ,__ , ..... A preliminary fit (Figure 2) using MGM was also MGM Fit to Bandong (LI_6) done on LL6 chondrite Bandong. The model requires olivine and two pyroxenes (a low-calcium and a high- 0£ calcium). Inferred modal abundances derived from 0 this model are withing 10% of those derived from m • normative calculations. These results are consistent _ -0.2 with those of McFadden et al. [6] where both low- and high-Ca pyroxenes are needed to produce the Band I .1 centers consistent with those of ordinary chondrites. -j -0.4 Low-CalciumPyroxene Conclusions: We are re-examing the regions de- f'med by different meteorite classes in Band Area Ratio High-CalciumPyroxene Olivine plots. These data will allow us to better determine the compositions of asteroids fi'om remote sensing. "0,6 i I i , * , I .... I , , , , I _ i , , I References: [1] Trombka J. I. et al. (2000) Science, 0.5 1,O 1.5 2.0 2.5 Wavelength(pm) 289, 2101-2105. [2] Veverka J. et al. (2000) Science, Figure 2. MGM fit to LL6 chondrite Bandong. 289, 2088-2097. [3] Jarosewich E. (1990)Meteoritics, 25, 323-337. [4] Gaffey M. J. (1976) JGR, 8I, 905- 920. [5] Gaffey M. J. et al. (1993) Icarus, 106, 573-602.